Kieselalgen (Diatomee) sind eukaryotische einzellige Algen die hochspezifische Proteine (sogenannte Silaffine) erzeugen, um ‘nanopatterned’ Silica-Zellwände herzustellen. Diese Proteine zeigen geringe oder gar keine Homologie innerhalb der Diatomeen Gattung und sind ausgiebig (extensiv) posttranslatorisch modifiziert. Zum Unterschied zu konventioneller Modifikation (z.B. Phosphorylierung und Glykosylierung) weisen Lysinreste von Silaffinen einige Polyaminketten mit sehr heterogenen molekularen Strukturen auf. Diese Modifikationen sind spezifisch für Kieselalgen und spielen somit hypothetisch eine Rolle in der Biosilica-Synthese. Allerdings sind Lysin Polyamin Modifikationen, modifizierte Proteine und modifizierte Stellen kaum charakterisiert. Um diese Frage zu beantworten entwickelten wir eine Methode Polyamine zu quantifizieren und die Position von Polyamin-Modifikationen in engverwandte Proteine zu identifizieren (in morphologisch unterschiedliche Diatomeen Thalassiosira pseudonana, T. oceanica und Cyclotella cryptica). Wir zeigten, dass das Gesamtmuster von Polyaminender phylogenetischen Nähe dieser Kieselalgenarten folgt und dass diese Polyaminmodifikationen an Konsensusstellen sogar in Proteinen auftraten, die keine Sequenzähnlichkeit zeigten.:CONTENTS Summary Zusammenfassung List of figures List of tables Abbreviations 1 Introduction 1.1 Diatoms 1.2 Diatom biosilica 1.2.1 Biosilicification in nature 1.2.2 Diatom biosilica structure and cell cycle 1.2.3 The cell biology of biosilica morphogenesis 1.3 The role of polyamine PTMs in diatom biosilicification 1.3.1 Identifying biomolecules associated with diatom biosilica 1.3.2 PTM complexity of biosilica-associated proteins 1.3.3 Lysine ε-polyamine PTMs in biosilica-associated proteins 1.4 Mass spectrometry in PTM discovery 1.4.1 Modification-specific proteomics 1.4.2 Analysis of polyamine-modified lysines by MS 1.4.3 Fractionation of proteins and peptides prior to MS 1.4.4 MS/MS analysis in modification-specific proteomics 1.4.5 Bioinformatics tools for modification-specific proteomics 1.5 Rationale of the thesis 2 Aim of the thesis 3 Results and discussion 3.1 A method for analysis of ε-polyamine PTMs 3.1.1 Establishing a method to analyse ε-polyamines 3.1.2 Method applicability for lysine PTM profiling 3.1.3 Profiling of lysine PTMs in silaffin-3 3.2 Profiling lysine PTMs in biosilica extracts 3.2.1 Lysine PTM profile and characteristic fragments 3.2.2 Elucidation of phosphopolyamine structures 3.2.3 LysinePTMprofilesofAFSMextracts 3.2.4 Comparison of AFIM and AFSM profiles in T. pseudonana 3.2.5 Phylogenetic relationship across three diatom species 3.3 PTM localization and discovery of consensus motifs 3.3.1 Multiple protease strategy for mapping lysine PTMs 3.3.2 Selection of deprotection technique 3.3.3 Mapping lysine PTMs on tpSil3 using iterative search strategy 3.3.4 Deconvolution of raw MS/MS spectra 3.3.5 PTM mapping by polyamine-specific fragments 3.3.6 Identification of consensus motifs harboring lysine PTMs 4 Conclusions and Outlook 5.1 Synthesis of polyamine standards 5.2 Isolation of biosilica-associated proteins 5.3 Expression of tpSil3 from synthetic gene 5.4 HCl hydrolysis 5.5 AQC-derivatization of amino acids and polyamines 5.6 LC-MS/MS analysis of QAC-derivatives 5.7 Amino acid measurement using UV-detection 5.8 Direct infusion MS/MS analysis 5.9 Acetylation of phosphopolyamines 5.10 31P-NMR measurements 5.11 Deglycosylation with TFMS 5.12 Treatment with HF·pyridine soluble complex 5.13 Anhydrous HF-treatment 5.14 Protein analysis by GeLC-MS/MS 5.15 Proteomics data processing A Appendix B Bibliography Acknowledgments Publications Declaration / Erklärung / Diatoms are eukaryotic unicellular algae that employ highly specialized proteins called silaffins for making nanopatterned silica-based cell walls. These proteins share little or no homology across diatom species and are extensively post-translationally modified. Apart from conventional modifications (e. g., phosphorylation and glycosylation) lysine residues of silaffins bear polyamine chains with highly heterogeneous molecular structure. The latter appear to be specific for silicifying organisms