A TRACER AIDED STUDY ON SILICON CHEMISTRY IN BIOLOGICAL ... · PDF fileA tracer aided study on...

A TRACER AIDED STUDY ON SILICON CHEMISTRY

IN BIOLOGICAL SYSTEMS

A tracer aided study on silicon chemistry in biological systems

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof.dr.ir. J.T. Fokkema, voorzitter van het College voor Promoties,

in het openbaar te verdedigen op 14 december 2009 om 10.00 uur

door Helena Josephina BRASSER

scheikundig ingenieur geboren te ‘s‐Gravenhage

Dit proefschrift is goedgekeurd door de promotor: Prof. dr. H.Th. Wolterbeek Samenstelling promotiecommissie: Rector Magnificus, voorzitter Prof.dr. H.Th. Wolterbeek, Technische Universiteit Delft, promotor Prof.dr. B. Markert, International Graduate School Zittau, Duitsland Prof.dr. W.W.C. Gieskes, Rijksuniversiteit Groningen Prof.dr.ir. J.J. Heijnen, Technische Universiteit Delft Prof.dr.ir. M. de Bruin, Technische Universiteit Delft Prof.dr.ir. M.C.M. van Loosdrecht, Technische Universiteit Delft dr. G.C. Krijger, Universiteit Wageningen © 2009 The author and IOS Press All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without prior permission from the publisher. ISBN 978‐1‐60750‐083‐4 Keywords: silicon, silicon chemistry, tracer, Si‐31, Baker’s yeast, diatom, biofilm Published and distributed by IOS Press under the imprint Delft University Press Publisher IOS Press Nieuwe Hemweg 6b 1013 BG Amsterdam The Netherlands tel: +31‐20‐688 3355 fax: +31‐20‐687 0019 email: [email protected] www.iospress.nl www.dupress.nl LEGAL NOTICE The publisher is not responsible for the use which might be made of the following information. PRINTED IN THE NETHERLANDS

mailto:[email protected]�

http://www.iospress.nl/�

http://www.dupress.nl/�

iv

Table of contents 1. Introduction..........................................................................................................1

1.1 General properties of silicon..........................................................................1 1.2 Silicon in living systems.................................................................................2 1.2.1 Sponges and protozoa..........................................................................3 1.2.2 Diatoms................................................................................................3 1.2.3 Higher plants........................................................................................5 1.2.4 Higher animals......................................................................................5 1.2.5 Overview...............................................................................................7 1.3 Thesis outline................................................................................................8 1.4 References..................................................................................................10

2. Preparation of 31Si labelled silicate: a radiotracer for silicon studies in biosystems..........................................................................................................19

2.1 Introduction................................................................................................20 2.2 Experimental...............................................................................................21 2.2.1 Chemicals...........................................................................................21 2.2.2 Preparation of the radionuclide...........................................................21 2.2.3 Removal of phosphate, including 32P from the radionuclide solution...22 2.2.4 Analyses.............................................................................................23 2.2.5 Gel formation.....................................................................................23 2.2.6 Paper chromatography.......................................................................23 2.2.7 Interaction of 32P with non‐radioactive Si............................................24 2.3 Results and discussion.................................................................................24 2.3.1 Production of the radionuclide............................................................24 2.3.2 Removal of 32P from the irridiated target material...............................25 2.3.3 Investigation of the chemical form of 31Si............................................28 2.4 Conclusions.................................................................................................31 2.5 References..................................................................................................32

3. The Influence of Silicon on Cobalt, Zinc and Magnesium in Baker's Yeast Saccharomyces cerevisiae...................................................................................35

3.1 Introduction.................................................................................................36 3.2 Materials and Methods................................................................................36 3.2.1 Chemicals and labware........................................................................36 3.2.2 Organism and culture procedures........................................................37 3.2.3 Analyses..............................................................................................38 3.2.4 65Zn and 60Co uptake in yeast cells......................................................38 3.3 Results.........................................................................................................38 3.3.1 Influence of silicate on the growth rate of Baker’s yeast in standard medium..............................................................................................38 3.3.2 Influence of silicate on metal levels in Baker’s yeast............................39 3.3.3 Influence of silicon on growth rate at toxic or deficient levels of metals............................................................................................40 3.3.4 Influence of silicon on zinc and cobalt uptake in the cell......................41 3.3.5 Comparison of the effect of silicate and germanium on baker’s yeast..42

v

3.4 Discussion and conclusions..........................................................................43 3.5 References..................................................................................................46

4. On the beneficial role of silicon to organisms: a case study on the importance of silicon chemistry to metal accumulation in yeast................................................49

4.1 Introduction................................................................................................50 4.2 Materials and Methods................................................................................51 4.2.1 Chemicals and labware.......................................................................51 4.2.2 Organism and culture procedures.......................................................51 4.2.3 Analyses.............................................................................................52 4.2.4 31Si accumulation in yeast cells...........................................................52 4.2.5 Determination of 31Si release..............................................................52 4.2.6 Determination of the Freundlich parameters 1/n and K.......................53 4.2.7 Determination of the depolymerization rate constant.........................53 4.3 Results........................................................................................................54 4.3.1 Determination of experimental conditions and non‐biological components in silicate‐cell interaction................................................54 4.3.2 Determination of silicate adsorption on the cell wall...........................56 4.3.3 Interaction of metal ions with adsorbed silicate...................................58 4.4 Discussion...................................................................................................60 4.5 References..................................................................................................62

5. Systematic compartmental analysis for describing observed 31Si‐labeled silicic acid uptake during diatom valve formation – A mathematical approach.............65

5.1 Introduction................................................................................................66 5.2 Materials and Methods................................................................................68 5.2.1 Chemicals for radioactivity studies......................................................68 5.2.2 Chemical analyses of medium and cells..............................................68 5.2.3 Organisms and culture conditions.......................................................69 5.2.4 Experimental procedures....................................................................69 5.2.5 Mathematical approaches...................................................................70 5.3 Results.........................................................................................................72 5.3.1 Silicic acid uptake................................................................................72 5.3.2 Mathematically defined number of observable compartments in P. laevis...............................................................................................73 5.3.3 Comparative mathematical assessment of model B and the SIT‐ mediated silicic acid uptake model in P. laevis.....................................74 5.3.4 Mathematical assessment of alternative silicic acid uptake models in P. laevis...............................................................................................76 5.3.5 Kinetic parameters of (macro)pinocytosis‐mediated silicic acid uptake in P. laevis............................................................................................77 5.3.6 Compartmental analyses for other diatom species..............................80 5.4 Discussion...................................................................................................81 5.5 References..................................................................................................83 5.6 Supplement 1: mass transfer over the Nernst layer......................................85 5.6.1 Reference...........................................................................................85 5.7 Supplement 2: mathematical formulation of the models.............................86

vi

5.7.1 Determination of the number of compartments (Model A‐E)..............86 5.7.1.1 Model A: the cell as black box (no separate compartments in the cell).................................................................................87 5.7.1.2 Model B: two cellular compartments and an unidirectional flux to the final compartment.....................................................87 5.7.1.3 Model C: three cellular compartments and an unidirectional flux to the final compartment...............................................88 5.7.1.4 Model D: two cellular compartments and a bidirectional flux between all cellular compartments (i.e. model B with efflux from the final compartment included)........................89 5.7.1.5 Model E: three cellular compartments and a bidirectional flux between all cell compartments (i.e. model C with efflux from the final compartment included).......................89 5.7.2 Description of silicic acid transport by means of SITs (model F), transport vesicles (model G) and macropinocytosis mediated transport (model H)............................................................................90 5.7.2.1 Model F: Silicon transporters (SITs) mediated uptake and transport...............................................................................90 5.7.2.2 Model G: transport vesicles mediated uptake and transport..91 5.7.2.3 Model H: uptake and transport by means of macropinocytosis.................................................................93 5.7.3 References..........................................................................................94 5.8 Supplement 3: mathematical comparison of the transport vesicle mechanism (model G) with model B...............................................................................95 5.8.1 References..........................................................................................97

6. A new method to study heterogeneous binding and precipitation of silicate and phosphate in heterotrophic biofilms...................................................................99

6.1 Introduction...............................................................................................100 6.2 Material and Methods................................................................................101 6.2.1 Biofilm growth conditions.................................................................101 6.2.2 Metal analysis...................................................................................102 6.2.3 31Si and 32P tracers.............................................................................102 6.2.4 Radioactive 31Si and 32P treatment of the biofilms.............................103 6.2.5 Scanning and autoradiography procedures.......................................103 6.2.6 Image treatment...............................................................................104

vii

6.3 Results.......................................................................................................106 6.3.1 Effect of growth conditions on metal binding in biofilms...................106 6.3.2 Optimization of the autoradiographic method..................................108 6.3.3 Time dependant uptake behaviour....................................................109 6.3.4 Silicate and phosphate spatial distribution........................................110 6.4 Discussion..................................................................................................113 6.5 References.................................................................................................115

7. Evaluation and outlook......................................................................................117 7.1 The 31Si‐silicate tracer.................................................................................117 7.2 Evaluation of the studies on yeast, diatoms and biofilms............................118 7.2.1 Subjected to silicon chemistry: yeast cells and biofilms......................118 7.2.2 Diatoms: active “practitioners” of silicon chemistry...........................119 7.3 Outlook......................................................................................................119 7.3.1 Further research................................................................................119 7.3.2 Towards a new definintion of silicon essentiality...............................120 7.4 References.................................................................................................121

Summary..............................................................................................................123

Samenvatting.......................................................................................................127

Dankwoord...........................................................................................................131

Curriculum vitae....................................................................................................133

List of publications................................................................................................135

1

Chapter 1

Introduction 1.1 General properties of silicon Silicon (Si) is the second most abundant element (27.7 %) in the earth crust after oxygen (46.6 %).1 It is present as a major compound of many rocks and soil minerals, and in natural waters it exists in dissolved form in concentrations ranging from 0.07 to 0.2 mmol/L.1-5 In a natural environment silicon is present as silica (SiO2∙xH2O), metal silicate (MxSiO3 or MxSiO4 where M=metal) or dissolved in water as silicic acid (Si(OH)4).1,6-10 Dissolution and precipitation of siliceous compounds is an ongoing process in nature. On the one hand, silicic acid is dissolved into water as a result of the weathering of silicate containing rocks, a process that is enhanced by microorganisms and lichens.11-16 On the other hand, silicic acid in water can precipitate with metal ions to solid compounds.1 Also in this process microorganisms play a role, because the cell wall of certain bacteria can act as a starting point for metal silicate precipitation.17-19 Silicon is a group 14 element, together with carbon, germanium, tin and lead.6,20 Its three stable isotopes, 28Si, 29Si and 30Si, have abundances of 92.23, 4.67 and 3.10 %, respectively.20 Silicon in a natural environment is always present in the oxidized form, and redox reactions involving silicon do not occur in nature,6 which can explain why the role as a cofactor in enzymes has not been identified so far. Biological processes often take place in an aqueous environment, so the behaviour of silicic acid in water is of high importance. At neutral pH (at which most biological processes take place) 97% of the silicic acid molecules are in the undissociated form. An increase in pH will cause the compound to dissociate under the formation of silicate ions (eq.1‐4). The pK values of the consecutive dissociation reactions are 9.9, 11.8, 12 and 12 respectively.6,20

4 3 1: ( ) ( )step Si OH Si OH O H− ++ 1

23 2 2 2: ( ) ( )step Si OH O Si OH O H− − ++ 2

2 3

2 2 3 3: ( ) ( )step Si OH O Si OH O H− − ++ 3

3 43 4: ( )step Si OH O SiO H− − ++ 4

In the scientific literature regarding silicon biology the words “silicon”, “silicic acid” and “silicate” are often used amongst each other and refer to silicic acid in the undissociated as well as the dissociated form. A well known feature of silicic acid is its capability to polymerize to oligomers and higher polymers, and its polymerization behaviour depends on the concentration in water. In low concentrations it exists in the monomeric state, but when its concentration exceeds 2 mmol/L oligomerization is initiated (eq. 5):

Chapter 1: Introduction

2

5 Depending on its concentration also trimers and higher linear, circular and branched oligomers can be formed (“polysilicate”). Eventually, at very high concentrations a network of polymeric silicate is formed (silicagel and silica, SiO2.xH2O). Polymerized silicic acid will depolymerize to silicic acid monomers when the concentration drops below 2 mmol/L.6,7 Both the polymerization and depolymerization processes depend on concentration, temperature, salt strength and pH.21-24 Monomeric and polymeric silicic acid can adsorb to several surfaces,25 which can be described by a Freundlich isotherm.26 Furthermore monomeric and polymeric silicic acid can interact with metal ions under the formation of silicates like MxSiO4 (orthosilicate) and MxSiO3 (metasilicate). Metal silicates, except for sodium metasilicate (Na2SiO3) and potassium metasilicate (K2SiO3), are insoluble in water and will precipitate when formed.20,27,28 Polysilicate can also form complexes with metals,6 and is also able to interact with several organic compounds that are important in biology, like proteins and sugars which are present almost everywhere in all biological systems (fig. 1).29-32

Figure 1: Examples of polysilicate complexes. Left: polysilicate‐metal complex,6 and right: silicate‐sugar complex.30

The ability to interact with metals and other compounds that are important in biology offers many opportunities for influences on biological processes. Metal ions are present in all organisms, as are organic compounds like sugars and proteins.33 Silicon‐metal interactions are found in plants and animals, as are silicon interactions with proteins in sponges and diatoms. It is likely that silicon chemistry itself has enough potentials to influence biological processes, without the involvement of enzymes or binding sites, because it offers many opportunities for interaction with several biological compounds. Possibly physicochemical events like adsorption could also play a role. 1.2 Silicon in living systems Silicon may have played a significant role in the origin of life. It was hypothesized that certain silicate minerals have provided the right circumstances for the genesis of the first cells.34-36 For example, certain weathered feldspars have a cell‐like honeycomb structure on a micrometer scale, and the surface of these minerals contain organophilic and catalytic properties which enable the accumulation and


3

conversion of complex organic compounds. These accumulated organic compounds could have been converted to compounds that are important for life such as oligopeptides and nucleic acids.35 In this way the formation of primitive cells inside the feldspar honeycomb structure might have been realized.34-36 Nowadays, silicon is important for many organisms, and its role has been investigated for several decades.37-40 It is an essential element for diatoms, some protozoa, sponges and several higher plants. These organisms often contain silicon in high amounts and frequently possess metabolic pathways to handle silicon. In many other organisms, such as higher animals and most higher plants, silicon also plays a beneficial or essential role, but underlying biochemical mechanisms have not been discovered so far. This means that the evidence for Si requirement is only circumstantial for these organisms.1,41-47 The role of silicon in sponges and protozoa, diatoms, higher plants, and higher animals is discussed below in more detail. 1.2.1 Sponges and protozoa Many protozoa and sponges use silicon for the formation of the cytoskeleton, spicules and tests.48 These silicate structures support the cell and function as protective means for the organism. Some protozoa and sponges build structures from silica particles collected from the environment such as sand grains and diatom shells. These particles are cemented together to the desired structure. It is clear that these organisms must possess biochemical means necessary for the recognition and handling of silicon compounds.48,49 Other protozoa and sponges take up silicic acid from the water and deposit it as amorphous hydrated silica in the cell or at the cell surface as spicules, plates, needles and other structures.50,51 The silicic acid is deposited on a central protein filament called silicatein produced by specialized cells close to the growing spicule, test or other structure.50,52 Silicatein has a highly regular structure and directs the polymerization of silicic acid to silica in both a chemical and spatial way.52,53 Silicic acid uptake is very efficient,34,35 and the expression of genes encoding for silicatein are strongly enhanced by the presence of silicic acid in the water.50,51,54 The discovery of silicatein has given a stong impulse to the research on silicon in biological systems and the production of artificial silica structures.55 1.2.2 Diatoms Diatoms are eukaryotic photosynthetic algeae (order Bacillariophyceae) and a major component of the phytoplankton community.56 An important feature of a diatom is its silicified cell wall called frustule; a beautiful structure divided in two overlapping halves that surrounds the cell as a “glass box” (Fig. 2). The diatoms capability to master silicon chemistry is interesting for technology to gain more knowledge for the industrial production of complex silica structures.57,58 The frustule consists of amorphous hydrated silica and protects the cell against predators and other dangers.56


4

Figure 2: The centric diatom Coscinodiscus wailesii (left) and the pennate diatom Pleurosira laevis (right).

Apart from the frustule production, the diatom cell requires silicon for the synthesis of DNA and chlorophyll.59-63 Most diatoms take up silicon as undissociated silicic acid from the surrounding water, but some exceptions exist.61,64 The uptake rate and quantities vary during the cell cycle, especially depending on the amount of silicon that is needed for the production of new frustules for daughter cells.65 In case of silicon deficiency the entire cell cycle stops.66,67 Like in sponges and protozoa silicon uptake is very efficient and, depending on the species, the affinity constant for silicic acid ranges from 0.2 to 100 μmol/L, which indicates active uptake.61,68-70 During the production of a daughter frustule silicic acid is transported to a specialized membrane vesticle, better known as the silica deposition vesticle or SDV.71 The new frustule is produced within the SDV by deposition of polymerized silicic acid particles. In vitro this process is promoted by certain peptides (silaffins) on which silica particles are deposited.72,73 The interaction of silicic acid with this protein directs the polymerization process and results in the formation of a new frustule. Diatoms control the silicic acid polymerization process to enable the production of very complex structures. This control over silicon chemistry within the cells is also necessary for the transport of high amounts of silicic acid through the cell without the occurence of unwanted chemical side reactions, such as polymerization in other parts of the cell. It is obvious that silicon chemistry, especially the polymerization of silicic acid, plays a key role in frustule growth. Silicon chemistry is also of importance in the transport process of silicic acid through the cell to the SDV. To build the daughter frustule high amounts of silicic acid are needed, being transported in bulk quantities through the cell. Because silicic acid has the tendency to polymerize when its concentration exceeds a certain level, the presence of high amounts of this compound in the cell offers a risk to the cell.74,75 Several proteins that transport silicic acid through the cytoplasm have been identified,76 but it is not clear how the cell prevents spontaneous polymerization of these high amounts of silicic acid needed for frustule formation.59


5

1.2.3 Higher plants Silicon is essential or at least beneficial for most higher plants.77,78 They contain silicon amounts ranging from 0.1% to 15% (w/w) of their dry matter depending on plant species and tissue,79,80 Silicic acid is taken up from the soil by the roots and is transported to several tissues where it is deposited as amorphous silica.81 The uptake is probably regulated by a control mechanism.82-84 Silica precipitates give structural mechanical strength to plants, for example horsetails (Equisitacea).85 The use of silicon for this purpose is an easy way to reduce the energy and carbon consuming lignin production.44,45,86,87 Moreover silicon improves water economy of the plant, enhances wear tolerance, provides resistance against heat and freezing,78 and protects against predators and pathogens like insects, herbivores and fungi.44,88-

90 In some plant species silicon enhances the activities of enzymes that are involved in fungal cell breakdown. Other plants encounter fungal infections by silicifying entire infected cells to stop the fungal growth.91-93 Apart from the protection against several stresses as described above, silicon is also involved in the alleviation of metal toxicity.45 Many plants use silicic acid for the formation of silicate‐metal complexes that are less toxic then the uncomplexed metals. An excess of aluminium for example is removed by the formation of aluminium‐silicate compounds that are deposited in root cell walls. These chemically stable complexes ensure a permanent removal of aluminium from the plant system.94-97 Toxic levels of zinc are reduced by the formation of a zinc‐silicate complex which is eventually transported to the vacuole. In this location the unstable zinc‐silicate complex slowly degrades, leaving silicate precipitations behind. The zinc accumulates in the vacuole of the cell bound to a yet unknown compound.81,98 Some plants take up more silicon during high manganese stress to enable a more equal spreading of manganese over the leaves, thereby decreasing the toxic effects.44,99 Silicon is also involved in indirect mechanisms that alleviate metal stress. In case of sodium stress caused by an excess of salt in the environment silicon enhances the uptake of potassium which in its turn counteracts the toxic effects of sodium. 100-102 Because of the beneficial effects described above much research has been carried out on the use of silicon as a fertilizer. Currently, silicate fertilization is common use for the growth of rice, sugarcane and bahiagrass.103-106 Silicon chemistry and especially silicon‐metal interactions seem to be importance, but most studies were performed in a phenomenological way so many mechanisms are not elucidated yet. 44,45 1.2.4 Higher animals Silicon is an essential element for higher animals including man, but until today no silicon handling enzymes or other ways of biochemical involvement have been discovered. Rich sources of silicon are unrefined grains of high fibre content, cereals, root vegetables, drinking water and beer.41-43 The human daily requirement ranges from 5 ‐ 50 mg/day,107 and an excess is easily excreted by the kidneys.108-114 Silicon is found in many tissues and organs including liver, spleen, kidney, bone and connective tissue,43,115-118 and its concentration in the human blood serum ranges from 140 to 280 μg/L.43,116,119-122 The essentiality is clear from observations that silicon deficiency causes many problems in physiological functions of the body.114,123-

125 Some studies suggest that the essentiality of silicon is based on the


6

counteraction of aluminium toxicity either by bonding with aluminium compounds or by facilitating the excretion of aluminium from the body,8,120,123,126-130 although this effect was not always found.131 Deficiency symptoms have not been identified in humans, probably as a result of its omnipresence in the environment.112 Silicon plays a major role in bone and collagen formation (involving connective tissue, skin, arterial walls, and cartilage), cell proliferation (including the immune system, bone growth and wound healing), and several hormonal and enzymatic activities.43,116,117,132-141 Silicon is present in active growth areas in bone and plays an important role in the formation, calcification, and mineral composition of bone.125 Furthermore it has a positive influence on the arterial wall lipid metabolism and it protects against atheromatous lesions (lipid precipitates on the arterial wall).132,142 Some studies suggest that silicon can counteract Alzheimer’s disease to a certain extend.143,144 As stated above, an important functions of silicon is its positive influence on collagen and bone formation.108,125,145 Collagen is present in bone, cartilage, skin, tendons, arterial walls and other connective tissue. In case of silicon deficiency many processes involving collagen and collagen formation are impared8,120,146 resulting in a decline of collagen content in several tissues, including skin, bone and cartilage,43,135,137-139 and in a depressed wound healing.140,141 Silicon deficiency affects the concentration of several minerals in bone and causes a decreased growth (Fig. 3). It also lowers the mineralization rate leading to abnormalities like smaller, thinner and less flexible bones.43,115,137-139,147-150

Figure 3: Photograph of 4‐week‐old chicks on silicon supplemented diet (left) and on low‐silicon diet (right),115*

* Picture from: EM Carlisle, Silicon in Bone Formation in Silicon and Siliceous Structures in Biological Systems, eds. TL Simpson and E Volcani (1981) Springer‐Verlag, New York. Chapter 4, page 78, figure 4‐5, with kind permission of Springer Science and Business Media.


7

Silicon also affects cell proliferation, which is an important aspect in the function of the immune system, wound healing and bone growth.43,117,134,140,141,151 DNA synthesis and cell division regulation are probably affected by silicon as well,117 but the mechanisms behind these phenomena are not well understood yet. The effect of silicon on processes involving bone and collagen production and other processes originates in its ability to alter several enzymatic activities. Enzymes involved in bone and collagen formation are enhanced by silicon,140,145,150 whereas the opposite effect takes place in the antioxidant enzymes superoxide dismutase, catalase and glutathione peroxidase. The latter enzymes are crucial for the removal of free oxygen radicals from the cell, and the depression of these enzymatic activities probably plays a role in the cause of diseases like silicosis.152 The question arises how silicon can affect enzymatic activities and other bioprocesses, since no specific binding sites for silicon have been found in animals. Silicon compounds do not show any redox reactions in a natural environment,6 which makes it unlikely that they can function as active sites in enzymes. The role of silicon might be explained by its effect on the concentration of several essential (trace) elements like calcium, phosphorus, potassium, copper, zinc, iron, manganese and magnesium in serum, bone, cartilage and other tissues.108,138-141,149-

151,153-157 Many of these elements play a role as a cofactor in several essential enzymes. Copper is a cofactor in an enzyme involved in cartilage synthesis,158 in an antioxidant enzyme,159 and in enzymes that play a role in cholesterol and lipid metabolism.107,160,161 Zinc plays a role in wound healing and bone mineralization162 and processes involving carbohydrates, lipids, proteins and nucleic acids.160,163 Manganese has a key role in wound healing and bone formation,141 and calcium, phosphorus, potassium, magnesium and zinc probably play a role in lymphocyte proliferation.141 And, finally, calcium and phosphorus are the main components of bone.115 An alteration of metal concentrations in the body induced by silicon could influence the performance of specific enzymes. Therefore, a better understanding is needed about the interaction between silicon and metal ions in the body. 1.2.5 Overview The omnipresence of silicon in the ecosystem forces living organisms to use this element. Although much is known about silicon in biology, many processes remain unclear. Most organisms, like the higher animals, need silicon as an essential element, but only a few biochemical means to handle silicon have been discovered so far in the form of the enzymes silicatein and silaffin in sponges and diatoms. A role as enzyme cofactor has not been found (yet). So it seems that the influence of silicon on several processes in many organisms cannot be explained by silicon biochemistry alone. This means that physical‐chemical mechanisms rather than a biochemical one could be involved. To get a deeper insight in the influence of silicon in biological processes a good understanding of silicon chemistry is crucial.


8

1.3 Thesis outline The aim of the research described in this thesis was to investigate the role of silicon chemistry in biological processes. Emphasis was layed on the role of silicon in metal metabolism, on in vivo silicon polymerization reactions, and on a possible application in biotechnology. The studies described in this thesis focused on the behaviour and uptake kinetics of silicic acid in several biological systems. Radiotracer techniques are highly suitable for these kind of investigations. A radiotracer that is chemically identical to natural silicon compounds is useful and provides an easy detection of the compound with a detection limit well below those of other techniques. Radiotracers can also be used for visualization of uptake kinetics by autoradiography. In chapter 2 the development of a suitable silicon tracer was discussed. This study was focused on the production and chemical characterization of a tracer of high specific activity that can be produced in a nuclear reactor. This tracer was used in other studies that are described in this thesis, for instance in uptake kinetics studies in yeast and diatom cells, and in the development of a novel autoradiography method. In chapter 3 and 4 it was investigated if silicon chemistry can play a role in biology without the involvement of enzymes, binding sites etc. The accumulation of silicon and its interactions with metals was investigated in Saccharomyces cerevisiae (baker’s yeast), which commonly serves as a model for the eukaryotic cell.164 Yeast does not have the drawback of higher animals in which it is difficult to determine whether the observed phenomena reflect a direct interaction or a more complex series of events in the body. These studies focused on possible mechanisms of silicon chemistry in biological processes. Three tracers (31Si, 65Zn and 60Co) were used for the measurement of uptake rates. In chapter 5 a study on silicon uptake mechanisms in diatoms was described. As is stated above silicic acid polymerization is used by diatoms and sponges for the formation of several silica structures. For frustule formation in diatoms a large amount of silicic acid has to be transported through the cell, and the polymerization process offers a risk when it takes place in an unwanted location of the cell. This means that diatoms must possess a mechanism to safely handle this transport. It was hypothesized that this bulk uptake and transport followed a mechanism that involved pinocytosis events.165 In this thesis this mechanism was investigated and compared to earlier proposed mechanisms. Three models were designed to describe different silicon uptake mechanisms. The experimental results, obtained by tracer aided uptake determinations, were processed by compartmental analysis to investigate these mechanisms in the view of silicon transport through the cell during frustule formation. Compartmental analysis is a mathematical approach to obtain information about the fate of a compound in a closed biological or chemical system.166 The studied system is divided into “compartments” such as pools of a compound in cell organelles or different chemical forms of the compound. The time dependent behaviour of the compound in each compartment is mathematically described, based on first order transport kinetics. Experimental data are used to calculate the transport rate constants in the model. Compartmental analysis has


9

been successfully used to describe e.g. the accumulation of technetium in plants,167 the transport of compounds in solid‐liquid systems,168 and the speciation of calcium in milk.169 Silicic acid adsorption on organic compounds could yield some applications in biotechnology. Systems like biofilms (communities of microorganisms that grow on several surfaces) possess a structural matrix of proteins and other (polymeric) compounds that is excreted by the microorganisms in the biofilm170,171 and that can interact with silicon compounds. Biofilms are used in several biotechnological processes like vinegar production172 and water purification,173 but can also be used in biogrouting (the stabilisation of ground works in civil engineering).174 Currently it is investigated if silicon precipitation can play a positive role in the latter application. For this purpose the deposition of silicic acid on the matrix of a biofilm is investigated in chapter 6. This process could influence the mechanical strength of the biofilm, and in this way could find an application in the biotechnological process of biogrouting. A novel autoradiography technique is developed in this study. In chapter 7 the results are evaluated and an outlook is given for further investigations.


10

1.4 References Reference List

1. Ehrlich HL Geomicrobiology, third edition, revised and expanded; 3rd ed.; Marcel Bakker Inc.:

New York, 1996. 2. Aston S Natural water and atmospheric chemistry of silicon. In Silicon geochemistry and

biogeochemistry, Aston S, Ed.; Academis Press Inc.: London, 1983; pp 77‐100. 3. Tréguer P; Nelson DM; van Bennekom AJ; DeMaster DJ; Leynaert A; Quéguiner B The silica

balance in the world ocean ‐ a reestimate. Science, 1995, pp 375‐379. 4. Conley DJ Riverine contribution of biogenic silica to the oceanic budget. Limnology and

Oceanography, 1997, pp 774‐777. 5. Willén E Planktonic diatoms ‐ An ecological review. Archiv für Hydrobiologie, 1991, pp 69‐

106. 6. Iler RK The chemistry of silica; John Wiley & Sons: New York, 1979. 7. Petzold A; Hinz W Silikatchemie, Einführung in die Grundlagen; VEB Deutsche Verlag für

Grundstoffindustrie: Leipzig, 1978. 8. Birchall JD; Bellia JP; Roberts NB On the mechanisms underlying the essentiality of silicon ‐

interactions with aluminium and copper. Coordination Chemistry Reviews, 1996, pp 231‐240. 9. Farmer VC Sources and speciation of aluminium and silicon in natural waters. In Silicon

Biochemistry, Evered D, O'Connor M, Eds.; John Wiley & Sons Ltd.: Chichester, 1986; pp 4‐23.

10. Bold HC; Wynne MJ Introduction to the algae. Stucture and reproduction; Prentice‐Hall, INC.: Englewood Cliffs, 1978.

11. Bennett PC; Rogers JR; Choi WJ Silicates, silicate weathering, and microbial ecology. Geomicrobiology Journal, 2001, pp 3‐19.

12. Maurice PA; Vierkorn MA; Hersman LE; Fulghum JE; Ferryman A Enhancement of kaolinite dissolution by an anearobic Pseudomonas mendocina bacterium. Geomicrobiology Journal, 2001, pp 21‐35.

13. Ehrlich HL How microbes influence mineral growth and dissolution. Chemical Geology, 1996, pp 5‐9.

14. Crawford RH; Floyd M; Li CY Degradation of serpentine and muscovite rock minerals and immobilization of cations by Penicilllium spp. Phyton‐Annales Rei Botanicae, 2000, pp 312‐315.

15. Barker WW; Welch SA; Banfield JF Biogeochemical weathering of silicate minerals. Reviews in Mineralogy, 1997, pp 391‐428.

16. Chen J; Blume HP; Beyer L Weathering of rocks induced by lichen colonization ‐ a review. Catena, 2000, pp 121‐146.

17. Urrutia MM; Beveridge TJ Formation of fine grained metal and silicate precipitates on a bacterial surface (Bacillus‐subtilis). Chemical Geology, 1994, pp 261‐280.

18. Fortin D; Ferris FG Precipitation of iron, silica and sulfate on bacterial cell surfaces. Geomicrobiology Journal, 1998, pp 309‐324.

19. Mera MU; Beveridge TJ Mechanism of silicate binding to the bacterial‐cell wall in Bacillus subtilis. Journal of Bacteriology, 1993, pp 1936‐1945.

20. Lide DR Handbook of chemistry and physics; 83rd edition ed.; CRC Press: Boca Raton, 2002. 21. Dietzel M Dissolution of silicates and the stability of polysilicic acid. Geochimica et

Cosmochimica Acta, 2000, pp 3275‐3281. 22. Dietzel M; Usdowski E Depolymerization of soluble silicate in dilute aqueous solutions.

Colloid and Polymer Science, 1995, pp 590‐597. 23. Dietzel M Polysilicic acid and the dissolution of silicates. Mineralogical Magazine, 1998, pp

385‐386. 24. Baumann H Polymerisation und Depolymerisation der Kieselsaüre unter verschiedenen

Bedingungen. Kolloid Zeitschrift, 1959, pp 28‐35. 25. Dietzel M; Bohme G Adsorption and stability of polymeric silica. Chemie der Erde ‐

Geochemistry, 1997, pp 189‐203. 26. Adamson AW; Gast AP Physical chemistry of surfaces; 6th edition ed.; New York, 1997.


11

27. Schwieger W; Heyer W; Wolf F; Bergk KH Zur Synthese von kristallinen Metallsilicathydraten mit Schichtstruktur. Zeitschrift für anorganische und allgemeine Chemie, 1987, pp 204‐216.

28. Marshall WL; Warakomski JM Amorphous silica solubilities ‐ II. Effects of aqueous salt solutions. Geochimica et Cosmochimica Acta, 1980, pp 915‐924.

29. Kinrade SD; Hamilton RJ; Schach AS; Knight CTG Aqueous hypervalent silicon complexes with aliphatic sugar acids. Journal of the Chemical Society ‐ Dalton Transactions, 2001, pp 961‐963.

30. Lambert JB; Lu G; Singer SR; Kolb VM Silicate complexes of sugars in aqueous solution. Journal of the American Chemical Society, 2004, pp 9611‐9625.

31. Sahai N Calculation of 29Si NMR shifts of silicate complexes with carbohydrates, amino acids, and MuHicarboxylic acids: Potential role in biological silica utilization. Geochimica et Cosmochimica Acta, 2004, pp 227‐237.

32. Kinrade SD; Del Nin JW; Schach AS; Sloan TA; Wilson KL; Knight CTG Stable five‐ and six‐coordinated silicate anions in aqueous solution. Science, 1999, pp 1542‐1545.

33. Stryer L Biochemistry; W.H. Freeman and Company: San Francisco, 1981. 34. Smith JV; Arnold FP; Parsons I; Lee MR Biochemical evolution III: Polymerization on

organophilic silica‐rich surfaces, crystal‐chemical modeling, formation of first cells, and geological clues. Proceedings of the National Academy of Sciences of the United States of America, 1999, pp 3479‐3485.

35. Smith JV Biochemical evolution. I. Polymerization on internal, organophilic silica surfaces of dealuminated zeolites and feldspars. Proceedings of the National Academy of Sciences of the United States of America, 1998, pp 3370‐3375.

36. Parsons I; Lee MR; Smith JV Biochemical evolution II: Origin of life in tubular microstructures on weathered feldspar surfaces. Proceedings of the National Academy of Sciences of the United States of America, 1998, pp 15173‐15176.

37. Bendz G; Lindqvist I Biochemistry of silicon and related problems; Plenum Press: New York ‐ London, 1978.

38. Evered D; O'Connor M Silicon biochemistry; John Wiley & Sons Ltd.: Chichester, 1986. 39. Simpson TL; Volcani BE Silicon and siliceous structures in biological systems; Springer Verlag:

New York, 1981. 40. Voronkov MG; Zelchan GI; Lukevitz E Silizium und Leben; Akademie‐Verlag: Berlin, 1975. 41. Nielsen FH Nutritional‐requirements for boron, silicon, vanadium, nickel, and arsenic ‐

current knowledge and speculation. FASEB Journal, 1991, pp 2661‐2667. 42. Nielsen FH Ultratrace elements in nutrition: current knowledge and speculation. Journal of

Trace Elements in Experimental Medicine, 1998, pp 251‐274. 43. Carlisle EM Silicon as an essential trace element in animal nutrition. In Silicon biochemistry,

Evered D, O'Connor M, Eds.; 1986. 44. Epstein E Silicon. Annual Review of Plant Physiology and Plant Molecular Biology, 1999, pp

641‐664. 45. Epstein E The anomaly of silicon in plant biology. Proceedings of the National Academy of

Sciences of the United States of America, 1994, pp 11‐17. 46. Nielsen FH Evolutionary events culminating in specific minerals becoming essential for life.

European Journal of Nutrition, 2000, pp 62‐66. 47. Exley C Silicon in life: A bioinorganic solution to bioorganic essentiality. Journal of Inorganic

Biochemistry, 1998, pp 139‐144. 48. Bovee EC Distribution and forms of siliceous structures among protozoa. In Silicon and

siliceous structures in biological systems, Simpson TL, Volcani BE, Eds.; Springer Verlag: New York, 1981; pp 233‐279.

49. Reincke T; Barthel D Silica uptake kinetics of Halichondra panicea in Kiel Bight. Marine Biology, 1997, pp 591‐593.

50. Imsiecke G; Steffen RT; Custodio M; Borojevic R; Muller W Formation of spicules by sclerocytes from the fresh‐water sponge Ephydatia‐muelleri in short‐term cultures in‐vitro. In Vitro Cellular & Developmental Biology ‐ Animal, 1995, pp 528‐535.

51. Uriz MJ; Turon X; Becerro MA Silica deposition in Demosponges: spiculogenesis in Crambe crambe. Cell and Tissue Research, 2000, pp 299‐309.


12

52. Shimizu K; Cha JN; Stucky GD; Morse DE Silicatein alpha: Cathepsine L ‐ like protein in sponge biosilica. Proceedings of the National Academy of Sciences of the United States of America: Genetics, 1998, pp 6234‐6238.

53. Cha JN; Shimizu K; Zhou JZ; Christiansen SC; Chmelka BF; Stucky GD; Morse DE Silicatein filaments and subunits from a marine sponge direct the polymerization of silica and silicones in vitro. Proceedings of the National Academy of Sciences of the United States of America. Biochemistry, 1999, pp 361‐365.

54. Krasko A; Lorenz B; Batel R; Schröder HC; Müller IM; Müller WEG Expression of silicatein and collagen genes in the marine sponge Suberites domuncula is controled by silicate and myotrophin. European Journal of Biochemistry, 2000, pp 4878‐4887.

55. Morse DE Silicon biotechnology: harnessing biological silica production to construct new materials. Trends in Biotechnology, Jun, 1999, pp 230‐232.

56. Crawford RM The siliceous components of the diatom cell wall and their morphological variation. In Silicon and siliceous structures in biological systems, Simpson TL, Volcani BE, Eds.; Springer Verlag: New York, 1981; pp 129‐156.

57. Morse DE Silicon biotechnology: harnessing biological silica production to construct new materials. Trends in Biotechnology, Jun, 1999, pp 230‐232.

58. Gordon R; Losic D; Tiffany MA; Nagy SS; Sterrenburg FAS The Glass Menagerie: diatoms for novel applications in nanotechnology. Trends in Biotechnology, 2009, pp 116‐127.

59. Scala S; Bowler C Molecular insights into the novel aspects of diatom biolology. Cellular and Molecular Life Sciences, 2001, pp 1666‐1673.

60. Kristiansen S; Hoell EE The importance of silicon for marine production. Hydrobiologia, 2002, pp 21‐31.

61. Martin‐Jezequel V; Hildebrand M; Brzezinski MA Silicon metabolism in diatoms: implications for growth. Journal of Phycology, 2000, pp 821‐840.

62. Wang LJ; Wang YH; Li M; Zhang WX; Zhang FS; Li TJ Biogenic nanostructured silica. Progress in Natural Science, 2002, pp 401‐406.

63. Round FE; Crawford FM; Mann DG The diatoms. Biology & morphology of the genera; Cambridge University Press: Cambridge, 1990.

64. Del Amo Y; Brzezinski M The chemical form of dissolved Si taken up by marine diatoms. Journal of Phycology, 1999, pp 1162‐1170.

65. Brzezinski MA; Conley DJ Silicon deposition during the cell ‐ cycle of Thalassiosira‐Weissflogii (Bacillarioophyceae) determined using dual rhodamine‐123 and propidium iodide staining. Journal of Phycology, 1994, pp 45‐55.

