Diatoms are unicellular photosynthetic algae that play an essential role in global ecology.

They build intricate, nanostructured silica cell walls (frustules) under benign conditions,

i.e neutral pH and 18 ◦C. Understanding this mechanism could enable  synthesis of

customized structured silica-based materials under benign conditions. However, molecular machinery involved in diatom bio-silicification, and the interactions of these different elements remains poorly understood. This dissertation investigates the atomic and molecular basis of bios-silicification using omics-based tools to annotate proteins involved in this process. Genetic engineering and heterologous expression were used to investigate selected candidates, while solid-state nuclear magnetic resonance (NMR) spectroscopy was used to characterize the atomic interactions involved in silica formation.



First, the transcriptional response of the increase in bio-silicification of a pennate diatom Entomoneis vertebralis, when aluminum was added was characterized. With 10 µM added aluminum, the biogenic silica content of the cell wall increased approximately twofold, accompanied by coordinated changes in silicon, carbon, and nitrogen metabolism. Increased transcription of nitrate transporter genes were observed despite the downregulation of nitrogen assimilation pathways. Additionally, transcription of carbon-fixation genes also increased. Amino acid composition-based motif analysis, and structure-based analyses identified differentially expressed genes encoding proteins with features characteristic of silicification-associated proteins. One of the protein families which share a conserved structural core despite divergent sequences, was identified to be highly downregulated. Additionally, protein motif analysis identified a silicadin looking protein in this non-model diatom. Overall, this chapter provided a transcriptomic approach to identify silicification associated proteins in non-model diatom, along with understanding how diatoms react towards increase in aluminum.



Second, a proteomic approach was developed to spatially resolve proteins within the

frustule of Thalassiosira pseudonana using a series of increasingly harsh chemical treatments. Integrated with comparative genomics and published transcriptomic data, this

approach identified 746 candidate bio-silicification genes. 72%  of 746 genes were detected only in the surface treatment, which is removal of lipid layer, exposing the proteins that are associated with it. It also revealed a previously unknown protein, SILVA (Silica Vesicle Acidification Associated Protein), which resides only in the siliceous layer of the frustule and is conserved across all diatom genomes examined. SILVA is hypothesized to promote assembly of the V-type ATPase and thereby modulate pH inside the silica deposition vesicle.



Third, the interactions and biosynthesis of long-chain polyamines (LCPAs) were examined. DNP (Dynamic Nuclear Polarization)-enhanced solid-state NMR of T. pseudonana frustules showed that phosphorylated moieties of frustule-associated proteins interact strongly with LCPAs, participate in strong, short hydrogen bonds, and lie closer to LCPAs than to the protein backbone. Additionally, Diatoms produce species-specific long-chain polyamines, which are hypothesized to be synthesized via novel fusion of proteins involved in polyamine biosynthesis pathway. Comparative genomics showed that pennate and centric diatoms have evolved distinct sets of S-adenosylmethionine decarboxylase–spermidine synthase fusion proteins proposed to synthesize LCPAs, with one fusion protein shared between the two clades. Heterologous expression of an Entomoneis vertebralis fusion protein in Escherichia coli produced DAP, which could be a possible intermediary step to produce long chain polyamine.



Finally, gene delivery methods were optimized to validate gene candidates such as SILVA, including optimized electroporation of Chaetoceros muelleri, protoplasting to achieve higher efficiency and setting foundation work to CRISPR/Cas9 knockout of LCPA-associated genes in C. muelleri.



Together, these findings provide insights into the molecular machinery underlying diatom bio-silicification, including differences in silicification-associated proteins across diatom species and their potential roles in frustule formation. This work identifies candidate genes and establishes experimental methods to investigate their functions. These findings lay the groundwork for future studies of diatom bio-silicification and the bioinspired synthesis of structured silica-based materials.