and therefore hypothesized to play a key role in biosilica synthesis. However, polyamine modifications of lysines, modified proteins, and modification sites remain poorly characterized. To address these questions, we developed a method to quantify polyamines and identify sites of polyamine modifications in proteins from phylogenetically closely related, yet morphologically distinct diatoms Thalassiosira pseudonana, T. oceanica, and Cyclotella cryptica. We demonstrated that the overall pattern of polyamines followed the phylogenetic proximity across these diatom species and showed that polyamine modifications occurred at consensus sites even in proteins showing no sequence similarity.:CONTENTS Summary Zusammenfassung List of figures List of tables Abbreviations 1 Introduction 1.1 Diatoms 1.2 Diatom biosilica 1.2.1 Biosilicification in nature 1.2.2 Diatom biosilica structure and cell cycle 1.2.3 The cell biology of biosilica morphogenesis 1.3 The role of polyamine PTMs in diatom biosilicification 1.3.1 Identifying biomolecules associated with diatom biosilica 1.3.2 PTM complexity of biosilica-associated proteins 1.3.3 Lysine ε-polyamine PTMs in biosilica-associated proteins 1.4 Mass spectrometry in PTM discovery 1.4.1 Modification-specific proteomics 1.4.2 Analysis of polyamine-modified lysines by MS 1.4.3 Fractionation of proteins and peptides prior to MS 1.4.4 MS/MS analysis in modification-specific proteomics 1.4.5 Bioinformatics tools for modification-specific proteomics 1.5 Rationale of the thesis 2 Aim of the thesis 3 Results and discussion 3.1 A method for analysis of ε-polyamine PTMs 3.1.1 Establishing a method to analyse ε-polyamines 3.1.2 Method applicability for lysine PTM profiling 3.1.3 Profiling of lysine PTMs in silaffin-3 3.2 Profiling lysine PTMs in biosilica extracts 3.2.1 Lysine PTM profile and characteristic fragments 3.2.2 Elucidation of phosphopolyamine structures 3.2.3 LysinePTMprofilesofAFSMextracts 3.2.4 Comparison of AFIM and AFSM profiles in T. pseudonana 3.2.5 Phylogenetic relationship across three diatom species 3.3 PTM localization and discovery of consensus motifs 3.3.1 Multiple protease strategy for mapping lysine PTMs 3.3.2 Selection of deprotection technique 3.3.3 Mapping lysine PTMs on tpSil3 using iterative search strategy 3.3.4 Deconvolution of raw MS/MS spectra 3.3.5 PTM mapping by polyamine-specific fragments 3.3.6 Identification of consensus motifs harboring lysine PTMs 4 Conclusions and Outlook 5.1 Synthesis of polyamine standards 5.2 Isolation of biosilica-associated proteins 5.3 Expression of tpSil3 from synthetic gene 5.4 HCl hydrolysis 5.5 AQC-derivatization of amino acids and polyamines 5.6 LC-MS/MS analysis of QAC-derivatives 5.7 Amino acid measurement using UV-detection 5.8 Direct infusion MS/MS analysis 5.9 Acetylation of phosphopolyamines 5.10 31P-NMR measurements 5.11 Deglycosylation with TFMS 5.12 Treatment with HF·pyridine soluble complex 5.13 Anhydrous HF-treatment 5.14 Protein analysis by GeLC-MS/MS 5.15 Proteomics data processing A Appendix B Bibliography Acknowledgments Publications Declaration / Erklärung

Les travaux présentés dans ce manuscrit portent sur l’étude des propriétés photoniques de quatre espèces de diatomées du genre Coscinodiscus : C. wailesii, C. specie, C. radiatus et C. radiatus-cf. Les diatomées sont des micro-algues unicellulaires constituées d’une cellule unique encapsulée dans une matrice inorganique de silice poreuse appelée frustule. Nous avons effectué des cultures de ces espèces, puis extrait les frustules afin de les caractériser, tant du point de vue de leur composition que de leur morphologie, par microscopie électronique et tomographie RX. Le frustule est constitué de trois couches de silice (le foramen, le cribrum et le cribellum) présentant une structure poreuse hiérarchique, s’étendant du micromètre au nanomètre, et formant ainsi une matrice tridimensionnelle complexe pouvant présenter des propriétés photoniques. Le foramen et le cribrum présentent des réseaux de pores de périodicité importante dont les dimensions caractéristiques (taille des pores, pas du réseau…) sont de l’ordre de grandeur des ondes lumineuses. Les propriétés