66. Reeves CD; Volcani BE Role of silicon in diatom metabolism. Patterns of protein phosphorylation in Cylindrotheca fusiformis during recovery from silicon starvation. Archives of Microbiology, 1984, pp 291‐294.

67. Sullivan CW; Volcani BE Role of silicon in diatom metabolism III. The effects of silicic acid on DNA polymerase, TMP kinase and DNA synthesis in Cylindrotheca fusiformis. Biochimica et Biophysica Acta, 2003, pp 212‐229.

68. Sarthou G; Timmermans KR; Blain S; Tréguer P Growth physiology and fate of diatoms in the ocean: a review. Journal of Sea Research, 2005, pp 25‐42.

69. Azam F; Hemmingsen BB; Volcani BE Role of silicon in diatom metabolism V. Silicic acid transport and metabolism in the heterotrophic diatom Nitzschia alba. Archives of Microbiology, 1974, pp 103‐114.

70. Bhattacharyya P; Volcani BE Sodium dependant silicate transport in the apochlorotic marine diatom Nitzschia alba. Proceedings of the National Academy of Sciences of the United States of America, 1980, pp 6386‐6390.

71. Pickett‐Heaps JD; Schmid MM; Edgar LA The cell biology of diatom valve formation. In Progress in Phycological Research, volume 7, Round FE, Chapman DJ, Eds.; Biopress: Bristol, UK, 1990.

72. Kröger N; Deutzmann R; Sumper M Polycationic peptides from diatom biosilica that direct silica nanosphere formation. Science, 1999, pp 1129‐1132.

73. Kröger N; Deutzmann R; Sumper M Silica precipitating peptides from diatoms ‐ The chemical structure of silaffin‐1A from Cylindrotheca fusiformis. Journal of Biological Chemistry, 2001, pp 26066‐26070.


13

74. Kröger N; Sumper M The biochemistry of silica formation in diatoms. In Biomineralization ‐ From biology to biotechnology and medical application, Baeuerlein E, Ed.; Wiley‐VCH Verlag GmbH: Weinheim, Germany, 2000; pp 151‐170.

75. Hildebrand M Silicic acid transport and its control during cell wall silification in diatoms. In Biomineralization ‐ From biology to biotechnology and medical application, Baeuerlein E, Ed.; Wiley‐VCH Verlag GmbH: Weinheim, Germany, 2000; pp 171‐188.

76. Hildebrand M; Dahlin K; Volcani BE Characterization of a silicon transporter gene family in Cylindrotheca fusiformis: sequences, expression analysis, and identification of homologs in other diatoms. Molecular and General Genetics, 1998, pp 480‐486.

77. Epstein E; Bloom AJ Mineral nutrition of plants: Principles and perspectives; 2nd edition ed.; Sinauer: Sunderland, MA, 2005.

78. Liang YC; Sun WC; Zhu YG; Christie P Mechanisms of silicon‐mediated alleviation of abiotic stresses in higher plants: a review. Environmental Pollution, 2007, pp 422‐428.

79. Sangster AG; Parry DW Ultrastructure of silica deposits in higher plants. In Silicon and siliceous structures in biological sysems, Simpson TL, Volcani BE, Eds.; Springer Verlag: New York, 1981; pp 383‐408.

80. Sangster AG; Hodson MJ Silica in higher plants. In Silicon biochemistry, Evered D, O'Connor M, Eds.; 1986.

81. Neumann D; Figueiredo C De A novel mechanism of silicon uptake. Protoplasma, 2002, pp 59‐67.

82. de Bakker NVJ; Hemminga MA; van Soelen J The relationship between silicon availability, and growth and silicon concentration of the salt marsh halophyte Spartina anglica. Plant and Soil, 1999, pp 19‐27.

83. Rafi MM; Epstein E Silicon absorption by wheat (Triticum aestivum L.). Plant and Soil, 1999, pp 223‐230.

84. Ma JF; Tamai K; Yamaji N; Mitani N; Konishi S; Katsuhara M; Ishiguro M; Murata Y; Yano M A silicon transporter in rice. Nature, 2006, pp 688‐691.

85. Kaufman PD; Dayanandan P; Takeoka Y; Bigelow WC; Jones JD; Iler R Silica in shoots of higher plants. In Silicon and siliceous structures in biological sysems, Simpson TL, Volcani BE, Eds.; Springer Verlag: New York, 1981; pp 409‐449.

86. Wang LJ; Nie Q; Li M; Zhang FS; Zhuang JQ; Yang WS; Li TJ; Wang YH Biosilicified structures for cooling plant leaves: A mechanism of highly efficient midinfrared thermal emission. Applied Physics Letters, 2005, p 194105.

87. Trenholm LE; Duncan RR; Carrow RN; Snyder GH Influence of silica on growth, quality, and wear tolerance of seashore Paspalum. Journal of Plant Nutrition, 2001, pp 245‐259.

88. Cotteril JV; Watkins RW; Brennon CW; Cowan DP Boosting silica levels in wheat leaves reduces grazing by rabbits. Pest Management Science, 2007, pp 247‐253.

89. Guevel MH; Menzies JG; Belanger RR Effect of root and foliar applications of soluble silicon on powdery mildew control and growth of wheat plants. European Journal of Plant Pathology, 2007, pp 429‐436.

90. Heine G; Tikum G; Horst WJ The effect of silicon on the infection by and spread of Pythium aphanidermatum in single roots of tomato and bitter gourd. Journal of Experimental Botany, 2007, pp 569‐577.

91. Green S; Bailey KL; Tewari JP The infection process of Alternaria cirsinoxia on Canada thistle (Cirsium arvense) and host structural defence responses. Mycological Research, 2001, pp 344‐351.

92. Dann EK; Muir S Peas grown in media with elevated plant‐available silicon levels have higher activities of chitinase and beta‐1,3‐glucanase, are less susceptible to a fungal leaf spot pathogen and accumulate more foliate silicon. Australasian Plant Pathology, 2002, pp 9‐13.

93. Rodrigues FA; Vale FXR; Datnoff LE; Prabhu AS; Korndorfer GH Effect of rice growth stages and silicon on sheath blight development. Phytopathology, 2003, pp 256‐261.

94. Liang YC; Yang CG; Shi HH Effects of silicon on growth and mineral composition of barley grown under toxic levels of aluminium. Journal of Plant Nutrition, 2001, pp 229‐243.

95. Hara T; Gu MH; Koyama H Ameliorative effect of silicon on aluminium injury in the rice plant. Soil Science and Plant Nutrition, 1999, pp 929‐936.

96. Rahman MT; Koyama H; Hara T Effect of aluminium and silicon in solution on the growth of rice (Oryza sativa L.) suspension cells. Soil Science and Plant Nutrition, 1999, pp 693‐700.


14

97. Cocker KM; Evans DE; Hodson MJ The amelioration of aluminium toxicity by silicon in higher plants: Solution chemistry or an in plants mechanism. Physiologia Plantarum, 1998, pp 608‐614.

98. Neumann D; zur Nieden U Silicon and heavy metal tolerance of higher plants. Phytochemistry, 2001, pp 685‐692.

99. Nogueira MA; Cardoso EJBN; Hampp R Manganese toxicity and callose deposition in leaves are attenuated in mycorrhyzal soybean. Plant and Soil, 2002, pp 1‐10.

100. Liang YC Effects of silicon on enzyme activity and sodium, potassium and calcium concentration in barley under salt stress. Plant and Soil, 1999, pp 217‐224.

101. Gunes A; Inal A; Bagci EG; Pilbeam DJ Silicon‐mediated changes of some physiological and enzymatic parameters symptomatic for oxidative stress in spinach and tomato in sodic‐B toxic soil. Plant and Soil, 2007, pp 103‐114.

102. Gunes A; Inal A; Bagci EG; Coban S; Sahin O Silicon increases boron tolerance and reduces oxidative damage of wheat grown in soil with excess boron. Biologia Plantarum, 2007, pp 571‐574.

103. Savant NK; Korndorfer GH; Datnoff LE; Snyder GH Silicon nutrition and sugarcane production: a review. Journal of Plant Nutrition, 1999, pp 1853‐1903.

104. Ando H; Kakuda K; Fujii H; Suzuki K; Ajiki T Growth and canopy structure of rice plants grown under field conditions as effected by Si application. Soil Science and Plant Nutrition, 2002, pp 429‐432.

105. Esser KB Can the application of fused calcium silicate to rice contribute to sustained yields and higher pest resistance? Outlook on Agriculture, 2002, pp 199‐201.

106. Matichenkov VV; Calvert DV; Bocharnikova EA Effect of Si fertilization on growth and P nutrition of bahiagrass. Soil and Crop Science Society of Florida Proceedings, 2001, pp 30‐36.

107. Klevay LM Cholesterotropic and cuprotropic chemicals. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 180‐183.

108. Calomme MR; VandenBerghe DA Supplementation of calves with stabilized orthosilicic acid ‐ Effect on the Si, Ca, Mg, and P concentrations in serum and the collagen concentration in skin and cartilage. Biological Trace Element Research, 1997, pp 153‐165.

109. Reffitt DM; Jugdaohsingh R; Thompson RPH; Powell JJ Silicic acid: its gastrointestinal uptake and urinary excretion in man and effects on aluminium excretion. Journal of Inorganic Biochemistry, 1999, pp 141‐147.

110. van Dyck K; van Cauwenbergh R; Robberecht H; Deelstra H Bioavailability of silicon from food and food supplements. Fresenius Journal of Analytical Chemistry, 1999, pp 541‐544.

111. Dobbie JW; Smith MJB Urinary and serum silicon in normal and uraemic individuals. In Silicon biochemistry, Evered D, O'Connor M, Eds.; 1986.

112. Jugdaohsingh R; Anderson SHC; Tucker KL; Elliott H; Kiel DP; Thompson RPH Dietary silicon intake and absorption. American Journal of Clinical Nutrition, 2002, pp 887‐893.

113. Popplewell JF; King SJ; Day JP; Ackrill P; Fifield LK; Cresswell RG; Di Tada ML; Liu K Kinetics of uptake and elimination of silicic acid by a human subject: A novel application of Si‐32 and accelerator mass spectrometry. Journal of Inorganic Biochemistry, 1998, pp 177‐180.

114. Martin KR The chemistry of silica and its potential health benefits. Journal of Nutrition Health & Aging, 2007, pp 94‐98.

115. Carlisle EM Silicon in bone formation. In Silicon and siliceous structures in biological systems, Simpson TL, Volcani BE, Eds.; Springer‐Verlag: New York, 1981; pp 69‐94.

116. Voronkov MG Silicon in living systems. In Silicon Chemistry, Corey ER, Corey JY, Gaspar PP, Eds.; 1988.

117. Henrotte JG; Viza D; Vich JM; Gueyne J Le rôle régulateur du silicium dans la division cellulaire. Comptes Rendus de l'Académie des Sciences. Série III, 1988, pp 525‐528.

118. Charnot Y; Peres G Silicon, endocrine balance and mineral metabolism. In Biochemistry of silicon and related problems, Bendz G, Lindqvist I, Eds.; 1978; pp 269‐280.

119. Lugowski SJ; Smith DC; Lugowski JZ; Peters W; Semple J A review of silicon and silicone determination in tissue and body fluids ‐ a need for standard reference materials. Fresenius Journal of Analytical Chemistry, 1998, pp 486‐488.

120. Birchall JD Silicon ‐ aluminum interactions and biology. Colloid Chemistry of Silica ‐ Advances in Chemistry Series, 1994, pp 601‐615.


15

121. Dhaese PC; Shaheen FA; Huraib SO; Djukanovic L; Polenakovic MH; Spasovski G; Shikole A; Schurgers ML; Daneels RF; Lamberts LV; van Landeghem GF; Debroe ME Increased silicon levels in dialysis patients due to high‐silicon content in the drinking‐water, inadequate water‐treatment procedures, and concentrate contamination ‐ a multicenter study. Nephrology Dialysis Transplantation, 1995, pp 1838‐1844.

122. Bisse E; Epting T; Beil A; Lindinger G; Lang H; Wieland H Reference values for serum silicon in adults. Analytical Biochemistry, 2005, pp 130‐135.

123. Birchall JD The role of silicon in biology. Chemistry in Brittain, 1990, pp 141‐144. 124. Tacke R Milestones in the biochemistry of silicon: From basic research to biotechnological

applications. Angewandte Chemie ‐ International Edition, 1999, pp 3015‐3018. 125. Jugdaohsingh R Silicon and bone health. Journal of Nutrition Health & Aging, 2007, pp 99‐

110. 126. Birchall JD; Exley C; Chappell JS; Phillips MJ Acute toxicity of aluminium to fish eliminated in

silicon‐rich acid waters. Nature, 1989, pp 146‐148. 127. Edwardson JA; Moore PB; Ferrier IN; Lilliy JS; Newton GWA; Barker J; Templar J; Day JP

Effect of silicon on gastrointestinal absorption of aluminum. Lancet, 1993, pp 211‐212. 128. Belles M; Sanchez DJ; Gomez M; Corbiella J; Domingo JL Silicon reduces aluminum

accumulation in rats: Relevance to the aluminum hypothesis of Alzheimer disease. Alzheimer Disease & Associated Disorders, 1998, pp 83‐87.

129. Perry CC; Keeling‐Tucker T Model studies of the precipitation of silica in the presence of aluminium; implications for biology and industry. Journal of Inorganic Biochemistry, 2000, pp 331‐339.

130. Kerr DNS; Ward MK Aluminium and its role in biology. In Metal ions in biological systems, Sigel H, Sigel A, Eds.; Dekker: New York, Basel, 1988; pp 217‐258.

131. Elliot MA; Edwards HM Some effects of dietary aluminium and silicon on broiler‐chickens. Poultry Science, 1991, pp 1390‐1402.

132. Najda J; Gminski J; Drozdz M; Flak A The effect of silicon (Si) on lipid parameters in blood‐serum and arterial‐wall. Biological Trace Element Research, 1991, pp 235‐247.

133. Najda J; Gminski J; Drozdz M; Zych F The influence of inorganic silicon (Si) on pitutary‐thyroid axis. Biological Trace Element Research, 1993, pp 101‐106.

134. Bedu O; Goy J; Roufy J; Loeper J Action of silicon on cultured lymphocytes. Medical Science Research, 1991, pp 317‐318.

135. Carlisle EM A metabolic role for silicon in cartilage growth. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 128‐133.

136. Seaborn CD; Nielsen FH Further studies of the effect of silicon on collagen‐metabolism. FASEB Journal, 1995, p A448.

137. Seaborn CD; Nielsen FH Germanium, silicon and their interaction affect bone composition. FASEB Journal, 1994, p A430.

138. Seaborn CD; Nielsen FH Silicon deprivation and arginine and cysteine supplementation affect bone collagen and bone and plasma trace mineral concentrations in rats. Journal of Trace Elements in Experimental Medicine, 2002, pp 113‐122.

139. Seaborn CD; Nielsen FH Cysteine, silicon and their interaction affect bone composition in the rat. FASEB Journal, 2001, p A972.

140. Seaborn CD; Nielsen FH Silicon deprivation decreases collagen formation in wounds and bone, and ornithine transaminase enzyme activity in liver. Biological Trace Element Research, 2002, pp 251‐261.

141. Seaborn CD; Briske‐Anderson M; Nielsen FH An interaction between dietary silicon and arginine affects imune function indicated by Con‐A‐induced DNA sythesis of rat splenic T‐lymphocytes. Biological Trace Element Research, 2002, pp 133‐142.

142. Loeper J; Goy J; Fragny M; Troniou R; Bedu O Study of fatty acids in atheroma induced in rabbits by an atherogenic diet with or without silicon i.v. treatment. Life Sciences, 1988, pp 2105‐2112.

143. Fasman GD; Moore CD The solubilization of model Alzheimer tangles ‐ reversing the beta‐sheet conformation induced by aluminum with silicates. Proceedings of the National Academy of Sciences of the United States of America, 1994, pp 11232‐11235.


16

144. Fasman GD; Perczel A; Moore CD Solubilization of beta‐amyloid‐(1‐42)‐peptide ‐ reversing the beta‐sheet conformation induced by aluminum with silicates. Proceedings of the National Academy of Sciences of the United States of America, 1995, pp 369‐371.

145. Arumugam MQ; Ireland DC; Brooks RA; Rushton N; Bonfield W Orthosilicic acid increases collagen type 1 mRNA expression in human bone‐derived osteoblasts in vitro. Bioceramics, Vol 16,Key Engineering Materials, 2004, pp 869‐872.

146. Schwarz K Significance and function of silicon in warm blooded animals. Review and outlook. In Biochemistry of silicon and related problems, Bendz G, Lindqvist I, Eds.; Plenum Press: New York, 1978; pp 207‐230.

147. Carlisle EM Essentiality and function of silicon. In Biochemistry of silicon and related problems, Bendz G, Lindqvist I, Eds.; 1977; pp 231‐253.

148. Izu A; Kumai T; Tohno Y; Tohno S; Minami T; Yamada G; Yamada MO Silicon intake to vertebral columns of mice after dietary supply. Biological Trace Element Research, 2006, pp 297‐316.

149. Seaborn CD; Nielsen FH Dietary silicon affects acid and alkaline‐phosphatase and calcium‐45 uptake in bone of rats. Journal of Trace Elements in Experimental Medicine, 1994, pp 11‐18.

150. Seaborn CD; Nielsen FH Dietary silicon and arginine affect mineral element composition of rat femur and vertebra. Biological Trace Element Research, 2002, pp 239‐250.

151. Keeting PE; Oursler MJ; Wiegand KE; Bonde SK; Spelsberg TC; Riggs BL Zeolite‐A increases proliferation, differentiation, and transforming growth‐factor‐beta production in normal adult human osteoblast‐like cells in vitro. Journal of Bone and Mineral Research, 1992, pp 1281‐1289.

152. Najda J; Goss M; Gminski J; Weglarz L; Siemianowicz K; Olszowy Z The antioxidant enzymes activity in the conditions of systemic hypersilicemia. Biological Trace Element Research, 1994, pp 63‐70.

153. Ott EA; Asquith RL Trace mineral supplementation of yearling horses. Journal of Animal Science, 1995, pp 466‐471.

154. Najda J; Gminski J; Drozdz M; Danch A The action of excessive, inorganic silicon (Si) on the mineral metabolism of calcium (Ca) and magnesium (Mg). Biological Trace Element Research, 1993, pp 107‐114.

155. Perry CC; Keeling‐Tucker T Aspects of the bioinorganic chemistry of silicon in conjunction with the biometals calcium, iron and aluminium. Journal of Inorganic Biochemistry, 1998, pp 181‐191.

156. Najda J; Gminski J; Drozdz M; Danch A Silicon metabolism ‐ The interrelations of inorganic silicon (Si) with systemic iron (Fe), zinc (Zn) and copper (Cu) pools in the rat. Biological Trace Element Research, 1992, pp 185‐195.

157. Lyu K; Nathanson D; Chou L Effects of silicon, calcium and phosphorus on human osteoblast culture. Journal of Dental Research, 2000, p 220.

158. Robins SP; Milne G; Stewart P The effects of copper deficiency on the lysine‐derived, pyridinium crosslinks of rat bone collagen. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 42‐45.

159. Saito K; Saito T; Ito K; Fujimoto S; Sasaki T Interaction of copper, zinc, manganese, and selenium to superoxide dismutase, catalase, and glutathione peroxidase in stomach cancer. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 805‐807.

160. Cunnane SC Contrasting effects of zinc and copper on the composition of long chain fatty acids in tissue lipids of the rat. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 70‐75.

161. Davies NT; Sarkozy P The effects of copper deficiency on lipid peroxidation in rat liver microsomes. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 39‐42.

162. Tenegrup I; Ågren MS; Bonderson L; Forsgren A; Hallmans G; Söderberg T The effect of zinc and occlusion on the healing of open wounds. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 87‐89.


17

163. Falchuk KH; Vallee BL Zinc and chromatin structure, composition and function. In Trace Elements in Man and Animals, Mills CF, Bremner I, Chesters JK, Eds.; Commonwealth Agricultural Bureaux: Slough, Farnham Royal, UK, 1985; pp 48‐55.

164. Walker GM Yeast physiology and biotechnology; John Wiley & Sons: Chichester, 1998. 165. Vrieling EG; Sun QY; Tian M; Kooyman PJ; Gieskes WWC; van Santen RA; Sommerdijk

NAJM Salinity‐dependant diatom biosilicification implies an important role of external ionic strength. Proceedings of the National Academy of Sciences of the United States of America, 2007, pp 10441‐10446.

166. Shipley RA; Clark RE Tracer methods for in vivo kinetics; Academic Press: New York, London, 1972.

167. Krijger GC; Harms AV; Leen R; Verburg TG; Wolterbeek B Chemical forms of technetium in tomato plants; TcO4

‐, Tc‐cysteine, Tc‐glutathione and Tc‐proteins. Environmental and Experimental Botany, 1999, pp 69‐81.

168. Kolar ZI Tracing of interfacial transport of matter in solid‐liquid systems in dynamic equilibrium. Journal of Trace and Microprobe Techniques, 1990, pp 103‐120.

169. Kolar ZI; Verburg TG; van Dijk HJM Three kinetically different inorganic phosphate entities in bovine casein micelles revealed by isotopic exchange method and compartmental analysis. Inorganic Biochemistry, 2002, pp 61‐66.

170. Marshall KC Interfaces in microbial ecology; Harvard University Press: Cambridge, 1976. 171. Gjaltema A Biofilm development: growth versus detachment. Technische Universiteit Delft,

1996. 172. Boletti FT Manufacture of vinegar. In Microbiology; a textbook on microorganisms general and

applied., third revised edition ed.; Marshall CE, Ed.; Churchill: London, 1921; pp 636‐648. 173. Bryers JD; Characklis WG Biofilms in wastewater treatment. In Biofilms, Characklis WG,

Marshall KC, Eds.; John Wiley & Sons: New York, 1990; pp 671‐696. 174. Stocks‐Fischer S; Galinat JK; Bang SS Microbiological precipitation of CaCO3. Soil Biology &

Biochemistry, 1999, pp 1563‐1571.


18

Chapter 2

Preparation of 31Si labelled silicate:

a radiotracer for silicon studies in biosystems

Abstract

No‐carrier added 31Si was produced by 31P(n,p)31Si in the nuclear reactor of the Delft University of Technology. Several methods were investigated to remove the side product 32P, and the chemical form of 31Si was investigated. All methods except one gave good results. Anion exchange with Dowex resin gave the best results in activity concentration (3.8 ∙ 106 Bq/ml, with only 3.2 ∙ 103 Bq/ml 32P as by product) and specific activity (As > 17 TBq/g). The product is suitable for biological systems.*

* This chapter is reprinted with kind permission of Wiley‐Blackwell, Oxford, United Kingdom: “H.J. Brasser, G. Gürboğa, J.J. Kroon, Z.I. Kolar, H.T. Wolterbeek, K.J. Volkers, G.C. Krijger (2006) Preparation of 31Si lebelled silicate: a radiotracer for silicon studies in biosystems. Journal of Labelled Compounds and Radiopharmaceuticals 47: 867‐882”.

Chapter 2: Tracer preparation and characterization

20

2.1 Introduction

Silicon (Si) is the second most abundant element (25.7 % w/w) in the earth’s crust, being exceeded only by oxygen (49.2 %).1 Its three stable isotopes, 28Si, 29Si and 30Si, have natural abundances of 92.23, 4.67 and 3.10 %, respectively. Silicon is undoubtedly important for many living organisms and it may have played a significant role in the origin of life.2 It is considered to be beneficial or essential to some plants, higher animals and humans, although only a few functions of silicon have been unravelled so far.3,4 In biosystems and natural waters silicon is always in the chemical form of silicate (ortho, meta or higher polymerized). At low concentrations (< 5 mmol/l) it exists in the form of orthosilicate (Si(OH)4), while above this concentration polymerization will take place to “metasilicate” (orthosilicate oligomers). At very high concentrations formation of silica gel ((SiO3

2‐

)n) will occur. Metasilicate can depolymerize to orthosilicate when diluted to concentrations below 5 mmol/l.5,6

A convenient way for measuring transport rates in biosystems, including man, is by means of radioactive tracers that are chemically and physically identical to the compound/substance of interest. For silicon two radionuclides are suitable to be used as labels in such tracers, namely 31Si (t1/2 = 2.62 h, decays to 31P (stable)) and 32Si (t1/2 = 160 y, decays to 32P (radioactive)).7 Both radionuclides emit β‐ radiation, with maximum energies of 0.225 MeV (0.07%) and 1.49 MeV (99.9%) for 31Si and of 0.213 MeV (100%) for 32Si. In addition, 31Si emits 1.27 MeV γ‐rays (0.07%).8‐10 Instead of Si radionuclides, radioactive Ge (68Ge or 71Ge) has been used to study the behaviour of Si,11,12 but this is probably not representative under all circumstances. 32Si can be produced by either the 37Cl(p,2pα)32Si reaction or the reaction sequence 31P(n,γ)32P(n,p)32Si using an accelerator or a high neutron flux nuclear reactor respectively.13 32Si labelled compounds have been used for example in studies of silicon in marine food webs and for human and animal uptake kinetics.14‐17 However, the use of 31Si is advantageous in short‐term experiments, because (a) it is relatively easy to produce in high yields using a relatively low neutron flux reactor, (b) it does not decay to 32P which can interfere with radioactivity measurements and (c) it results in far less radioactive waste. In addition, double tracer experiments using both 31Si and 32Si would become feasible. 31Si can be produced by reactor thermal neutron irradiation of silicon or its compounds by the 30Si(n,γ)31Si nuclear reaction with a cross section (σ) of 0.108 barn. It can also be produced by fast neutron irradiation of phosphorus or its compounds by the 31P(n,p)31Si nuclear reaction having a threshold energy of 0.7328 MeV and a σ for fast neutrons of 0.034 barn.7,8,17,18 Preparation of 31Si from 31P using the (n,p) route is accompanied by the simultaneous production of 32P (t1/2 = 14.26 d) due to the presence of thermal neutrons. The involved nuclear reaction 31P(n,γ)32P has a σ of 0.16 barn. 32P emits β‐ radiation with maximum energy of 1.710 MeV.7,8 For doping experiments 31Si is often produced by the 30Si(d,p)31Si reaction using an accelerator,19,20 but no information is given on a purity check of the produced nuclide. We will focus on the production of 31Si in a nuclear reactor. The physical data of the mentioned radionuclides are summarized in Table 1.


21

Table 1: Preparation of the radionuclides 31Si, 32Si and 32P and their physical data,8‐10 *recalculated value.10

Nuclide 31Si 32Si 32P Half‐live time 2.62 h 160 y 14.62 d Radiation type, energies

β‐ 0.225 MeV (0.07%) β‐ 1.49 MeV (99.9%) γ 1.27 MeV (0.07%).

β‐ 0.213 MeV (100%) β‐ 1.710 MeV (100 %)

Reaction, cross section

30Si(n,γ)31Si (0.108 barn) 31P(n,p)31Si (0.0312 barn)

37Cl(p,2pα)32Si (0.42 barn) 31P(n,γ)32P(n,p)32Si (0.16 resp. 0.12* barn)

31P(n,γ)32P (0.16 barn)

Our present study aims at producing 31Si via the 31P(n,p)31Si reaction followed by radiochemical processing and chemical characterization of the final product for use as a label for biological studies with high specific activity. Note that silicon is present as a contaminant in water and in many chemicals, which influences the specific activity of the final product. In several studies 31Si was produced by the 31P(n,p)31Si 21‐23 reaction. However the specific activity was not given,21 or the silicon concentration was not determined 21,22 which makes the found specific activity questionable. Also the chemical form of the tracer was not investigated.21‐23 As target material 31P free of silicon impurities is preferred instead of 30Si because of the possibility to produce no carrier added 31Si, and because the natural abundance of the target material is 100%, so no enrichment is needed, in contrast to 30Si which has a low abundance in natural Si, or is very expensive in enriched form. The use in biosystems implies the final radiotracer solution should have a pH that matches the system to study. It should have a chemical form identical to the natural non‐radioactive form, and it should not contain toxic or harmful substances.

2.2 Experimental 2.2.1 Chemicals

The following chemicals were used: phosphoric acid (H3PO4, Fluka 76922), silver acetate (CH3COOAg, Fluka 85140), barium carbonate (BaCO3, Fluka 11729), lanthanum nitrate (La(NO3)3, Merck 5326), and sodium metasilicate (Na2SiO3, Aldrich 30,781‐5). All chemicals were at least of analytical grade, and silicon was not on the list of impurities 2.2.2 Preparation of the radionuclide

The target material, 100 mg (1 mmol) orthophosphoric acid (H3PO4) was solid packed in small polyethylene tubes which were sealed at both ends. These tubes were placed in polyethylene “rabbits” and pneumatically transported to a position close to the core of the 2 MW swimming pool research reactor of the Reactor Institute Delft, Delft University of Technology, The Netherlands. The targets were neutron irradiated for 5.0 h. The neutron fluxes at the target position were about 5.1 × 1012 cm‐2s‐1 for thermal and 3.7 × 1011 cm‐2s‐1 for fast neutrons.


22

31Si and 32P activities were measured in 5 ml (diluted) sample using a LKB liquid scintillation counter (β‐ radiation, 98% detection efficiency or Čerenkov light, 39% detection efficiency). For determination of β‐ radiation 15 ml Ultima Gold LSC cocktail was added to the sample. Čerenkov light detection was used in cases when chemical conditions (like low pH) gave problems with the LSC cocktail. Moreover, 31Si was also determined in a well‐type Ge‐Li‐detector (γ radiation). The radioactive decay of the samples was followed over time in order to calculate the initial 31Si and 32P activities at the end of the target bombardment. Moreover, half‐lives were checked to ascertain for the absence of contaminations by mathematical fitting the half‐live times to the decay of the radionuclides over time using MicroMath Scientist software. 2.2.3 Removal of phosphate, including 32P from the radionuclide solution

Phosphate (including 32P) was removed from the target material solution by precipitation with barium carbonate (Method 1), silver acetate (Method 2) or lanthanum nitrate (Method 3), or by anion exchange separation with Dowex resin (Method 4) or prefabricated columns (Method 5). Method 1. The irradiated H3PO4 was dissolved in 2.0 ml deionized water and transferred to a plastic centrifuge tube. To the solution 100 μl 1.0 mol/l HCl was added, followed by 0.35 g (1.7 mmol, 10% excess) solid BaCO3 causing CO2 release and a white precipitate, probably barium hydrogen phosphate (BaHPO4, Ks = 10‐6.74).24 After 5 min. CO2 escape had stopped and the solution was filtered (membrane filter, pore size 0.45 μm). The pH was increased by the addition of 150 μl 1 mol/l NaOH to lower the Ks of barium phosphate in order to precipitate the remaining barium and phosphate ions, followed by filtration to remove the precipitate particles. Method 2. The irradiated H3PO4 was dissolved in 2.0 ml deionized water and transferred to a plastic centrifuge tube. 0.60 g (3.3 mmol) solid silver acetate was added to the radioactive solution causing a yellow precipitate, probably silver orthophosphate (Ag3PO4, Ks = 10‐16.1).1 After 5 min. the suspension was filtered (membrane filter, pore size 0.45 μm), and 0.03 g (0.5 mmol) NaCl (solid) was admixed to the filtrate in order to react with the remaining silver ions and form AgCl (Ks = 10‐9.75).1 A final filtration removed the precipitated particles. Method 3. The irradiated H3PO4 was dissolved in 3.0 ml deionized water and transferred to a plastic centrifuge tube. To the solution 200 μl 1 mol/l NaOH was added, followed by 0.40 g (1.2 mmol) solid lanthanum nitrate La(NO3)3 yielding a white precipitate (lanthanum orthophosphate LaPO4, Ks = 10‐25.75).25 The suspension was filtered (membrane filter, 0.45 μm) and 100 μl 1 mol/l NaOH was added to the filtrate to react with the remaining La ions and form La(OH)3 (Ks = 10‐20.06).26 A second filtration yielded a filtrate free of precipitate particles. Method 4. The irradiated target material (100 mg) was dissolved in 2.0 ml water. Dowex 8x2 anion exchange resin was soaked in ultrapure water, washed with 8.0 mol/l HNO3 and rinsed with water until neutral pH. A plastic column (Bio‐Rad Econo‐


23

pac) was filled with 10 ml resin slurry. The resin was loaded with 10 ml 5.0 mol/l NaCl solution and rinsed with 10 ml 100 mmol/l TRIS in ultrapure water (pH 5.0). The solution of the irradiated target material was adjusted to pH 5.0, added to the top of the resin and eluted with 10 ml 100 mmol/l TRIS in ultrapure water of the same pH. The first 4 ml eluent was discarded, the following 6 ml yielded the purified product. Method 5. The irradiated target material (100 mg) was dissolved in 2.0 ml water. Prefabricated anion resin columns (Accell™ Plus QMA) were purchased from Waters (Etten Leur, The Netherlands). Before use the columns were rinsed with 5 mol/l NaCl (4 ml/ml bed volume) and water of the desired pH (8 ml/ml bed volume). The solution of the irradiated target material was adjusted to the desired pH with 1 mol/l NaOH and added to a column (max. 1 ml solution/g resin). The eluent was kept for activity determination. 2.2.4 Analyses

Silicate concentrations were determined with inductively coupled plasma optical emission spectrometry (ICP‐OES, Perkin Elmer OES Optima 4300DV). The spectrometer was calibrated with Merck CertiPUR silicon standard solution (1703). X‐ray diffraction measurements were performed on a Bruker‐AXS type D5005 diffractometer, equipped with a Huber CuKa1 incident‐beam monochromator and Braun Position Sensitive Detector. The measurement range was 15‐70 degrees 2‐theta with a stepsize of 0.039 degrees and a counting time per step of 1 s. The specimen consisted of a thin layer of powder on a Si single crystal wafer with orientation <510> (a 'zero‐background’ substrate). 2.2.5 Gel formation

A 5.0 ml 1.5 mol/l Na2SiO3 solution was added to a tube together with 100 µl 31Si/32P solution (refined by Method 1). Gel formation was started by adding 10 ml 0 ‐ 3.0 mol/l HCl solution to the tube. The final concentrations in the tube were 500 mmol/l Na2SiO3 and 0 – 2.0 mol/l HCl. It is crucial the HCl solution is slowly added against the wall of the tube without producing any turbulence or mixing in the liquid. Disturbing the liquid, e.g. by shaking, retards the start of the polymerization reaction for several hours. After HCl addition the tubes were left undisturbed for 15 minutes to allow the gel formation reaction to start. Then the tubes were shaken to yield a homogeneous mixture of gel and liquid and allowed to stand for another 15 minutes to complete the polymerization reaction. After centrifugation at 1500 g the supernatant was added to a counting vial for activity determination. The gel pellet was dispersed in 10 ml water and the mixture was added to a counting vial for radioactivity determination. The volume and density of gel and supernatant were determined by weighing. 2.2.6 Paper chromatography

Paper chromatography was carried out on tracer solution at pH 7‐12. 31Si solution (refined by Method 1) was added to chromatography paper (Whatman, Maidstone, England), allowed to dry, and eluted in 2‐propanol (70%), water (10%), 20% trichloroacetic acid (20%) and 25% ammonia (0.3%) mobile phase as described by


24

Hettler27 until the front had reached about 75 % of the paper. After drying the paper was put on a linear transport system and the elution profile of the tracer (31Si and 32P) was determined using a GM detector. After complete decay of 31Si (after at least two days) 32P activity was determined in the elution profile to calculate the original 31Si activity in the separate spots. 2.2.7 Interaction of 32P with non‐radioactive Si 32P solution was prepared by irradiating 100 mg H3PO4 as described above followed by dissolving the target material in 2.0 ml deionized water. The solution was kept for at least 2 days allowing 31Si to decay completely. Subsequently 2.9 ml solution containing 50 mg/ml H3PO4 and 0, 1.0 or 10 mmol/l Na2SiO3 was added to a counting vial together with 100 μl 32P solution. Plastic counting vials were used because of the good heat and pressure resistance. The vials were closed and incubated at 20 or 115 oC for 5 hours. After incubation the liquid was added to plastic centrifuge tubes and possible loss of water by evaporation was compensated for. 1.0 ml sample was used for radioactivity determination, and precipitation with barium carbonate was carried out on the remaining 2.0 ml liquid. Separation of precipitate and solution was performed by centrifugation at 1500 g. The supernatant and pellet were added to counting vials for radioactivity determination. The volume of the supernatant was determined by weighing.

2.3 Results and discussion 2.3.1 Production of the radionuclide

Phosphoric acid was irradiated as described in the experimental section. The 31P(n,p) 31Si and 31P(n,γ) 32P reactions resulted in the production of 31Si and 32P as determined using both a liquid scintillation counter (31Si and 32P) and a well‐type Ge‐Li‐detector (31Si). An amount of 4.3 ± 0.8 MBq 31Si per mmol P was formed, accompanied by 5.7 ± 1.1 MBq 32P under the conditions in the reactor. To determine the presence of radionuclide contaminations (apart from 31Si and 32P) half‐live times were checked by mathematical fitting of the decay data (Figure 1). The found half‐live values, 2.67 ± 0.05 hours and 13.98 ± 0.16 days, respectively, agreed well with the literature.7,8 This indicates the absence of major radionuclide contaminations.


25

Figure 1: The 31Si and 32P activity in irradiated target material over time (■ not purified, ▼ purified by Method 1, lines: half life time fit of data)

The 31Si/32P ratio’s can be modified by the use of neutrons with different energies. In addition, shielding material with a high thermal neutron capture cross‐section (e.g. cadmium or boron) during the irradiation can be used to reduce the occurrence of the 31P(n,γ)32P reaction and to increase the 31Si/32P ratio. However, this would slightly reduce the 31Si yield and is only required when 32P interferes with the experiments.

2.3.2 Removal of 32P from the irradiated target material

Five methods to remove 32P from the irradiated target material solution have been investigated: three using chemical precipitation (Method 1 – 3) and two using anion exchange (Method 4 and 5). Purification of the irradiated material by chemical precipitation of phosphate (including 32P) was carried out as described in the experimental section. Precipitation of 32P with barium carbonate (Method 1) and silver acetate (Method 2) gave good results (Table 2).

Table 2: Properties of 31Si tracer obtained by different purification methods. The remaining activity after purification is given as fraction of the original produced nuclide.

Method Nuclide Activity concentration

Remaining activity at t = 0

[Si] Specific activity

Bq/ml % mg/l TBq/g 1. precipitation Si‐31 3.2 ± 0.8 ∙ 106 74 ± 13 0.67 ± 0.10 4.8 ± 1.4 (BaCO3) P‐32 2.8 ± 1.4 ∙ 104 1.1 ± 0.2 2. precipitation Si‐31 1.8 ± 0.05 ∙ 106 66 ± 22 0.59 ± 0.06 3.0 ± 0.08 (Ag‐acetate) P‐32 7.1 ± 0.8 ∙ 103 0.25 ± 0.09 3. precipitation Si‐31 5.4 ± 2.9 ∙ 103 0.5 ± 0.3 < 10‐3 > 5 (La(NO3)3) P‐32 6.7 ± 2.5 ∙ 102 0.04 ± 0.01 4. anion exchange Si‐31 3.8 ± 0.1 ∙ 106 43.9 ± 1.1 0.21 ± 0.01 17.8 ± 0.4 (Dowex resin) P‐32 3.2 ± 0.3 ∙ 103 0.03 ± 0.003 5. anion exchange Si‐31 1.1 ± 0.1 ∙106 91.0 ± 1.1 5.14 ± 0.09 0.21 ± 0.02 (Accell™ Plus Q) P‐32 6.1 ± 1.9 ∙ 103 0.4 ± 0.1


26

Table 3: Precipitating salts in refining methods 1 ‐ 3 as determined by X‐ray diffraction Method After addition of Formed precipitates Method 1 BaCO3 + HCl BaHPO4 and BaCO3 NaOH BaCO3 Method 2 Ag‐acetate Ag3PO4 and Ag‐acetate NaCl AgCl Method 3 La(NO3)3 + NaOH LaPO4 NaOH no precipitate

Removal of 32P as barium hydrogen phosphate (Method 1) or silver phosphate (Method 2) was successful although 1.1 % (Method 1) or 0.36 % (Method 2) of it remained in solution, calculated as percentage of 32P at the end of neutron bombardment. 73 % (Method 1) or 70 % (Method 2) of the initially present 31Si remained in the purified solution, which make these procedures suitable for the production of 31Si labelled silicate. The specific activity of 31Si amounted to 4.8 ± 1.4 TBq/g (Method 1) or 3.0 ± 0.08 TBq/g (Method 2). Probably non‐radioactive silicon was present as a contaminant in the chemicals used. Whether the removal of 32P by Method 1 could be improved was investigated by means of centrifugation instead of filtration to remove small precipitate particles that can pass the filter, but no difference was found. Sequential precipitation with fresh barium carbonate or silver acetate did not lead to higher recoveries of 32P. Precipitation with lanthanum nitrate (Method 3) removed more than 99.95% of the original 32P as lanthanum phosphate. However, 99.5% of the 31Si was co‐precipitated, which clearly makes this method unsuitable for the preparation of the tracer (Table 2). X‐ray diffraction analysis revealed the presence of phosphate in all precipitates from the first precipitation step of Method 1, 2 and 3. In case of Method 1 and 2 the original salt (BaCO3 or Ag acetate) is also precipitated, probably as a result of pH change (Table 3). In Figure 2 a typical X‐ray diffraction spectrum is shown.

