Introduction

Diatoms are ubiquitous in seawater and freshwater environments, with the number of species estimated to be between 104–105 1. The complexity and the precision at which the cell wall, the frustule, is synthesised, at both the micro- and nano-scale, is a paradigm among material chemists for the controlled assembly of nanostructured materials2,3. Over the past decade there has been a surge of interest in altering the chemistry of the diatom frustule while preserving the intricate architecture. A number of processes have described the use of frustules as sacrificial templates for the generation of non-siliceous diatom replicas4,5,6,7. Despite the successful use of these processes, an emerging area of particular interest is the alteration of both diatom structure8 and chemical composition9,10,11 in culture.

Advances in chemical manipulation of the living diatom require an understanding of the biomineralization processes that underlie the formation of the intricate valve architecture. The first biomolecules indicated in diatom silica formation are the silaffins12,13 and long-chain polyamines14, proteins shown to induce in-vitro precipitation of silica from silicic acid12. More recently, TiO2 precipitation has also been induced by silaffins over a wide range of pHs using TiBALDH as a precursor15. It was hypothesized that substitution of the natural silica source of diatoms, Na2SiO3, with the Ti-based precursor, TiBALDH, will allow incorporation of Ti into the frustule of the centric diatom Thalassiosira weissflogii (T. weissflogii).

The range of proposed applications for the diatom frustule spans across many disciplines including; catalysis16, separation science17,18,19, optics20,21 and drug delivery22,23,24. It is always the cleaned harvested diatom that attracts attention and the possibility of harnessing the living diatom has not yet been fully explored.

This manuscript describes a method to alter the chemical composition of the living diatom T. weissflogii via Ti substitution. A high level of Ti incorporation is achieved via multiple dosing of cultures with concentrations of TiBALDH that satisfy the criteria of non-cytotoxicity and solubility. The chemical modification is not associated with alterations to the pore architecture of the diatom. However, minor changes to the rib structure are observed. Finally, irradiation of TiBALDH-modified diatoms with UV light led to the decay of Escherichia coli (E. coli) in co-culture demonstrating a novel photocatalytic activity. This property was also indicated when degradation of the photodegradable dye was observed following incubation with cleaned TiBALDH-modifed T. weissflogii exposed to UV light.

Results

T. weissflogii growth profiles in the presence of TiBALDH

In this study design it was essential that the concentration of TiBALDH added to the culture medium meets the following balance; (i) it does not adversely affect the growth profile of T. weissflogii and (ii) it does not precipitate in the culture medium. A series of cultures were supplemented with TiBALDH at concentrations ranging from 0.2 mM to 2 mM. Cell density and precipitate formation were monitored daily. TiBALDH concentrations lower than 1 mM were non-toxic to T. weissflogii (Figure 1a). Concentrations above 200 μM resulted in the formation of precipitate in the culture media over time. A comparison of the growth profile of T. weissflogii cultured in the presence of 200 μM Na2SiO3 versus 200 μM TiBALDH shows a similar pattern (Figure 1b) indicating that TiBALDH is not detrimental to T. weissflogii growth.

Figure 1
figure 1

TiBALDH associated growth profiles of T. weissflogii.

(a) Increased starting concentrations of TiBALDH did not increase the growth pattern of T. weissflogii. Concentrations above 0.2 mM caused the appearance of precipitate in the culture and concentrations above 1 mM adversely affected growth of T. weissflogii. Two-way ANOVA revealed an effect of time and treatment. Bonferroni post-hoc analysis revealed statistical significance ** p < 0.01 (n = 4). (b) T. weissflogii grown in the presence of TiBALDH only does not adversely affect growth patterns (n = 3). (c) Addition of TiBALDH every 48 hours extends exponential growth of T. weissflogii compared to a single dose of TiBALDH (n = 6). The decrease observed in cell density at 48-hour intervals is due to sampling of cultures following addition of Na2SiO3 or TiBALDH. (d) Sodium azide inhibits growth of T. weissflogii. In the absence of sodium azide cultures undergo a greater than 2-fold increase in cell density. One-way ANOVA revealed a significant effect of treatment F(5,15) = 6.787, p = 0.0052 (n = 3). Student Newman Keuls post hoc analysis (p < 0.05) showed statistical difference between Control vs Na2SiO3, Control vs TiBALDH, Na2SiO3 vs Na2SiO3 + Azide. Data are presented as mean ± sem in all panels.