optiques des frustules ont alors été étudiées expérimentalement par imagerie spectrale et théoriquement par simulations numériques par la méthode des éléments finis. Nous avons ensuite mis en évidence un effet de concentration de la lumière visible transmise sur l’axe de la diatomée à quelques dizaines de micromètres derrière le frustule. Cet effet s’avère dépendre de la longueur d’onde incidente ainsi que de l’orientation du frustule. Finalement, nous avons montré, par approche théorique et simulations numériques, que cet effet de concentration de la lumière le long de l’axe est majoritairement dû au foramen et que les autres couches n’ont qu’un effet marginal. Une telle micro et nano-structuration complexe de la matière est actuellement inaccessible à l’échelle industrielle. Cette structure présente cependant des effets de concentration et de filtration des ondes lumineuses d’un grand intérêt dans de nombreux domaines (photovoltaïque, cosmétique, peinture…) et on peut donc envisager raisonnablement une utilisation directe des diatomées. / The work presented in this thesis deals with the study of the photonic properties of four diatom species of the genus Coscinodiscus: C. wailesii, C. specie, C. radiatus et C. radiatus-cf. Diatoms are unicellular microalgae consisting of a single cell encapsulated in an inorganic matrix of porous silica called frustule. We performed cultures of these species, then extracted the frustules to characterize them, both in terms of their composition and their morphology, by electron microscopy and X-ray tomography. The frustule consists of three silica layers (the foramen, cribrum and cribellum) with a hierarchical porous structure, ranging from micrometer to nanometer scale, and thus forming a complex three-dimensional matrix that may have photonic properties. The foramen and cribrum have periodic pores networks whose characteristic dimensions (pore size, lattice constant…) are of the order of light wavelengths. The optical properties of the frustules were then studied experimentally by spectral imaging and theoratically by numerical simulations by finite element method. We demonstrated a concentration effect of visible light transmitted on diatom’s axis a few tens of micrometers behind the frustule. This effect appears to depend on the incident wavelength as well as the orientation of the frustule. Finally, we have shown by theoratical approach and numerical simulations, that light concentration effect along the axis is mainly due to the foramen and the rest of layers have only a marginal effect. Such complex micro and nano-structuring of matter is currently impossible on an industrial scale. This structure, however, has optical effects of light concentration and filtration of great interest in many fields (photovoltaics, cosmetics, paint…) and it is therefore reasonnable to consider a direct use of diatoms.

Morphogenesis of the intricate patterns of diatom silica cell walls is a protein-guided process, yet to date only very few such silica biomineralization proteins have been identified. Therefore, it is currently unknown whether all diatoms share conserved proteins of a basal silica forming machinery, and whether unique proteins are responsible for the morphogenesis of species-specific silica patterns. To answer these questions, we extracted proteins from the silica of three diatom species (Thalassiosira pseudonana, Thalassiosira oceanica, and Cyclotella cryptica) by complete demineralization of the cell walls. Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) analysis of the extracts identified 92 proteins that we name ‘soluble silicome proteins’ (SSPs). Surprisingly, no SSPs are common to all three species, and most SSPs showed very low similarity to one another in sequence alignments. In-depth bioinformatics analyses revealed that SSPs could be grouped into distinct classes based on short unconventional sequence motifs whose functions are yet unknown. The results from the in vivo localization of selected SSPs indicates that proteins, which lack sequence homology but share unconventional sequence motifs may exert similar functions in the morphogenesis of the diatom silica cell wall.