Anion exchange was investigated as a method for phosphate removal from the irradiated target material. Silicic acid has its first pK value at 9.66 and phosphate at 2.12.1 As long as the pH is kept between 3 and 9 silicic acid will not be dissociated and will pass through the resin whereas dissociated phosphate will interact with the anion exchange resin. Above pH 9 silicate will also start to dissociate and interact with the anion exchange resin. This principle has been investigated with Dowex anion exchange resin in laboratory made columns (Method 4) and with a commercially available prefab column (Method 5) as described in the experimental section.


27

Figure 2: Typical X‐ray diffraction spectrum of LaPO4.

Anion exchange purification at pH 5 resulted in 43.9 % recovery of 31Si and 99.7 % removal of 32P (Method 4), and in 91.0 % recovery of 31Si and 99.6% removal of 32P (Method 5). The obtained specific activity was 17.8 ± 0.4 TBq/g (Method 4) and 0.21 ± 0.02 TBq/g (Method 5). Solution refined by Method 5 contained a rather high amount of Si (5.14 ± 0.09 mg/l). The not purified solution contained 0.010 ± 0.001 mg/l Si, which indicates the column itself contains Si as a contaminant which is released during elution. Method 4 yielded a solution containing 0.21 ± 0.01 mg/l Si. Probably Si is removed during pretreatment with 8 M HNO3. It is not possible to treat Method 5 columns with 8 M HNO3, but probably the Si contamination can be reduced by increasing the volumes of NaCl solution and water during pretreatment. The 31Si yield (Method 5) could be improved to 96.7% by rinsing the column with water pH 5 (volume equal to bed volume) after elution. It was not possible to remove the remaining 32P by a second elution (Method 5). The results are summarized in Table 2. The influence of the pH has been investigated for pH 2‐8 (Method 5). At pH 2 20% of 32P remains in solution, above pH 4 99.7‐99.9 % of 32P is removed. Above pH 6 the recovery of 31Si shows a declining tendency (Figure 3). This is an indication 31Si has the silicate chemical form. In the next two experiments the chemical form of the tracer was further investigated.


28

Figure 3: Recovery of 31Si (�) and 32P (∆) from the target solution as a function of pH after treatment with anion exchange (Method 5) (n=3, ± sd).

It was not possible to completely remove 32P by any method. The results of Method 3 indicate 32P exists for at least 99.95% as orthophosphate. The precipitation product, lanthanum orthophosphate, has a very low solubility (Ks = 10‐25.75),25 so only minute amounts of orthophosphate can remain in solution. Possibly the remaining 32P does not exist in phosphate form, but is chemically altered during irradiation.

The 31P(n,p)31Si reaction suggests the possibility of carrier free 31Si preparation. But the results show all purified tracer solutions contain non‐radioactive silicon, which is probably introduced as contaminants of the used chemicals. When the silicon content is kept low by the use of ultrapure chemicals high specific 31Si activities can be obtained.

2.3.3 Investigation of the chemical form of 31Si

When a radiotracer is used to investigate the role of silicon in biosystems it is important that the chemical form is identical to that of natural silicon. In natural waters and biosystems silicon is usually present in the form of silicate.5 In dilute aqueous solutions (< 5 mmol/l) silicate molecules exist in the form of silicic acid, Si(OH)4 or orthosilicate (SiO(OH)3‐), and in higher concentrations as "metasilicate" (oligomers of orthosilicate). In highly concentrated solutions polymerization (gel formation) occurs. The gel formation process is started when a concentrated silicate solution (> 0.1 mol/l) is brought into contact with a strong acid.5,6 This results in a gel phase with a silicate concentration higher than in the starting solution, and a water phase (supernatant) with lower silicate concentration than in the starting solution. In this experiment the ability of silicate molecules to form a gel is used to investigate the chemical form of 31Si. If 31Si is in the silicate form it will participate in the chemical reaction of the polymerization process and will be incorporated in the gel matrix. As a result the 31Si activity per ml gel (Agel) will be higher than the activity per ml starting solution (A0), and the activity per ml supernatant (Asup) will be lower than A0. In other words: Agel/A0 > 1 and Asup/A0 < 1. If 31Si is not in the silicate form the activity will be equally distributed over gel and supernatant. The results can be compared with 32P present in the tracer solution, which can not participate in the gel


29

forming process. The activity of 32P will be equally distributed over gel (trapped in the gel matrix) and supernatant (Agel/A0 = Asup/A0 = 1). The experiment was conducted as described in the experimental section. The amount of gel that is formed depends on the final concentration of HCl (Figure 4). Activities of 31Si and 32P were determined before addition of HCl (to calculate A0), and after gel formation (to calculate Agel and Asup), and the ratios Agel/A0 and Asup/A0 were calculated for both nuclides (Figure 5).

Figure 4: Gel volume obtained in 0.5 mol/l Na2SiO3 as a function of HCl concentration (n=4, ± sd).

The results show Agel/A0 >1 and Asup/A0 < 1 for 31Si, especially when more acid is added. This indicates the participation of 31Si in the gel forming process, indicating 31Si is in the silicate form. For 32P both Agel/A0 and Asup/A0 are about 1, resulting from an equal distribution of the nuclide over gel phase and supernatant.

Figure 5: Activity (A) per ml gel or per ml supernatant after gel formation as compared to the activity (A0) per ml at the start (■ 31Si gel, ● 31Si supernatant, ∆ 32P gel, � 32P supernatant, n=4, ± sd).

To confirm the silicate form of 31Si as suggested by the above experiment its behaviour is further investigated using paper chromatography. Silicic acid (Si(OH)4) has a pK value of 9.66,1 which means the molecules are non‐dissociated at low and


30

neutral pH. At pH 8‐12 silicate will be partly dissociated, and above pH 12 dissociation will be complete. The dissociated and non‐dissociated form can be separated by paper chromatography as described in the experimental section. Non‐dissociated silicate will move faster during elution than the dissociated form, resulting in a spot for the non‐dissociated form with high RF and one for the dissociated form with low RF. The pH determines the degree of dissociation of the silicate molecules, and hence the size of the two spots. 31Si solution (pH 7‐12) was added to chromatography paper, as described in the experimental section. After development two radioactive spots were found (spot 1, RF 0.1 and spot 2, RF 0.9) containing an amount of 31Si dependent on the pH of the solution. An increase in pH results in a higher activity of 31Si in spot 1 and a decrease of activity in spot 2. The fraction of dissociated and non‐dissociated silicate is calculated and the results are summarized (Figure 6). These results are in accordance with the results of the gel formation experiment, and confirm the silicate chemical form of the 31Si tracer.

Figure 6: Activity of 31Si in elution profile as fraction of the total activity of 31Si as a function of pH, determined using paper chromatography (�RF 0.1 dissociated form, ∆RF 0.9 non‐dissociated form, n=3, ± sd). It is possible that during neutron irradiation of the target material the formed 31Si reacts with the phosphoric acid as a result of the heating up,28 influencing the chemical properties of the 31Si nuclide. To investigate whether this is feasible 32P is brought into contact with 0, 1 or 10 mM silicate and incubated for 5 hours at room temperature (20 oC) or at 115 oC as described in the experimental section. The high temperature is chosen to cover for a possible rise of temperature in the reactor during irradiation, the incubation time is equal to the irradiation time in the tracer production. After incubation the solutions were treated with Method 1, and the fraction of 32P that remained in solution was determined. The results showed no influence of silicate concentration or temperature on the fraction 32P that remained in solution (Table 4). In all cases 1.4 to 1.5 % of 32P was not precipitated, which equals the amount found in Method 1 as described above. So it is unlikely a phosphorus‐silicon complex is formed during irradiation.


31

Table 4: influence of temperature and silicate concentration on the fraction of not precipitated 32P. ns: not significant (P>0.5)

[H3PO4] [Na2SiO3] Fraction 32P in supernatant T‐test

g/ml mmol/l 115 oC (%) 20 oC (%) 0.05 0 1.50 ± 0.08 1.44 ± 0.05 ns 0.05 1 1.45 ± 0.10 1.41 ± 0.07 ns 0.05 10 1.48 ± 0.13 1.38 ± 0.01 ns

2.4 Conclusions All refining methods are suitable for tracer production, except Method 3 (lanthanum nitrate precipitation), which resulted in 99.5 % loss of 31Si, probably by co‐precipitation. The results show the specific activity is determined by the presence of silicon as a contaminant in chemicals and materials. Methods 4 and 5 (anion exchange) have the advantage of the highest removal of 32P, but Method 4 is more time consuming than the precipitation methods. Probably this drawback can be overcome by applying a vacuum to increase the elution speed. Method 5 is a fast method with the highest recovery of 31Si and removal of 32P. However its specific activity is lowest of all methods, and the costs are relatively high compared to the other methods. If chemical precipitation is used Method 1 (barium carbonate precipitation) is preferred over Method 2 (silver acetate precipitation) because barium is less toxic than silver. Moreover acetate, the counter ion of silver stays in solution and can be consumed by many organisms as a substrate and interfere with the experiments, whereas carbonate, the counter ion of barium escapes as CO2 during precipitation. The presence of a minor amount of 32P in the tracer should not be a problem as long as it does not interfere with the experiments (especially when phosphate concentrations in the media are relatively high). However it is mandatory to do more activity determinations on one sample to correct for the contribution of 32P to the signal. The experiments show 31Si has the silicate chemical form, which means the tracer is chemically identical to natural silicon and can be used for studies of biological systems. In conclusion, the purification Methods 1, 4 and 5 yield 31Si tracer of high activity and high specific activity that are usable in biosystem studies. The choice depends on the needs of the experiment. When a high activity per volume is needed Method 1 is preferred, whereas Method 4 yields the highest specific activity, but is more time consuming.


32

2.5 References 1. Lide LR (chief ed) (2002) CRC Handbook of Chemistry and Physics, 83rd ed. CRC Press LLC: Boca

Raton Florida. 2. Smith JV, Arnold FP, Parsons I, Lee MR (1999) Biochemical evolution III: Polymerization on

organophilic silica‐rich surfaces, crystal‐chemical modeling, formation of first cells, and geological clues, P Nati Ac Sci Unit 96: 3479 ‐ 3485.

3. Roperto F, Borzacchiello G, Ungaro R, Galati P (2000). Silicate pneumoconiosis in hens, J Co Path 122: 249 ‐ 254. DOI: 10.1053/jcpa.1999.0367

4. Miura Y, Nakai K, Sera K, Sato M (1999) Trace elements in sera from patients with renal disease, Nucl Inst Meth Phy B 150: 218 ‐ 221. DOI: 10.1016/S0168‐583X(98)01028‐3

5. Iler RK (1979) The chemistry of silica, Solubility, Polymerization, Colloid and Surface Properties, and Biochemistry. John Wiley & Sons Ltd: New York.

6. Petzold A, Hinz W (1978) Silikatchemie, Einführung in die Grundlagen. VEB Deutsche Verlag für Grundstoffindustrie: Leipzig. (in german)

7. Moller P, Nix JR, Myers WD, Swiatecki WJ (1995) Atomic Data and Nuclear Data Tables 59: 185 ‐ 381. DOI: 10.1006/adnd.1995.1002

8. Firestone RB, Shirley VS (eds) (1996) Table of isotopes, 8th ed. John Wiley & Sons Ltd: New York. 9. Pfennig G, Klewe‐Nebenius H, Seelman‐Eggebert W (1998) Karlsruher Nuklidkarte, 6th ed 1995,

rev 1998. Forschungszentrum Karlsruhe GmbH: Karlsruhe. 10. Forberg S (1972) Nuclear‐reactor production of carrier‐free Si‐32, Radiochim Acta 18: 194 ‐197. 11. Matics S, Frank WFJ (2000) Diffusion of Ge‐71 in the amorphous ceramic Si28C36N36, J Non Crys

Soli; 266: 830 ‐ 834 Part B. DOI: 10.1016/S0022‐3093(99)00850‐9 12. Tallberg P, Koski‐Vahala J, Hartikainen H (2002) Germanium‐68 as a tracer for silicon fluxes in

freshwater sediment, Water Res 36: 956 ‐ 962. DOI: 10.1016/S0043‐1354(01)00312‐8 13. Hofmann HJ, Bonani G, Suter M, Wölfli W, Zimmermann D, Von Gunten HR (1990) A new

determination of the half‐life of Si‐32, Nucl Inst Meth Phy B 52: 544 ‐ 551. DOI: 10.1016/0168‐583X(90)90474‐9

14. Brzezinski MA, Phillips DR (1997) Evaluation of 32Si as a tracer for measuring silica production rates in marine waters, Limn Ocea 42: 856 ‐ 865.

15. Popplewell JF, King SJ, Day JP, Ackrill P, Fifield LK, Cresswell RG, Di Tada ML, Liu K (1998) Kinetics of uptake and elimination of silicic acid by a human subject: A novel application of Si‐32 and accelerator mass spectrometry, J Inorg Biochem 69: 177‐ 180. DOI: 10.1016/S0162‐0134(97)10016‐2

16. Taylor GA, Pullen GRL, Keith AB, Edwardson JA (1992) Ge‐68 as a possible marker for silicon transport in rat‐brain, Neurochem Res 17: 1181 ‐ 1185. DOI: 10.1007/BF00968396

17. Mehard CW, Volcani BE (1975) Similarity in uptake and retention of trace amounts of 31‐silicon and 68‐germanium in rat tissues and cell organelles, Bioinorg Chem 5: 107 ‐ 124. DOI: 10.1016/S0006‐3061(00)80055‐1

18. Geraldo LP, Dias MS, Koskinas MF (1992) Average neutron cross section measurements in U‐235 fission spectrum for some threshold reactions, Radiochim Act 57: 63 ‐ 67.

19. Laitinen P, Strohm A, Huikari J, Nieminen A, Voss T, Grodon C, Riihimaki I, Kummer M, Aysto J, Dendooven P, Raisanen J, Frank W (2002) Self‐diffusion of Si‐31 and Ge‐71 in relaxed Si0.20Ge0.80 layers, Phys Rev Letters 89: art. no. 085902. DOI: 10.1103/PhysRevLett.89.085902

20. Salamon M, Strohm A, Voss T, Laitinen P, Riihimaki I, Divinski S, Frank W, Raisanen J, Mehrer H (2004) Self‐diffusion of silicon in molybdenum disilicide, Philosoph Mag 84: 737 ‐ 756. DOI: 10.1080/14786430310001641966

21. Ichikawa F, Sato T (1970) Studies of behaviour of carrier‐free radioisotopes. 4. Some behaviours of carrier‐free silicon‐31 from view point of radiocolloid formation, Radiochim Acta 13: 69 ‐ 70.

22. Koskinas MF, Dias MS (1993) Measurement of the gamma‐ray probability per decay of Si‐31, Appl Rad Isotopes 44: 1209 ‐ 1211. DOI: 10.1016/0969‐8042(93)90066‐J

23. Berlyne GM, Shainkinkestenbaum R, Yagil R, Alfassi Z, Kushelevsky A (1986) Distribution of silicon‐31‐labeled silicic‐acid in the rat, Bio Trac Elem Res 10: 159 ‐ 162.

24. Olmsted J, Williams G (1998) Chemistry, the Molecular Science, 2nd ed. California State University: Fullerton.


33

25. Liu XW, Byrne RH (1998) Comprehensive investigation of yttrium and rare earth element complexation by carbonate ions using ICP‐Mass spectrometry, J Solut Chem 27: 803 ‐ 815. DOI: 10.1023/A:1022677119835

26. Diakonov II, Ragnasrdottir KV, Tagirov BR (1998) Standard thermodynamic properties and heat capacity equations of rare earth hydroxides: II. Ce(III)‐, Pr‐, Sm‐, Eu(III)‐, Gd‐, Tb‐, Dy‐, Ho‐, Er‐, Tm‐, Yb‐, and Y‐hydroxides. Comparison of thermodynamical and solubility data, Chem Geol 151: 327 ‐ 347. DOI: 10.1016/S0009‐2541(98)00088‐6

27. Hettler H (1959) Paper chromatography of inorganic phosphorus compounds, Chromatogr Rev 1: 225 ‐ 245.

28. Makart H (1967) Untersuchungen an Siliciumphosphaten, Helv Chim Acta 50: 339 ‐ 405.


34

Chapter 3

The Influence of Silicon on Cobalt, Zinc and Magnesium in Baker’s Yeast Saccharomyces cerevisiae

Abstract Silicon (Si, as silicate) is involved in numerous important structure and function roles in a wide range of organisms, including man. Silicate availability influences metal concentrations within various cell‐ and tissue‐types but, as yet, clear mechanisms for such influence have been discovered only within the diatoms and sponges. In this study, the influence of silicate on the intracellular accumulation of metals was investigated in baker's yeast Saccharomyces cerevisiae. It was found that at concentrations up to 10 mM silicate did not influence the growth rate of S. cerevisiae within a standard complete medium. However, an 11 % growth inhibition was observed when silicate was present at 100 mM. Intracellular metal concentrations were investigated in yeast cultures grown without added silicate (‐Si) or with the addition of 10 mM silicate (+Si). Decreased amounts of Co (52%), Mn (35%) and Fe (20%) were found within +Si grown yeast cultures as compared to ‐Si grown ones, while increased amounts of Mo (56%) and Mg (38%) were found. Amounts of Zn and K were apparently unaffected by the presence of silicon. +Si enhanced yeast growth rate in a low Zn2+, but decreased growth rate under conditions of low Mg2+, and did not alter the growth rates in high Zn2+ and Co2+ media. +Si doubled the uptake rate of Co2+, but did not influence that of Zn2+. We propose that a possible explanation for these results is that polysilicate formation at the cell wall changes the cell wall binding capacity for metal ions. The toxicity of silicate was compared to germanium (Ge, as GeO2), a member of the same group of elements as Si (group 14). Hence, Si and Ge are chemically similar, but silicate starts to polymerize to oligomers above 5 mM while Ge salts remain as monomers at such concentrations. Ge proved to be far more toxic to yeast than Si and no influence was found of Si on Ge toxicity. We propose that these results relate to differences in cellular uptake.*

* This chapter is reprinted with kind permission from Springer Science+Business Media: “The Influence of Silicon on Cobalt, Zinc and Magnesium in Baker's Yeast Saccharomyces cerevisiae. H.J. Brasser, G.C. Krijger, T.G. van Meerten, H.T. Wolterbeek, Biol. Trace Elem. Res. 2006, 112: 175‐190, 4 figures.”

Chapter 3: Silicon influence on metals in yeast

36

3.1 Introduction Silicon is the second most abundant element within the earth’s crust. In natural waters it is present in the form of orthosilicate (Si(OH)4) at low concentrations (< 5 mM). Above this concentration, polymerization is initiated, forming metasilicate (orthosilicate oligomers) and, at very high concentrations, silica ((SiO3

2‐)n). Metasilicate, when diluted to concentrations below 5 mM, may depolymerize to produce orthosilicate.1,2 In both plants and animals, silicon participates in essential structural and functional roles. For example, in plants silicon‐based compounds are involved in providing mechanical strength and protection against drought, fungi and metal toxicity.3 In higher animals, silicon compounds influence the production of bone and cartilage, lipid metabolism and DNA synthesis, and may affect enzymatic activities. Furthermore, both in plants and higher animals, silicon can influence the concentration of essential metals within tissues. However, despite these important roles for silicon, only in the diatoms and sponges is there any substantial degree of understanding of the mechanisms underlying such functions and very little is currently known in relation to the higher animals.4‐9 Silicon, as silicate, may directly influence cellular metal concentrations by a chemical mechanism, either inside or outside the cell. The latter situation is found in some plants that cope with toxic levels of zinc by formation of a zinc‐silicate precipitate, which can be removed from the tissue.10,11 Silicon may also act indirectly by influencing bioprocesses in cells which in turn influence metal levels. In higher animals it is difficult to investigate whether the influence of silicon on metal levels in the body reflects a direct interaction or a more complex series of events in the body. A single cell simplifies this situation somewhat in that it is more likely to react directly to changes in its external environment. Yeasts are good candidate single‐cell organisms for such studies as they are important in their own right in the biotechnology industry and, in research, commonly serve as models for eucaryotic cells.12 Therefore, we chose to study the influence of silicon on metal‐uptake and homoeostasis in Saccharomyces cerevisiae (baker’s yeast), and to assess the chemical or biological role of silicon in metal metabolism.

3.2 Materials and Methods 3.2.1 Chemicals and labware

All chemicals were of analytical grade and obtained from BDH (Amsterdam, The Netherlands) or Aldrich (Zwijndrecht, The Netherlands). All solutions were prepared in ultrapure water (18.2 MΩ/cm, Millipore Milli‐Q, Billerica, MA, USA). Potassium silicate (K2SiO3) stocks, 0.1 M, were freshly prepared. The maximum silicate concentration supplied in the growth medium was 100 mM. Higher concentrations were not used because of problems in adjusting the pH, and the likelihood of SiOx gel formation. GeO2 stock solutions were prepared as 40 mM (saturation concentration). Precautions were taken to avoid silicate contamination from dust, chemicals and water: 1) all used labware was made of polycarbonate or polypropylene, 2) labware was rinsed with 1 M HCl, then rinsed with demineralized water and finally washed


37

with ultrapure water, and 3) a laminar flow cabinet was used to prevent dust contamination and to ensure sterile conditions during sampling and inoculation. 3.2.2 Organism and culture procedures

The yeast strain used was Saccharomyces cerevisiae CEN.PK 113‐7D, wild type (provided by Kluyverlab for Biotechnology, University of Technology Delft, The Netherlands), and it was cultured in standard inorganic mineral medium (pH 6.5) according to Verduyn et al.,13 containing the following compounds per liter: (NH4)2SO4, 5g; KH2PO4, 3g MgSO4.7H2O, 0.5g; EDTA, 15 mg; ZnSO4.7H2O, 4.5 mg; CoCl2.6H2O, 0.3 mg; MnCl2.4H2O, 1 mg; CuSO4.5H2O, 0.3 mg; CaCl2.2H2O, 4.5 mg; FeSO4.7H2O, 3 mg; NaMoO4.2H2O, 0.4 mg; H3BO4, 1 mg; and KI, 0.1 mg;. Final vitamin concentrations per liter were: biotin, 0.05 mg; calcium pantothenate, 1 mg; nicotinic acid, 1 mg; inositol, 25 mg; thiamine HCl, 1 mg; pyridoxine HCl, 1 mg; and para‐aminobenzoic acid, 0.2 mg. The carbon source consisted of 10 gl‐1 glucose. Culture procedures and sampling were carried out under sterile conditions in a laminar flow cabinet. The organism was grown aerobically, in batch culture within 500 ml Erlenmeyer flasks containing 100 ml standard medium, in an incubator (New Brunswick Scientific, Edison, NJ, USA), maintained at 180 ‐ 200 rpm and 30 oC. To obtain exponentially growing cells, experimental cultures were inoculated from a stationary‐phase preculture. All experiments were carried out on cells within the exponential growth phase. The growth rate of the culture was determined from the time‐dependent increase of the optical density (OD660). To obtain dose‐response curves for silicate or GeO2, potassium silicate or GeO2 stocks, respectively, were added to the standard growth medium (0‐100 mM silicate; 1 µM‐35 mM GeO2) and the growth rate was determined, as above. To obtain high Zn2+ or Co2+ concentrations ZnCl2 or CoCl2 stocks (0.1 M) were added to the standard growth medium. Cultures low in Zn2+ or Mg2+ were obtained by inoculation of a flask containing Zn2+ or Mg2+ deficient medium from a standard medium preculture. The desired amount of metal in the culture liquid was obtained by adjusting the inoculation volume (1 – 5 ml), in some cases accompanied by the addition of extra salts from a stock solution to prevent large inoculation volumes. When very low concentrations were desired (nM range), a “two‐preculture” procedure was used. A flask containing Zn2+ deficient medium was inoculated with 1 ml culture liquid from a standard medium preculture. After 6 to 12 hours of growth the experimental flasks (also containing Zn2+ deficient medium) were inoculated (1 – 5 ml) from this culture. To obtain biomass for elemental analysis by INAA, cells were harvested after 6 hours of growth, when the culture was still in the exponential growth phase. The 100 mL batch culture was centrifuged, the cells were washed twice in 10 ml TRIS‐HCl buffer (0.1 M, pH 6.5) and resuspended in 5 ml TRIS‐HCl buffer at 4 oC. The cell suspension was frozen at –50 oC and freeze‐dried. The biomass of 45 cultures (100 ml) was harvested and homogenized after freeze‐drying to obtain a representative sample. To obtain biomass for Si analysis, cells were harvested after 6 hours of growth, when the culture was still in the exponential growth phase. The culture liquid was centrifuged, the cells were washed in ultrapure water and the cell pellet was digested with 2 ml concentrated HNO3 at 60 oC for 1‐2 hours.14 After cooling, the


38

samples were made up to 10 ml with ultrapure water. Si concentrations were determined by ICP‐OES (see below). An internal standard was added to the samples. 3.2.3 Analyses The metal content of freeze‐dried yeast cells was determined with instrumental neutron activation analysis (INAA) in the nuclear reactor of the Reactor Institute Delft of the University of Technology in Delft, The Netherlands, following methods as described by Blaauw.15 All elements present in the standard medium were analyzed (i.e. K, Mn, Co, Mg, Fe, Zn, Mo). Standard reference material NBS1577b (bovine liver) was used as reference material for INAA. Inductively coupled plasma optical emission spectrometry (ICP‐OES, Perkin Elmer OES Optima 4300DV, Boston, USA) was used to determine Si, Co, Zn and Ge concentrations in the culture liquid and Si concentration in digested yeast cells. For calibration, Merck CertiPUR standard solutions 1703 (Amsterdam, The Netherlands) were used. Culture liquid samples were taken directly after inoculation and yeast cells were removed, by centrifugation, prior to analysis. 3.2.4 65Zn and 60Co uptake in yeast cells 60Co tracer (37 MBq, t1/2 = 5.27 years, γ = 1173 and 1333 keV) was purchased from NEN Life Science Products Inc. (Boston, USA). The specific activity was 7.44 GBq 60Co per mg Co. 65Zn tracer (218 MBq, t1/2 = 244 days, βmax 330 keV, γ = 511 and 1115 keV) was bought from PerkinElmer Life and Analytical Sciences (Boston, USA). The specific activity was 0.11 GBq 65Zn per mg Zn. 100 μl Zn or Co tracer solution was added to 100 ml of the standard growth medium. The solution was mixed well and allowed to equilibrate for 15 minutes prior to inoculation from a standard yeast preculture. To determine the tracer uptake, 10 ml samples were taken from the culture at regular time intervals during growth and the cells pelleted by centrifugation. The cells were washed once in 5 ml standard medium (4 oC), centrifuged, resuspended in 5 ml standard medium within a vial for radioactivity determination. Content of 65Zn and 60Co in the samples was measured on a Wallac 1480 WizardTM 3" Gamma Counter (Turku, Finland) and related to OD660 to calculate mol Zn or mol Co per gram dry wheight (DW).

3.3 Results 3.3.1 Influence of silicate on the growth rate of Baker’s yeast in standard medium

To determine the general affect of silicate levels on growth of S. cerevisiae, a dose‐response curve was obtained by supplementing a standard complete growth medium with potassium silicate (K2SiO3). The results show that the growth rate is unaffected by Si concentrations up to 10 mM (Fig. 1).


39

Figure 1: Influence of Si on growth rate (mu) of Baker’s yeast. Values are means of 3 to 6 cultures. A slight inhibition of growth was found above 10 mM range with a maximum of 11% growth inhibition at 100 mM silicate as compared to cultures grown with no additional Si (P = 0.003). The lowest achievable silicate concentration under normal lab conditions was 5.3 μM ± 1.5 μM as measured in the standard complete medium. At this concentration, no indications for deficiency symptoms of silicate (i.e. a decline of the growth rate) were found as compared to standard medium. Based on these results, further experiments were carried out using "no silicate added" (i.e. 5.3 μM, the measured silicate content) and 10 mM‐supplemented silicate. These conditions will be referred to as ‐Si and +Si. 3.3.2 Influence of silicate on metal levels in Baker’s yeast

To investigate the influence of silicate on metal concentrations in yeast, an elemental analysis was carried out on yeast cells cultured in a standard medium (i.e. ‐Si) and compared with +Si cultures. The results are summarized in Table 1. The corresponding elemental analysis in the standard reference material was consistent with the certified values (<5%). Cells grown in +Si‐standard medium contained over 9‐fold higher levels of Si than those grown –Si. Higher silicate corresponded with significant decreases in the cellular levels of Co (52 % decrease), Mn (35 %), Fe (20 %) and K (5%) and increases Mo (56 %) and Mg (38 %). Zn was not affected by silicate addition. To check for the possibility of depletion of the medium with respect to one or more of the metals, the fraction of metals consumed from the medium was calculated (Table 1).


40

Table 1: Metal content of yeast cells in ‐Si and + Si cultures. Si is determined by ICP‐OES, the other elements by INAA, as described in Materials and Methods. * significant (P < 0.05), ** significant (P < 0.01), ns not significant. Average of 3 samples measured by INAA

element metal concentration Increase consumed (µg/g DW) (%) (%) ‐ Si + Si +Si/‐Si ‐Si +Si Si 389 ± 141 3692 ± 321 949 ** 3.95 0.20 Co 0.87 ± 0.01 0.42 ± 0.034 ‐52 ** 1.42 0.48 Mn 14.97 ± 0.25 9.67 ± 0.46 ‐35 ** 16.99 11.04 Fe 98.00 ± 2.00 78.0 ± 2.65 ‐20 * 19.66 15.69 Zn 122.7 ± 2.31 115.3 ± 4.0 ‐6 ns 14.50 13.63 K 15433 ± 231 14600 ± 100 ‐5 * 0.22 0.21 Mg 1237 ± 50 1710 ± 50 38 * 3.42 4.76 Mo 1.14 ± 0.09 1.77 ± 0.05 56 ** 8.56 13.56

The fraction of silicate consumed by the organism was calculated using the silicate medium concentrations as measured by ICP‐OES. All metals in the standard medium, including silicate were found to be present in surplus at the moment the cells were harvested. 3.3.3 Influence of silicon on growth rate at toxic or deficient levels of metals

From the data shown in Table 1, it is apparent that silicate addition influenced the intracellular levels of several metals. However, this was not reflected in the growth rate of the yeast culture when silicate was supplemented to the standard complete medium (Fig. 1). Therefore it was investigated whether the influence of silicate on intracellular metal levels had an affect on growth rate when the yeast was cultured under conditions of metal deficiency or toxicity. Cobalt, zinc and magnesium were chosen for this further study. The results are summarized in Table 2. Table 2: Influence of silicon on growth rate of Baker's yeast under different conditions. Standard medium contains 16 µM Zn2+, 1.2 µM Co2+ and 4.2 mM Mg 2+. * significant (P< 0.005), ** significant (P<0.0005), ns not significant. Average of 5 (standard medium, Co2+ tox, Mg2+ low), 4 (Zn2+ def), or 3 determinations (Zn2+ tox). # It was not possible to get a deficient culture

Conditions growth rate (h‐1) ‐ Si + Si standard medium 0.356 ± 0.011 0.354 ± 0.006 ns Co2+ tox (0.3 mM) 0.213 ± 0.012 0.204 ± 0.010 ns Co2+ def # n.m. n.m. Zn2+ tox (5 mM) 0.199 ± 0.019 0.198 ± 0.011 ns Zn2+ def (10 nM) 0.083 ± 0.025 0.121 ± 0.024 * Mg2+ low (0.05 mM) 0.355 ± 0.025 0.315 ± 0.005 **

Yeast cultured in standard medium (no silicate) supplemented with 0.3 mM Co2+ or 5 mM Zn2+ showed 50% growth inhibition. Despite the finding that intracellular Co levels were reduced (approximately 2‐fold) in the presence of supplementary silicate (Table 1), +Si medium did not alleviate the toxic effects of high Co2+. In the case of yeast growth‐rate in a Mg2+‐deficient medium, the addition of supplementary


41

silicate was associated with a decrease in growth rate; this seems counter‐intuitive in the light of the enhanced intracellular Mg levels (Table 1). +Si appeared to have no influence on Zn levels within the cell (Table 1). Si‐supplementation of high Zn2+ cultures produced no change in growth rate compared those in high Zn but without additional silicate. Silicate addition of Zn2+ low cultures accelerated growth rate. 3.3.4 Influence of silicon on zinc and cobalt uptake in the cell

From the data described above, it was apparent that metal concentrations accumulated within yeast cells are influenced by silicate addition to the medium. To follow the change in intracellular accumulation of Co, as influenced by +Si, the radiotracer 60Co was used. The uptake of 65Zn was also followed +/‐Si , in this case no affect of silicate had been found on the cells ability to accumulate this metal. Figures 2 and 3 show the yeast uptake of Co2+ or Zn2+ per g DW during exponential growth and the calculated rates are shown in Table 3.

Figure 2: Zn2+ uptake per gram dry weight as determined with 65Zn uptake per gram DW in –Si and +Si standard media. Table 3: Co and Zn uptake per g dry weight as determined with 60Co and 65Zn uptake per OD in –Si and +Si standard medium

metal

uptake rate 10‐9 mol/gDW/h

ratio +Si/‐Si

‐Si +Si Co2+ 2.87 ± 0.15 5.46 ± 0.44 1.9 Zn2+ 304 ± 21.3 293 ± 15.6 1.0


42

Figure 3: Co2+ uptake per gram dry weight as determined with 60Co uptake per gram dry weight in –Si and +Si standard media. Potassium (counter ion of silicate in the stock solution) did not influence Zn2+ or Co2+ uptake under these conditions (data not shown). In addition, blank determinations (no yeast) were performed to rule out precipitation of silicate with Co2+ or Zn2+ in the medium. No precipitates were detected (data not shown). This means the 60Co and 65Zn uptake as determined in this experiment can be ascribed to uptake by the biomass itself and not to possible precipitates that would end up in the same fraction as the yeast cells. 3.3.5 Comparison of the effect of silicate and germanium on baker’s yeast

Silicon and germanium (Ge) belong to the same group of elements in the periodic table (group 14) and, hence, are chemically similar. However, the biological importance of Si has not similarly been found for Ge (as germanate, dissolved GeO2). It is established in the literature16 that several yeast species can take‐up Ge into the cell, and this uptake is associated with toxic effects such as a decline in growth rate. This is marked contrast, therefore, to our findings of a lack of silicate toxicity (Fig. 1). We therefore confirmed the toxicity of Ge under the conditions and yeast strain used in this paper and, furthermore, investigated the interaction of silicate and Ge with respect to the toxic response. The effect of Ge was determined by a dose‐response curve (see Materials and Methods). Toxic effects of Ge are clearly visible as a decline in growth rate (Fig. 4) with 35 mM resulting in 70 ‐ 90% inhibition of growth rate.


43

Figure 4: Influence of Ge on yeast growth rate (mu) and influence of Si on Ge toxicity. Values are means of three cultures. The interaction of silicate with Ge toxicity was studied using growth rate experiments in +Si and ‐Si cultures at three different Ge concentrations (3.0, 7.7 and 35 mM Ge). Figure 4 shows that the addition of silicate had no significant influence on the growth rate (P>0.4) for any of the Ge concentrations tested.