Nutrient depletion of either Na2SiO3 or TiBALDH in the culture media leads to a prolonged stationary phase (Figure 1b). Hence; a multiple dosing strategy of adding either Na2SiO3 or TiBALDH at 48 hour intervals was investigated to prolong the increased growth. The growth profile of T. weissflogii was similar using either precursor; furthermore an extension of increased growth is observed (Figure 1c).

A complete understanding of TiBALDH-associated growth of T. weissflogii is limited as the genetic manipulations required to fully investigate the mechanistic pathways are not developed. A specific inhibitor of Ti uptake by T. weissflogii or any diatom does not currently exist. Thus, an indirect method of investigating TiBALDH-associated growth involved monitoring growth in the presence of sodium azide, a respiratory inhibitor, that has been shown to inhibit silica uptake and arrest cell division25,26. Increases in cell density were reduced significantly when either T. weissflogii or TiBALDH-modified T. weissflogii were cultured in the presence of 10 mM sodium azide (Figure 1d).

Ti content of T. weissflogii frustule

EDX-SEM analysis confirmed that Ti was incorporated into the frustule of TiBALDH-modified T. weissflogii (Figure 2). The Ti content was dependent on the number of TiBALDH doses that the culture received. A single addition of TiBALDH at the time of inoculation resulted in a maximum of 8.8 pg Ti/valve at 48 hours and then decreased to ca. 3 pg Ti/valve in the stationary phase (Figure 2a). The replenishment of the TiBALDH precursor in the culture media every 48 hours revealed consistency of Ti content over the period of culturing with maximum content of 13.8 pg Ti/valve after 192 hours following multiple addition of TiBALDH (Figure 2b). ICP-MS analysis corroborates this result revealing a Ti content of 14.2 ± 5.1 pg Ti per valve. In addition, the quantity of biogenic silica in the TiBALDH-modified T. weissflogii frustule is reduced in comparison to the unmodified diatom, with values of SiO2 per frustule of 28 pg and 51 pg respectively.

Figure 2
figure 2

Investigation of the effect of TiBALDH dosing on Ti content in frustule reveals increases in Ti content is achieved by multiple dosing of cultures every 48 hours.

EDX-SEM analysis was performed on the valve of T. weissflogii to quantify the ratio of Ti:Si allowing the Ti content per valve to be calculated. Cultures were treated with (a) a single addition or (b) multiple additions of TiBALDH. Data are presented as median and IQR with minimum and maximum values (n = 12). EDX-SEM spectra of (c) T. weissflogii showing a Si signal and (d) TiBALDH-modified T. weissflogii showing both a Si and Ti signal. (e) Incorporation of titanium into the frustule of T. weissflogii results in a gradient of titanium content across the girdle view. Data are represented as mean ± sem (n = 9). t-test revealed a statistical difference between Girdle 1 and Girdle 2. (f) Schematic illustrating the division of upper and lower girdle for analysis of Ti gradient across the girdle view by EDX-SEM analysis.

EDX-SEM measurements across the girdle band revealed the existence of a gradient in Ti content (Figure 2e). The weight ratio of Ti:Si gave values of 0.24:1 compared 0.34:1 between bands. The distinction between bands is illustrated schematically in Figure 2f. EDX-TEM analysis of 90 nm thick cross sections of T. weissflogii and TiBALDH-modified T. weissflogii further confirmed the presence of Ti within the frustule of the modified diatom (Figure 3a & 3c). These observations suggest that Ti incorporation is growth associated as depicted in Figure 4. Furthermore, EDX-SEM analysis confirmed the absence of Ti in the frustule of T. weissflogii cultured in the presence of TiBALDH and sodium azide (Supplementary Figure 1). XPS analysis performed on a bulk sample of the TiBALDH-modified diatom confirmed the chemical form of Si and Ti present as SiO2 and TiO2 (Figure 3b & 3d).

Figure 3
figure 3

Qualitative analysis of titanium in TiBALDH-modified T. weissflogii.

(a & c) EDX-TEM spectra and micrograph of 90 nm cross sections of fixed T. weissflogii and TiBALDH-modified T. weissflogii. Arrows indicate cell organelles encased within the frustule (b & d) XPS spectra of T. weissflogii and TiBALDH-modified T. weissflogii. The chemical form of Si and Ti present is SiO2 and TiO2.

Figure 4
figure 4

Incorporation of Ti into the diatom during frustule synthesis.