3.4 Discussion and conclusions We were unable to detect an affect of depleted silicate levels on yeast growth, thus, no deficiency effects were apparent at levels as low as 5.3 μM ± 1.5 μM silicate. However, the limitation of the system meant that we were unable to decrease measurable silicate levels below this value and therefore we cannot discount the possibility of deficiency symptoms in a further depleted medium. With respect to improving the system to allow this further depletion, our analyses showed that silicate was below the limits of detection (ca. 0.03 µM) in the ultrapure water and the vitamin/glucose solution. Thus, it appears most likely that the contaminant is present within one or more chemicals of the culture medium. Extra purified analytical grade chemicals will be required, in combination with clean‐room conditions, to achieve, effectively, null silicate levels (Si contamination of the chemicals is hard to determine due to signal‐noise ratio caused by the chemicals). However, we can conclude that no symptoms of silicate deficiency will occur under normal lab conditions and this was not, therefore, a factor in any of the experiments reported here. Ge and Si share many chemical similarities, and Ge radiotracers are used in biology as a probe for Si.17 Based on the chemical similarities it might be predicted that Si toxicity to yeast would be similar to that of Ge, whereas, our results show that yeast growth is far more sensitive to the toxic effects of Ge (ca. 90% growth‐inhibition at 35 mM Ge compared to 11% at 0.1 M Si) (Figs. 1 and 4). No influence was found of supplementary silicate on Ge toxicity (Fig. 4). This is consistent with the results in Table 2 (high Zn2+ and Co2+), where no influence was found of silicate addition on metal toxicity. In Table 1 it was shown cells grown in +Si contain over 8 times more Si than cells grown in –Si, and, therefore, it is apparent that silicate, like Ge16 is readily taken up


44

by yeast cells. However, the biochemical behaviour of silicate and Ge differs substantially, especially in the higher concentration range (> 5 mM). One key difference is the ability of silicate to form polymers above ca. 5 mM, whereas Ge remains as monomers.2 It is likely that silicate oligomers are physically inhibited from passing through the cell membrane. It is possible, therefore, that the higher Si content measured in the yeast grown in +Si medium in fact reflects the polymeric form adhereing to the external cell surface, rather than internal levels. In this scenario, the difference in biochemical behaviour between Ge and silicate in the higher concentration range may in fact simply reflect a difference in uptake into the cell interior. Silicate affects metal concentrations in yeast cells (Table 1) and the manner of this influence varies for each metal which can be subdivided in three categories: 1. a decrease in cellular content by +Si (Co, Mn, Fe); 2. no influence on cellular content by +Si (Zn, K). 3. an increase in cellular content by +Si (Mg, Mo). In the first category (1), it was expected that silicate addition would alleviate the toxic effects of these particular metals because the metal concentrations in the cell were decreased, relative to those found in the absence of supplementary Si. However, this prediction was not verified by the results (Table 2, Table 3, Fig. 3). In category 2, no influence of +Si, either on growth rates or Zn uptake rate was found, as expected, with the exception of low Zn2+ growth rates (Table 2, Table 3, Fig 2). In the case of Mg (category 3), supplementary silicate meant that the apparent Mg content of the cell was increased. At apparent odds with this, silicate addition negatively influenced the growth rate at low Mg2+ (Table 2). As already mentioned, it cannot be ruled out that, rather than the supplementary silicate being taken up intracellulary, it in fact becomes complexed with the cell wall. Different studies showed that complex formation between silicate oligomers and several saccharides can occur.18‐20 The concentration in the +Si cultures was 10 mM, well above the starting concentration of oligomerization of orthosilicate ions (around 5 mM) (2), so silicate oligomers were present in +Si cultures. It is plausible that silicate oligomers within the medium complex with cell wall polysaccharides resulting in a relatively stable layer of polysilicates associated with the cell wall. If so, it is likely that this Si‐complex would interfere with the processes of the cell wall. The cell wall consists of polysaccharides (a glucan‐chitin layer bound to the cell membrane and an outer layer of mannoproteins) and plays an important role in uptake of metals.12 High quantities of metal ions are adsorbed to the mannoprotein layer and, in lesser amounts, to the glucan‐chitin layer. The metal ion adsorptive capacity of the cell wall is merely determined by the degree of dissociation of the negatively charged functional groups. The affinity of the cell wall to metals is element specific: Zn2+ > Co2+ ≈ Mn2+ > Mg2+.21‐25 Polysilicates can bind with metal ions in a chelate‐like way (Fig. 5). The stability of the complex is element specific: Mo2+ and Mg2+ form a stable, Zn2+ and K+ an unstable complex.2,26‐28


45

Figure 5: Polysilicate metal chelate‐like complex2

It is possible that a layer of polysilicates on the cell wall could interfere with the normal metal adsorption process by shielding negatively charged groups, thereby altering the normal metal binding capacity of the cell wall. The formation of a layer of polysilicates could cause diffusion limitation for metal ions through the cell wall by interference with the cell wall's ability to bind metals, and could alter the capacity of the cell wall for metal ions, and directly influence the metal concentration in the cell wall and the biomass. This mechanism could provide an explanation for the results found in this study. Category 1. The rate of Co2+‐uptake in +Si cultures was approximately twice that found in –Si cultures but, paradoxically, elemental analysis showed a decline of Co levels within the cell. An explanation for this observation can be proposed if it is assumed that a polysilicate layer over the cell wall shields the charged groups, thereby lowering the affinity for Co2+. Thus, availability of silicon for polysilicate formation would result in reduced cell‐wall levels of Co2+ which, in turn, would cause the yeast cell to experience a lower Co2+ concentration in the periplasmic space. Our data indicated that any such decrease in Co2+ was not profound enough to trigger deficiency symptoms such as a decrease in growth rate, but may have been responsible for the enhanced cellular uptake rate of Co2+, presumably induced in response to the cell’s unfulfilled biochemical demands for Co2+. A precedent for such an effect is found in a report of metal uptake in erythrocytes. Here, when the metal concentration in the cell was lowered, the uptake rate was enhanced and no symptoms of metal deficiency in the cell were observed.29 Hence, this scenario appears compatible with the decline of Co in the biomass in combination with the enhanced Co2+ uptake rate, as observed in the +Si yeast cultures (Tables 1 and 3, Fig. 3). Category 2. The formation of polysilicate‐metal complexes may have an effect on metal levels in the cell that depends on the dissociation rate of the metal‐polysilicate complexes. In this scenario we propose that a layer of metasilicate oligomers is associated with the cell wall and that further complexes are created through the scavenging of metal ions from the medium. With regard to metal availability to the cell, this can be advantageous or disadvantageous, depending on the stability of the complex. Thus, if the complex dissociates easily, the free metal ions are released to create a concentrated micro‐environment very close to the cell wall, thereby facilitating uptake. This scenario appears compatible with our observation of a positive influence of silicate on the growth rate of yeast cells growing in a Zn2+ deficient medium. Category 3. The formation of a relatively stable polysilicate complex with Mg2+ may offer an explanation for the apparently contradictory observations of enhanced Mg content of the cell in +Si medium (Table 1) and yet reduced growth rate of cultures grown under conditions of low Mg2+ (Table 2) in the presence of +Si. Thus, the


46

normal affinity of the cell wall for Mg2+ is assumed to be low but, if stabilized as a Mg‐polysilicate complex, this cell‐wall component could explain the enhanced levels detected by elemental analysis. In this scenario where the Mg‐polysilicate complex is relatively stable, the silicate layer around the cell acts as a trap for Mg2+ ions. In +Si standard medium, i.e. conditions of excess Mg2+, there would most likely remain sufficient free Mg2+ to sustain normal growth. In low Mg2+ medium, however, this complex‐formation could restrict the levels of free Mg2+ available for intracellular uptake, so eliciting the deficiency symptom of reduced growth rate. This hypothesis implies that the Mg‐polysilicate layer was sufficiently stable to be retained by the cell‐wall throughout the washing procedures used during cell harvest. The growth rate of the yeast culture in media supplying toxic levels of Co2+, Zn2+ and germanate was found not to be influenced by the addition of silicate. In line with the above proposals, we suggest that, for these particular metals, the layer of polysilicate is readily saturated, even at the higher silicate concentrations. Therefore, levels of Co2+, Zn2+ and germanate remain above the toxic threshold both in our ‐Si and + Si media. There exist several reports of interactions between organisms and silicate oligomers or polysilicates. These include the precipitation of polysilicates on the cell wall of certain bacteria and these precipitations were found not to comprise the viability of the organism.30 There is growing evidence that silicon does not work directly on cellular bioprocesses but, rather, interferes with the metal concentrations of the intimate external environment experienced by the cell or organism. The data reported in this paper adds further weight to the argument for the importance of these mechanisms. If, indeed, silicon is involved in such a straight‐forward chemical reactions, then the influence of silicon as a regulator of metal concentrations in the cell or tissue could be extremely wide‐spread. Clearly, it is important to pursue this proposal further by studying, in more detail, cell biochemistry in relation to the formation and stability of silicate layers.

3.5 References 1. A. Petzold and W Hinz (1978) Silikatchemie, Einführung in die Grundlagen, VEB Deutsche Verlag

für Grundstoffindustrie, Leipzig (in german) 2. R.K. Iler (1979) The chemistry of silica, John Wiley & Sons, New York 3. E. Epstein (1999) Silicon, Annu. Rev. Plant Physl. 50, 641‐664 4. D. Evered and M. O'Connor (1986) Silicon biochemistry, Wiley, Chechester 5. T.L. Simpson and B.E. Volcani (1981) Silicon and siliceous structures in biological systems,

Springer Verlag, New York 6. J.D. Birchall (1990) The role of silicon in biology, Chem. Brit. 26, 141‐144 7. C.D. Seaborn and F.H. Nielsen (2002) Silicon deprivation and arginine and cysteine

supplementation affect bone collagen and bone plasma trace mineral concentrations in rats, J. Trac. Elem. Exp. Med. 15, 113‐122

8. V. Martin‐Jezequel, M. Hildebrand and M.A. Brzezinski (2000) Silicon metabolism in diatoms: implications for growth, J. Phyc. 36, 821‐840

9. W.E.G. Muller, A. Krasko, G. Le Pennec and H.C. Schröder (2003) Biochemistry and cell biology of silica formation in sponges, Microscop. Res. Tech. 62, 368‐377

10. D. Neumann and U. zur Nieden (2001) Silicon and heavy metal tolerance of higher plants, Phytochemistry 56, 685‐692

11. D. Neumann and C. de Figueiredo (2002) A novel mechanism of silicon uptake, Protoplasma 220, 59‐67


47

12. G.M. Walker (1998) Yeast physiology and biotechnology, Wiley, Chichester 13. C. Verduyn, E. Postma, A. Scheffers and J.P. van Dijken (1992) Effect of benzoic acid on the

metabolic fluxes in yeast: a continuous‐culture study on the regulation of respiration and alcoholic fermentation, Yeast 8, 501‐517

14. L. Bisconti, M. Pepi, S. Mangani and F. Baldi (1997) Reduction of vanadate to vanadyl by a strain of Saccharomyces cerevisiae, Biometals 10, 239‐246

15. M. Blaauw (1993) The holistic analysis of gamma‐ray spectra in instrumental neutron activation analysis, PhD Thesis, University of Technology, Delft, The Netherlands

16. M.I. van Dyke, H. Lee and J.T. Trevors (1989) Germanium toxicity in selected bacterial and yeast strains, J. Indus. Microbiol. 4, 299‐306

17. G.A. Taylor, G.R.L. Pullen, A.B. Keith and J.A. Edwardson (1992) Ge‐68 as a possible marker for silicon transport in rat‐brain, Neurochem. Res. 17, 1181‐1185

18. S.D. Kinrade, R.J. Hamilton, A.S. Schach and C.T.G. Knight (2001) Aqueous hypervalent silicon complexes with aliphatic sugar acids, J. Chem. Soc. Dalt. T. 7, 961‐963

19. S.D. Kinrade, J.W. Del Nin, A.S. Schach, T.A. Sloan, K.L. Wilson and C.T.G.Knight (1999) Stable five‐ and six‐coordinated silicate anions in aqueous solution, Science 285, 1542‐1545

20. J.B. Lambert, G. Lu, S.R. Singer and V.M. Kolb (2004) Silicate complexes of sugars in aqueous solution, J. Am. Chem. Soc. 126, 9611‐9625

21. D. Brady, A.D. Stoll, L. Starke and J.R. Duncan (1994) Chemical and enzymatic extraction of heavy‐metal binding polymers from isolated cell‐walls of Saccharomyces cerevisiae, Biotech. Bioeng. 44, 297‐302

22. E.G. Davidova and S.G. Kasparova (1992) Adsorption of metals by yeast cell walls, Microbiology 61, 716‐719

23. H. Sentenac and C. Grignon (1981) A model for predicting ionic equilibrium concentrations in cell‐walls, Plant Physiol. 68, 415‐419

24. P.R. Norris and D.P. Kelly (1979) Accumulation of metals by bacteria and yeasts, Dev. Ind. Microbiol. 20, 299‐308

25. C. White and G.M. Gadd (1987) The uptake and cellular distribution of zinc in Saccharomyces cerevisiae, J. Gen. Microbiol. 133, 727‐737,

26. G.B. Alexander (1954) The polymerization of monosilicic acid, J. Am. Chem. Soc. 76, 2095‐2096 27. W. Schwieger, W. Heyer, F. Wolf and K.‐H. Berg (1987) Zur Synthese von kristallinen

Metallsilikathydraten mit Schichtstruktur, Z. Anorg. Chem. 548, 204‐216 (in german) 28. W.L. Marshall and J.M. Warakomski (1980) Amorphous silica solubilities – II. Effect of aqueous

salt solutions at 25 oC, Geochim. Cosmochim. Acta 44, 915‐924 29. J. de Kok, C. van der Schoot, M. Veldhuizen and H.T. Wolterbeek (1993) The uptake of zinc by

erythrocytes under near‐physiological conditions, Biol. Trac. Elem. Res. 38, 13‐26 30. J.B. Fein, S. Scott and N. Rivera (2002) The effect of Fe on Si adsorption by Bacillus subtilis cell

walls: insights into non‐metabolic bacterial precipitation of silicate minerals, Chem. Geol. 182, 265‐273


48

Chapter 4

On the beneficial role of silicon to organisms: a case study on the importance of silicon chemistry

to metal accumulation in yeast

Abstract Silicon is involved in numerous important structural and functional roles in a wide range of organisms, including diatoms, plants and humans, but clear mechanisms have been discovered only in diatoms and sponges. Silicate availability influences metal concentrations within various cell‐ and tissue‐types, but a mechanism has not been discovered so far. In an earlier study on Baker’s yeast Saccharomyces cerevisiae it was proposed that a chemical mechanism, rather than a biological one, is important. In the present study the interaction of silicon with Baker’s yeast is further investigated by studying the influence of zinc and magnesium on Si‐accumulation both at a low and a high silicate concentration in the medium. Si‐accumulation fitted well with Freundlich adsorption and Si‐release followed depolymerization kinetics, indicating that silicate adsorbs to the surface of the cell rather than being transported over the cell membrane. Subsequently, adsorbed silicate interacts with metal ions and, therefore, alters the cell’s affinity for these ions. Since several metals are nutritional, these Si interactions can significantly change the growth and viability of organisms. In conclusion, the results show that chemistry is important in Si and metal accumulation in Baker’s yeast, and suggest that similar mechanisms should be studied in detail in other organisms to unravel essential roles of Si.∗

∗ This chapter is reprinted with kind permission from Springer Science+Business Media: “On the beneficial role of silicon to organisms: a case study on the importance of silicon chemistry to metal accumulation in yeast. H.J. Brasser, G.C. Krijger and H.T. Wolterbeek, Biol. Trace Elem. Res. 2008, 125: 81‐95, 7 figures.”

Chapter 4: Silicon chemistry in yeast

50

4.1 Introduction Silicon (Si) is the second most abundant element (25.7 % w/w) in the earth’s crust, being exceeded only by oxygen (49.2 %).1 In natural waters silicon is present as silicic acid, in concentration ranging from 0.07 to 0.2 mmol/L.2‐5 Silicic acid is undissociated at neutral pH. At low concentrations (< 2 mmol/L) it has the form of orthosilicic acid (Si(OH)4), but above this concentration polymerization is initiated, forming orthosilicate oligomers. Polymeric silicate depolymerizes to orthosilicic acid when diluted to concentrations below 2 mmol/L.6,7 Silicon is omnipresent in the ecosystem, and during the evolution life had to cope with its presence. It may have played a significant role in the origin of life8 and it is considered to be essential or beneficial to various organisms, including humans, although only a few functions of silicon have been unraveled so far. In both plants and animals, silicon participates in essential structural and functional roles. For example, in plants silicon‐based compounds are involved in providing mechanical strength and protection against drought, pathogens and metal toxicity.9‐11 In higher animals, silicon compounds influence the production of bone and cartilage, lipid metabolism and DNA synthesis, and affect several enzymatic activities. Furthermore, both in plants and higher animals, silicon can influence the concentration of essential metals within tissues.12‐16 Despite these important roles for silicon only in the diatoms and sponges the underlying mechanisms are well understood.17,18

Very little is known of biological silicate interactions in higher animals, and despite many efforts a biological process or silicate binding site has not been found so far.12‐16 This could be due to the focus on biological processes (e.g. in enzymes or binding sites) so far, while the indirect role of silicon via (physical‐) chemical processes could be far more important.19 Namely, silicate molecules in solution show a rich chemistry, including adsorption, (de)polymerization reactions and complex formation with metal ions.7,20‐24 Probably silicate could play a role in organisms by its chemical interactions with metal ions. Silicate‐metal complexes could influence the availability of metals that are important for the organism (e.g. as an enzymatic cofactor). The effects on several biological processes could then be ascribed to the availability of important metal ions. To investigate this hypothesis the single cell organism Baker’s yeast was used. Yeast is the most common model for eukaryotic cells, it is easy to handle and culture, and it lacks the many additional complicating experimental parameters of a higher organism.25 The study on Baker’s yeast revealed that silicate affects metal concentrations and metal uptake rates in the cell, and that under certain conditions the growth rate was affected. It was proposed that silicate adsorbed on the cell wall and influenced the metal availability for the cell by the formation of silicate‐metal complexes.26 In the present paper this hypothesis is further investigated. We focus on this possibly crucial indirect (physical‐) chemical role of silicon in living organisms rather than a biological one. The accumulation of silicate in yeast cells is investigated, and it is determined whether this process can be described by a biological or a (physical‐) chemical process. Finally the interaction of silicate with metal ions is studied. Radiotracer techniques are used to study the accumulation of silicate in yeast cells.


51

4.2 Materials and Methods 4.2.1 Chemicals and labware

All chemicals were at least of analytical grade and obtained from BDH (Amsterdam, The Netherlands) or Aldrich (Zwijndrecht, The Netherlands). All solutions were prepared in ultrapure water (18.2 MΩ/cm, Millipore Milli‐Q, Billerica, MA, USA). Precautions were taken to avoid silicate contamination from dust, chemicals and water: 1) all used labware was made of polycarbonate or polypropylene, 2) labware was rinsed with 1 mol/L HCl, then rinsed with demineralized water and finally washed with ultrapure water, and 3) a laminar flow cabinet was used to prevent dust contamination and to ensure sterile conditions during sampling and inoculation. Despite all precautions it was not possible to obtain a silicate free cell culture. No‐Si‐added media contain 1 μmol/L silicate. 31Si‐silicate (t1/2 2.62 h, β‐ 1.49 MeV) was produced in the nuclear reactor of the Reactor Institute Delft, University of Technology Delft, The Netherlands. No‐carrier‐added 31Si‐silicate solution (specific activity 4.8 TBq/g) was prepared by purification using a chemical precipitation reaction with barium carbonate as described earlier.27

4.2.2 Organism and culture procedures

The yeast strain Saccharomyces cerevisiae CEN.PK 113‐7D, wild type (kindly provided by the Kluyver laboratory for Biotechnology, University of Technology Delft, The Netherlands), was cultured in standard mineral medium (pH 6.5),28 containing the following compounds per liter: (NH4)2SO4, 5g; KH2PO4, 3g; MgSO4.7H2O, 0.5g; EDTA, 15 mg; ZnSO4.7H2O, 4.5 mg; CoCl2.6H2O, 0.3 mg; MnCl2.4H2O, 1 mg; CuSO4.5H2O, 0.3 mg; CaCl2.2H2O, 4.5 mg; FeSO4.7H2O, 3 mg; NaMoO4.2H2O, 0.4 mg; H3BO4, 1 mg; and KI, 0.1 mg;. Final vitamin concentrations per liter were: biotin, 0.05 mg; calcium pantothenate, 1 mg; nicotinic acid, 1 mg; inositol, 25 mg; thiamine HCl, 1 mg; pyridoxine HCl, 1 mg; and para‐aminobenzoic acid, 0.2 mg. The carbon source consisted of 10 g/L glucose. Culture procedures and sampling were carried out under sterile conditions in a laminar flow cabinet. The organism was grown aerobically, in batch culture within 500 mL Erlenmeyer flasks containing 100 mL standard medium, in an incubator (New Brunswick Scientific, Edison, NJ, USA), maintained at 180 ‐ 200 rpm and 30 oC. The biomass of the culture was determined from the optical density at 660 nm (OD660) of the culture (1 OD660 unit corresponded with 0.39 g/L dry weight). Silicate was added from a potassium silicate (K2SiO3) stock, 0.1 mol/L, which was freshly prepared. High Zn2+ concentrations in the growth medium were obtained by addition of Zn2+ from a 10 mmol/L ZnCl stock solution. Low Mg2+ cultures were obtained by inoculation of the experimental flasks (containing 80 mL Mg2+‐deficient medium) with 20 mL from a normal preculture. Cultures low in Zn2+ were obtained in two steps by inoculation of a flask (containing 100 mL Zn2+‐deficient medium) with 1 mL culture liquid from a normal preculture. After 24 hours of growth the experimental flasks (containing 80 mL Zn‐deficient medium) were inoculated with 20 mL from this culture. Some experiments were carried out on dead cells, which were obtained by γ‐irradiation of a culture with a 10 kGy dose (60Co source, Gammacell® 220 Excel, MDS


52

Nordion). Viability of the cells was examined by determination of the growth rate and by methylene blue staining. 4.2.3 Analyses

Inductively coupled plasma optical emission spectrometry (ICP‐OES, Perkin Elmer OES Optima 4300DV, Boston, USA) was used to determine Si, Co and Zn concentrations in the culture liquid. For calibration, Merck CertiPUR standard solutions 1703 (Amsterdam, The Netherlands) were used. Culture liquid samples were taken directly after inoculation and yeast cells were removed, by centrifugation (600 g, 10 min), prior to analysis. The activity of 31Si was measured on a LKB liquid scintillation counter (β‐ radiation) using Ultima Gold XR liquid scintillation counting cocktail. Filters containing radioactive samples were dissolved in 10 mL liquid scintillation counting (LSC) cocktail within a counting vial prior to analysis. The volume of liquid samples was adjusted to 5 mL with demineralized water and mixed with 15 mL LSC counting cocktail prior to analysis. 4.2.4 31Si accumulation in yeast cells

No carrier‐added 31Si tracer solution (0.1‐0.5 mL, 50‐250 kBq, 4.8 TBq/g) was added to 80 mL standard medium. The solution was mixed well and allowed to equilibrate for at least 15 minutes prior to inoculation. The flasks were inoculated with 20 mL from a standard no‐Si added yeast preculture (OD660 2‐3). To determine the 31Si tracer accumulation, 10 mL samples were taken from the culture. Samples were taken after 120 minutes of incubation, unless stated otherwise. The samples were filtered over a 0.45 µm membrane filter and washed with 2.5 mL standard medium containing 5 mmol/L Si. The filters and a 1 mL sample of the culture suspension were kept for determination of amounts of radioactivity. These amounts of radioactivity were related to OD660 and silicate concentration in the samples to calculate the silicate accumulation in mol/g DW. A correction was applied for 31Si adsorption on the filter. 4.2.5 Determination of 31Si release

Yeast cells were cultured in medium containing 10 mmol/L silicate and 31Si (1 MBq) as described above. After 4 hours the biomass of the culture was harvested by centrifugation (20 min, 600 g), washed once in medium containing 10 mmol/L silicate. At t=0 the cells were resuspended in 50 mL standard no‐Si added medium (not containing 31Si). At regular intervals a 1.5 mL sample of the culture was put in an eppendorf vial and centrifuged (1 min, 1800 g). The supernatant (1 mL) and a sample of the suspension (1 mL) were kept for activity determination. The activity was related to the silicate concentration in the in the original culture and to OD660 of the suspension to calculate the silicate concentration in the biomass (Ccell in mol/L) and in the release medium (Cmed in mol/L). The decrease of the silicate concentration in the biomass Ccell in time was fitted to pseudo 2nd order depolymerization kinetics (see below).


53

4.2.6 Determination of the Freundlich parameters 1/n and K

The adsorption of a compound on a surface can be described by a Freundlich isotherm.29 The process is defined by two parameters: 1/n as the order of the adsorption process, and Kads as the affinity of the adsorbed compound for that specific surface (defined as the adsorption in 1 mol/L solute).

1n

ads medx K Cm= ⋅ (1)

with x as the amount of solute adsorbed by mass m of a solid (in this case the amount of silicate per mass of dry weight of yeast cells), Cmed as the solute concentration (in this case the silicate concentration in the medium), 1/n as the order of the adsorption reaction, and Kads as the affinity (adsorption at 1 mol/L solute). The equation can be rewritten as:

1log log logads med

x K Cm n= + ⋅

Plotting log adsorption against log concentration yields a straight line with a slope 1/n and an abscissa log Kads. 4.2.7 Determination of the depolymerization rate constant

The depolymerization of polysilicate is described as earlier22,24 and is determined by the depolymerization rate constant Kd,0 (in L/(mol∙s)). Kd,0 is dependent of the pH and the temperature:

10 11.61,0

115265.614 10 exp( 39.32)dK pHT

−= ⋅ ⋅ ⋅ − +

with temperature T in K. Kd,0 is influenced by the presence of ions in the solution. The rate constant corrected for ions in the solution Kd is calculated following: ,0 (exp( (exp( ) exp( ))))d d i iK K A a B a C= ⋅Π ⋅ ⋅ − ⋅

with ai as the activity of ion i. The constants A, B and C are summarized in Dietzel22 for different ions. At pH 6.5, 30 oC and standard medium composition the Kd value is 25.2 L/(mol∙min) (0.421 L/(mol∙s)). The initial monosilicate and polysilicate concentration (Cmed,i and Ccell,i) in the solution also influence the depolymerization rate constant. Two correction factors αM and αP have to be introduced to obtain the rate constant K that is actually measured in the solution: d M PK K= ⋅α ⋅α with αM and αP:

,

,

0.8053 1med iM

cell i

CC

α = ⋅ +

,exp(2.708 0.894 ln( ))P cell iCα = − ⋅

with Ccell,i in mg SiO2/L. The polysilicate depolymerization reaction follows pseudo 2nd order kinetics and is expressed as:


54

,

tot tottot

cell cell i

C CK t CC C

= ⋅ ⋅ + (2)

with t as time (min), Ctot as the total silicate (mol/L) (= total silicate present in cell suspension), Ccell as the polysilicate concentration (mol/L) (as polysilicate on the cell wall), and K as the above described rate constant (L/(mol∙min)). Rearranging formula (2) (division by Ctot) yields:

,

1 1

cell cell i

K tC C

= ⋅ + (3)

The decrease of Ccell in time is fitted in formula (3) using linear regression resulting in a K value (slope) and 1/Ccell,i value (abscissa). The obtained K value is valid for the initial mono‐ and polysilicate concentrations for that particulate experiment. To compare the obtained K value with the Kd value for the given pH, temperature and medium composition (25.2 L/(mol∙min)) the obtained K value has to be divided by αM and αP to yield the experimental value for Kd.

4.3 Results 4.3.1 Determination of experimental conditions and non‐biological components in silicate‐cell interaction

To determine whether silicate accumulation in yeast cells (i.e. in the cell or on the cell wall) takes place, and to choose the experimental conditions for the other experiments the silicate accumulation is followed in time at two silicate concentrations: 1 μmol/L (i.e. no‐Si added, no silicate polymerization) and 5 mmol/L (high concentration, silicate polymerization) (Fig. 1). Silicate accumulation by yeast cells is observed in 1 μmol/L as well as in 5 mmol/L silicate. In 1 μmol/L the silicate accumulation is linear in time over the time course studied, whereas in 5 mmol/L Si the silicate accumulation has the form of a saturation curve. Possibly the accumulation in 1 μmol/L silicate also has the form of a saturation curve at longer time intervals, but due to the decay rate of the 31Si tracer it was not possible to investigate this any further. At short incubation times (< 30 minutes) the errors in the measurements are large, so longer incubation times are preferred. On the other hand it is not possible to take the incubation time too long, because of the decay rate of the 31Si tracer. Based on these results 120 minutes incubation is chosen for all further experiments. At this time the accumulation reaches its saturation point in 5 mmol/L.


55

Figure 1: Silicate accumulation (in mol per g dry weight) in time by Baker’s yeast (Saccharomyces cerevisiae). The cells were cultured in A) 1 μmol/L silicate, and in B) 5 mmol/L silicate, and the silicate accumulation in the cells was followed in time using a 31Si‐silicate tracer. Note that the y‐axis in A) runs from ‐5∙10‐11 to 3∙10‐10. All datapoints are averages (n ≥ 3, ± sd). To investigate whether biology plays a role, the Si‐accumulation is determined in dead cells and at 0 oC (living cells). Under these conditions all biological processes come to a complete standstill, so any observed accumulation can be ascribed to a non‐biological process. Dead cells were obtained by γ‐irradiation of a culture with a 10 kGy dose followed by microscopic examination (see Materials and Methods section). The sufficient dose was determined by irradiation of a culture followed by the determination of the growth rate (Fig. 2) and by microscopic examination (methylene blue staining).

Figure 2: Influence of the γ‐dose on the growth rate of Baker’s yeast. A yeast culture was irradiated with different γ‐radiation doses from a 60Co source. Samples were taken to determine the growth rate and the viability of the irradiated cells. A 10 kGy dose completely stopped the growth while microscopic examination revealed that the dead cells were still intact. The size of the irradiated cells (10 kGy) was increased about 1.5 times in diameter. The accumulation in dead cells (30 oC) and in living cells (0 oC) is investigated in 1 μmol/L silicate and compared to accumulation in living cells at 30 oC (standard medium and conditions), 120 minutes incubation. Because dead cells showed an increase in diameter the silicate accumulation in dead cells can also provide some information on the influence of the cell size on the process. The silicate


56

concentration of 1 μmol/L ensures that no silicate polymerization takes place. The experiment is repeated in 5 mmol/L silicate to investigate for possible additional effects of silicate polymerization in the medium (Fig. 3). Silicate accumulation is observed in dead cells and at 0 oC. Lowering the temperature to 0 oC does not influence silicate accumulation in 1 μmol/L silicate, which is an indication that the silicate interaction with the cell is likely of non‐biological origin. In 5 mmol/L silicate the silicate accumulation is reduced with 73 %. This result may indicate that both biological accumulation and silicate chemistry should also be taken into account at this silicate concentration.

Figure 3: Silicate accumulation in Baker’s yeast cultured in A) 1 μmol/L silicate, and in B) 5 mmol/L silicate. Yeast is cultured for 2 hours at 30 oC and at 0 oC (living cells), and at 30 oC (dead cells). The silicate accumulation in the cells is determined with the use of a 31Si‐silicate tracer. All datapoints are averages (n ≥ 3, ± sd). The accumulation in dead cells is increased compared to normal living cells (147% in 1 μmol/L Si, and 48% in 5 mmol/L Si). It can be calculated that the increase of the silicate accumulation in dead cells corresponds reasonably well with the increase of the cell surface area (increase of cell diameter by a factor 1.2‐1.7). Although based on these results the existence of a biological process can not be ruled out the existence of a chemical interaction of silicate with the cell becomes more likely. This is further investigated in the next experiment. 4.3.2 Determination of silicate adsorption on the cell wall

The above described results indicate that the interaction of silicate with the cell could take place by a chemical mechanism. Moreover the silicate accumulation in dead cells indicates that the surface area of the cell possibly plays a role in the interaction of silicate with the cell. In this experiment it is investigated whether silicate accumulation in the cell can be described by an adsorption mechanism. Yeast cells are incubated for 120 minutes at different silicate concentrations, and the silicate accumulation is determined (Fig. 4).


57

Figure 4: Silicate accumulation in yeast cells at different silicate concentrations. Yeast is cultured in normal media for 2 hours and the silicate accumulation in the cells is determined with the use of a 31Si‐silicate tracer. All datapoints are averages (n=3‐5 ± sd). Note that the silicate concentration is in mol/L and the adsorption in mol/gDW. The lines represent the Freundlich adsorption isotherms for the concentration regimes Si < 2 mmol/L and Si > 2 mmol/L The silicate accumulation falls apart in two regimes. Under 2 mmol/L silicate (“low Si”) the accumulation increases linear with the concentration, while above 2 mmol/L (“high Si”) the accumulation increases dramatically. It is remarkable that the transition between the two regimes at 2 mmol/L silicate coincides with the concentration orthosilicate starts to polymerize.7 This makes it plausible that a chemical mechanism plays a role in silicate accumulation. The linear response in low Si media supports the assumption of an adsorption mechanism, and makes a biological uptake mechanism unlikely. Subsequently it is investigated if the accumulation behaviour can be described by an adsorption mechanism. A linear adsorption isotherm can be well described by a Freundlich isotherm (formula 1, Materials and Methods section). The adsorption characteristics of the Freundlich adsorption are determined by the slope of the line (1/n) and the abscissa at 1 mol/L (Kads). Both accumulation regimes in Fig. 4 are shown as Freundlich isotherms, and the parameters 1/n and Kads are calculated and summarized in Table 1. Table 1: Values of Freundlich adsorption parameters, as calculated from silicate accumulation in yeast cells at different silicate concentrations

Regime 1/n Kads ‐ mol/gDW

Si < 2 mmol/L 1.01 ± 0.03 1.92 ± 0.06 ∙ 10‐4 Si > 2 mmol/L 4.04 ± 0.46 2.70 ± 0.63∙ 104

The two regimes have different adsorption characteristics as can be seen in Table 1. The value of 1/n (order of the adsorption reaction) equals 1 in low Si media, which means the adsorption is proportional to the silicate concentration. In high Si media the value of 1/n increases to 4. A remarkable increase of several orders of magnitude is observed for the affinity Kads in high Si media. It is obvious the mechanisms differ in Si < 2 mmol/L and Si > 2 mmol/L. The possibility of an adsorption mechanism is further investigated. Cells grown in high Si media should contain a layer of silicate oligomers adsorbed to the cell’s


58

surface if adsorption takes place. When these cells are brought in no‐Si added media the silicate layer on the cell wall should start to depolymerize. When the cells are grown in 31Si containing media this depolymerization process will be visible as a release of 31Si from the cell biomass. Chemical depolymerization follows chemical pseudo 2nd order kinetics,22,24 so the release rate of 31Si from the cell was calculated using chemical depolymerization kinetics.

Figure 5: Typical diagram of silicate release from a yeast cell suspension, 1/Ccell is shown against time. The cells were cultured for 4 hours in 10 mmol/L Si in the presence of a 31Si‐silicate tracer, harvested, and put in no Si‐added medium. At different time intervals samples were taken and the silicate concentration in the supernatant of the medium Cmed was determined to calculate Ccell. The slope yields the depolymerization rate constant of polysilicate adsorbed on the cell wall (see Materials and Methods). Cells grown in 10 mmol/L silicate and 31Si were put in no‐Si added media and the silicate release was determined. The data were fitted to pseudo 2nd order kinetics (formula 3, Materials and Methods section), and the depolymerization rate constant Kd was calculated and compared to the theoretical depolymerization rate constant under the given culture conditions. Fig. 5 shows a typical example of silicate release from yeast cells. The silicate release results fitted well in pseudo 2nd order kinetics, and yielded a rate constant of 22.0 ± 9.3 L/(mol∙min) (weighted means, n=3). This value is in agreement (P = 0.48) with the theoretical Kd value of 25.2 L/(mol∙min) belonging to standard medium. From the results it can be concluded that pseudo 2nd order depolymerization kinetics can be applied for silicate release from yeast cells. These findings confirm silicate adsorption on the cell surface. 4.3.3. Interaction of metal ions with adsorbed silicate

Earlier it was found silicate interacts with several metals in the yeast cell.26 In this experiment it is investigated whether interaction of metals can be observed with the adsorbed silicate op the cell wall. The influence of the medium components zinc and magnesium on the Freundlich parameters 1/n and Kads of silicate adsorption was determined. Both zinc and magnesium are able to form a complex with silicate, and zinc has a profound influence on silicate polymerization.7,20‐24 So it can be expected that both zinc and magnesium will influence the affinity Kads, and that zinc will influence the order of the adsorption reaction 1/n in high Si media. The following


59

media are investigated: 0.03 μmol/L, or 0.1 mmol/L Zn2+ (Mg2+ as in standard media) and 0.54 mmol/L Mg2+ (Zn2+ as in standard media), and the results are added to the above described experiment in standard media (16 μmol/L Zn2+, 4.2 mmol/L Mg2+). The cultures were inoculated from a 1 μmol/L silicate (= no‐Si added) preculture, the silicate accumulation was determined at different silicate concentrations, and the Freundlich parameters 1/n (slope) and Kads (abscissa at 1 mol/L Si) were calculated from the results and summarized in Figures. 6 and 7.

Figure 6: Freundlich adsorption parameter 1/n for silicate adsorption on Baker’s yeast. A) dependency of Zn2+ concentration, and B) dependency of the Mg2+ concentration. Yeast was cultured for 2 hours at different silicate concentrations and the silicate accumulation in the cells was determined with use of a 31Si‐silicate tracer. From the results the freundlich parameter 1/n is determined.

In 0.03 μmol/L Zn2+ (high Si media) the amount or size of oligomers was clearly increased and caused clogging of the filter. For this reason the sample volume was lowered from 10 to 1 mL. Fig. 5 shows that 1/n is not affected by the Zn2+ or Mg2+ concentration in low Si medium, and has the value 1. This means that the adsorption of silicate is a first order process which is not influenced by Zn2+ or Mg2+ ions. In high Si media the value of 1/n increases to 3.5‐5.2, and is affected by the Zn2+ concentration but not by the Mg2+ concentration. The affinity Kads (Fig. 6) is influenced by the Zn2+ and Mg2+ concentration. In low Si media Kads increases with increasing Zn2+ or Mg2+ concentration. In high Si media Kads decreases dramatically with increasing Zn2+ concentration, but increases with increasing Mg2+ concentration.


60

Figure 7: Freundlich adsorption parameter Kads for silicate adsorption on Baker’s yeast (in mol/g dry weight). Left: dependency of Zn2+ concentration, and Right: dependency of Mg2+ concentration. Note that in Left Kads for [Si] > 2 mmol/L is displayed on a logarithmic scale. Yeast is cultured for 2 hours at different silicate concentrations and the silicate accumulation in the cells is determined with use of a 31Si‐silicate tracer. From the results the freundlich parameter Kads is determined.

4.4 Discussion Silicate is accumulated by living cells at 30 oC as well as at 0 oC and by dead cells (30 oC) (Figs. 1 and 3). At 0 oC the accumulation is reduced in high Si media. The temperature affects the polymerization rate and the size of silicate oligomers.7,30 The decreased polymerization rate results in smaller oligomers, so accumulation of the same amount of molecules will result in a lower observed 31Si accumulation. The activation energy of silicate polymerization is 60 kJ/mol.30 It can be calculated that the polymerization rate of silicate decreases with 90 % when the temperature is lowered from 30 to 0 oC, which is in good agreement with the observed decrease in accumulation. The increased accumulation in dead cells can be explained by the increased surface area of the cells. Probably the swelling of the cells makes it possible more adsorption sites can be reached by the silicate molecules in the solution. Furthermore, it is feasible that the dead cell’s membrane becomes permeable allowing orthosilicate and (to a lesser extend) silicate oligmers to diffuse into the cell. The silicate accumulation in yeast shows a linear dependency of the silicate concentration up to 2 mmol/L. Above this concentration the accumulation shows a dramatic increase (Fig. 4). These two regimes correspond with the two regimes in silicate chemistry: < 2 mmol/L (low Si, silicate exists as orthosilicate), and > 2 mmol/L (high Si, silicate starts to polymerize).7 The silicate accumulation in both regimes can be described by an adsorption mechanism, and the Freundlich isotherm fits well for this purpose. The linear response makes a biological uptake mechanism unlikely. The release of silicate from the cell wall follows silicate depolymerization kinetics and supports the assumption of silicate adsorption on the cell wall.


61

The two concentration regimes have different adsorption characteristics as can be seen in Table 1 and Figures 6 and 7. In low Si the value of 1/n equals 1, regardless of Zn2+ and Mg2+ in the medium, but it increases to 3.5‐5.2 in high Si. It is striking the value of 1/n (order of the adsorption reaction) equals 1 in low Si. This means that the adsorption is proportional to the silicate concentration as a first order process (one silicate molecule per adsorption site). In high Si media the order of the adsorption reaction 1/n increases to 3.5‐5.2, which means 3 to 5 silicate molecules per adsorption site. Since 1/n reflects the order of the adsorption process29 it is likely 1/n shows the degree of polymerization. Note that the silicate concentration in the experiments is expressed in mmoles monosilicate, the degree of polymerization is not taken into account. Zn2+ and Mg2+ influence the adsorption process (Figs. 6 and 7). In 0.03 μmol/L Zn2+ (high Si) media the degree of polymerization of silicate was higher than in other media, which was clearly visible during the experiment (clogging of the filter). A high value of 1/n (5.2) was obtained under these circumstances. Zn2+ ions are known to influence the (de)polymerization of silicate.22,24 Mg2+ does not affect the polymerization of silicate, and this explains why 1/n remains constant at different Mg2+ concentrations. The adsorption affinity Kads differ orders of magnitude in low Si and high Si media, which is a clear indication two chemically entirely different processes take place. Obviously the existence of silicate oligomers in high Si results in a much higher silicate adsorption on the cell wall. Kads increases with the Zn2+ and Mg2+ concentration, except for Zn2+ in high Si media. Low Zn2+ strongly enhances Kads in high Si media, which can be explained by the increased polymerization of silicate under these conditions (Fig. 6). The increase of Kads by Mg2+ (low and high Si) and by Zn2+ (low Si) could be explained by the formation of stable metal‐monosilicates (low Si) or Mg‐silicate complexes (high Si). The Zn‐silicate complex in high Si is unstable, and this could also explain the decrease in Kads with increasing Zn. More research is needed on this. It can be concluded that the adsorbed layer of silicate on the cell wall can interact with metal ions, and in doing so can play a role in the interaction of metal ions with the cell itself. Probably the adsorbed silicate has an influence on the cell’s affinity for metal ions which is entirely dependent of chemistry. The results of earlier work26 can be explained by this. The formation of a stable Mg‐silicate complex on the cell wall could cause Mg2+ ions to be “trapped” on the cell wall, reducing the availability of Mg2+ ions for the cell. The increased adsorption of silicate oligomers (in low Zn) and the instability of the Zn‐silicate complex on the other hand could help the cell “harvesting” Zn2+ ions that are easily released again near to the cell membrane. This could cause an increased availability of Zn2+. The observed effects in this study could be explained by this. The adsorbed silicate can also shield several compounds in the cell wall itself. From the above it is expected that silicate can interfere with biological processes by adsorption to yeast cells and subsequent chemical interaction with metal ions. Most likely, similar consequences of Si‐chemistry can also occur in other organisms than yeast. In Bacillus subtilis for example, silicate precipitation on the cell wall was observed previously.31 And although animal cells do not possess a cell wall in general, the animal cell membrane contains many glycolipids and glycoproteins32


62

that could serve as an adsorption site for silicate molecules. Since complex formation between silicate oligomers and several saccharides can occur33‐35 it is likely that silicate can adsorb on the animal cell membrane as well. In higher animals and humans silicate affects metal concentrations in many tissues and several enzymatic activities. Presumably silicate interacts with metal ions in the tissues in a chemical way and hence affects the availability of the metals. Therefore silicon chemistry should be incorporated carefully in future hypotheses and investigations on the roles of silicon in living organisms including humans.

4.5 References 1. Lide LR (ed) (2002) CRC Handbook of Chemistry and Physics, 83rd ed. CRC Press LLC, Boca

Raton Florida. 2. Aston S (ed) (1983) Silicon geochemistry and biogeochemistry. Academic Press Inc., London,

1983. 3. Tréguer P, Nelson DM, van Bennekom AJ, DeMaster DJ, Leynaert A, Quéguiner B (1995) The

silica balance in the world ocean – a reestimate. Science 268:375‐379 DOI 10.1126/science.268.5209.375

4. Conley DJ (1997) Riverine contribution of biogenic silica to the ocean budget. Limnol Oceanogr 42:774‐777

5. Willén E (1991) Planctonic diatoms – An ecological review. Arch Hydrobiol 69(suppl. 89):69‐106 6. Petzold A, Hinz W (1978) Silikatchemie, Einführung in die Grundlagen. VEB Deutsche Verlag für

Grundstoffindustrie, Leipzig (in German) 7. Iler RK (1979) The chemistry of silica. John Wiley & Sons, New York 8. Smith JV, Arnold FP, Parsons I, Lee MR (1999) Biochemical evolution III: Polymerization on

organophilic silica‐rich surfaces, crystal‐chemical modeling, formation of first cells, and geological clues. Proc Natl Acad Sci USA 96:3479 – 3485

9. Epstein E (1999) Silicon. Annu Rev Plant Physiol 50:641‐664 DOI 10.1146/annurev.arplant.50.1.641

10. Neumann D, zur Nieden U (2001) Silicon and heavy metal tolerance in higher plants. Phytochemistry 56:685‐692 DOI 10.1016/S0031‐9422(00)00472‐6

11. Neumann D, de Figueiredo C (2002) A novel mechanism of silicon uptake. Protoplasma 220:59‐67 DOI 10.1007/s00709‐002‐0034‐7

12. Evered D, O'Connor M (eds) (1986) Silicon biochemistry. Wiley, Chechester 13. Simpson TL, Volcani BE (1981) Silicon and siliceous structures in biological systems. Springer

Verlag, New York 14. Birchall JD (1990) The role of silicon in biology. Chem Brit 26:141‐144 15. Seaborn CD, Nielsen FH (2002) Silicon deprivation and arginine and cysteine supplementation

affect bone collagen and bone and plasma trace mineral concentrations in rats. J Trac Elem Exp Med 15:113‐122. DOI 10.1002/jtra.10011

16. Valerio P, Pereira MM, Goes AM, Leite MF (2004) The effect of ionic products from bioactive glass dissolution on osteoblasts proliferation and collagen production. Biomaterials 25:2941‐2948. DOI 10.1016/j.biomaterials.2003.09.086

17. Martin‐Jezequel V, Hildebrand M, Brzezinski MA (2000) Silicon metabolism in diatoms: implications for growth. J Phycol 36:821‐840. DOI 10.1046/j.1529‐8817.2000.00019.x

18. Müller WEG, Krasko A, Le Pennec G, Schröder HC (2003) Biochemistry and cell biology of silica formation in sponges. Microscop Res Tech 62:368‐377 DOI 10.1002/jemt.10402

19. Exley C (1998) Silicon in life: A bioinorganic solution to bioorganic essentiality. J Inorg Biochem 69:139‐144

20. Schwieger W, Heyer W, Wolf F, Berg KH (1987) Zur Synthese von kristallinen Metallsilicathydraten mit Schichtstruktur. Z Anorg Chem 548:204‐216 (in German)

21. Marshall WL, Warakomski JM (1980) Amorphous silica solubilities – II. Effect of aqueous salt solutions. Geochim Cosmochim Acta 44:915‐924


63

22. Dietzel M, Usdowski E (1995) Depolymerization of soluble silicate in dilute aqueous solutions. Colloid Polym Sci 273:590‐597 DOI 10.1007/BF00658690

23. Dietzel M, Böhme G (1997) Adsorption und Stabilität von polymerer Kieselsäure. Chem Erde – Geochem 57:189‐203 (in German)

24. Dietzel M (2000) Dissolution of silicates and the stability of polysilicic acid. Geochim Cosmochim Acta 64:3275‐3281 DOI 10.1016/S0016‐7037(00)00426‐9

25. Walker GM (1998) Yeast physiology and biotechnology. Wiley, Chichester 26. Brasser HJ, Krijger GC, van Meerten TG, Wolterbeek HT (2006) Influence of silicon on cobalt,

zinc, and magnesium in Baker’s yeast Saccharomyces cerevisiae. Biol Trace Elem Res 112:175‐190

27. Brasser HJ, Gürboğa G, Kroon JJ, Kolar ZI, Wolterbeek HT, Volkers KJ, Krijger GC (2006) Preparation of 31Si‐labeled silicate: a radiotracer for silicon studies in biosystems. J Labelled Compd Radiopharm 47:867‐882 DOI 10.1002/jlcr.1096

28. Verduyn C, Postma E, Scheffers A, van Dijken JP (1992) Effect of benzoic acid on the metabolic fluxes in yeast: a continuous‐culture study on the regulation of respiration and alcoholic fermentation. Yeast 8:501‐517

29. Adamson AW, Gast AP (1997) Physical chemistry of surfaces, 6th ed. John Wiley & Sons, Inc., New York

30. Coudurier M, Baudru B, Donnet JB (1971) Étude de la polycondensation de l’acide disilicique. III. – Influence de la concentration et de la température sur la cinétique et le mécanisme de la polycondensation de l’acide disilicique. Relation avec la texture des produits formés. Bull Soc Chim France 9:3161‐3165 (in French)

31. Fein JB, Scott S, Rivera N (2002) The effect of Fe on Si adsorption by Bacillus subtilis cell walls:insight into non‐metabolic bacterial precipitation of silicate minerals. Chem Geol 182:265‐273

32. Stryer L (1981) Biochemistry. W.H. Freeman and Company, San Francisco 33. Kinrade SD, Hamilton RJ, Schach AS, Knight CTG (2001) Aqueous hypervalent silicon

complexes with aliphatic sugar acids. J Chem Soc Dalt T 7:961‐963 34. Kinrade SD, Del Nin JW, Schach AS, Sloan TA, Wilson KL, Knight CTG (1999) Stable five‐ and

six‐coordinated silicate anions in aqueous solution. Science 285:1542‐1545 35. Lambert JB, Lu G, Singer SR, Kolb VM (2004) Silicate complexes of sugars in aqueous solution.

J Am Chem Soc 126:9611‐9625


64

Chapter 5

Systematic compartmental analysis for describing observed 31Si‐labeled silicic acid uptake during diatom valve formation – A mathematical approach

Abstract Valve formation in diatoms requires bulk uptake and transport of silicic acid (Si(OH)4) for silica deposition inside the silica deposition vesicle (SDV). Uptake of silicon via silicon transporters (SITs) with subsequent intracellular transport requires a controlled mechanism to stabilize the high amounts of reactive silicon species to prevent autopolymerization and simultaneously direct these species towards the SDV; together this pathway should meet the pace at which valve formation occurs. In this study silicic acid uptake was studied during valve formation in synchroneously dividing cells of the diatoms Coscinodiscus wailesii, Navicula pelliculosa, N. salinarum, and Pleurosira laevis using 31Si(OH)4. The experimental data were correlated to systematically derived mathematical models for a compartmental analysis of the possible uptake/transport pathways; including those for both SITs‐ and (macro)pinocytosis‐mediated uptake and transport. Our study indicates that the experimental data on silicon uptake during valve formation matched best with the model that describes (macro)pinocytosis‐mediated uptake. This process not only explains observed surge uptake at high demands for silicon, but apparently infers that in diatoms a pathway exists in which SITs apparently are not involved.

Chapter 5: Silicon uptake in diatoms

66

5.1 Introduction Diatoms (Bacillariophycea) are unicellular eukaryotic photosynthesizing algae that are widely abundant in aquatic habitats on Earth. The most distinctive feature of the diatom cell is its frustule, an exoskeleton of amorphous silica that surrounds the cell.1,2 The silicon required for the formation of the solid silica of the different frustule parts (i.e. valves and girdle bands) has to be taken up by the cell and transported to or translocated in the cellular compartment in which the siliceous parts are formed. During diatom growth each daughter cell produces a new hypovalve and one or more girdle bands before cell separation occurs and the division cycle repeats. This new hypovalve forms in a specialized compartment, the silicon deposition vesicle (SDV), and expectedly is exocytosed once it has been completed.1 The molecular mechanisms that drive frustule formation in diatoms have been studied quite extensively over the last decades from both a biological and physico‐chemical direction. So far, it is also clear that silica formation integrates silica chemistry3‐5 involved in the polymerization and condensation of silicic acid (i.e. concentration of the precursors, pH and ionic strength) with molecular biological pathways (e.g. 6, 7) on how structure‐directing and silica precipitating peptides are produced and delivered to the SDV. At present, however, the interplay of biological and chemical aspects in diatom frustule formation is hardly assessed. Special attention has been paid to the uptake and transport mechanisms of silicon (e.g. 8 and references therein) and it is clear that cell physiology determines uptake kinetics, being dependent on cellular and external concentrations of silicic acid and especially growth rate.9,10 Beside the need for silicon for frustule formation – an obligatory process in diatom cell division1,11 – this element also acts as essential trace element in other cellular compartments such as chloroplasts, mitochondria, and microsomes.3,12

The preferred chemical structure for silicon in uptake is silicic acid (Si(OH)4).13 Based on various uptake studies it has been determined that silicon uptake in general meets the criteria of Michaelis Menten kinetics with initially three distinctive mechanisms, depending on availability of silicic acid and physiological status of the cell; these are: i) surge uptake, ii) internally controlled uptake, and iii) externally controlled uptake (9 and references therein). Most recently it was proposed that silicon uptake is controlled by two mechanisms, namely: uptake of silicon at lower silicic acid concentrations mediated by silicon transporters (SITs) and a diffusion‐mediated route at higher silicon concentrations in the environment.8 An important feature of silicic acid in aqueous solution is its ability to polymerize to condensed silica and in particular when concentrations exceed 2 mmol/L.14 The silica polymerization reaction and physico‐chemical properties of the solid silica is affected by addition of surfactants (e.g. organic compounds such as peptides and proteins), temperature, pH, and ionic strength.14,15 It has been demonstrated that in diatoms several of such abiotic parameters (pH, ionic strength) steer the chemical reaction conditions inside SDV.4,5 It is assumed that on one side a decreasing pH in


67

the SDV upon valve maturation enhances finalization of the silica polymerization reaction and simultaneously prevents the silica to dissolve.4 Ionic strength determines the degree of condensation of the silica and affects the nanostructure of the solid silica formed.5 Both processes occur inside the SDV and possibly contribute to overcome the necessity to utilize cellular transport routes during which unwanted polymerization of silicon pools could occur.3,5 The exact localization of silicic acid or low molecular weight silica precursors is under debate; at present intracellular silicon pools have not been detected except for the SDV and the random distribution of very low silicon levels inside the cell.3,12

The absence of distinctive intracellular silicon pools seems contradictive to the presence of SITs, of which several have been identified and characterized for diatoms.16‐19 Despite the need for high silicon concentration inside the SDV during the rapid process of valve formation,20,21 it is striking that SIT genes become specifically expressed following this major silicification event.22 Therefore, it is not easy to ascribe uptake of sufficient amounts of silicon to SITs only. Recent 29Si‐NMR investigations revealed that silicon in the diatom cell (excluding the silica from the frustule) is present in a precondensed silica sol‐like form,23 while during valve formation organo‐silicon interactions occur.21 However, the spatio‐temporal localization of these precondensed organo‐silica species again remained elusive. Based on studies of high membrane activity in dividing diatom cells24 and the importance of ionic strength on the biosilica nanostructure,5 an alternative silicon uptake pathway was proposed, namely (macro)pinocytosis eventually combined with use of transport vesicles. Such a route could account for the quick uptake of external fluid to accommodate the need for silicon during valve formation, which may well have been recognized as diffusion mediated uptake if external silicon concentrations are sufficient.7 It is clear that a threshold concentration is a prerequisite to initiate valve formation to complete cytokinesis.25 Noteworthy is that in plants a quite similar silicon uptake process has been ascribed to pinocytosis.26 Fluorescent probing of diatoms also indicated that at the very early stages of valve formation (less than 10 min) a quite large weakly fluorescent intracellular vesicle is formed20,27 that in the following stages seem to be compressed towards the cleavage furrow and simultaneously fluoresces brighter due to concentration and condensation of the silica. To facilitate the rapid 2‐D valve completion,20,21 flattening and possibly simultaneous molding of the SDV could well explain the high membrane activity during cell division.24

The aim of the present study is to further investigate the kinetic properties of silicic acid uptake during valve formation for the two aforementioned pathways, i.e. i) SITs‐mediated silicon uptake and transport and ii) pinocytosis‐mediated silicon uptake. To achieve this, silicic acid uptake measurements were combined with a so‐called compartmental analysis. This method is a mathematical approach to obtain information about the fate of a compound in a closed biological or chemical system.28 Based on physiological knowledge and/or speculation the studied system is divided into well‐defined compartments such as pools of a compound in a cell organelle, a vesicle, or different chemical forms of the compound (e.g. monomer vs. polymer). The temporal behaviour of the compound in each compartment is


68

mathematically described based on first order transport kinetics. Experimental data (e.g. silicon uptake) are used to calculate the transport rate constants in the model. By formulating and testing multiple models the best fitting one can be assessed statistically for its significance.28 This compartmental analysis approach has been successfully used to describe the accumulation of technetium in plants,29 mass transport in solid‐liquid systems ,30 and the speciation of calcium in milk.31

In order to perform a compartmental analysis, using parameters for silicon uptake during the rapid 2‐D valve formation, a reliable sensitive technique is required. Silicic acid uptake in diatoms has been studied using three types of radiotracers, being 31Si (e.g. 12), 32Si (e.g. 32) and 68Ge (e.g. 33, 8 and references therein). Of these, 68Ge is a suitable analog for silicon and has been used to mimic silicon uptake kinetics in silicon limited cells; on its own, however, germanium is certainly not preferred13 and even is toxic at higher concentrations. Both 31Si and 32Si have the advantage over 68Ge that they are chemically identical to natural silicon, although these nuclides differ enormously in stability (t½ 31Si = 2.62 h vs t½ 32Si = 160 yrs) and β radiation energy. The short half life time of 31Si allows an experimental time frame that suffices in studying diatom valve formation; in addition, it decays within 48 hours without radioactive waste.34 Fresh 31Si can be readily produced without the presence of natural (non radioactive) silicic acid35 and it decays to the stable isotope 31P, whereas 32Si decays to the radioactive isotope 32P.34 For relative short‐term processes such as initial valve formation and 2‐D valve completion – a period of less than 60 min20,21 – and a preferred high specific activity (and thus low detection limit) freshly prepared 31Si‐silicic acid was used in determination of silicic acid uptake constants for the subsequent compartmental analysis.

5.2 Materials and Methods 5.2.1 Chemicals for radioactivity studies.

All chemicals for preparation and analysis of the radiotracer were at least of analytical grade and obtained from Aldrich (Zwijndrecht, The Netherlands). All solutions were prepared in ultrapure Milli‐Q grade water (18.2 MΩ/cm; Millipore, Billerica, MA, USA). For a sensitive analysis of silicon uptake, the silicon isotope 31Si as silicic acid (t1/2 2.62 h, β‐ 1.49 MeV) was produced in the nuclear reactor of the Reactor Institute Delft, Delft University of Technology, The Netherlands. No‐carrier‐added (i.e. without the presence of natural Si) 31Si solution (specific activity 4.8 TBq/g) was prepared followed by purification of 31Si‐silicic acid with a barium carbonate‐induced chemical precipitation reaction.35 This tracer was successfully used earlier in silicic acid uptake studies in Baker’s yeast.36

5.2.2 Chemical analyses of medium and cells

Inductively coupled plasma optical emission spectrometry (ICP‐OES, OES Optima 4300DV, Perkin Elmer, Boston, USA) was used to determine Si concentrations in the culture medium. For calibration, Merck CertiPUR standard solution 1703 (Amsterdam, The Netherlands) was used. Cells were directly removed from every


69

sample by mild centrifugation (100 rpm, 10 min, Jouan CR4‐11) and ICP‐OES measurements were carried out immediately on the cell free supernatants. The activity of 31Si was measured on a LKB liquid scintillation counter (Tri‐Carb 2750TR/LL, Packard, Meriden, USA) for determining the β‐ radiation, using liquid scintillation counting (LSC) cocktail (Ultima Gold XR™, Perkin Elmer, Boston, USA). 5.2.3 Organisms and culture conditions

The experiments were carried out on the following diatom species: the pennate diatoms Pleurosira laevis (FDCC L1451), Navicula salinarum (FDCC L1262), both from the Fresh Water Diatom Culture Collection Loras College, Navicula pelliculosa (CCMP 543), and the centric diatom Coscinodiscus wailesii (CCMP 2513), both from the Provasoli‐Guillard National Centre for Culture of Marine Phytoplankton, Bigelow Laboratory for Ocean Sciences. The cells were exponentially grown in 500 mL batch cultures in 2 L erlenmeyer flasks using artificial seawater.37 Permissive salinities, in practical salinity units (PSU), were applied and for two species (P. laevis and N. salinarum) were lowered to induce formation more robust valves of condensed less hydrated silica (5) without affecting growth rates. The following salinities were used: 9 PSU for P. laevis, 20 PSU for N. salinarum, and 33.7 PSU for both N. pelliculosa and C. wailesii. Silicic acid concentrations under standard culture conditions were 400 μmol/L for both P. laevis and C. wailesii, 200 μmol/L for N. salinarum, and 100 μmol/L for N. pelliculosa. Temperature (16 ºC) and light conditions (35 +/‐ 5 μmol photons∙m‐2∙s‐1 under a 16: 8 hrs day/night cycle) were constant during growth. In order to determine the uptake of silicic acid during the process of valve formation, the cells were incubated for at least 48 hrs in Si‐free medium to induce immediate synchroneous valve formation when silicon is replenished.11,20,21

5.2.4 Experimental procedures

For analysis of 31Si uptake in diatom cells 0.1‐0.5 ml freshly produced 31Si tracer solution was divided over polypropylene experimental tubes to sterile growth medium and silicic acid stock solution to a final volume of 0.5‐1 mL. The solution was well mixed and allowed to equilibrate for at least 15 minutes (at 16 oC) prior to replenishing Si‐synchronized cells with labeled silicon. To initiate synchroneous valve formation aliquots of 10‐40 mL Si‐depleted culture were brought into the culture tubes to establish a range of conditions with different silicic acid concentrations. The kinetic parameters were determined at standard culture conditions for synchroneously dividing cells to which the aforementioned species‐specific silicon concentrations were added. Additionally, these parameters were determined for synchroneously dividing P. laevis cells, being exposed to medium concentrations of 3, 10, 20, 50, 100 and 200 μmol/L. During the experiments the culture tubes were regularly mixed gently to ensure homogeneous culture conditions. Just prior (t = 0*) and following repletion of silicon, 500 μL samples were taken over a period of ~ 5‐6 hrs. (i.e. at t = 0, 2, 5, 10, 30, 60, 150, 240 and 360 min). Three subsamples for cell counting (1 mL), silicon determination (2‐5 mL), and total 31Si (0.1‐1 μL) were taken and immediately processed for further analysis; a laminar flow cabinet was used to prevent


70

contamination by dust and to ensure clean semi‐sterile conditions during silicon replenishment and sampling. The cell density of N. pelliculosa and N. salinarum was determined by cell counting by use of a Bürker Türk counting chamber. The cells of P. laevis and C. wailesii were too large for this type of counting chamber, the cell density was determined by pipetting 10 μL culture liquid on a microscope slide and counting all cells which were present. The other samples were filtered over a 0.45 µm membrane filter (ME25, Whatman, Dessel, Germany, 25 mm diameter) and washed with 2.5 mL sterile medium, containing 5 mmol/L silicic acid. The activity of the 31Si‐tracer of the sample (including the filter) was determined. Prior to analysis, the sampled filters containing radioactive samples (including the filters) were dissolved in 10.0 ml LSC cocktail within a LSC glass counting vial (20 mL). The volume of liquid samples was adjusted to 5.0 ml with MilliQ grade water and mixed with 15.0 ml LSC cocktail prior to analysis. In addition, 1.0 ml of the culture suspension was also used for determining the total 31Si‐tracer activity. A correction was applied for silicic acid adsorption on the filter. The measured activity was related to the cell density and silicic acid concentration of the cultures in order to calculate the uptake of silicic acid per cell in mol cell‐1. These results were used to calculate the initial uptake rate in mol cell‐1 min‐1 by linear regression on the silicic acid uptake per cell in time over the first 30 minutes. 5.2.5 Mathematical approaches

Compartmental analysis was performed with MicroMathScientist® software for time resolved uptake of silicic acid in diatom cells. The models that were studied in detail were based on mass transport between different compartments within a closed system. In line with Shipley28 the transport of a compound from one compartment to another is regarded as a first order kinetics process in two directions (i.e. forwards and backwards). The different transport processes between the compartments are described by differential equations for each compartment. All equations are combined to form a model (Table 1) that describes the entire transport process of the compound through the whole system. A strict focus was kept on the process of valve formation in diatoms and the formulation of the models was based on the current knowledge of the expected compartments involved for either: i) an internal transport route of silicic acid and/or silica precursors following uptake of silicic acid via SITs,8 ii) a macropinocytosis‐mediated silicon uptake eventually combined with transport vesicle mediated transport.5 Silicic acid uptake and transport by means of SITs was described as a 4 compartment system with the medium, the cytoplasm, the SDV and finally the solid silica of the valve itself as the important compartments for silicic acid transport for valve formation (Table 1, Model F). In addition two alternative models describing silicic acid uptake and transport by (macro)pinocytosis and transport vesicle mediated transport5 were formulated (Table 1, Models G and H). Model G describes the uptake of medium by pinocytosis and subsequent transport to the SDV by transport vesicles with the following compartments: the medium, the transport vesicles, the SDV and the silica of the growing valve (this model is denoted as the transport vesicle‐mediated uptake mechanism). In model H macropinocytosis for the formation of a larger vesicle is described, this vesicle is considered to be the initial SDV. This model lacks the transport vesicle compartment of the previous


71

model G and consists of the medium, the SDV and the valve silica (denoted as the macropinocytosis‐mediated uptake mechanism). Table 1: The defined models based on current knowledge the presence of distinctive compartments model No. of

compartments Compartments Remarks

in cell in total Models for determination of number of compartments

A 1 2 Medium and cell the cell is a black box B 2 3 Medium, 2 cell compartments unidirectional flux to last compartment C 2 3 Medium, 3 cell compartments unidirectional flux to last compartment D 3 4 Medium, 2 cell compartments bidirectional flux to last compartment E 3 4 Medium, 3 cell compartments bidirectional flux to last compartment

Models based on physiological knowledge and speculation

F 3 4 Medium, cytoplasm, SDV, silica

SIT‐mediated transport

G 3 4 Medium, transport vesicles, SDV, silica

transport vesicle‐mediated transport

H 2 3 Medium, SDV, silica macropinocytosis mediated transport

To determine whether all compartments play a role in silicic acid uptake and transport five additional models (A–E) were formulated with different numbers of cell compartments involved, varying from 2 to 4 compartments (table 1). Due to the rapid valve formation20,21 and because of acidification of the SDV upon valve maturation4 it is likely that silicon efflux from the valve does not occur and all silicic acid and/or silica precursors will polymerize to solid silica in the final compartment, the siliceous valve; consequently, an unidirectional flux to the last compartment was is implemented in models B and C. The detailed mathematical description of each model is provided in the supplements. The outcome of the models was fitted to the experimental data for silicon uptake using MicroMath Scientist® software and the residual sum of squares was determined. Also simulations were performed with this software. A statistical F‐test was carried out on closely related models and the P value of this comparison was calculated; if this P value was < 0.05 the most complicated model was applied, whereas in other cases the simpler model was preferred. After fitting the models to experimental data followed by simulation the models yield values for transport rate constants and silicon contents in the separate compartments. If the volume of a compartment is known the silicon concentration inside that particular compartment can be calculated ( /i i iC N V= ). It is also possible to calculate the compartment volume that is required to reach a certain concentration in that compartment ( /i i iV N C= ). If the volume of the entire cell is known the percentage of cell volume that is occupied by the compartment can be calculated.


72

5.3 Results 5.3.1 Silicic acid uptake

Silicic uptake (in mol cell‐1) was determined in the course of valve formation for C. wailesii, N. pelliculosa, N. salinarum and P. laevis, for which 31Si was provided to silicon starved cells (Fig 1). The uptake was examined over a fixed time frame (up to 6 hrs) for synchroneously dividing cells under standard culture conditions. For P. laevis silicic acid uptake was determined at two different replenishment concentrations of 31Si‐labelled silicic acid (Fig. 1 A).

Figure 1: Silicic acid uptake (in mol cell‐1) under standard culture conditions by synchroneously growing diatoms cells in the first stages of valve formation (4‐6 hrs) following silicic acid addition to Si‐limited cells. A) Silicic acid uptake regimes for Pleurosira laevis cells for replenishment with either 456 (■) or 491 (○) μmol/L silicic acid. Two distinct uptake regimes can be recognized by the slopes of the curves (lines) in the first hour and beyond 2 hrs. B) Silicic acid uptake by Coscinodiscus wailesii (456 μmol/L silicic acid), C) Navicula pelliculosa (140 μmol/L silicic acid), and D) Navicula salinarum (317 μmol/L silicic acid). Similar datasets were obtained for P. laevis for other silicic acid medium concentrations (not shown).


73

Emphasis was on the rapid initial stage of valve formation, because 2‐D completion of the valve already occurs within 2 hrs.20,21 For every species two distinct uptake regimes could be identified in the overall time frame that was analysed. It was clear that rapid uptake of silicon occurred in the first hour of valve formation followed by a steadily decreasing uptake rate in the following 4 hrs (Fig. 1). The occurrence of two distinct uptake regimes confirmed that more than one compartment in the cell is involved in the silicification process.28 For every species the collected datasets were used for further mathematical evaluation of the uptake mechanism in view of the defined models. For P. laevis several datasets of time‐dependent uptake were collected for different silicic acid concentrations. The initial uptake rates (in mol cell‐1 min‐1) which were calculated from these datasets, increased with the silicic acid concentration in the medium, reaching a maximal value of 1.7∙10‐14 mol cell‐1 min‐1 at 200‐300 μmol/L (Fig. 2).

Figure 2: Initial silicic acid uptake rate (in mol (cell∙min)‐1) of Pleurosira laevis during the first 30 minutes of valve formation for different silicic acid concentrations in the medium, calculated from the silicic acid uptake in time (in mol cell‐1). At higher silicic acid concentrations the initial uptake rate declined steeply. From these initial uptake rates it was determined that the diffusion of silicic acid from the bulk liquid through the water layer around the cell – the so‐called Nernst layer38 – into the cell itself was not rate limiting (see S1 for a detailed explanation). Consequently, this diffusion process formed no barrier and was excluded from the models. 5.3.2 Mathematically defined number of observable compartments in P. laevis.

Based on diatom cell physiology and current knowledge of valve morphogenesis it was concluded that involvement of at least three cellular compartments would be sufficient to describe silicic acid uptake and transport for valve formation by means of SITs (Table 1, model F), especially in view of the demand for high concentrations of silicic acid for a unidirectional transformation of silicic acid into solid silica. If three cellular compartments were involved also three distinct uptake regimes were expected.28 Three uptake regimes, however, cannot be clearly distinguished in the determined silicic acid uptake patterns, but could be present. To assess this better, a


74

more detailed mathematical analysis of the obtained data was required. In order to determine how many cell compartments could be discerned mathematically, the results of the uptake experiments for P. laevis were subjected to a comprehensive compartmental analysis. For this, five different models were formulated each including a different number of cell compartments (Table 1, models A‐E). These models were mathematically fitted (see also supplement S2) against the acquired experimental data for silicic acid uptake to assign the best fitting one. The magnitude of the transport rate constants of the best fitting model then was determined. Multiple trials (n=6) for fitting the models indicated that a model with just two cellular compartments and medium (model B) always resulted in the best fit (P<0.05). Although it was assumed that efflux from the final compartment (the solid silica) did not occur during rapid valve formation20,21 and as a consequence of the acidification of the SDV,4 the potential efflux from the final compartment was incorporated (Table 1, models D and E). Fitting results clearly confirmed our assumption, because the similar models without efflux from the final compartment strongly outperformed the ones that included efflux. Moreover, it appeared that there was no reverse flux from the final compartment in either model D or E, because the value of the corresponding transport rate constant was (nearly) zero. This strongly suggested that all silicic acid and/or silica precursors were used in valve construction and remained present as solid silica. The results assigned model B performing as the best and consequently three compartments were determined to be involved in the process of silicon uptake during valve formation. These expectedly were comprised of the medium and two cellular compartments, which as such well coincided with the observed two uptake regimes (Fig. 1). The number of observed uptake regimes, however, did not give a conclusive answer about the validity of the SITs‐mediated uptake mechanism (with three cellular compartments involved), because only two uptake regimes were actually observed. Since biological processes proceed relatively fast it may have been possible that one of the compartments was not observed distinctively in the experimental set up, being a so‐called “invisible” compartment. In compartmental analyses such a phenomenon is known and explained as following: i) one compartment reaches the state of equilibrium (i.e. becomes saturated) before the first measurement and the concentration of the target component inside that specific compartment does not alter during the experiment, ii) one compartment is very small compared to the others and for this reason cannot be distinguished from the others, or iii) the compartment is not involved in the whole transport process .28 To test the validity of the SIT‐mechanism it was of importance to determine whether invisible compartments could occur in this silicon uptake mechanism. 5.3.3 Comparative mathematical assessment of model B and the SIT‐mediated silicic acid uptake model in P.laevis

Our compartmental analysis of silicic acid uptake during valve formation revealed that only two cellular compartments and the medium could be mathematically assigned (Table 1, model B), whereas three cellular compartments were expected in a SIT‐mediated uptake mechanism. In principle it would be possible that one compartment has been ‘invisible’ in our analysis of the silicic uptake regimes, as a


75

result of a very fast uptake kinetics in the first 1‐2 min after silicon replenishment, or a result of its small size or very low silicon content. In any case an invisible compartment would also have to display fast kinetics, due to the high amount of silicon that has to be transported to the SDV during valve formation. Still, if this would have occurred, it was not possible to directly correlate model B (Table 1) mathematically to the one for SITs‐mediated silicon uptake (Table 1, model F) since more rate constants are involved and because the rate constant k3 of model B cannot be directly assigned to one of the constants in the model F. In order to assess whether a cellular compartment could not be observed and to investigate how silicon uptake via model B could relate to SITs‐mediated uptake, the algorithms for these mechanisms were subjected to a closer investigation (in particular equations 2.12‐2.15 of model F in supplement S2). Firstly, the magnitudes of the parameters and constants in model B were determined, whereas the silicic acid concentration in the medium (Cm) and the cell density (Cx) in these equations were chosen equal to those in the experiments. The fitting and simulations of model B provided the order of magnitude of the transport rate constants (kup, kef and k3) and of the silicon contents (N1 and N2) in all compartments (Table 2). Table 2: magnitudes of the experimental and kinetic parameters and variables for P. laevis

Parameter / variable symbol min.‐max. value unity Si conc. medium Cm 10‐6 ‐ 10‐3 mol/L cell density Cx 107 ‐ 108 cell/L uptake rate constant kup 10‐11 ‐ 10‐10 L/cell∙min

efflux rate constant kef 0 ‐ 10‐2 min‐1 third model B rate constant k3 0 ‐ 10‐2 min‐1

Si contents in all compartments N1,2 0 ‐ 10‐12 mol/cell The experiments were carried out on silicon starved cells for which the silicic acid concentration was assumed to be zero prior to replenishment of silicic acid. Under this condition it was expected that at the start of the experiment silicic acid was not present in any cellular compartment that is involved in valve morphogenesis. In view of this it is obligatory that all differential equations display a positive value at the beginning of the experiment and this limits the combinations of the transport rate values. All determined parameters (Table 2) were substituted in all applicable equations (i.e. eq. 2.12‐2.15, supplement S2), with varying values for either kcs, ksc and ksv to determine which rate constant values yielded positive values. This occurred when 0.01 < kcs < 0.1 min‐1, ksc = 0.01 min‐1, and ksv = 0.01 min‐1. All other combinations yielded a negative value for one or more differential equations and thus were rejected. Because the saturation time is characteristic for each compartment, it can be used to assign any “invisible” compartment. The saturation time was calculated from the values of dNx/dt (see eq. 2.12‐2.15, supplement S2) and the maximal silicon contents of the compartments (Ni, Table 3) via:

, max /char satdNt Ndt

= (1)

The characteristic saturation time (tchar,sat) of all compartments involved displayed an order of magnitude of at least 100 minutes, indicating that the silicic acid content


76

in the applicable compartments was still changing during valve formation and expectedly would be observed. Therewith, it was unlikely that the model for SITs‐mediated uptake contained any “invisible” compartment and that this mechanism could not accommodate massive silicon transport for silica deposition during valve formation. In other words, the SITs‐mediated silicon uptake mechanism did not agree with the observed rapid 2‐D formation of the valve and alternative routes had to be assessed. Table 3: Calculated magnitudes of minimum and maximum rates dN/dt (mol/(cell∙min)) and dVt/dt (L/(cell∙min)) and characteristic saturation times for models F and G for P. laevis.

compartment rate, eq. nr. min.‐max value char. saturation time (min)

Model F: SIT‐mediated transport cytoplasm dNc/dt, 2.13 0 ‐ 10‐14 > 100 SDV dNs/dt, 2.14 0 ‐ 10‐14 > 100 valve dNv/dt, 2.15 0 ‐ 10‐14 > 100 Model G: transport vesicle‐mediated transport tr. vesicles dVt/dt, 2.20 10‐11 ‐ 10‐10 0.1 ‐ 1 SDV dNs/dt, 2.18 0 ‐ 10‐14 > 100 valve dNv/dt, 2.19 0 ‐ 10‐14 > 100 Model H: macropinocytosis‐mediated transport SDV dNs/dt, 2.21 0 ‐ 10‐14 > 100 valve dNv/dt, 2.22 0 ‐ 10‐14 > 100

5.3.4 Mathematical assessment of alternative silicic uptake models in P. laevis.

The outcome of our analysis so far was not satisfactory and initiated us to define alternative pathways. In view of the demand for high silicic acid concentrations for the unidirectional transformation into solid silica during valve formation the following mechanisms were selected: i) transport vesicle‐mediated uptake and intracellular transport of silicic acid towards the SDV, containing three cellular compartments (Table 1, model G) and ii) (macro)pinocytosis‐mediated uptake of silicic acid from the medium, containing two cellular compartments (Table 1, model H) and where the initial vesicle transforms to become the SDV. Since the model for transport vesicle‐mediated silicon uptake contains three cellular compartments, this mechanism only applies when it contains an invisible compartment to match with the observed dual uptake regime (Fig. 1). Therefore, the number of observable compartments in the model was assessed similarly as just explained by varying the values for kts, ksv and Vt. The actual total volume of the transport vesicles (Vt) in this model was unknown, but this volume logically never could exceed the total cell volume, being estimated at 7.4∙10‐10 L/cell.39 The values of the parameters were substituted in equations 2.16‐2.19 (supplement S2) to calculate the order of magnitude of the transport rates (i.e. dN/dt and dVt/dt). A positive value for the differential equations was only obtained for kts > 10 min‐1, ksv = 0.01 min‐1, and Vt < 10‐11 L/cell. The saturation times for the different compartments were calculated from the values of dNx/dt (eq. 2.16‐2.19, supplement S2) and the maximal Si contents of the compartments (Ni) or the maximal total transport vesicle volume (Vt, Table 3) via:


77

, max ,max/ or / tchar sat t

dVdNt N Vdt dt

= (2)

The saturation time of the transport vesicle compartment appeared to be very low (0.1‐1 min). This means that at the applied experimental intervals to determine silicon uptake this rapid saturation could have been missed at the very early stage (i.e. the first 5 min) and at much later stages when 3‐D valve thickening occurred (> 60 min). Mathematically, however, the algorithms for a mechanism based on transport vesicle‐mediated silicon uptake matches only with those of the mechanism including two cellular compartments (Table 1, model B) at high saturation rates of transport vesicle compartments (Supplement S3). This outcome prompted us to further focus on (macro)pinocytosis‐mediated silicon uptake (Table 1, model H). In this mechanism it is hypothesized that a larger vesicle is formed intracellularly, which subsequently is compressed towards the cleavage furrow by means of cellular and membrane activity24,40 while simultaneously water is exported41 so that silicic acid concentration reaches the threshold for autopolymerization to solid silica.14 In this mechanism two distinct cellular compartments (vesicle plus solution and the solid silica of the valve) are involved, matching exactly mathematically with model B (Table 1). Physiologically a transport vesicle‐mediated and the (macro)pinocytosis‐mediated uptake mechanisms are quite well related, but only if saturation of transport vesicles proceeds rapidly similar mathematical principles apply. In diatoms a combination of both eventually may occur to enable the 2‐and 3‐dimensional expansion of the initial SDV,3,5 allowing the cell to better steer the process. 5.3.5 Kinetic parameters of (macro)pinocytosis‐mediated silicic acid uptake in P. laevis.

Our comparative mathematical assessment revealed that silicic acid uptake during valve formation under standard culture conditions could be described by the model that involves (macro)pinocytosis events (model H, Table 1). In fact, this agreed well with various observations such as: i) the nanostructural variation of the biosilica induced by external salinity,5 ii) the fluorescent probing to monitor the rapid 2‐D development of new hypovalves,20,21 and iii) the high membrane activity in dividing diatoms cells,24 in particular when compression of a initial SDV occurs as is suggested by fluorescence characteristics of the probe PDMPO20,27 and activity of aquaporins to export water from the SDV.41 In order to determine whether the model matched well with our experimental data, both were fitted and the kinetic parameters and rate constants were calculated for different silicic acid concentrations in the medium (Fig. 3a‐c). The value of all transport rate constants decreased at silicic acid concentrations above 200 μmol/L, which may imply that diatoms have reached the optimal uptake rate and not necessarily need to acquire more silicic acid; a reasonable explanation when one considers that a surplus of silicic acid would require intracellular storage, that never has been confirmed. The transport rate constants kef and ksv clearly obey the Michaelis Menten kinetics at lower silicic acid concentrations in the medium. However, for uptake of silicic acid (kup) this was less obvious, although Michaelis Menten kinetics could be expected (Table 4). This kup also displayed a high affinity for silicic acid as was deduced from the low Km value (2.0 μmol/L). The rate constant for valve formation (ksv) revealed a much higher Km value (12 μmol/L), whereas the Km value of the efflux rate constant


78

kef even was higher than any of the other rate constant (22 μmol/L). Using the obtained rate constants the concentration of silicon species in the SDV (Ns) and valve (Nv) were calculated (Fig. 4).

Figure 3: Dependency of calculated rate constants in Pleurosira laevis on different silicic acid concentrations in the medium. A) rate constant for uptake of silicic acid (kup) from the medium, B) rate constant for efflux of silicic acid (kef) from cell, and C) rate constant for transformation of silicic acid to solid valve silica (ksv). The lines represents the Michaelis Menten kinetics. Table 4: Michaelis Menten constants of the rate constants in macropinocytosis mediated silicic acid uptake (model H) for P. laevis

Michaelis‐Menten constants rate constant

transport process Km (μmol/L) kmax

kup medium to SDV 2.0 ± 1.7 1.6 ± 0.16 ∙ 10‐10 L∙(cell∙min)‐1 kef efflux to medium 22 ± 10 3.5 ± 0.57 ∙ 10‐2 min‐1 ksv SDV to valve 12 ± 11 3.1 ± 0.73 ∙ 10‐2 min‐1


79

Figure 4: A typical result (MicroMath Scientist® plot) obtained by fitting total silicic acid uptake per cell (■) to model H and by simulation with the rate constants determined for synchroneously dividing cells of Pleurosira laevis at medium concentration of silicic acid of 490 μmol/L. The straight line represents the fitted total silicon in cell, the dashed line Si in SDV, the dotted line the silicon that is deposited in the valve. All concentrations are at the level of 10‐15 mol Si/cell. For all datasets similar results were obtained (not shown). This exercise showed that the silicic acid content in SDV (Ns) reached a maximal value between 1‐2 hrs, well coinciding with the rapid valve formation as determined by fluorescent probing.20,21 The maximal silicic acid content in the SDV (Ns, Fig. 5a) depended on the silicic acid concentration in the medium and appeared to increase from zero to maximal values of in the range of 3.2∙10‐13 mol/cell; the latter occurring when 200‐300 μmol/L silicic acid was present in the medium. At higher silicic acid concentrations in the medium the maximal content of the SDV decreased, agreeing with the decrease in the experimentally determined rate constant values (Fig. 3).