(a) The diatom is composed of an epitheca and hypotheca that fit together like a petri dish. Each theca is composed of the valve face (epivalve or hypovalve) and the valve mantle. The overlap region of the valves is surrounded by structures referred to as girdle bands and this region is referred to as the cingulum. A cross section of the diatom reveals that the pores present on the external valve face penetrate to the internal face. (b) Incorporation of Ti into both valves of the diatom frustule requires a minimum of two cell cycle divisions. The diatom cell cycle consists of an S phase where DNA synthesis occurs, followed by a gap in time (G2 phase). Upon silica uptake, two daughter cell protoplasts are synthesized in the silica deposition vesicle (SDV) of the mother cell before mitosis (M phase). This is followed by another gap (G1 phase) where daughter cells separate and growth via expansion of the girdle band occurs. (c) Depletion of TiBALDH in the culture media will lead to a decrease in Ti content. Failure to replenish TiBALDH in the culture media will result in decreased Ti content with successive cell divisions. Decaying diatoms or dissolution of silica from living diatoms in the culture media may provide a source of silica for uptake enabling continued cell divisions.

Architectural parameters of TiBALDH-modified T. weissflogii

T. weissflogii is a centric diatom measuring 10–15 μm in diameter with a central ring of fultoportulae and ribs that radiate to the periphery of the diatom (Figure 5). The valve surface is decorated with pores, with an average perimeter of 125 nm, that increase in density from the centre of the diatom to the periphery (Figure 6). Conversely, the rib-like structures become less dense at the periphery. Scanning electron microscopy (SEM) micrographs revealed that the overall frustule morphology of T. weissflogii was preserved under multiple additions of TiBALDH over 192 hours (Figure 5). The resemblance observed by SEM was supported by images generated by transmission electron microscopy (TEM) with the pores decorating the valve face appearing unaltered (Figure 6a & 6b). However, further analysis was required to elucidate whether the architecture of the TiBALDH-modified frustules were completely analogous to the morphology described for the T. weissflogii cultured in Na2SiO3. For this purpose, the valve surface was subdivided into four concentric regions and extensive characterization of the pore size and radial pore distribution using TEM images was performed. Neither the pore size nor the pore density across the valve of TiBALDH-modified frustules were significantly different to those treated with Na2SiO3 (Figure 6c & 6d). The architectural parameters were further analysed using atomic force microscopy (AFM) analysis (Figure 7 & 8). The density of pores within Regions 2 and 3 of the valve surface were not statistically different between Na2SiO3 and TiBALDH diatoms (Figure 6b). A preliminary investigation comparing data generated by TEM and AFM was performed to ensure the accuracy of the data collected by AFM. The spacing from pore edge-to-edge from three separate sections of a TiBALDH-modified diatom in Regions 2 and 3 quantified from TEM data was 72 ± 3 nm, 73 ± 3 nm and 68 ± 3 nm (Supplementary Figure 2). The average spacing from pore valley-to-valley in the same regions quantified from AFM data was 68 ± 3 nm (Supplementary Figure 2) in agreement with data from TEM. Further AFM analysis revealed that TiBALDH-modified diatoms did not differ from diatoms cultured in Na2SiO3 in parameters of pore valley-to-valley distance, pore depth, ironed surface area, spacing of the ribs, or rib height (Figure 8a–8e). However, there was an increase in the rib width (Figure 8f) in TiBALDH-modified diatoms as compared to those cultured in Na2SiO3. Previous studies have reported that the morphology of the precipitates obtained in vitro exhibit significant differences depending on whether Na2SiO3 or TiBALDH was used as a precursor12,15. The type of silaffin involved in the precipitation15, the secondary structure of the silaffin27, the amino acid sequence of the long-chain polyamine28 and the peptide sequence29 are hypothesized to play a role in the morphologies of the precipitates obtained in vitro. Whether these factors influence the rib morphology of the frustules obtained in the culture medium that contains TiBALDH is unclear and further investigation will be required to elucidate this issue.

Figure 5
figure 5

The characteristic hierarchical architecture of the diatom is conserved in TiBALDH-modified T. weissflogii.

Scanning electron microscopy images of T. weissflogii show that T. weissflogii is a centric diatom with highly intricate protrusions on the valve surface. Ribs radiate from the centre of the structure, underneath which the valve is decorated with openings. The images shown here are diatoms grown in the presence of (a) Na2SiO3 or (b) & (c) TiBALDH collected at 192 hours post inoculation following multiple addition of Na2SiO3 or TiBALDH at 48 hour intervals.