Figure 5: The calculated maximal maximum amount of Si in the SDV (Ns in mol cell‐1) in Pleurosira laevis depending on silicic acid concentration in the medium (A), and the determined cell volume (in %) occupied by SDV (B). For the latter, the volumes where calculated for silicon concentrations inside the SDV at 2 mmol/L.


80

It was investigated whether the amount of silicic acid in the SDV (Ns in mol/cell) was sufficient to initiate silica polymerization, i.e. the silicic acid concentration inside the SDV should reach the required threshold of 2 mmol/L.14 The volume of the SDV correlating with this concentration was calculated. With this information the percentage of cell volume occupied by the SDV could then be estimated (Fig 5B). To maintain the silicon concentration at the saturation level of 2 mmol/L an SDV volume reaching up to 22% of the cell volume (200 μmol/L silicic acid in the medium) is required. Due to the decreased silicic acid uptake at increasing medium concentrations (Fig. 2 and 5A) a lower SDV volume is sufficient (Fig. 5B). Apparently, the calculated silicon concentrations of the SDV (Ns) and its volume dependency for silica deposition yielded plausible conditions to achieve valve formation in P. laevis. Noteworthy is that in probing valve morphogenesis, the fluorescence intensity increases at higher silica concentrations,27 while in parallel the shape of the SDV, and potentially its volume, changes upon compression towards the cleavage furrow.20,21 5.3.6 Compartmental analyses for other diatom species

To confirm whether the approach for P. laevis applied to other species, we performed the same compartmental analysis on C. wailesii, N. pelliculosa, and N. salinarum, using the obtained experimental data on silicic acid uptake. By assessing the number of cell compartments being involved in silicic acid uptake during valve formation, it appeared that the mechanism comprised of a maximum of two cellular compartments (SDV and the solid silica of the valve, Table 1, models B and H) was favoured. Similar to P. laevis, the involved rate constants (kup, kef and ksv), the maximal silicon content in the SDV (Ns), and the volume of the SDV were calculated (Table 5). Table 5: Rate constants and other parameters for N. pelliculosa, N. salinarum and C. wailesii describing silicic acid uptake by macropinocytosis (model H) under standard culture conditions

Organism N. pelliculosa N. salinarum C. wailesii cell volume L/cell 7.86∙10‐15 1.37∙10‐12 2.40∙10‐8 [Si] medium μmol/L 140 317 456 kup L∙(cell∙min)‐1 2.53∙10‐16 2.66∙10‐14 9.42∙10‐10 kef min‐1 0.00739 0.0188 0.0656 ksv min‐1 0.00813 0.0253 0.00565 Ns,max mol/cell 1.58∙10‐18 1.66∙10‐16 5.76∙10‐12 vol SDV (2 mmol/L Si) % cell vol. 10.0 6.05 12.0

Again the uptake rate constant (kup) and the silicic acid content in the SDV (Ns) depended on the cell volume comparable to what has been determined for P. laevis. For the three species the rate constants for efflux (kef) and transport to the valve (ksv) were in the same order of magnitude as was observed in P. laevis (Table 5). The SDV volumes varied somewhat between species and occupied respectively 6.1‐12% of the cell volume ([Si(OH)4] in SDV = 2 mmol/L), which demonstrated that the obtained results are plausible for silica polymerization and thus valve formation. In view of cell physiology, the (macro)pinocytosis‐mediated uptake of silicic acid for valve formation proved to be the best performing one, in particular when soluble


81

silicic acid inside this vesicle, or potentially the premature SDV, concentrates towards saturation levels during compression and 2‐D orientation upon maturation.

5.4 Discussion and Conclusions To our best knowledge a detailed compartmental analysis for studying uptake kinetics of silicic acid during valve formation in diatoms has not been reported. This combined experimental and mathematical approach allowed us to assess the number of compartments involved by correlating uptake parameters of 31Si‐labelled silicic acid to a series of models. Our data revealed that during valve formation uptake mechanisms in which two cellular compartments are involved matched best to the experimentally identified uptake regimes. Irrespective to the presence of SITs16,19,22,42 it was not possible to assign SIT‐mediated silicon uptake and transport as the major mechanism in diatom valve formation. The SITs‐mediated uptake as proposed by Thamatrakoln8 would require intracellular transport and involvement of an additional compartment and only applies when this compartment cannot be discerned due to its rapid processing and saturation. Because SIT expression follows the initial stage of valve formation19 it seems unlikely that they take part in the first 1‐2 hrs when the 2‐D valve structure is deposited.20,21

Based on the models two mechanisms for uptake of silicic acid during valve formation are favoured. One involves a process of pincocytosis combined with rapidly saturating vesicles that form the SDV. The other assigns (macro)pinocytosis‐mediated silicon uptake and subsequent transformation of the vesicle and its content of silicic acid to respectively become the SDV and the solid silica. The latter mechanism agrees with silicon uptake that is described in plants26 and recently has been proposed to explain the effect of ionic strength on the nanostructure of diatom biosilica.5 (Macro)pinocytosis possibly coincides with mechanisms to change the vesicles into the primary SDV that upon maturation quickly becomes compressed towards the cleavage furrow. The observations for a high cellular and membrane activity in dividing diatoms cells24,40 as well as the suggested dehydration of the SDV by aquaporins41 and simultaneous silica concentration as seen in fluorescent probing20,21,27 all support such a SDV molding pathway. This pathway surely would explain a simple cost‐effectively mechanism to efficiently contain and concentrate silicic acid at the proper location, not requiring any intracellular transport and/or stabilization to prevent autopolymerization. This also is in line with observations that diatom silicification proceeds rather independent to other metabolic processes.9,10 Yet, further research is required to confirm this (macro)pinocytosis process. The (macro)pinocytosis‐mediated uptake outperformed the other mechanisms, well matched our experimental data, and could account for uptake of sufficient silicic acid to produce at least the initial 2‐D base of the new hypovalve. The transport rate constants for silicic acid uptake as well as its transport inside the cell obeyed the Michaelis‐Menten kinetics (Fig. 3 and Table 4) and this agreed well with previous observations (8, 9 and references in both). The affinities for the silicic acid uptake and intracellular transport processes clearly differed as was displayed in the Km values of


82

the rate constants. The high affinity for silicic acid of the uptake process – which relates to the cell’s absolute need for silicic acid – was determined from the low Km value of the rate constant kup (2.0 ± 1.7 μmol/L Si; Table 4) and should enable the cell to take up silicic acid very efficiently. The Km value of silicic acid transformation inside the SDV to solid silica of the valve (ksv) was clearly higher (12 ± 11 μmol/L Si), suggesting that a threshold value has to be reached to initiate valve formation. In this respect, one should notice that a threshold of silicic acid controls the expression of DNA polymerase for DNA transcription in diatom cell division25 and of interest is to find out whether these transcription and valve formation thresholds have a concerted mode action to fully control the cell division process in diatoms and to assure that cell division only proceeds when sufficient silicon is available. The efflux of silicic acid from the cells occurred only at very high external silicic acid concentrations and only under these conditions the affinity of the efflux rate constant (kef) was estimated higher than the other rate constants. Elevated efflux rates are expected to ensure that a surplus of silicic acid is not built up in the cell and should be considered a feed back mechanism. Our compartmental analysis suggests that the SDV volume is of importance in view of the increasing internal silicic acid concentration (Fig. 5b and Table 5) and most probably is adjusted or molded by cellular activity during 2‐D development of the new hypovalve. Molding of the SDV has been described (1, 3 and references in both), whereas a high membrane activity is observed during diatom cell division.24 The calculated vesicle volumes to enable silica polymerization to form the new hypovalve are quite reasonable and even quite well correlate with the observed pale fluorescent vesicles that are observed in studying valve morphogenesis.20,21,27 Upon maturation this vesicle, potentially the initial SDV, is directed and flattened towards the cleavage furrow. If simultaneously water is expelled41 the SDV constituents concentrate allowing silicic acid to polymerize when the proper saturation level has been reached.14 This concentration is supported by the intensity change of the compound used in fluorescent probing.27 It should also be noted that the saturation level not only depends on the presence of other ions, but on their concentrations as well. Both steer the polymerization reaction13,14 and could account for the salinity effect on the nanostructure of the biosilica.5 (Macro)pinocytosis‐mediated uptake enables the fast silicon uptake kinetics which has been referred to as surge uptake (8 and references therein) and well explains the high demand for silicic acid to form new hypovalves. With respect to cell biology and physiology diatoms may have evolved separate pathways for uptake and transport of silicic acid allowing the cell to distinctively use one for rapid uptake of larger amounts of silicic acid during valve formation. Another one could be responsible for silicic acid uptake/transport for use of silicon in different processes. Presence of silicon inside cytoplasm and organelles (e.g. mitochondria, microsomes, chloroplasts) as well as its role in controlling DNA replication indicates that silicon is constitutively present.12,25 Since SITs become expressed when valve formation has progressed largely,19 it also seems unlikely that these transporters are involved in valve formation when bulk uptake is essential. In line with observations of Thamatrakoln8 SITs‐mediated silicon transport would account for uptake of silicon at low concentrations and/or small quantities, making it tempting to suggest that SITs are used to control intracellular silicon concentrations


83

at locations other than the SDV. In contrast, (macro)pinocytosis could largely account for silicic acid uptake during valve formation. Further research should indentify the components of the molecular machinery involved to see whether they are specifically activated and controlled prior, during and following valve formation.

5.5 References 1. Pickett‐Heaps JD, Schmid AMM, Edgar LA (1990) The cell biology of diatom valve formation.

Progr Phycol Res 7:1‐168. 2. Round FE, Crawford FM, Mann DG (1990) Diatoms: The biology and morphology of the genera

(Cambridge Univ Press, Cambridge, UK). 3. Gordon R, Drum RW (1994) The chemical basis of diatom morphogenesis. Int Rev Cytol 150: 243‐

272. 4. Vrieling EG, Gieskes WWC, Beelen TPM (1999) Silicon deposition in diatoms: Control by the pH

inside the silicon deposition vesicle. J Phycol 35:548‐559. 5. Vrieling EG, Sun Q, Tian M, Kooyman PJ, Gieskes WWC, van Santen RA, Sommerdijk NAJM

(2007) Salinity‐dependent diatom biosilicification implies an important role of external ionic strength. Proc Natl Acad Sci USA 104:10441‐10446.

6. Lopez PJ, Descles J, Allen AE, Bowler C (2005) Prospects in diatom research. Curr Opin Biotech 16:180‐186.

7. Wong Po Foo C, Huang J, Kaplan DL (2004) Lessons from seashells: silica mineralization via protein templating. Trends biotech 22:577‐585.

8. Thamatrakoln K, Hildebrand M (2008) Silicon uptake in diatoms revisited: A model for saturable and nonsaturable uptake kinetics and the role of silicon transporters. Plant Physl 146:1397‐1407.

9. Martin‐Jézéquel V, Hildebrand M, Brzezinski MA (2000) Silicon metabolism in diatoms: implications for growth. J Phycol 36:821‐840.

10. Claquin P, Martin‐Jezequel V (2002) Uncoupling of silicon compared with carbon and nitrogen metabolisms and the role of the cell cycle in continuous cultures of Thalassiosira pseudonana (Bacillariophyceae) under light, nitrogen, and phosphorus control. J Phycol 38:922‐930

11. Coombs J, Volcani BE (1968) Studies on the biochemistry and fine structure of silica‐shell formation in diatoms. Chemical changes in the wall of Navicula pelliculosa during its formation. Planta 82:280‐292.

12. Mehard CW, Sullivan CW, Azam F, Volcani BE (1974) Role of silicon in diatom metabolism. 4. Subcellular‐localization of silicon and germanium in Nitzschia‐alba and Cylindrotheca‐fusiformis. Phys Plant 30:265‐272.

13. Del Amo Y, Brzezinski MA (1999) The chemical form of dissolved Si taken up by marine diatoms. J Phycol 35:1162–1170.

14. Iler RK (1979) The chemistry of silica (John Wiley & Sons, New York). 15. Brinker CJ, Scherer GW (1990) Sol‐Gel science (Academic, New York). 16. Hildebrand M, Volcani BE, Gassmann W, Schroeder JI (1997) A gene family of silicon

transporters. Nature 385:688‐689. 17. Hildebrand M (2000) Silicic acid transport and its control during cell wall silicification in diatoms.

in Biomineralization; from Biology to Biotechnology and medical Application, ed Baeuerlein E (Wiley‐VCH Verlag GmbH, Weinheim, Germany), pp. 171‐188.

18. Hildebrand M (2003) Biological processing of nanostructured silica in diatoms. Progr Org Coat 47:256‐266.

19. Thamatrakoln K, Hildebrand M (2007) Analysis of Thalassiosira pseudonana silicon transporters indicates distinct regulatory levels and transport activity through the cell cycle. Eukaryotic Cell 6:271‐279.

20. Hazelaar S, van der Strate HJ, Gieskes WWC, Vrieling EG (2005) Monitoring rapid valve formation in the pennate diatom species Navicula salinarum (Bacillariophyceae). J Phycol 41:354‐358.

21. Heredia A, van der Strate HJ, Delgadillo I, Basiuk VA, Vrieling EG (2008) Analysis of organo‐silica interactions during valve formation in synchroneously growing cells of the diatom Navicula pelliculosa. ChemBioChem 9:573‐584.


84

22. Thamatrakoln K, Alverson AJ, Hildebrand M (2006) Comparative sequence analysis of diatom silicon transporters: Toward a mechanistic model of silicon transport. J Phycol 42:822‐834.

23. Gröger C, Sumper M, Brunner E (2008) Silicon uptake and metabolism of the marine diatom Thalassiosira pseudodana: Solid‐state Si‐29 NMR and fluorescence microscopic studies. J Structural Biol 161:55‐63.

24. Kühn S, Brownlee C (2005) Membrane organisation and dynamics in the marine diatom Coscinodiscus wailesii (Bacillariophyceae). Bot Mar 48:297‐305.

25. Okita TW, Volcani BE (1977) The deoxyribonucleic acid polymerases from the diatom Cylindrotheca fusiformis. Partial purification and characterization of four distinct activities. Biochem J 167:601‐610.

26. Neumann D, De Figueirdo C (2002) A novel mechanism of silicon uptake. Protoplasma 220:59‐67. 27. Shimizu K, Del Amo Y, Brzezinski MA, Stucky GD, Morse DE (2001) A novel fluorescent silica

tracer for biological silicification studies. Chem Biol 8:1051‐1060. 28. Shipley RA, Clark RE (1972) Tracer methods for in vivo kinetics – Theory and applications,

(Academic Press, New York, London). 29. Krijger GC, Harms AV, Leen R, Verburg TG, Wolterbeek B (1999) Chemical forms of technetium

in tomato plants; TcO4‐, Tc‐cysteine, Tc‐glutathione and Tc‐proteins. Env Exp Bot 42:69‐81.

30. Kolar ZI, (1990) Tracing of interfacial transport of matter in solid‐liquid systems in dynamic equilibrium. J Trac Micropro Techn 8:103‐120.

31. Kolar ZI, Verburg TG, van Dijk HJM (2002) Three kinetically different inorganic phosphate entities in bovine casein micelles revealed by isotopic exchange method and compartmental analysis. J Inorg Biochem 90:61‐66.

32. Shipe RF, Brzezinski MA (1999) A Study of Si Deposition Synchrony in Rhizosolenia (Bacillariophyceae) Mats Using a Novel 32Si Autoradiographic Method. J Phycol 35:995‐1004.

33. Azam F, Hemmings BB, Volcani BE (1973) Germanium incorporation into silica of diatom cell‐walls. Arch Microbiol 92:11‐20.

34. Firestone BR, Shirly VS (1996) Table of isotopes, 8th ed. (John Wiley & Sons, New York). 35. Brasser HJ, Gürboğa G, Kroon JJ, Kolar ZI, Wolterbeek HT, Volkers KJ, d Krijger GC (2006)

Preparation of 31Si‐labelled silicate: a radiotracer for silicon studies in biosystems. J Label Comp Radiopharm 47:867‐882.

36. Brasser HJ, Krijger GC, Wolterbeek HT (2008) On the beneficial role of silicon to organisms: A case study on the importance of silicon chemistry to metal accumulation in yeast. Biol Trace Elem Res 125:81‐95.

37. Veldhuis MJW, Admiraal W (1987) Influence of phosphate‐depletion on the growth and colony formation of Phaeocystis‐pouchetii. Mar Biol 95:47‐54.

38. van ‘t Riet K, Tramper J (1991) Basic bioreactor design. (Marcel Dekker inc., New York). 39. Snoeijs P, Busse S, Potapova M (2002) The importance of diatom cell size in community analysis.

J Phycol 38:265‐272. 40. Schmid AMM (1986) Wall morphogenesis in Coscinodiscus Wailesii Gran et Angst II: cytoplasmic

events of wall morphogenesis. in Proceedings of the 8th International Diatom Symposium, ed Richard M (Koeltz, Königstein, Germany), pp 293‐314.

41. Grachev MA, Annenkov VV, Likhoshway YV (2008) Silicon nanotechnologies of pigmented heterokonts. Bioessays 30:328‐337.

42. Armbrust EV, Berges JA, Bowler C, Green BR, Martinez D, Putnam NH, Zhou SG, Allen AE, Apt KE, Bechner M, et al. (2004) The genome of the diatom Thalassiosira pseudodana: Ecology, evolution, and metabolism. Science 306:79‐86


85

5.6 Supplement 1: mass transfer over Nernst layer The cell is surrounded by a laminar water layer known as the Nernst layer, and silicic acid that is taken up has to pass this layer. Mass transport inside the Nernst layer is determined by diffusion, so it is of importance to assess whether this process is rate limiting for uptake kinetics. Following van ‘t Riet & Tramper1 the mass transfer flux by diffusion (F) over the boundary layer to the cell (in mol/cell/s) is calculated: ( )water cellF k A C C k A C= ⋅ ⋅ − = ⋅ ⋅Δ (1.1) with mass transfer coefficient k (m∙s‐1), area of the cell A (5∙10‐8 m2), silicic acid concentration (in mol∙m‐3) in the medium Cmedium and in the cell Ccell. The value of k is calculated with the Sherwood number (Sh):

k dShD⋅

= (1.2)

with d as the diameter of the cell, and D as the diffusion coefficient in water (1∙10‐9 m2∙s‐1). The value of k is calculated for the utmost limiting condition; i.e. when convection in the liquid is absent (thus no stirring or mixing). In this case the Sherwood number has the value of 2. This yields a k value:

2 Dkd⋅

= (1.3)

When flux F is calculated with this k value the following transport rates over the Nernst layer were calculated. These values were compared to experimental observed uptake rates (table S1.1). Table S1.1: Silicic acid transport rates over the Nernst layer when medium convection is absent (column 4) compared to experimental observed uptake rates (column 5) in Pleurosira laevis, Navicula pelliculosa, N. salinarum and Coscinodiscus wailesii.

Organism diameter cell

concentration gradient ΔC

diffusion transport rate

observed uptake rate

m μmol/L mol∙(cell∙min)‐1 mol∙(cell∙min)‐1 P. laevis 1∙10‐4 1 6 ∙ 10‐14 3 ∙ 10‐17 10 6 ∙ 10‐13 8 ∙ 10‐16 100 6 ∙ 10‐12 2 ∙ 10‐15 1000 6 ∙ 10‐11 1 ∙ 10‐14 N. pelliculosa 2∙10‐6 100 2 ∙ 10‐13 2 ∙ 10‐20 N. salinarum 1∙10‐5 100 1 ∙ 10‐12 4 ∙ 10‐18 C. wailesii 4∙10‐4 100 1 ∙ 10‐11 4 ∙ 10‐13

It was calculated that the silicic acid flux F per cell is at least 25 times higher for all organisms than the observed uptake rate per cell, so diffusion over the Nernst layer is not rate limiting. Consequently, silicic acid uptake should be considered the active uptake of the cell itself. This means that presence of the Nernst layer does not have to be included in the compartmental analysis for which models for silicic acid uptake were defined. 5.6.1 Reference

1. van ‘t Riet K, Tramper J (1991) Basic bioreactor design. (Marcel Dekker inc., New York).


86

5.7 Supplement 2: mathematical formulation of the models All models are based on first order kinetics mass transport between different compartments within a closed system,1 and are valid for Si‐limited diatom cells that do not contain intracellular silicic acid pools at the moment the uptake experiments were started (t=0). After replenishment the cells start to take up silicic acid and the different compartments of the cell will start to be filled. The whole uptake and transport system expectedly consists of the medium and the different cellular compartments (Table S2.1). Table S2.1: The defined models based on current knowledge the presence of distinctive compartments model No. of

compartments Compartments Remarks

in cell in total Models for determination of number of compartments

A 1 2 Medium and cell the cell is a black box B 2 3 Medium, 2 cell compartments unidirectional flux to last compartment C 2 3 Medium, 3 cell compartments unidirectional flux to last compartment D 3 4 Medium, 2 cell compartments bidirectional flux to last compartment E 3 4 Medium, 3 cell compartments bidirectional flux to last compartment

Models based on physiological knowledge and speculation

F 3 4 Medium, cytoplasm, SDV, silica

SIT‐mediated transport

G 3 4 Medium, transport vesicles, SDV, silica

transport vesicle‐mediated transport

H 2 3 Medium, SDV, silica macropinocytosis mediated transport

Note that the medium is regarded as a separate compartment. Silicic acid uptake of from the medium is determined by the transport rate constant kup, and efflux from the cell to the medium by kef (all models). The defined models were separated into two categories: category 1 (Models A‐E) is used to determine the number of experimentally observable compartments in the cell, and category 2 (Models F‐H) describe silicic acid uptake and transport during valve formation based on current physiological data. Synchroneously dividing diatoms generally need at least about 8 hours to finish valve formation and cell separation;2‐5 within this time frame the cell density does not alter so that the growth rate does not affect the models. 5.7.1 Determination of the number of compartments (Model A – E)

To determine the number of compartments that can be experimentally discerned, five models (A – E) were formulated, comprising 1 to 3 cellular compartments; thus 2 – 4 compartments when the medium is included (Table 1/S2.1). In models B and C the assumption was that silicic acid in the final cellular compartment transforms fully into solid valve silica6 and as such is not transported back to another compartment of the cell. The transport towards the final compartment therefore is considered a unidirectional flux. Nevertheless, to confirm this assumption, models D and E were formulated that each include a reflux from the final compartment.


87

5.7.1.1 Model A: the cell as black box (no separate compartments in the cell)

→ kup medium ← kef

cell

Parameters Cm silicic acid concentration in the medium (mol/L) Cx cell density of the culture (cell/L) Ntotal mol Si taken up by the entire cell (mol/cell) Transport rate constants kup from medium to cell (L/(cell∙min)) kef from cell to medium (min‐1) Algorithms to determine transport from and into the compartments involved:

Medium: mef total x up m x

dC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.1)

Cell: totalup m ef total

dN k C k Ndt

= ⋅ − ⋅ (2.2)

5.7.1.2 Model B: two cellular compartments and an unidirectional flux to the final compartment

→ kup

→ k3 medium ← kef

comp 1

comp 2

Additional parameters and transport rates in comparison to model A: N1 mol Si in compartment 1 (mol/cell) N2 mol Si in compartment 2 (mol/cell) k3 from compartment 1 to compartment 2 (min‐1) In model B the total amount of silicic acid in the cell Ntotal consists of the total amount of silicic acid in the compartments 1 and 2 (N1 and N2): 1 2totalN N N= + Algorithms to determine transport from and into the compartments involved:

Medium: 1m

ef x up m xdC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.3)

Cellular compartment 1: 13 1( )up m ef

dN k C k k Ndt

= ⋅ − + ⋅ (2.4)


88

Cellular compartment 2: 23 1

dN k Ndt

= ⋅ (2.5)

5.7.1.3 Model C: three cellular compartments and an unidirectional flux tothe final compartment

→ kup → k3

→ k5 medium ← kef

comp 1 ← k4

comp 2

comp 3

Additional parameters and transport rates in comparison to model B N3 mol Si in compartment 3 (mol/cell) kup from medium to compartment 1 (L/(cell∙min)) k4 from compartment 2 to compartment 1 (min‐1) k5 from compartment 2 to compartment 3 (min‐1)

In model C the total amount of silicic acid in the cell (Ntotal) consists of the total

amount of silicic acid in compartments 1, 2 and 3 (N1, N2 and N3):

1 2 3totalN N N N= + + Algorithms to determine transport from and into the compartments involved:

Medium: 1m


= ⋅ ⋅ − ⋅ ⋅ (2.3)

Cellular compartment 1: 14 2 3 1( )up m ef

dN k C k N k k Ndt

= ⋅ + ⋅ − + ⋅ (2.6)

Cellular compartment 2: 23 1 4 2 5 2

dN k N k N k Ndt

= ⋅ − ⋅ − ⋅ (2.7)

Cellular compartment 3: 35 2

dN k Ndt

= ⋅ (2.8)


89

5.7.1.4 Model D: two cellular compartments and a bidirectional flux between all cellular compartments (i.e. model B with efflux from the final compartment included)

→ kup → k3 medium ← kef

comp1

← k4

comp2

Additional parameters and transport rates in comparison to model B k4 efflux from compartment 2 to compartment 1 (min‐1) The total amount of silicic acid in the cell Ntotal consists of the total amount of silicic acid in compartments 1 (N1) and 2 (N2): 1 2totalN N N= + Algorithms to determine transport from and into the compartments involved:

Medium: 1m


= ⋅ ⋅ − ⋅ ⋅ (2.3)


dN k C k N k k Ndt

= ⋅ + ⋅ − + ⋅ (2.6)

Cellular compartment 2: 23 1 4 2

dN k N k Ndt

= ⋅ − ⋅ (2.9)

5.7.1.5 Model E: three cellular compartments and a bidirectional flux between all cell compartments (i.e. model C with efflux from the final compartment included)

→ kup → k3 → k5 medium

← kef comp 1

← k4 comp 2

← k6 comp 3

Additional parameters and transport rates in comparison to model C k6 from compartment 3 to compartment 2 (min‐1)

The total amount of silicic acid in the cell (Ntotal) consists of the total amount of silicic acid in compartments 1, 2 and 3 (N1, N2 and N3): 1 2 3totalN N N N= + + Algorithms to determine transport from and into the compartments involved:

Medium: 1m


= ⋅ ⋅ − ⋅ ⋅ (2.3)


dN k C k N k k Ndt

= ⋅ + ⋅ − + ⋅ (2.6)


90

Cellular compartment 2: 23 1 6 3 4 5 2( )dN k N k N k k N

dt= ⋅ + ⋅ − + ⋅ (2.10)

Cellular compartment 3: 35 2 6 3

dN k N k Ndt

= ⋅ − ⋅ (2.11)

5.7.2 Description of silicic acid transport by means of SITs (model F), transport vesicles (model G) and macropinocytosis mediated transport (model H)

Three mechanisms were defined for transport of silicic acid from the medium to the SDV based on current physiological data. These were: i) transport by means of silicon transporters (SITs; model F), ii) transport vesicle‐mediated uptake and intracellular transport (model G), and iii) (macro)pinocytosis‐ mediated uptake (model H). The siliceous valve inside the SDV in all models was considered as a chemically different type of compounds (solid silica) and distinguished as a distinct compartment. The transport process from SDV to the solid valve silica in fact represents the silica polymerization reaction that occurs inside the SDV. The silicic acid concentration in the SDV should be high enough to initiate the polymerization reaction.7 Moreover, the SDV appears to acidify in the course of valve formation, which favours the polymerization of free silicic acid and inhibits dissolution of silica.5 Therewith it is reasonable to assume that silicic acid in the SDV is fully transformed to solid silica that not dissolves (efflux of silicic acid); the efflux from solid valve silica to silicic acid in the SDV thus is excluded in the models. 5.7.2.1 Model F: Silicon transporters (SITs) mediated uptake and transport

This mechanism describes the active uptake of silicic acid from the medium over the cell wall into the cytoplasm and transport (assumingly in an yet unknown stabilized form8‐11) to the SDV where the valve is formed. This model describes a four‐compartments system comprised of the medium, the cytoplasm, the SDV and the siliceous valve. It should be noted that the mathematical description of the SIT‐mediated uptake mechanism is identical to model C.

→ kup → kcs

→ ksv medium ← kef

cytoplasm ← ksc

SDV

valve

Parameters Cm silicic acid concentration in medium (mol/L) Cx cell density of culture (cell/L) Ntotal mol Si in entire cell (mol/cell) Nc mol Si in cytoplasm compartment (mol/cell) Ns mol Si in SDV compartment (mol/cell) Nv mol Si in valve compartment (mol/cell)


91

Transport rate constants kup from medium to cytoplasm (uptake) (L/(cell∙min)) kef from cytoplasm to medium (efflux) (min‐1) kcs from cytoplasm to SDV (min‐1) ksc from SDV to cytoplasm (min‐1) ksv from SDV to valve (min‐1)

The total amount of silicic acid in the cell (Ntotal) consists of the amount of silicic acid in the cytoplasm, the SDV and the siliceous valve (Nc, Ns and Nv):

total c s vN N N N= + + Algorithms to determine transport from and into the compartments involved:

Medium: mef c x up m x

dC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.12)

Cytoplasm: ( )cup m sc s ef cs c

dN k C k N k k Ndt

= ⋅ + ⋅ − + ⋅ (2.13)

SDV: scs c sc s sv s

dN k N k N k Ndt

= ⋅ − ⋅ − ⋅ (2.14)

Valve: vsv s

dN k Ndt

= ⋅ (2.15)

5.7.2.2 Model G: transport vesicles mediated uptake and transport

According to this mechanism silicic acid is taken up by pinocytosis and transported to the SDV by means of transport vesicles. Morphological studies have revealed presence of vesicles that fuse to the expanding SDV,12 which were initially denoted as silicon transport vesicles (STVs). In addition an elevated membrane activity has been noticed in dividing diatoms cells,13,14 which could enable molding of the SDV and its fusion with vesicles. During molding, water may be expelled.15 In this whole process, the silicic acid concentration inside the transport vesicles is considered equal to that of the medium. Since any potential efflux mechanism involved is not assigned, an efflux from the SDV (if existing) to the medium has been described by the aspecific first order transport rate constant kef. The final model describes the following compartments: the medium, the transport vesicles, the SDV and the siliceous valve.

→ kup → kts

transport vesicles → ksv medium ← kef

SDV

valve


92

Additional or altered parameters and transport rates in comparison to model F Nt mol Si in transport vesicle compartment (mol/cell) kup from medium to transport vesicles (overall rate constant) (L/cell/min) kef from SDV to medium (aspecific rate constant) (min‐1) kts from transport vesicles to SDV (min‐1) ksv from SDV to valve (min‐1)

The total amount of silicic acid in the cell (Ntotal) consists of the amount of silicic acid in transport vesicles, the SDV and the siliceous valve (Nt, Ns and Nv):

total t s vN N N N= + + Algorithms to determine transport from and into the compartments involved:

Medium: mef s x up m x

dC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.16)

Transport vesicles: tup m ts t

dN k C k Ndt

= ⋅ − ⋅ (2.17)

SDV: ( )sts m t ef sv s

dN k C V k k Ndt

= ⋅ ⋅ − + ⋅ (2.18)

Valve: vsv s

dN k Ndt

= ⋅ (2.19)

The silicic acid concentration in the transport vesicles (Ct) equals the silicic acid concentration in the medium (Cm). With Vt as the total volume of the transport vesicles (in L/cell) Nt can be rewritten: t m tN C V= ⋅ This yields the differential equations:


dC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.16)

Transport vesicles: ( )t sup ts ef x up x t

m

dV Nk k k C k C Vdt C

= − + ⋅ ⋅ − ⋅ ⋅ (2.20)

SDV: ( )sts m t ef sv s

dN k C V k k Ndt

= ⋅ ⋅ − + ⋅ (2.18)

Valve: vsv s

dN k Ndt

= ⋅ (2.19)


93

5.7.2.3 Model H: uptake and transport by means of macropinocytosis

This mechanism describes the uptake of a large amount of silicic acid containing medium by macropinocytosis. It is hypothesized that this large vesicle (assumingly the initial SDV) is transformed into the SDV by compression of this vesicle towards the cleavage furrow in line with the observed elevated cellular (membrane) activity in SDV moding13,14 and fluorescent probing studies.2,3,16 During compression, silicic acids concentrates ‐ potentially when water is expelled15 ‐ until the threshold for silica polymerization is reached and silica is form. The increase in silica agrees with an increase in fluorescence intensity.2,3,16 The resulting model contains the following compartments: the medium, the SDV (originating from the pinocytosis vesicle) and the siliceous valve. Note that this model is mathematically identical to model B (Supplement 2.1).

→ kup

→ ksv medium ← kef

SDV

valve

Parameters, transport rates and constants Cm silicic acid concentration in medium (mol/L) Cx cell density of culture (cell/L) Ntotal mol Si in entire cell (mol/cell) Ns mol Si in SDV compartment (mol/cell) Nv mol Si in valve compartment (mol/cell) kup from medium to SDV (macropinocytosis) (L/cell/min) kef from SDV to medium (aquaporins ?) (min‐1) ksv from SDV to valve (min‐1) Ntotal consists of the amount of silicic acid in SDV and valve (Ns and Nv respectively):

total s vN N N= + Algorithms to determine transport from and into the compartments involved:


dC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.16)

SDV: ( )sup m ef sv s

dN k C k k Ndt

= ⋅ − + ⋅ (2.21)

Valve: vsv s

dN k Ndt

= ⋅ (2.22)


94

5.7.3 References 1. Shipley RA, Clark RE (1972) Tracer methods for in vivo kinetics – Theory and applications,

(Academic Press, New York, London). 2. Hazelaar S, van der Strate HJ, Gieskes WWC, Vrieling EG (2005) Monitoring rapid valve

formation in the pennate diatom species Navicula salinarum (Bacillariophyceae). J Phycol 41:354‐358.


4. Hazelaar S (2006) Nanoscale architecture; the role of proteins in diatom silicon biomineralization. PhD Thesis, University of Groningen, The Netherlands. http://irs.ub.rug.nl/ppn/296120626

5. van de Poll WH, Vrieling EG, Gieskes WWC (1999) Location and expression of frustulins in the pennate diatoms Cylindrotheca fusiformis, Navicula pelliculosa and Navicula salinarum (Bacillariophyceae). J Phycol 35:1044‐1053

6. Vrieling EG, Gieskes WWC, Beelen TPM (1999) Silicon deposition in diatoms: Control by the pH inside the silicon deposition vesicle. J Phycol 35:548‐559.

7. Iler RK (1979) The chemistry of silica (John Wiley & Sons, New York). 8. Hildebrand M, Volcani BE, Gassmann W, Schroeder JI (1997) A gene family of silicon

transporters. Nature 385:688‐689. 9. Hildebrand M (2000) Silicic acid transport and its control during cell wall silicification in

diatoms. in Biomineralization; from Biology to Biotechnology and medical Application, ed Baeuerlein E (Wiley‐VCH Verlag GmbH, Weinheim, Germany), pp. 171‐188.

10. Hildebrand M (2003) Biological processing of nanostructured silica in diatoms. Progr Org Coat 47:256‐266.

11. Thamatrakoln K, Hildebrand M (2007) Analysis of Thalassiosira pseudonana silicon transporters indicates distinct regulatory levels and transport activity through the cell cycle. Eukaryotic Cell 6:271‐279.

12. Pickett‐Heaps JD, Schmid AMM, Edgar LA (1990) The cell biology of diatom valve formation. Progr Phycol Res 7:1‐168.

13. Kühn S, Brownlee C (2005) Membrane organisation and dynamics in the marine diatom Coscinodiscus wailesii (Bacillariophyceae). Bot Mar 48:297‐305.

14. Schmid AMM (1986) Wall morphogenesis in Coscinodiscus Wailesii Gran et Angst II: cytoplasmic events of wall morphogenesis. in Proceedings of the 8th International Diatom Symposium, ed Richard M (Koeltz, Königstein, Germany), pp 293‐314.

15. Grachev MA, Annenkov VV, Likhoshway YV (2008) Silicon nanotechnologies of pigmented heterokonts. Bioessays 30:328‐337.

16. Shimizu K, Del Amo Y, Brzezinski MA, Stucky GD, Morse DE (2001) A novel fluorescent silica tracer for biological silicification studies. Chem Biol 8:1051‐1060.

http://irs.ub.rug.nl/ppn/296120626�


95

5.8 Supplement 3: mathematical comparison of the transport vesicle mechanism (model G) with model B It was observed that the silicon uptake model with only two compartments in the cell revealed the best fit to the experimental data for uptake of silicon. Mathematically it was demonstrated that the transport vesicle compartment displayed very fast uptake kinetics (Table S3.1). Table S3.1: Calculated magnitudes of minimum and maximum rates dN/dt (mol/(cell∙min)) and dVt/dt (L/(cell∙min)) and characteristic saturation times for models F and G for P. laevis.

compartment rate, eq. nr. min.‐max value char. saturation time (min)

Model F: SIT‐mediated transport cytoplasm dNc/dt, 2.13 0 ‐ 10‐14 > 100 SDV dNs/dt, 2.14 0 ‐ 10‐14 > 100 valve dNv/dt, 2.15 0 ‐ 10‐14 > 100 Model G: transport vesicle‐mediated transport tr. vesicles dVt/dt, 2.20 10‐11 ‐ 10‐10 0.1 ‐ 1 SDV dNs/dt, 2.18 0 ‐ 10‐14 > 100 valve dNv/dt, 2.19 0 ‐ 10‐14 > 100 Model H: macropinocytosis‐mediated transport SDV dNs/dt, 2.21 0 ‐ 10‐14 > 100 valve dNv/dt, 2.22 0 ‐ 10‐14 > 100

The question remained whether the uptake model (model G) for transport vesicle‐mediated silicon uptake matched with model B. In practice, this means that the algorithmic structure in the equations for both models should be similar and consequently N1 and N2 in model B (Suppl. 2; equations 2.3‐2.5) must respectively match Ns and Nv in model G (Suppl. 2; equations 2.16, 2.18 and 2.19). At the moment the transport vesicle compartment becomes saturated, there is no time dependent behavior in the experimental time frame and as such dVt/dt (Suppl. 2; equation 2.20) drops to zero so this equation can be rewritten as:

upt

sts ef x up x

m

kV Nk k C k C

C

=+ ⋅ ⋅ − ⋅

(3.1)

Subsequently, Vt is substituted into the equations of the model for transport vesicle‐mediated uptake (model G) in which then 3 compartments remain involved (i.e. the medium, the SDV and the valve). The resulting equations were then compared to both models:


96

Medium compartment:

model G: mef s x up m x

dC k N C k C Cdt

= ⋅ ⋅ − ⋅ ⋅ (2.16)

model B: 1m


= ⋅ ⋅ − ⋅ ⋅ (2.3)

Transport vesicles compartment (saturated, not time dependent):

model G: up mt t m

sts ef x up x

m

k CN V C Nk k C k C

C

⋅= ⋅ =

+ ⋅ ⋅ − ⋅ (3.2)

SDV compartment / model B cell compartment 1:

model G: ( )s tsup m ef sv s

sts ef x up x

m

dN k k C k k NNdt k k C k CC

= ⋅ ⋅ − + ⋅+ ⋅ ⋅ − ⋅

or simplified: constant ( )sup m ef sv s

dN k C k k Ndt

= ⋅ ⋅ − + ⋅ (3.3)

model B: 13 1( )up m ef

dN k C k k Ndt

= ⋅ − + ⋅ (2.4)

Valve compartment / model B cell compartment 2:

model G: vsv s

dN k Ndt

= ⋅ (2.19)

model B: 23 1

dN k Ndt

= ⋅ (2.5)

With comparing the equations of models B and G, the similarity in structure of the algorithms was obvious for two compartments; the medium and the siliceous valve. The silicic acid contents N1 (eq. 2.3) and N2 (eq. 2.5) of the cell compartments 1 and 2 are assigned similar to respectively Ns (eq. 2.16) and Nv (eq. 2.19), whereas the constant k3 (eq. 2.5) matches with ksv (eq. 2.19). Also the algorithmic similarity can be observed for the SDV compartment in case the constant in equation 3.3 becomes 1, in fact when:

sts ef x up x

m

Nk k C k CC

⋅ ⋅ − ⋅ (3.4)

The value kts then can be calculated via equation 3.1 for any chosen value of Vt. The mathematical evaluation of the observed silicic acid uptake shows that the transport vesicle model (model G) with three cellular compartments mathematically matches with the observed model B (two cellular compartments), but only in that case that for the transport vesicle compartments a very fast transport kinetics applies. In view of diatom valve formation and the pace at which the valve is formed in its 2‐D direction1,2 this energetically demanding process would not be likely. However, sequential pinocytosis steps could occur during 3‐D thickening of the valve, which proceeds less rapidly,1,2 allowing formation of the much more detailed nanostructures.3,4 As such transport vesicles cannot be excluded during diatom


97

valve formation; however, in massive uptake of silicic acid in the initial stages (the completion of the 2‐D valve structure) their role expectedly is less important. 5.8.1 References


2. Hazelaar S, van der Strate HJ, Gieskes WWC, Vrieling EG (2005) Monitoring rapid valve formation in the pennate diatom species Navicula salinarum (Bacillariophyceae). J Phycol 41:354‐358.

3. Pickett‐Heaps JD, Schmid AMM, Edgar LA (1990) The cell biology of diatom valve formation. Progr Phycol Res 7:1‐168.

4. Gordon R, Drum RW (1994) The chemical basis of diatom morphogenesis. Int Rev Cytol 150: 243‐272.


98

Chapter 6

A new method to study heterogeneous binding and precipitation of silicate and phosphate in heterotrophic biofilms

Abstract: Biologically mediated precipitation is currently being applied to improve soil properties for civil engineering purposes. Biofilms have developed mechanisms to accumulate nutrients and organic substrates in their structural extracellular polymeric substances (EPS) matrix. This is supposed to increase the substrate availability. It is expected that the binding of ions by the EPS can result in formation of precipitants and crystals in the biofilm. Here we present a new autoradiography method using the radioactive isotopes 31Si and 32P that allowed us to image the spatial distribution of silicate and phosphate binding in biofilms. The fast silicon uptake kinetics allowed us to quantify the 31Si signal, for 32P this was not possible. Using this method it was shown that both radioisotopes were bound heterogeneously by the biofilm. In addition, the metal concentrations in the growth medium affected the biofilm structure as well as the metal binding characteristics of the biofilm. The relation between binding of silicates and crystallization of silicate is discussed.

Chapter 6: Silicon binding on biofilms

100

6.1 Introduction Biologically mediated precipitation of inorganic minerals can be applied for many civil engineering purposes.1,2 For example, micro‐organisms are used in ground works to improve soil properties like mechanical strength and stiffness (biogrouting), or permeability (biosealing or bioclogging).3‐5 Precipitation can be driven purely by physical and chemical processes, though many examples of biological enhanced precipitation processes are found in nature.6‐8 Best known examples of biological enhanced precipitation processes are carbonate structures like corals, bivalves, exoskeletons in coccolithophores, stromatolites and trombolites, but also bone and skeleton structures result from biological induced precipitation.6,9 In many of these systems the biological component is able to control the crystal structure and composition to a certain extent. Several ways in which micro‐organisms contribute to the precipitation process are identified: 10,11 1) producing one of the precipitating ions (e.g. carbonate), 2) changing the local environmental conditions, like pH or redox potential or 3) by acting as a crystal nucleus or actively binding ions from the bulk solution. In contrast to carbonate, the use of silicate precipitates for biogrouting is hardly investigated and poorly understood, while silicate (quartz) is the most common cement in naturally cemented sandstones.12 The potential of silicate precipitates to apply for biogrouting seems high, due to a lower solubility and higher resistance to abrasion of quartz as compared to calcite.13‐15 Biological enhanced silicate precipitation is found in several organisms (diatoms and sponges). Silicate binding and precipitation in these systems is mainly controlled by variation of charged molecule groups in proteins 16,17 and these organisms are able to control for skeleton and frustule construction. Silicified microbial mats are another example of silicate precipitation on biological surfaces 18,19 and it has been suggested that microorganisms play a role in the precipitation of dissolved silicates. Cell walls of certain bacteria can serve as template for precipitation of several metal silicates.20 In addition, biofilms can induce precipitation albeit in a less controlled and indirect manner then in diatoms and sponges. Namely, extracellular polymeric substances (EPS) play a role in binding metals and counterions, thereby creating good sites for crystal nucleation or local supersaturation resulting in precipitation. EPS is a matrix produced by microorganisms that can be used for attaching and structuring biofilms. The EPS matrix contains mainly polymeric sugars, but also nucleic acids and proteins.21 The polymeric sugars can be composed of different types of monosaccharides, and saccharide composition may be regulated by environmental conditions as well as by the population composition in a biofilm. The EPS composition will partly determine the binding affinity for different micronutrients like silicate, phosphate and metal ions. For silicate precipitation it has been suggested that methylated groups in EPS, and phosphate and methylated groups in proteins, can serve as potential condensation nuclei. Also metal ions like iron in cell walls have been shown to enhance binding of silicate.20 Other factors that will determine the binding affinity may be temperature, pH, micronutrient concentration and diffusive transport.


101

Biofilms are rarely homogeneously structured and many studies have shown that biofilm structuring can be explained by self–organizing mechanisms.22 These mechanisms are mostly illustrated by simple models describing mass transport, cell growth, cell to cell signaling or chemotaxis.23‐25 Recently it has been suggested that biofilms may be considered as a unicellular structure in which microorganisms can be altruistic as long as it is beneficial for the performance of the total structure.26 Organization has been suggested as a way for a biofilms to cope with unfavorable environmental conditions, or to optimize the mass transport or its overall performance.27 Heterogeneity in structure may create variation in binding sites and thereby potentially improve the overall performance of the biofilm. Probably the affinity for nutrients is not uniformly distributed over the biofilm. To investigate the possibilities of biofilm use in ground works knowledge of the EPS silicate binding capacities, the influence of metals, and the role of heterogeneity is of importance. In the present study autoradiography is used to gain insight into the heterogeneity of binding of silicates and phosphates in biofilms grown under different concentrations of metals (Fe, Mg, Ca and Al). It was assumed that different metal availability would induce a variation in types of EPS expressed in the biofilm.28,29 We also investigated whether autoradiography data can be used for quantification of the binding of silicates in particular. For this, a new method to handle data and quantify the results was developed based on the use of radioisotopes. In this study a 31Si‐silicate tracer was used to study silicate binding on the biofilm. In addition phosphate (a biofilm nutrient) binding was investigated with 32P‐phosphate to be able to evaluate the new method with another tracer.

6.2 Material and methods 6.2.1 Biofilm growth conditions

Heterotrophic biofilms have been grown on microscope slides in a PVC incubator which allowed gas exchange with air. The incubator contained three physical separated incubation lanes. Every lane had a volume of 10 ml and contained three microscope slides. The lanes were connected to a circulation system and 200 ml medium was continuously circulated over the slides. Growth medium was refreshed daily. The metal concentrations in the growth media were different for the three lanes, but all media contained 5 mM NaCl, 10 mM NH4Cl, 10 mM NaNO3, 10mM KH2PO4, 25 mM Na2HPO4.7H2O, 0.1%ww D‐glucose, 0.05% ww NaAcetate and 0.05% ww NaSuccinate. Metal concentrations in the media of the three lanes are given in table 1. Since it is our goal to use silicate binding for grouting processes which require a high concentration of silicate, the pH of the medium was set to 9.5‐10 using a 1M NaOH. Nevertheless in these experiments no extra silicate was added during growth, and the silicate concentration in the medium was kept on 7.5 μmol/L. In this way it was easier to assess whether the new imaging method can work.


102

Table 1: Metal concentrations in the media for lane 1, 2 and 3 Metal Concentration (mmol/L) lane 1 lane 2 lane 3 Fe 0.8 1.6 4.0 Ca 0.8 4.0 20 Al 0.8 1.6 4.0 Mg 0.8 4.0 20

Heterotrophic biofilms were grown by initially adding an inoculum to all three lanes and allow settlement and attachment of microbial cells for three hours. The inoculum was derived from a sand column incubation flushed with a medium with the same composition as is used for the high metal lane. After three hours the inoculum was replaced by the medium and biofilms were allowed to develop for a period of 5 days at room temperature. After 5 days the Si and P binding experiments with the radioactive tracers were performed on every slide. The water content of the biofilms was used to estimate their volumes, and was determined by measurement of the difference between the wet weight of the biofilms per slide prior to the exposure of the radioactive material and the weight after drying. 6.2.2 Metal analysis

Inductively coupled plasma optical emission spectrometry (ICP‐OES, Perkin Elmer OES Optima 4300DV, Boston, USA) was used to determine Si, Fe, Ca, Al and Mg concentrations in the biofilm. Dried non‐radioactive biofilms were destructed (24 hours at 60o C) by pure HNO3,30 and the metal concentration was determined on the residue. Dry weight of the biofilms was determined by weighing before destruction to correct for biomass differences. For calibration, Merck CertiPUR standard solutions 1703 (Amsterdam, The Netherlands) were used. 6.2.3 31Si and 32P tracers 31Si in the form of 31Si‐silicate (t1/2 2.62 h, β‐ 1.49 MeV) accompanied by 32P in the form of 32P‐phosphate (t1/2 14.3 day, β‐ 1.7 MeV) was produced in the nuclear reactor of the Reactor Institute Delft of Delft University of Technology, Delft, The Netherlands. No‐carrier‐added 31Si‐labeled silicate solution (specific activity 4.8 TBq/g) accompanied by 32P‐phosphate (specific activity 2.7 MBq/g) was prepared by purification with a chemical precipitation reaction with barium carbonate as described earlier.31 By removing a part of the 32P‐phosphate during purification it was possible to prepare a tracer solution that contained both radionuclides in a good proportion, allowing double tracer experiments. The half‐lifes of 31Si and 32P are 2.62 h and 14.26 days (=342.24 h) respectively. This means that two measurements are obligatory to determine the activities of both tracers. The first measurement determines the total activity of both 31Si and 32P, and should take place soon after the preparation of the tracer solution. The second measurement must take place when 31Si has decayed to negligible radioactivity level (in this study after at least 48 hours). The activity of 32P at the time of the first measurement can be calculated from the activity at the second measurement by use of its half‐life. After subtraction, this yields the 31Si activity at the time of the first measurement. The activity in liquid samples of 31Si and 32P was determined on a


103

LKB liquid scintillation counter (β‐ radiation by liquid scintillation) using Ultima Gold XR liquid scintillation counting cocktail. The volume of the samples was adjusted to 5 ml with demineralized water and mixed with 15 ml LSC counting cocktail prior to analysis. In this paper all activities are recalculated to the end time of the tracer production in the nuclear reactor (t=0). The activities at t=0 are correlated (when possible) to the silicate and phosphate concentrations in the samples. The amounts 31Si and 32P do not significantly influence the concentrations of total Si and P in the mastermix, which were 7.5 μmol/L and 25 mmol/L, respectively. 6.2.4 Radioactive 31Si and 32P treatment of the biofilms

After rinsing the biofilms with demineralized water to wash away the metal containing media as much as possible, the biofilms were exposed to the radioactive master mix (mixture of metal free medium and tracer solution). Due to the fast decay rate of 31Si incubation times of 10 and 20 minutes (experiment 1 and 2, respectively) were chosen to allow a reasonable time for the tracer to interact with the biofilms, and to ensure that enough tracer activity was left for the autoradiography procedure. The incubator was gently shaken every 2 minutes for 15 seconds during this period, in order to allow a good exchange between the medium and the biofilm. Samples from the mastermix were taken prior to incubation and after incubation to determine the activity per mL. Overlaying water was removed directly after the treatment and the slides were rinsed again very shortly with non radioactive solution with the same composition as the mastermix. After this final washing, the slides were dried on air in a hood. This took approximately 45 minutes. 6.2.5 Scanning and autoradiography procedures

After drying biomass distribution images per slide were made using a black and white HP flat bed scanner scanning with a black background. The light reflection was used as indicator of the biomass density distribution. All images in this study have been scanned at 300 dpi, giving a pixel resolution of ~85μm pixel‐1. Scanning all images at the same resolution allowed a direct comparison of the different autoradiographical images. First an autoradiographic image (Imaget1, consisting of the signal of both 31Si and 32P) was made by exposure of the radioactive labeled biofilms to autoradiography screens of the Phosphor Imager system (Packard Cyclone Storage Phosphor System). This system is sensitive to both 31Si and 32P and can determine the spatial distribution of radioactivity. A second image (Imaget2) showing only 32P activity without any signal from 31Si was made at least 48 hours later, when the 31Si signal was undetectable due to its short half life time of 2.62 hour (see results section). Imaget2 was used to determine the 32P portion in Imaget1, after which it was possible to calculate the specific 31Si portion in Imaget1. The exposure times for both images were 3 hours. Quantitative calibration of the grey values (after background subtraction) was done with reference drops of dried mastermix with known 31Si and 32P activity. It was found that the autoradiographic plates were not equally sensitive for the different radiation coming from 31Si and 32P. Decay from 32P caused 2.77 higher grey values per Bq∙s than decay from 31Si. The relationship between decay (in cpm∙h,


104

determined from the activity of the mastermix drops by liquid scintillation) and the total grey value (corrected for pixel number) for both nuclides is given in Fig.1. The linear fit has been forced through zero. Using the exposure times the grey values of 32P in Imaget1 were recalculated from Imaget2 and subtracted from the observed grey values yielding 31Si decay in Imaget1.

Figure 1: Relationship between greyvalue on the autoradiography plates and decay (in cpm∙h) of the tracer solution measured with liquid scintillation counting. Linear regression for Si y=1248 X (r2= 0.984) and P: y =3457 X (r2=0.992). Both have P<0.05.

6.2.6 Image treatment

The flat bed images were 8‐bit black and white images and the autoradiographical images were 16 bit black and white images. Every scan contained several microscope slides per image. To be able to do calculations with the different images, the images were cropped gaining a biomass image, an autoradiography image at t1 and an autoradiography image at t2 per microscope slide, and the angle was changed such that every image fitted exactly in a 308 X 898 pixel box which is the size of a microscope slide (2.5 cm X 7.5 cm) at 300 dpi. Cropping and changes of the angle were performed manually in the imaging software program ImageJ (http://rsbweb.nih.gov/ij/). Initially we tried to overlay the different images with Geographical Information Software (arc‐view), using multiple reference points. This allowed us to overlay the images, but created extra information by interpolation, which made the calculations between images impossible. Therefore it was necessary to reposition the images per slide manually by simple rotation and cropping of the main images. Fortunately the slides have specific sizes and also clear visual structures that allowed us to overlay the different images. After generation of

http://rsbweb.nih.gov/ij/�


105

the different images the conversions are relatively simple since these conversions are all linear. The background grey value was subtracted from the autoradiographical imaging. After background correction it was possible to separate the 31Si signal from the 32P signal in the Imaget1 by: Imaget1‐ (Imaget2 x correction factor). The correction factor is dependent of the half‐life time of 32P (342.24 hrs) and the starting and ending time of the exposure of the Image t1 and Imaget2:

,1 ,1

,2 ,2

ln 2 ln 2exp exp342.42 342.24correction factor

ln 2 ln 2exp exp342.24 342.24

start end

start end

t t

t t

− −⎡ ⎤ ⎡ ⎤⋅ − ⋅⎢ ⎥ ⎢ ⎥⎣ ⎦ ⎣ ⎦=− −⎡ ⎤ ⎡ ⎤⋅ − ⋅⎢ ⎥ ⎢ ⎥⎣ ⎦ ⎣ ⎦

In this study the correction factor was 1.39115 (time in hours). After separation of the 31Si and 32P signal it was possible to relate these signals to the actual Si and P in the mastermix and biofilm yielding the uptake/exchange in the biofilm per pixel, provided that the entire system of biofilm and overlaying mastermix has reached an equilibrium state (no change in concentrations in time) for that particular element. The analysis of the average values and standard deviations of the images and the creation of false coloring of the images was performed by use of ImageJ software. In Fig. 2 an example of the entire procedure is shown.


106

Figure 2: An example of the method to calculate silicate and phosphate distribution images. Image 1 Silicate at t=1 is corrected for background (A) where after the P image (image 2 at t=2, after background 500 subtraction (B) and calculation to t=1 (C)) is subtracted (yielding D). Thereafter both Si t=1 and the P t=1 image are recalculated to t=0 (E and G) and converted (if possible) to concentration images by multiplication of a correction factor (yielding F and H). Thereafter the ratio of the Si and P images can be made, which indicates the variation in binding Si to P in the biofilm (I).

6.3 Results 6.3.1 Effect of growth conditions on metal binding in biofilms

Different medium concentrations of iron, magnesium and calcium were used to test the effect of metal concentration on the binding capacity for silicate and phosphate by the biofilm. The different metal concentrations resulted in different biofilm structures. The lower metal concentrations resulted in more homogenous and “slimy” biofilms (i.e. biofilms containing a major EPS matrix). The biofilms grown at the high metal concentrations formed much less slime and displayed a more heterogeneous distribution. The difference in slime (EPS) production is reflected by the water content of the biofilm. Since water is the major volume fraction of the biofilm the water amount is directly correlated to the biofilm volume. All experiments showed the same decreasing trend in biofilm volume with an increase in metal concentration: medium metal treatment caused a decline in biofilm volume of 66‐74 % and high metal treatment a decline of 84‐91 % compared to low metal treatment.


107

Three independent replications of the experiment (without radioactive Si and P labeling) showed an increase of the overall metal content in the biofilm with an increase in metal concentrations in the medium (Fig. 3A).

Figure 3: Average metal concentrations (n=3 slides) at the three different metal treatments (A) in December 2006 (striped bars, exp. 1 and 2) and in march 2007 (crossed bars, exp. 3), measured during three independent experiments and the relationship between metal concentration in the biofilm and the silicate concentration in the biofilm (B) in the same experiments (the December experiments indicated by open circles and triangles, the march experiment indicated by the black squares) . The differences between the experiments may originate from differences in the inoculum. The inoculum has been from the same origin for all experiments, but for the third experiment the inoculum had been stored in the fridge for three months, in the high metal concentration medium. The third experiment was performed the week before the labeling experiments. The difference between the experiments is indicated by the striped bars (exp. 1 and 2) and the crossed bars (exp. 3). There seems to be an inverse trend between the metal bound to the biofilm and the silicate bound per gram of biofilm, when two outlayers with extreme high metal concentrations were excluded. However, when these outlayers were included the relationship was lost (Fig. 3B).


108

6.3.2 Optimization of the autoradiographic method

The labeling experiments with the biofilm were performed in triplicate in two independent experiments. Since the tracer had to be prepared separately per experiment there was some difference in the 31Si and 32P activity concentrations. The ratio of the 31Si and 32P activities in the tracer solution on t=0 for experiment 1 and 2 was 30 and 47 respectively. This yielded a theoretical 31Si/32P ratio of the grey values of the mastermix on the autoradiographic plates on t = t1 of 2.65 and 4.12 respectively, taking into account the differences in sensitivity for both isotopes. Any deviation of this ratio indicates selective binding of one of the isotopes. In the calibration it was found that the 31Si/32P greyvalue ratios were not homogeneous for the area on which the master mix drops have dried (Fig. 4) indicating some differences in crystallization characteristics of Si and P.

Figure 4: Image of 31Si and 32P distribution at t1 on the autoradiography plates for 20, 15, 10 and 5 ul dried traces solution on a microscope slide as well as the ratio of the two. The calibration is made with the mastermix of the second experiment and should yield a theoretical 31Si/32P ratio of 4.12 at t1. Since the ratio’s were not normal distributed, we used the modes rather than the mean values. The modus of the ratio’s was 4.40, 4.05, 3.67 and 4.09 for respectively the 20, 15, 10 and 5 μl drops. All values were relatively close to the theoretical value of 4.12. Working with two radionuclides implied that two measurements were obligatory. In all cases the second measurement (LSC determination or image) was made at least 48 hours after t=0. It can be calculated that after 48 hours only 0.0003% of the original 31Si was still present, while 90.7% of 32P remained. In the mixtures of the two tracers as used in the experiments it can be calculated that the percentage of the total activity that can be ascribed to 31Si after 48 hours is 0.01‐0.02% for all experiments. So the assumption can be made that 31Si has decayed completely after


109

48 hours, and the observed activity of the second measurement can be totally ascribed to 32P. 6.3.3 Time dependent uptake behaviour

To have an indication of the rates of the processes the average uptake and/or exchange of silicate and phosphate per slide was compared between experiment 1 and experiment 2. Measurements on the mastermix showed that <6% of the active silicate and <21% of the active phosphate was bound by the biofilms. The grey values of the images were averaged over the entire slide and converted to 31Si and 32P activities (Fig. 5) for exp. 1 and 2 (10 resp. 20 minutes incubation time).

Figure 5: Average uptake per mL biomass (n=3) for silicic acid (in nmol/mL biofilm volume) and phosphate (in Bq/mL biofilm volume) for low, medium and high metal medium, incubated for 10 (striped bars) and 20 minutes (black bars). Data calculated from average grey values per slide. The error bars represent standard deviations. These signals showed that both silicate and phosphate uptake per biofilm volume increased with the metal concentration in the medium. There was no significant difference in silicate binding between a labeling period of 10 and 20 minutes indicating that silicate binding and exchange seems to be a fast phenomenon, and the equilibrium state was reached within 10 minutes. This means that Si concentration changes in biofilm and medium do not occur after 10 minutes. For this reason it is possible to relate the 31Si activity to natural Si in the mastermix and biofilm and the silicate uptake per mL biofilm could be calculated. Phosphate uptake on the other hand was different for the 10 and 20 minutes binding experiments, where the 20 minutes binding experiment yielded a higher phosphate binding. This indicates that P‐binding was still in progress after 20 minutes incubation, and the equilibrium state was not reached yet. Based on these


110

results it was not possible to determine the time scale in which the state of equilibrium will be reached, and it was not possible to quantify phosphate binding in the biofilm based on 32P activity, but our results indicate that a difference in kinetics eventually results in a lower Si/P ratio for the long incubation (Fig. 6).

Figure 6: Si/P ratios of the average Si and 32P uptake values of the different slides. A are the results of 10 minutes incubation and B shows the results of 20 minutes incubation. The errorbars indicate the standard deviation of the three slides per treatment. 6.3.4 Silicate and phosphate spatial distribution

To test whether the binding of both components was homogenously distributed over the biofilm autoradiography was used to image the distribution of 31Si and 32P Heterogeneous silicate and phosphate binding to the biofilms was observed for all three media types. Heterogeneity in binding could not be explained by variation in biofilm biomass distribution alone. Areas with high biomass resulted in higher labeling, but several strong labeled regions were found where the biomass content was low (Fig. 7). The Si/P ratio images also show that some regions within the biofilm bind Si better (green/blue zones) while other regions bind P better (yellow zones).


111

Figure 7: From left to the right the spatial distribution of biomass, bound Si and P, the ratio Si/P in biofilms grown in low, medium and high metal concentrations after 10 minutes incubation (experiment 1). The color bar in the biomass experiment indicates the relative biomass measured as grey value. The color bar in the Silicate and Phosphate images represents the binding of Si and P in respectively femtomol/pixel and mBq/pixel. The ratio Si/P is in femtomol/mBq.

Zones with different binding ratios can occur close to each other (0.3‐0.6 mm). It was investigated whether this technique could also be used for quantification of the results. The time dependent uptake behaviour showed that it was allowed to quantify the 31Si signal. This means that the actual Si concentration distribution in the biofilm can be calculated based on the overall Si concentration in the entire biofilm (Fig. 3B). For 32P on the other hand this was not allowed (explained above).


112

The relative spatial 32P distribution in Fig. 7 is based on assumed homogeneous uptake kinetics. The histogram curves of silicate indicate that low metal media resulted in a more heterogenous binding of silicates, followed by respectively the high metal and the medium metal medium. For phosphate it appears that the high metal media results in the most heterogeneous binding, followed by biofilms grown at respectively low metal and medium metal media (Fig. 8).

Figure 8: Histogram of the distribution of Si, P and the ratio Si/P in experiment 1 The ratios have been expressed as fmol Si/ Bq P for convenience. Black squares represent the low metal treatment, open circles the medium metal concentration and the open triangles the high metal concentration. It was also found that there were two optima in the Si histograms for the biofilms grown at low metal and high metal concentrations. The humps in the phosphate histograms indicate also two optima, but these are less pronounced. Two optima may result from different types of binding. The histograms of the Si/P images show that the binding was most equal (narrow peak) in the high metal treatment, while the low metal and medium metal curves were much wider indicating a bigger difference in binding characteristics in the biofilms growing under these treatments.


113

6.4 Discussion Our results show that silicate and phosphate directly bind to the biofilm matrix. Silicate precipitation on biological surfaces is a common feature in nature, like precipitation on several Bacillus species 20 and adsorption on yeast cell walls.32 This binding to individual cells, while biomass in biofilms is heterogeneously distributed already indicates that the binding of Si must be heterogeneous at the μm scale. In the case of EPS as a matrix for precipitation, the chemical composition of the EPS will determine the organization of the matrix and its binding capacity for both silicate and phosphate. Selective silicate binding to incorporated metal atoms (e.g. chelated Ca) or positive charged amine groups within the EPS matrix may be used to concentrate silicate or phosphate and thereby stimulate local crystal formation. The images show that different regions within a biofilm exist that bind Si or P better, probably as a result of a heterogeneous distribution of bacterial communities. The results also show differences between the binding kinetics of Si and P. This difference may be caused by a difference in binding characteristics within the biofilm, but may also result from fundamental differences between Si and P binding. Binding of silicate may be dependent on quick binding on available binding sites in the EPS matrix or cell walls. Phosphate on the other hand will also bind directly to specific binding sites, but in addition phosphate is expected to be taken up actively by growing cells by phosphate transporters. This active uptake may explain the ongoing P uptake per biovolume as found in the 20 minute incubation. The histograms show two optima for both silicate and phosphate binding, probably resulting from two different uptake mechanisms. Chemical precipitation as metal silicate or metal phosphate can explain one histogram optimum. Phosphate uptake as a nutrient by the organisms in the biofilm could explain the other optimum in the 32P histogram. For silicate adsorption on the EPS and cell walls may account for the other optimum. The narrow peak in the Si/P histogram could be a result of chemical precipitation of metal silicate or phosphate. For silicate it is allowed to convert the 31Si signal to an actual Si concentration in the biofilm, because the binding/exchange process appeared to be very fast (equilibrium within 10 minutes) while more than 93 % of the 31Si remained in the mastermix. Moreover the biofilms were pregrown in medium with almost the same silicate concentration as the mastermix. This means that a chemical equilibrium already existed between silicate in the biofilm and the preculture medium. Since the mastermix had almost the same concentration as the preculture medium a new chemical equilibrium could be established quite easily. The 32P signal on the other hand was changing during more then 20 minutes. This means that the equilibrium state was not reached yet, and that it is not possible to assign the 32P signal to a concentration. To determine the quantitative P uptake in the biofilms longer incubation times are needed. This has to be further investigated. It is clear that there is a relationship with silicate and phosphate binding and the metal treatment. Both the silicate and phosphate uptake per mL biofilm in high metal treatment were about six times higher than in low metal treatment. However, when silicate and metal concentrations in the biofilms were compared (Fig 3), the correlation between silicate concentration and the actual metal concentration in the biofilm seems to be negative. This difference may be explained by a difference in


114

metal content in the biofilms which is also different due to the clear difference variation in EPS production per treatment. In the experiment in which we compared the metal content with the Si content we did not add extra Si to the medium. Our experimental design was such that we looked for the highest binding capacity. In the silicate and phosphate binding experiment we tested whether it was possible to manipulate Si binding in soils and we took out the metals of the medium when we added the silicate. Highest binding was found in the high metal treatment. In principle almost every charged metal ion should be able to increase nucleation of silicate. The difference may be caused by the fact that in the high metal media, silicate is already precipitated in the media, thereby reducing the silicate concentrations to which the biofilms are exposed. The biofilms were grown under high pH regimes. The reason was that for biogrouting purposes high amounts of silicate are needed in the medium without the risk of gelation of crystallization. A reduction in pH due to metabolic activity will favor the precipitation of metal phosphates and silicates. Under lower pH regimes complex formation of (not dissociated) silicic acid with metal ions,33 and silicic acid adsorption directly on biological substances like yeast cell walls 32 could act as starting point for precipitation. Probably the EPS itself, which has a similar chemical structure as the yeast cell wall, can also act as precipitation matrix for silicate. The behavior under lower pH has to be further investigated. For the application of biologically mediated silicate precipitation it will probably be best to produce positively charged EPS (grown under high pH).34 Optionally we can create a negatively charged EPS surface, and charge this EPS with bivalent metal cations. Another option is the use of PEG or the production of specific proteins, in nature specially formed to stimulate silicate precipitation. These proteins are mostly formed by some bacteria, but also by all diatoms and sponges. In addition it also seems possible to grow bacteria which have iron bound to their cell wall which can function as condensation nuclei. This needs further investigation. This study shows that it is possible to combine autoradiographic methods and digital image handling, and that double tracer experiments are possible in spatial binding studies. The difference in half life times between 31Si and 32P allowed signal separation by imaging the sample directly after the experiment and consequently after at least 48 hrs. This enables the calculation of the individual contribution of 31Si and 32P to the grey values. In liquid samples it suffices to subtract the overall activity measuring results from each other after recalculation to the ending time of the tracer production in the reactor (t=0). But in imaging results the spatial distribution also plays a role. To be able to separate the imaging data for both 31Si and 32P signals it is required that the pixel positions of the images of the different time points match each other exactly. The spatial method is very sensitive, and can detect concentrations down to femtomol per pixel (89μm X 89 μm). The images show that the heterogeneity is big (Fig. 7), with values ranging between 2,5 to 25 femtomol Si per pixel within one treatment. When this method is used for quantification of the uptake it is obligatory that the incubation times are long enough for the system to reach an equilibrium state, and that the concentrations in biofilm and mastermix are known. In this study this was the case for silicate, but not for phosphate. To obtain reliable phosphate uptake


115

values further investigations are needed, like longer lasting experiments and phosphate efflux determinations. The spatial distribution method described here opens up opportunities to study precipitation in biofilms in more detail, not only for 31Si and 32P, but also for other radioisotopes. The sensitivity for other radiotracers depends especially on the beta energy, branching percentage, half‐life time and specific activity. When used for 32P alone, the method can be even more sensitive since purification to remove part of the 32P in order to create the right balance between the 31Si and the 32P isotope does not have to be applied. In this case it is better to produce the tracer at least two days before the experiment, allowing the complete decay of 31Si.

6.5 References 1. Ivanov V and Chu J (2008) Applications of microorganisms to geotechnical engineering for

bioclogging and biocementation of soil in situ, Reviews in Environmental Science and Biotechnology 7 (2), 139‐153.

2. DeJong JT, Mortensen BM, Martinez BC and Nelson DC (2009), Bio‐mediated soil improvement, Ecological Engineering In Press, Corrected Proof.

3. DeJong JT, Fritzges MB and Nusslein K (2006), Microbially Induced Cementation to Control Sand Response to Undrained Shear, Journal of Geotechnical and Geoenvironmental Engineering 132 (11), 1381‐1392.

4. Nemati M. and Voordouw G (2003), Modification of porous media permeability, using calcium carbonate produced enzymatically in situ, Enzyme and Microbial Technology, 33 (5), 635‐642.

5. Whiffin VS, van Paassen LA and Harkes MP (2007), Microbial Carbonate Precipitation as a Soil Improvement Technique, Geomicrobiology Journal 24 (5), 417‐423.

6. Lowenstam HA and Weiner S (1989), On Biomineralization. Oxford, Oxford University Press. 7. Stocks‐Fischer S, Galinat JK and Bang SS (1999), Microbiological precipitation of CaCO3, Soil

Biology and Biochemistry 31 (11), 1563‐1571. 8. Castanier S, Le Metayer‐Levrel G and Perthuisot J‐P (2000), Bacterial Roles in the Precipitation

of Carbonate Minerals in:. Microbial Sediments, Riding RE and SM Awramik (eds.), Springer, Berlin

9. Wood R (1999), Reef Evolution, Oxford, Oxford University Press. 10. Hammes F and Verstraete W (2002), Key roles of pH and calcium metabolism in microbial

carbonate precipitation, Reviews in Environmental Science and Biotechnology 1 (1), 3. 11. Braissant O, Cailleau G, Dupraz C and Verrecchia EP (2003), Bacterially Induced Mineralization of

Calcium Carbonate in Terrestrial Environments: The Role of Exopolysaccharides and Amino Acids. Journal of Sedimentary Research 73 (3), 485‐490.

12. McBride EF (1989), Quartz cement in sandstones: a review., Earth‐Science Reviews 26, 69‐112. 13. Dyke CG and Dobereiner L (1991), Evaluating the strength and deformability of sandstones,

Quarterly Journal of Engineering Geology and Hydrogeology 24 (1), 123‐134. 14. Goldstein RH and Rossi C (2002) Recrystallization in Quartz Overgrowths, Journal of

Sedimentary Research 72 (3), 432‐440. 15. Broz ME,, Cook RF and Whitney DL (2006), Microhardness, toughness, and modulus of Mohs

scale minerals, 91, 135‐142. 16. Kröger N, Deutzmann R and Sumper M (1999), Polycationic peptides from diatom biosilica that

direct silica nanosphere formation, Science 286, 1129‐1132. 17. Kröger N, Deutzmann R and Sumper M (2001), Silica precipitating peptides from diatoms ‐ The

chemical structure of silaffin‐1A from Cylindrotheca fusiformis, Journal of Biological Chemistry 276 (28), 26066‐26070.

18. Shipe RF and Brzezinski MA (1999), A study of Si deposition synchrony in Rhizosolenia (Bacillariophyceae) mats using a novel 32Si autoradiographic method, Journal of Phycology, 35, 995‐1004

19. Shipe RF, Brzezinski MA, Pilskaln C and Villareal TA (1999), Rhizosolenia mats: An overlooked source of silica production in the open sea, Limnology and Oceanography 44(5), 1282‐1292.


116

20. Fortin D and Ferris FG (1998), Precipitation of iron, silica and sulfate on bacterial cell surfaces, Geomicrobiology Journal 15 (4), 309‐324.

21. Sutherland IW (2001), Biofilm exopolysaccharides: a strong and sticky framework, Microbiology 147, 3‐9

22. Picioreanu C, Kreft JU, Klausen M, Haagensen JAJ, Tolker‐Nielsen T, Molin S (2007), Microbial motility involvement in a biofilm structure formation – a 3D modelling study, Water Science and Technology 55, 337‐343.

23. Van Loosdrecht MCM, Heijnen JJ, Eberl H, Kreft J, Picioreanu C (2002) Mathematical modelling of biofilm structures Antonie van Leeuwenhoek, International Journal of General and Molecular Microbiology 81 (1‐4), 245‐256.

24. Van Loosdrecht MCM, Picioreanu C, Heijnen JJ (1997) A more unifying hypothesis for biofilm structures FEMS Microbiology Ecology 24 (2), 181‐183.

25. Xavier JB, De Kreuk MK, Picioreanu C, Van Loosdrecht MCM Multi‐scale individual‐based model of microbial and byconversion dynamics in aerobic granular sludge (2007) Environmental Science and Technology 41 (18), 6410‐6417.

26. Kreft JU (2004) Biofilms promote altruism. Microbiology 150, 2751‐2760 27. Palkova Z (2004), Multicellular microorganisms: laboratory versus nature, EMBO Reports 5, 470‐

476. 28. Jang A, Kim SM, Kim SY, Lee SG, Kim IS (2001) Effect of heavy metals (Cu, Pb, and Ni) on the

composition of EPS in biofilms, Water Science and Technology 43 (6), 41‐48 29. Mikes J, Siglova M, Cejkova A, Masak J, Jirku V (2005) The influence of heavy metals on the

production of extracellular polymer substances in the processes of heavy metal ions elimination. Water Science and Technology 52 (10‐11) 151‐156

30. Bisconti L, Pepi M, Mangani S, and Baldi F (1997), Reduction of vanadate to vanadyl by a strain of Saccharomyces cerevisiae, Biometals 10, 239–246.

31. Brasser HJ, Gürboğa G, Kroon JJ, Kolar ZI, Wolterbeek HT, Volkers KJ, Krijger GC (2006), Preparation of 31Si labelled silicate: a radiotracer for silicon studies in biosystems, Journal of Labelled Compounds and Radiopharmaceuticals 47, 867‐882, DOI 10.1002/jlcr.1096

32. Brasser HJ, Krijger GC, Wolterbeek HT (2008), On the beneficial role of silicon to organisms: a case study on the importance of silicon chemistry to metal accumulation in yeast, Biological Trace Element Research, 125, 81‐95

33. Iler RK (1979), The chemistry of silica, John Wiley & Sons, New York. 34. Comte S, Guibaud G, Baudu M (2008), Biosorption properties of extracellular polymeric

substances (EPS) towards Cd, Cu and Pb for different pH values. Journal of Hazardous Materials 151 (1) 185‐193.

117

Chapter 7

Evaluation and outlook The aim of the studies described in this thesis was certainly fulfilled as more insights in the mechanisms behind the role of silicon and silicon chemistry in several organisms were gained. For this purpose a silicon tracer was developed (chapter 2) and used to investigate silicon behaviour in baker’s yeast (chapter 3 and 4), diatoms (chapter 5) and biofilms (chapter 6).

7.1 The 31Si‐silicate tracer For the determination of silicon uptake in organisms ICP‐OES (e.g. 1), or use of the stable isotopes 29Si or 30Si in combination with mass spectrometry (e.g. 2) can be applied, but these methods have the disadvantage of relatively high detection limits, requirement of relatively laborious sample destruction procedure, and/or large sample volumes. Radiotracers generally do not have this drawback. For silicon the radioactive isotopes 31Si, 32Si and the analogon 68Ge are avialable (e.g. 3‐5). The nuclide 31Si has the advantages over 32Si that it decays to a stable daughter isotope and it can be produced no‐carrier added. Furthermore it is chemically identical to natural silicon, in contrast to 68Ge. For the above reasons a no‐carrier added 31Si‐labelled silicate tracer was produced. This was performed via a 31P(n,p)31Si reaction using fast neutrons. A high specific activity (4.8 ± 1.4 TBq/g) was obtained. An amount of 4.3 ± 0.8 MBq/mmol P was formed, accompanied by 5.7 ± 1.1 MBq 32P. The yield can be enhanced by target irradiations in a high fast neutron flux facility in a high flux reactor (the Delft reactor is only 2 MW), which could be further investigated. The 31Si‐silicic acid tracer was proven to have a relatively high radionuclidic and radiochemical purity. The yield was sufficient for silicic acid uptake studies in yeast cells, diatoms and biofilms (Chapters 3‐6). Even in the case of yeast cells, which exhibit far less silicon uptake than diatom cells, uptake kinetic studies could be performed, although relatively large sample volumes were required (chapter 4). Results obtained using the radiotracer are in good agreement with results from ICP‐OES measurements. The radiotracer has the advantage over ICP‐OES determinations that it allows short sample time intervals, and in general smaller sample volumes. In addition, it can be used for autoradiographic purposes, as shown in biofilms (Chapter 6). The relatively short half‐life time of 31Si has the practical disadvantage that experiments, including the radioactivity measurements, have to be completed within about eight hours. For long lasting investigations the 31Si tracer should be combined with a longer living tracer like 32Si. An other (minor) drawback of the tracer is the side product 32P‐phosphate formed by thermal neutron activation via the 31P(n,γ)32P reaction. Probably the 32P impurity can be reduced by shielding of the thermal neutrons by cadmium foil or by irradiation in a facility that has a higher ratio between the fast neutron flux and the thermal neutron flux. A small 32P residue remained after chemical purification, and none of the methods investigated could further remove this residue. The problem of

Chapter 7: Evaluation and outlook

118

the remaining 32P can be dealt with by performing two measurements on the samples: one immediately after sampling, and the second at least two days later, when 31Si has decayed to undetectable levels. Based on these two measurements the original amount of 31Si present in the sample can be calculated. This procedure worked out very well in the experiments. In some cases the presence of 32P was an advantage, and made it possible to perform double tracer experiments where 32P was used in an experiment for the chemical characterization of the tracer (chapter 2), and for determination of phosphate uptake in biofilms (chapter 6).

7.2 Evaluation of the studies on yeast, diatoms and biofilms To investigate the role of silicon chemistry in biology a study was performed on Baker’s yeast (a model organism for the eukaryotic cell), diatoms and biofilms. Yeasts and biofilms have in common that no indications were found that they possess any biochemical means to handle silicon.6 The influence of silicon on bioprocesses in these organisms has therefore most likely a chemical origin. Diatoms on the other hand master silicon and its chemistry to a high extend and have developed mechanisms to cope with possible toxic effects of silicon accumulation. 7.2.1 Subjected to silicon chemistry: yeast cells and biofilms

In chapter 3 and 4 the influence of silicon on several processes in yeast was studied. Metals appeared to be affected by silicic acid (metal contents, metal uptake rates, and metal deficiency symptoms). The investigations indicated that silicic acid is probably not taken up by the cell and that the observed phenomena could be explained by a combination of (physico)‐chemical processes like adsorption and complex formation with metal ions and organic compounds in the cell wall. It was found that the adsorption characteristics varied with metal concentrations in the medium. The observed silicon‐metal interactions could be explained without any involvement of enzymes or binding sites. The existence of extracellular interactions of silicon with organic compounds and metals was also observed as metal dependent silicic acid deposition on the extracellular polymeric substances of biofilms (chapter 6). A PDMPO probe to visualize depositions of polymerized silicic acid in or on the cell7 can be usefull in further research, as are other techniques like X‐ray tomography or TEM‐EDX to investigate the surface structure of cells. The results show that plain (physical) chemistry can be as or sometimes even more important for biological processes as biochemical processes, depending on the organism and circumstances. It is likely that these types of silicon interactions take place in many or maybe all organisms, and that these processes are important in several biological processes. Silicon is omnipresent in the environment, and also in chemicals that are normally used for the composition of growth media it is often present as a contaminant.8 So a silicon deficiency is unlikely to occur, neither in vivo nor in vitro. Future studies should not only focus on the bioprocesses that are influenced by silicon, but also on possible underlying (physico‐)chemical mechanisms. For biotechnological purposes (extracellular) silicon interactions can possibly be useful. Biogrouting by means of biofilms loaded with silicic acid has already been


119

mentioned for use in civil engineering, for instance for enhancing the mechanical strength in ground works (e.g. 9). But other applications are also interesting to study, for example the deliberate steering of biotechnological processes in microbiological cultures. 7.2.2 Diatoms: active “practitioners” of silicon chemistry

Diatoms have an absolute need for silicon and they control silicon chemistry to such an extend that they are able to produce a very complex silica structure like the frustule. This makes diatoms interesting for technology to gain more knowledge for the industrial production of complex silica structures.10 Also the fact that diatom cells are able to transport high amounts of silicic acid through the cell membrane and the cytoplasm without the occurrence of any unwanted chemical reactions, especially polymerization in unwanted parts of the cell shows their ability to handle silicon. The study described chapter 5 in this thesis showed that compartmental analysis based on a pinocytosis mechanism could well describe the uptake kinetics measured. This result offers a simple explanation of how diatoms are able to fulfill their silicon needs without exposing the inner part of the cell to high silicic acid concentrations and undesired polymerization processes. Further research should be focused on molecular and (bio)physical‐chemical aspects of diatom biosilicification. Also some higher plants seem to possess a certain ability to handle silicon,11 and a mechanism similar to the transport vesicle mechanism in diatoms is found in some metal tolerant plants that use silicon for alleviation of metal toxicity.12 Probably this type of uptake and transport mechanism is used in nature for more purposes, and has to be further investigated.

7.3 Outlook Silicon is everywhere in nature, and so is silicon chemistry, but for a long time it was believed that silicon was a quite inert element in nature.13 But the results in this thesis show that silicon chemistry offers enough options to organisms to take advantage of it. It is remarkable that in the case of yeasts silicon seems to act by strict chemical pathways and outside the cell, and that it does not have a direct influence on biochemical processes. In some cases an intermediate step seems to be involved, for instance the influence on metals, to influence bioprocesses. In nature also other extracellular indirect chemical processes do occur. The bacterial oxidation of pyrite by use of extracellular ferric ions is such an example.14

7.3.1 Further reseach in higher animals

The results obtained in this thesis add to the accumulating evidence that silicon chemistry is omnipresent in nature and can interfere with several processes in living organisms. Literature on higher animals and humans shows evidence for the beneficial influence of silicon on many processes (e.g. 15, 16). But despite many years of research silicon binding sites or silicon handling enzymes have never been found in these organisms. The results described in this thesis shed new light on these processes and make it plausible that silicon chemistry plays a major role in nature. Silicon interactions with organic compounds and metals as found in baker’s yeast


120

could well take place in higher organisms as well, and its interference with several biological processes could be in a merely indirect way and analogous to yeast. Although Baker’s yeast serves as a model organism for the eukaryotic cell, the effects of silicon chemistry have to be investigated in animals as well, for instance in cell cultures. This means in practice that the research in higher animals should turn its focus to physical and chemical processes. Studies with PDMPO probing7 and other techniques can be usefull in here. Also the (physical)‐chemical interactions of silicon with other compounds and the chemical speciation of silicon compounds inside cells, body fluids and tissues are important aspects to investigate in vitro and in vivo. 7.3.2 Towards a new definintion of silicon essentiality

The evolution of life on earth takes place in an environment in which silicon and its chemistry is omnipresent, and it is likely that organisms have adapted to silicon chemistry and used its possibilities and opportunities to a certain extend. The use of silicon by nature seems to take place between two extremes: on the one end organisms passively undergo silicon chemistry like the yeasts, and on the other end the organisms that have a strong control over silicon chemistry like the diatoms. Probably higher plants and animals are found somewhere in between these extremes, depending on the species. For these organisms silicon could be at least beneficial. The use of silicon chemistry without the aid of enzymes or other biochemical means is an easy and less energy consuming way for an organism to enhance or decrease certain processes. It can be speculated that during the evolution a balance has been established between silicon, metal ions and other nutrients in the environment, food, body fluids and tissues. Silicon chemistry then plays a role in maintaining this balance, and a disturbance can cause silicon deficiency signs as described in literature (e.g. 17). No evidence is found that higher animals like mammals and humans are capable of an active handling of silicon chemistry like the diatoms and sponges can, but probably higher animals are capable of such indirect interactions. Establishing the essentiality of silicon for several organisms has been difficult since many years and is often based on circumstantial evidence.15 In a natural environment silicon compounds lack any redox behavior which makes them practically unsuitable as a cofactor in enzymes. Since research on biological processes is mainly focused on enzymatic reactions and interactions simple (physico)chemical processes are less studied. This may explain why despite many years of research only a few biological binding sites or bioorganical compounds containing silicon have been found (yet) except some plants and for “bulk consumers” like the diatoms, silicoflagellates, ciliates and some sponges. But when silicon chemistry itself is taken into account, enzymes and binding sites are probably not needed for silicon to do its job as an essential element. In this way silicon differs from many other essential elements like copper, zinc and iron. Possibly silicon essentiality is in particular shown in the organism’s ability to handle silicon chemistry. Much more research is needed on this subject, and unraveling the interactions between silicon chemistry and biological processes remains a major challenge.


121

7.4 References 1. Fischer AC, Steinebach OM, Timmermans KR, Wolterbeek HT (2007) A method for the

destruction and analysis of biogenic silicon in two Antarctic diatom species: Thalassiosira sp. and Chaetoceros brevis. Journal of Applied Phycology 19: 71‐77

2. Koster JR, Bol R, Leng MJ, Parker AG, Sloane HJ, Ma JF (2008) Effects of active silicon uptake by rice on Si‐29 fractination in various plant parts. Rapid Communications in Mass Spectrometry 23 (16) 2398‐2402

3. Mehard CW; Volcani BE (1975) Similarity in uptake and retention of trace amounts of 31‐silicon and 68‐germanium in rat tissues and cell organelles. Bioinorganic Chemistry 5: 107‐124

4. Shipe RF; Brzezinski M (1999) A study of Si deposition synchrony in Rhizosolenia (Bacillariophyceae) mats using a novel Si‐32 autoradiographic method. Journal of Phycology 35: 995‐1004

5. Azam F, Hemmings BB, Volcani BE (1973) Germanium incorporation into silica of diatom cell walls. Archiv für Mikrobiologie 92: 11‐20

6. Walker GM (1998) Yeast physiology and biotechnology. John Wiley & Sons, Chichester 7. Shimizu K, Del Amo Y, Brzezinski MA, Stucky GD, Morse DE (2001) A novel fluorescent silica

tracer for biological silicification studies. Chemistry & Biology 8 (11) 1051‐1060 8. Iler RK (1979) The chemistry of silica, John Wiley & Sons, New York 9. Ivanov V, Chu J (2008) Applications of microorganisms to geotechnical engineering for

bioclogging and biocementation of soil in situ. Reviews in Environmental Science and Biotechnology 7 (2) 139‐153

10. Morse DE (1999) Silicon biotechnology: harnessing biological silica production to construct new materials. Trends in Biotechnology 17 (6): 230‐232

11. Ma JF, Tamai K, Yamaji N, Mitani N, Konishi S, Katsuhara M, Ishiguro M, Murata Y, Yano M (2006), A silicon transporter in rice, Nature 440 (7084) 688‐691

12. Neumann D, De Figueiredo C (2002) A novel mechanism of silicon uptake, Protoplasma 220: 59‐67

13. Ehrlich HL (1996) Geomicrobiology, third edition, revised and expanded. Marcel Bakker Inc., New York

14. Boon M, Brasser HJ, Hansford GS, Heijnen JJ (1999) Comparison of the oxidation kinetics of different pyrites in the presence of Thiobacillus ferrooxidans or Leptospirillum ferrooxidans. Hydrometallurgy 53 (1): 57‐72

15. Seaborn CD, Nielsen FH (2002) Silicon deprivation and arginine and cysteine supplementation affect bone collagen and bone and plasma trace mineral concentrations in rats. Journal of Trace Elements in Experimental Medicine 15 (3): 113‐122

16. Najda J, Gminski J, Drozdz M, Danch A (1992) Silicon metabolism ‐ The interrelations of inorganic silicon (Si) with systemic iron (Fe), zinc (Zn) and copper (Cu) pools in the rat. Biological Trace Element Research 34 (2): 185‐195

17. Carlisle EM (1986) Silicon as an essential trace element in animal nutrition. in: Evered D; O'Connor M (eds.), Silicon biochemistry, John Wiley & Sons Ltd., Chichester.


122

123

Summary Thesis: A tracer aided study on silicon chemistry in biological systems

Silicon (Si) is omnipresent in nature, and it is involved in important but diverse roles in a broad range of organisms, including diatoms, higher plants and humans. Some organisms, like the diatoms, need high amounts of silicon, and master silicon chemistry to a high extend using several enzymes. Other organisms which need silicon as an essential trace element apparently do not have the capability to handle silicon by any biochemical means and it was hypothesized that silicon chemistry as such plays a major role. The aim of the research described in this thesis was to gain more insight in the mechanisms behind the role of silicon in several organisms and to investigate to what extend silicon chemistry can play a role in biological processes. This study focused on the chemical or biological role of silicon in metal metabolism, on the risks that are connected to silicon polymerization, and on a possible application in biotechnology. For this a study was performed on Baker’s yeast Saccharomyces cerevisiae which often serves as a model organism for the eukaryotic cell (chapters 3 and 4), diatoms (chapter 5) and biofilms (bacterial communities attached to a surface, chapter 6). A silicon tracer was developed (chapter 2) to aid the studies described in this thesis. No‐carrier added 31Si was produced by a 31P(n,p)31Si reaction by fast neutrons in the nuclear reactor of the Delft University of Technology. Several methods were investigated to remove the side product 32P. Anion exchange with Dowex resin gave the best results in total activity yield and specific activity, but precipitation with BaCO3 appeared to be the fastest and cheapest purification method, and sufficiently high yields were obtained as well. It was determined that the 31Si tracer was in the desired chemical form of silicic acid (Si(OH)4), and suitable to apply in biological systems. Since yeasts and biofilms do not possess any biochemical means to handle silicon, it is likely that any influence of silicon on these organisms has a chemical origin. This was investigated by studying the interaction of silicon and metals in these organisms. In chapter 3, the influence of silicon (as silicic acid Si(OH)4) on the growth rate and intracellular accumulation of a number of metals was investigated in Baker's yeast Saccharomyces cerevisiae, a model organism for the eukaryotic cell. It was found that the growth rate was not influenced by silicic acid up to concentrations of 10 mmol per liter growth medium and a slight growth inhibition was observed when silicate was present in an extremely high concentration of 100 mM. Intracellular metal concentrations were investigated in yeast cultures grown in normal culture medium without added silicate (‐Si) or with the addition of 10 mmol/L silicate (+Si). Decreased amounts of Co, Mn and Fe were found within +Si grown yeast cultures as compared to ‐Si grown ones, while increased amounts of Mo and Mg were found. Zn and K were apparently unaffected by the presence of silicon. +Si enhanced yeast growth rate under low Zn2+ conditions, but decreased growth rate under low Mg2+ conditions. +Si did not alter the growth rates in high

Summary

124

Zn2+ and Co2+ media. +Si doubled the uptake rate of Co2+, but did not influence that of Zn2+. It was proposed that these results could be explained by the formation of a polysilicate layer on the cell wall which changes the cell wall binding capacity for metal ions. The toxicity of silicic acid was compared to germanium (Ge, as GeO2), a member of the same group of elements as Si (group 14) and sometimes used in literature as a silicon analogue. Ge proved to be far more toxic to yeast than Si and no influence was found of Si on Ge toxicity. It was proposed that these results relate to differences in cellular uptake and that is not always possible to use Ge as a Si analogue. These results also indicated that a chemical mechanism, rather than a biological one, is important. This was further investigated by studying the influence of zinc and magnesium on Si‐accumulation at several silicate concentrations in the medium by use of 31Si(OH)4 (chapter 4). Si‐accumulation fitted well with Freundlich adsorption. Si‐release followed depolymerization kinetics, indicating that silicate adsorbs to the surface of the cell rather than being transported over the cell membrane. Subsequently, adsorbed silicate interacts with metal ions and, therefore, alters the cell’s affinity for these ions. Since several metals are nutritional, these Si interactions can significantly change the growth and viability of organisms. In conclusion, the results show that chemistry is important in Si and metal accumulation in Baker’s yeast, and suggest that similar mechanisms should be studied in detail in other organisms to unravel essential roles of Si. The capability of silicon to adsorb on organic substances was also investigated in biofilms, bacterial communities linked together by extracellular polymeric substances (EPS) to study a possible application of silicon chemistry in a biological system. Biofilms have developed mechanisms to accumulate nutrients and organic substrates in their EPS matrix, probably to increase the substrate availability. It may be expected that the binding of ions by the EPS can result in interaction with silicon in the biofilm. Probably this interaction can be used for applications in civil engineering (e.g. biogrouting of soil). To study this the spatial distribution of silicate and phosphate binding in biofilms under different metal conditions was investigated (chapter 6). For this a new autoradiography method using 31Si (as silicic acid) accompanied by 32P (as phosphate) was developed. The equilibrium in silicon uptake was reached within minutes, so it was possible to quantify the 31Si signal, but for 32P this was not possible. Using this method it was shown that both silicon and phosphate bound heterogeneously to the biofilm. In addition, the metal concentrations in the growth medium affected the biofilm structure as well as the silicate and phosphate binding characteristics of the biofilm. In contrast to yeast and biofilms, diatoms master silicon chemistry to a high extend. Here it was investigated how diatoms cope with high amounts of silicon during valve formation as polymerization at/in vital structures should be avoided (chapter 5). Silicic acid uptake using 31Si(OH)4 was studied during valve formation in synchroneously dividing cells of the diatom Pleurosira laevis and other diatoms. Valve formation in diatoms requires bulk uptake and transport of silicic acid to the silica deposition vesicle (SDV). Two earlier proposed mechanisms for silicic acid uptake and transport were investigated: 1) uptake of silicon via silicon transporters (SITs) with subsequent intracellular transport, and 2) (macro)pinocytosis‐mediated

Summary

125

uptake. The SITs mechanism requires a controlled mechanism to stabilize the high amounts of reactive silicon species to prevent autopolymerization and simultaneously direct these species towards the SDV, whereas this problem does not play a role in the (macro)pinocytosis‐mediated mechanism. Experimental data were correlated to systematically derived mathematical models for a compartmental analysis of the possible uptake/transport pathways, including those for both SITs‐ and (macro)pinocytosis‐mediated uptake and transport. This study indicates that the experimental data on silicon uptake during valve formation match best with the model that describes (macro)pinocytosis‐mediated uptake. This process not only explains observed surge uptake at high demands for silicon, but also suggests that another pathway exists in which SITs apparently are not involved. The study showed that the pinocytosis mechanism gave a good description of the uptake kinetics that were found in this study. This result offers a simple explanation for how the diatomic cell is able to fulfill its silicon needs without exposing the inner part of the cell to high silicic acid concentrations and the problems related to spontaneous polymerization. Further molecular and (bio)physical‐chemical research is needed on diatom biosilicification. The results described in this thesis shed new light on the role of silicon chemistry in several bioprocesses. The influence of silicon on bioprocesses in yeast is probably from chemical origin, resulting from interactions with organic compounds and metals. These interactions, which also occur in biofilms, could also take place in higher organisms as well, and could probably explain at least part of the influence of silicon in biological processes in higher organisms. This may explain why despite many years of research biological binding sites or bioorganical compounds containing silicon have never been found (yet) except for some plants and for bulk consumers like the diatoms and some sponges. But when silicon chemistry itself is taken into account, enzymes and binding sites are probably not needed for silicon to do its job as an essential element. Further research is required on this subject to get clear‐cut answers on this matter. ir. Heleen J. Brasser

Summary

126

127

Samenvatting Proefschrift: een tracer ondersteunde studie naar siliciumchemie

in biologische systemen Het element silicium (Si) komt overal in de natuur voor. Het is essentieel voor talloze levende wezens zoals kiezelwieren (diatomeeën), planten en zoogdieren, waar het een rol speelt in een scala aan biologische processen. Sommige organismen, zoals diatomeeën (die een extern kiezelskelet bezitten) hebben zeer veel silicium nodig, en beschikken over een set enzymen om het te hanteren. Maar bij veel andere organismen die ook silicium nodig hebben als essentieel element, zoals bijvoorbeeld de zoogdieren, is tot nu toe geen enkele siliciumbiochemie gevonden. De mechanismen die wel een rol spelen zijn (nog) niet opgehelderd. Een mogelijke verklaring kan zijn dat niet biochemie, maar gewone siliciumchemie een belangrijke rol speelt. Het onderzoek dat beschreven is in dit proefschrift richtte zich op deze hypothese. Er is onderzocht of siliciumchemie een rol kan spelen in verschillende biologische systemen, en wat de achterliggende mechanismen kunnen zijn. Nadruk is gelegd op de chemische en/of biologische rol van silicium in metaalmetabolisme, op de risico’s die verbonden zijn aan siliciumpolymerisatie, en op een mogelijke toepassing van de siliciumchemie in de biotechnologie. Hiertoe zijn de volgende organismen onderzocht: Bakkersgist Saccharomyces cerevisiae (een veel gebruikt modelorganisme voor de eukaryote cel, hoofstuk 3 en 4), diatomeeën (kiezelwieren, hoofdstuk 5), en biofilms (een laag bacteriën gehecht op een oppervlak, hoofdstuk 6). Om dit onderzoek uit te kunnen voeren is een siliciumtracer ontwikkeld die geproduceerd kon worden in de kernreactor van het Reactor Instituut Delft van de Technische Universiteit Delft (hoofdstuk 2). Door het bestralen van fosforzuur met snelle neutronen kon via een 31P(n,p)31Si reactie dragervrij radioactief silicium (31Si) gemaakt worden. Bij deze reactie ontstaat radioactief fosfor (32P) als bijproduct, en daarom zijn er verschillende methoden onderzocht om het 32P te verwijderen. Behandeling met de anionwisselaar Dowex gaf de beste resultaten op het gebied van activiteitsopbrengst en specifieke activiteit. Maar ook een neerslagreactie met bariumcarbonaat bleek een snelle en goedkope methode te zijn met een goede opbrengst. Onderzoek toonde aan dat de tracer de gewenste chemische vorm van kiezelzuur (Si(OH)4) bezat en geschikt was voor biologische systemen. Gisten en biofilms bezitten voor zover bekend geen specifieke siliciumbiochemie, en waarschijnlijk is de invloed van silicium op deze organismen van chemische oorsprong. Als siliciumchemie een rol speelt, kan dat tot uiting komen in de interactie van silicium en metalen in deze organismen. Dit is onderzocht in hoofdstuk 3 in de invloed van silicium (als kiezelzuur, Si(OH)4) op de groeisnelheid en de hoeveelheden metaal cellen van Bakkersgist Saccharomyces cerevisiae, een modelorganisme voor de eukaryote cel. De groeisnelheid bleek niet beïnvloed te worden door silicium tot een concentratie van 10 mmol per liter groeimedium, en er werd een lichte groeiremming geconstateerd bij de extreem hoge concentratie van 100 mmol per liter. De invloed van silicium op de metaalconcentraties in de cel is

Samenvatting

128

onderzocht in gistculturen zonder extra toegevoegd silicium (‐Si) en in culturen met 10 mmol/L silicium (+Si). In +Si culturen werden lagere concentraties kobalt, mangaan en ijzer in de cel gevonden dan in –Si culturen, terwijl de concentratie van molybdeen en magnesium juist toe nam. Zink en kalium werden niet beïnvloed door silicium. Gist onder zinkdeficiëntie bleek sneller te groeien in een +Si cultuur dan in een –Si cultuur, terwijl onder magnesiumdeficiëntie het omgekeerde het geval was. Er is geen invloed van silicium waargenomen op de groeisnelheid van culturen met hoge zink‐ en kobaltconcentraties. De opnamesnelheid van kobalt in de cel in standaard medium was verdubbeld in +Si culturen, terwijl die van zink onveranderd bleef. Als verklaring voor deze waarnemingen werd voorgesteld dat silicium als een laag polysilicaat adsorbeert op de buitenkant van de cel op de celwand. Deze laag zou de bindingscapaciteit van de celwand voor metaalionen beïnvloeden. Germanium (Ge) behoort tot de zelfde groep elementen in het periodiek systeem als silicium (groep 14) en wordt soms gebruikt als analogon of van silicium. De toxiciteit van silicium (als kiezelzuur) is vergeleken met die van germanium (als GeO2). Germanium bleek veel toxischer te zijn voor gist dan silicium. De aanwezigheid van silicium geen invloed bleek te hebben op de toxiciteit van germanium. De verklaring hiervoor werd gezocht in verschillen in opname in de cel. Germanium blijkt niet altijd geschikt te zijn als siliciumanalogon. Alle resultaten wezen er op dat veeleer een chemisch mechanisme een belangrijke rol speelt dan biochemisch of biologisch mechanisme. Dit is verder onderzocht met behulp van de 31Si(OH)4 tracer (hoofdstuk 4). De invloed van zink en magnesium op de siliciumophoping in of op de cel is bepaald bij verschillende siliciumconcentraties in het medium. De waargenomen siliciumophoping bleek goed te beschrijven te zijn met een Freundlich adsorptie‐isotherm. Vervolgens is de siliciumefflux vanuit de cel gevolgd, en deze bleek overeen te komen met de depolymerisatiekinetiek van polysilicaat. Dit alles wees erop dat silicium adsorbeert op de buitenkant van de cel, en niet wordt opgenomen in de cel. Geadsorbeerd silicium kan op zijn beurt reageren met metaalionen, en beïnvloedt daarmee de affiniteit van de cel voor deze metalen. Omdat veel metalen een belangrijke voedingsbron voor de cel zijn, kunnen siliciuminteracties van grote invloed zijn op de groei en levensvatbaarheid van het organisme. De conclusie is dat chemie een grote rol speelt in silicium en de metaalhuishouding van Bakkersgist. Mogelijk spelen soortgelijke mechanismen een rol in andere organismen voor silicium als essentieel element. Dit moet verder onderzocht worden. De eigenschap van silicium om te adsorberen op organische oppervlakken is verder onderzocht in biofilms met het oog op een mogelijke biotechnologische toepassing. Biofilms zijn bacteriën die aan elkaar en aan een oppervlak gehecht zijn met extracellulaire polymere verbindingen (“extracellular polymeric substances”, EPS). Biofilms hebben mechanismen ontwikkeld om voedingsstoffen en organische substraten in te vangen in de EPS‐matrix, waarschijnlijk om op deze manier de beschikbaarheid van de substraten voor de cellen te verhogen. De verwachting was dat de metalen in het EPS een interactie kunnen aangaan met silicium. Mogelijk kan deze interactie toegepast worden in de civiele techniek (bijvoorbeeld bij het biogrouten van grond). Dit is onderzocht door de ruimtelijke verdeling van silicium en fosfaat in biofilms te bepalen onder verschillende metaalcondities (hoofdstuk 6).

Samenvatting

129

Hiertoe is een nieuwe autoradiografiemethode ontwikkeld die gebruik maakt van 31Si (als silicaat) vergezeld van 32P (als fosfaat). De opname van silicaat bereikte binnen enkele minuten een evenwichtwaarde, en daarom kon in dit geval de opname gekwantificeerd worden. Bij fosfaat was dit niet mogelijk. Met deze methode kon worden aangetoond dat zowel silicium als fosfaat hetrogeen gebonden werden aan de biofilm. Bovendien bleek de metaalconcentratie in het groeimedium zowel de structuur van de biofilm als de bindingskarakteritieken voor silicaat en fosfaat te beïnvloeden. Diatomeeën hebben een externe kiezelskelet, en bij vorming hiervan worden grote hoeveelheden silicium door de cel getransporteerd. Dit proces is niet zonder risico, omdat silicium in hogere concentraties spontaan kan polymeriseren. Diatomeeën bezitten (in tegenstelling tot gisten en biofilms) een scala aan biochemische mogelijkheden om de siliciumchemie te beheersen, maar het was niet duidelijk hoe de cel omgaat het het risico van spontane polymerisatie. Om dit te onderzoeken is de siliciumopname in gesynchroniseerde cellen van de diatomee Pleurosira laevis en van andere diatomeeën gemeten tijdens de vorming van het kiezelskelet (hoofdstuk 5). Hierbij is de 31Si(OH)4 tracer is gebruikt. In het verleden zijn er twee mechanismen voorgesteld die de opname en het transport van silicium in de cel beschrijven: 1) opname en transport door middel van transporteiwitten (“silicon transporters” of SITs), en 2) opname door (macro)pinocytose waarbij de de diatomee een “bel” vloeistof opneemt in de cel en omgeeft met een membraan. In het eerste geval is er een vorm van stabilisatie nodig die er voor zorgt dat de grote hoeveelheden silicium in de cel niet spontaan gaan polymeriseren. In het tweede geval is dit niet aan de orde. Deze beide mechanismen zijn verder onderzocht. Experimentele data zijn gecorreleerd aan een aantal wiskundige modellen voor een compartimentenanalyse en om de mogelijke opname en transportroutes te onderzoeken. Ook van het SITs‐ en (macro)pinocytosemechanisme is een wiskundig model opgesteld. Uit dit onderzoek bleek dat de experimentele data van siliciumopname gedurende de vorming van het kiezelskelet het beste bleken te passen in het model voor opname via (macro)pinocytose. Met dit proces kan de grote, snelle siliciumopname (“surge uptake”) die optreedt bij hoge siliciumbehoefte verklaard worden. Bovendien wordt hiermee gesuggereerd dat er een alternatieve opnameroute bestaat waarbij geen SITs betrokken zijn. Het onderzoek toonde aan dat opname via pinocytose zeer goed de opnamekinetiek beschrijft zoals die gevonden is in de experimenten. Dit resultaat verklaart eenvoudig hoe de diatomeecel in staat is om te voorzien in zijn siliciumbehoefte zonder het inwendige van de cel bloot te stellen aan hoge concentraties silicium en de daarmee verbonden risico’s van spontane polymerisatie. Nader moleculair en (bio)fysisch‐chemisch onderzoek naar biosilicificatie in diatomeeën blijft nodig. De resultaten die beschreven zijn in dit proefschrift werpen een nieuw licht op de rol van de siliciumchemie in verschillende biologische processen. De invloed van silicium op biologische processen in gist is waarschijnlijk van chemische oorsprong, en resulteert uit interacties met organische verbindingen en metalen. Deze interacties, die ook bestaan in biofilms, zouden ook voor kunnen komen in hogere

Samenvatting

130

organismen, en zouden de invloed van silicium op biologische processen in deze organismen ten minste deels kunnen verklaren. Dit kan ook verklaren waarom er tot nu toe geen siliciumbiochemie is gevonden, behalve in enkele planten en grootverbruikers zoals de diatomeeën en de sponzen. Maar als er ook rekening gehouden wordt met de chemie zelf van silicium, dan zijn biochemische processen en enzymen waarschijnlijk niet nodig voor silicium om zijn rol als essentieel element te kunnen vervullen. Om hier duidelijke antwoorden op de verkrijgen is nader onderzoek nodig. ir. Heleen J. Brasser

131

Dankwoord

“Always keep an open mind” Dr. Lesley A. Robertson

Het beginnen met een promotieonderzoek na een ingewikkelde periode in mijn leven was een sprong in het diepe, maar ook een onverwachte kans. Daarom wil ik allereerst Bert Wolterbeek, mijn promotor, bedanken omdat hij mij deze kans bood. Het project was aanvankelijk heel vaag, maar gaandeweg bij onze gesprekken werd duidelijk dat de rol van het element silicium in de biologie het onderzoeksonderwerp moest worden. Bert heeft de gave om dingen vanuit heel verschillende invalshoeken te kunnen zien. Dat maakte het onderzoek zeer boeiend en afwisselend. Bert ziet ook kans om iets positiefs te maken van “hopeloze” resultaten. Al was ik na een gesprek met Bert altijd doodop, ik zag het wel weer helemaal zitten. Daarom Bert, hartelijk dank, ’t was hartstikke leuk! Gerard Krijger was mijn dagelijkse begeleider. Hij had veel goede ideeën en was altijd zeer geinteresseerd in de laatste resultaten. En daar werd dan stevig over gediscussieerd, soms op afspraak, maar meestal spontaan bij de koffietafel, op het lab of zomaar ergens in de gang. Dit was voor mij heel belangrijk om een goed beeld te krijgen van wat ik aan het onderzoeken was en waar ik nog beter naar moest kijken. Gerard, hartelijk dank, je was onmisbaar! Er was een hele club mede‐onderzoekers met wie ik samengewerkt heb. Engel Vrieling en Han van der Strate deden het diatomeeënproject van de Rijksuniversiteit Groningen. Han kwam voor dag en dauw uit Groningen naar Delft met een hoofd vol ideeën en een rugzak vol cultures en media. En zo konden we een hele serie experimenten uitvoeren. Engel heeft een zeer grote bijdrage geleverd aan het artikel, niet alleen inhoudelijk, maar ook met zijn mooie Engels waarmee hij ingewikkelde zaken zeer helder formuleerde. Het biofilmproject was in handen van Marc Staal, Leon van Paassen en Mark van Loosdrecht. Marc zat boordevol plannen en was een door de wol geverfde experimentator. We hadden soms stevige wetenschappelijke discussies, en die leidden vaak tot interessante inzichten. Het was leuk om samen te werken met jullie, en om samen een artikel te schrijven. Engel, Han, Marc, Leon en Mark, allen hartelijk dank voor de prettige samenwerking op het lab en bij het schrijven van onze artikelen. Zonder jullie zou dit proefschrift niet geschreven zijn. Zvonko Kolar, Kees Volkers en Gülşen Gürboğa waren coauteur van het tracer‐artikel. Zvonko heeft met zijn radiochemiekennis een belangrijke bijdrage hieraan geleverd. Hij werkte zeer nauwkeurig de hele tekst van het artikel door, waarbij hij alles aanwees wat nog verduidelijking behoefde. Zvonko, bedankt daarvoor. Kees Volkers heeft zeer veel werk verricht in de ontwikkeling van de tracer. Ook heeft hij vele malen de opzuivering van de tracer uitgevoerd en maakte hij er een sport van om dit weer een minuut sneller te doen. Kees, je werk was superbelangrijk, hartelijk dank. Gülşen, die als gast op de afdeling werkte, heeft met haar werk en ideeën een belangrijke bijdrage geleverd aan de karakterisering van de tracer. Daarom, Gülşen, heel hartelijk bedankt.

Dankwoord

132

Het onderzoek zou beslist minder goed verlopen zijn zonder de hulp van Christa Timmermans, Ehsan Marandi en Jorrit Heikamp, die hun stage liepen op onze afdeling. Allemaal hartelijk dank voor jullie bijdrage aan het onderzoek door jullie keiharde werk dat gepaard ging met een groot enthousiasme. Ook de bijdrage van Ralph Feld, Ou‐Zheng Hu en Ernst de Wiljes wil ik niet onvermeld laten. Zij hebben voor hun propedeuse‐eindproject onderzoek gedaan naar de zuiverheid van de tracer en de efficiency van de verschillende detectiemethoden. Jullie werk was meer dan welkom, en daarom hartelijk dank. Dank aan Tona Verburg die mij bijstond in mijn gevechten met Origin, Paint Shop en ander computergedoe, maar vooral voor de adviezen en hulp op gebied van modellering en statistiek. “Komt goed, meissie”, zei ze dan, en dan kwam het ook altijd goed. Tona, je bent geweldig. Heel belangrijk voor het onderzoek is de ondersteuning op het lab. Daarom dank aan Koos Kroon die een aantal experimenten heeft uitgevoerd, en die een onmisbare vraagbaak was voor alles op labgebied. Dank aan Thea van Meerten voor de INAA metingen, die in een aantal gevallen gepaard gingen met enige gevechten met gedroogde gistcellen, Olav Steinebach die een deel van het ICP‐OES werk heeft uitgevoerd, Marijke Luttik van Biotechnologie voor advies over het kweken van gist, en Niek van der Pers van Materiaalkunde voor de Röntgendiffractieanalyses. Verder zorgde Folkert Geurink er altijd voor dat alles op computergebied vlekkeloos verliep. Geen tracer zonder kernreactor, en vooral, geen tracer zonder een HOR operator die om vier uur ’s nachts mijn rabbits in de reactor schoot. Daarom, mannen van de HOR, hartelijk dank. Yvon Weijgertse, hartelijk dank voor al het regelwerk, en ook dat je je hoofd koel hield bij onverwachte dingen of als er heel veel tegelijk moest gebeuren. Zonder jou waren er beslist dingen fout gegaan. Een speciaal woord van dank voor Peter Bode, die onmiddelijk actie ondernam toen er op korte termijn een nieuw commissielid gezocht moest worden. Verder dank aan alle andere (oud‐)collega’s die ik nog niet genoemd heb voor de prettige sfeer en samenwerking. In het bijzonder bedank ik mijn kamergenotes Astrid van der Meer, met wie ik heel wat lief en leed gedeeld heb, Diane Abou and Anna Sevcenco, I really enjoyed your company! Halverwege 2002 overleed plotseling de man met wie ik elf jaar getrouwd was geweest. Hoewel wij anderhalf jaar daarvoor gescheiden waren na een ingewikkeld huwelijk, was dit toch een grote slag voor mij. De dagelijkse routine van werk, en het gezelschap van mijn collega’s heeft mij zeker door die moeilijke tijd heen geholpen. Het deed mij goed om te ervaren dat ik er ook bijhoorde toen ik niet zo goed in mijn vel zat. Bedankt iedereen! En “last but not least” mijn dank aan het privéfront: mijn ouders, Paul, Sada, Aukje, en al die andere familie, vrienden en kennissen die ik niet genoemd heb maar die wel interesse getoond hebben of mij op één of andere manier gesteund hebben. Ook jullie hartelijk dank!

133

Curriculum vitae Heleen Brasser werd geboren op 10 april 1964 in Den Haag. In 1982 behaalde zij het VWO diploma aan het Interconfessioneel Westland College te Naaldwijk, en in 1983 het VWO certificaat Natuurkunde via een staatsexamen. In datzelfde jaar begon zij met de studie Scheikundige Technologie aan de Technische Universiteit te Delft. Tijdens de studie liep zij stage bij het Drinkwater Leiding Bedrijf te Rotterdam waar zij onderzoek deed aan adsorptie van organische verbindingen aan actief koolfilters. Zij specialiseerde zich in de microbiologie en studeerde in 1991 af op het stikstof‐ en zwavelmetabolisme van de bacterie Thiosphaera Pantotropha. In 1991 tot 1996 deed ze de postdoctorale beroepsopleiding Biotechnologie aan het Institute for Postgraduate Biotechnology Studies Delft Leiden en specialiseerde zij zich in de Bioprocestechnologie. Het afstudeeronderwerp was bioleaching van pyriet door Thiobacillus ferrooxidans en Leptospirillum ferrooxidans. Hierna werd zij werkzaam als gastdocent microbiologie en aquatische chemie aan het Unesco‐IHE in Delft naast andere werkzaamheden. In 2001 begon zij met het promotieonderzoek waarvan dit proefschrift het resultaat is. In 2004 behaalde zij het diploma Stralingsdeskundigheid niveau 3. Naast het onderzoek was zij practicumdocent en begeleidde zij een aantal studenten, stagiaires en een gastmedewerker.

Curriculum vitae

134

135

List of publications H.J. Brasser, G.C. Krijger, H.T. Wolterbeek. On the beneficial role of silicon to organisms: A case study on the importance of silicon chemistry to metal accumulation in yeast. Biological Trace Elements Research 125: 81‐95 (2008) H.J. Brasser, G.C. Krijger, T.G. van Meerten and H.T. Wolterbeek. Influence of silicon on cobalt, zinc, and magnesium in Baker’s yeast Saccharomyces cerevisiae, Biological Trace Elements Research 112: 175‐190 (2006) H.J. Brasser, G. Gürboğa, J.J. Kroon, Z.I. Kolar, H.T. Wolterbeek, K.J. Volkers and G.C. Krijger. Preparation of 31Si‐labelled silicate: a radiotracer for silicon studies in biosystems, Journal of Labelled Compounds and Radiopharmaceuticals 47: 867‐882 (2006). DOI: 10.1002/jlcr.1096 M.J. Staal, L.A. van Paassen, H.J. Brasser, G. Krijger, M.C.M. van Loosdrecht. Can biological mediated silicate binding in biofilms be used for biogrouting? In H. Jonkers, J. Booster, L. van Paassen, M. van Loosdrecht (Eds.), BGCE 2008. 1st Int. Conf. BioGeoCivil Eng., 23‐25 June 2008, Delft, The Netherlands, (pp. 150‐153). TU Delft and Deltares, Delft. G.C. Krijger, K.J. Volkers, H.J. Brasser and Z.I. Kolar. Production of no‐carrier added 31Si for silicon‐uptake studies in yeast cells, Advances in Nuclear and Radiochemistry, 6th Int. Conf. Nucl. Radiochem, aug 29 – sep 3 2004 M. Boon, H.J. Brasser, G.S. Hansford, J.J. Heijnen. Comparison of the oxidation kinetics of different pyrites in the presence of Thiobacillus ferrooxidans or Leptospirillum ferrooxidans. Hydrometallurgy 53: 57‐72 (1999)

A TRACER AIDED STUDY ON SILICON CHEMISTRY IN BIOLOGICAL ... · PDF fileA tracer aided study on...

Documents

Transcript of A TRACER AIDED STUDY ON SILICON CHEMISTRY IN BIOLOGICAL ... · PDF fileA tracer aided study on...