Figure 6
figure 6

The architectural properties related to the pores on the valve surface of T. weissflogii are unaltered in TiBALDH-modified diatoms.

Transmission electron microscopy images of (a) T. weissflogii and (b) TiBALDH-modified T. weissflogii. Scale bar represents 2 μm. (c) Pore parameters of perimeter, width, length and area were not significantly different between Na2SiO3 and TiBALDH treated diatoms. (d) The pore distribution decreases in density from the periphery of the diatom to the centre of the diatom similarly for Na2SiO3 and TiBALDH treated diatoms. The schematic illustrates regions analysed on the valve surface with Region 4 represented in the centre of the valve. All data was generated from diatoms collected at 192 hours post inoculation following multiple addition of Na2SiO3 or TiBALDH at 0, 48, 96 and 144 hours post inoculation (n = 3 diatoms per culture, 3 cultures per treatment).

Figure 7
figure 7

Fine nodular features decorate the rib structure on the valve of T. weissflogii.

The surface topography of (a) T. weissflogii and (b) TiBALDH modified T. weissflogii imaged by AFM reveals the fine structural details of the valve surface showing nodes decorating the ribs and the textured surface underneath the ribs. Images were generated from diatoms collected at 192 hours post inoculation following multiple addition of Na2SiO3 or TiBALDH at 0, 48, 96 and 144 hours post inoculation. (c) Measurements were collected in Regions 2 and 3 as defined in Figure 6d. The panel of graphs represent profiles taken on the valve surface depicting criteria for measurements of (i) valley-to-valley distance (ii) pore depth (iii) rib-to-rib distance (iv) rib width (green) and height (red).

Figure 8
figure 8

Architectural parameters relating to the rib structure on the valve surface are altered in TiBALDH modified T. weissflogii.

Data are quantified from AFM images and represented as mean ± sem (n = 4 diatoms per treatment). (a) Distance between neighbouring pores measured from valley-to-valley (b) Pore depth (c) Ironed surface area of a 0.5 μm2 section on valve surface (d) Distance between neighbouring ribs measured from rib crest-to-crest (e) Rib height (f) Rib width, t-test revealed significant difference for Na2SiO3 vs TiBALDH *p < 0.05. Data were generated from diatoms collected at 192 hours post inoculation following multiple addition of Na2SiO3 or TiBALDH at 0, 48, 96 and 144 hours post inoculation.

Novel properties of TiBALDH-modified T. weissflogii

In the present study, co-cultivation of TiBALDH-modified T. weissflogii in the presence of Escherichia coli (E. coli) provided interesting results. The model bacterial strain of choice E. coli was used in this study. Cultures were inoculated in sterile de-ionised water so as to provide no nutrients to either diatom or bacteria. The abundance of either bacteria or diatom present was quantified over the time frame of the experiment. Co-cultures of (i) T. weissflogii with E. coli or (ii) TiBALDH-modified T. weissflogii with E. coli were studied with or without illumunation over a 24-hour period. Axenic cultures of (i) T. weissflogii, (ii) TiBALDH-modified T. weissflogii and (iii) E. coli were studied under identical experimental conditions to serve as a control.

Both unmodified and TiBALDH-modified diatoms were capable of preserving the number of bacteria over the first six hours irrespective of exposure to UV light (Figure 9). These results are in agreement with what is known in the field, that diatoms tend to promote the growth of bacteria30,31,32,33. A close inspection of the changes after the first six hours revealed that there was a significant decay in the number of colonies in co-cultures of E. coli with TiBALDH-modified T. weissflogii following exposure to UV light. Any decay of E. coli in the presence of TiBALDH-modified diatoms disappeared when maintained in the dark. The profile of E. coli colony counts in the presence of unmodified T. weissflogii did not change under either light or dark settings demonstrating a slight decrease between 12 and 24 hours post inoculation. The decay of E. coli in co-cultures with diatoms has no precedent and this is the first such observation reported to date. The decay of E. coli in the presence of TiBALDH-modified diatoms was expected to be due to the photocatalytic activity of TiO2 as it is absent in cultures maintained under dark settings.

Figure 9
figure 9

TiBALDH-modified T. weissflogii decreases E. coli density under illumination but not under dark conditions.

E. coli colony count expressed as a percentage of the starting colony count for co-cultures under illumination and dark conditions. Data are expressed as mean ± sem (n = 3).

The photocatalytic ability of cleaned TiBALDH-modified T. weissflogii frustules was also investigated. Degradation of methylene blue under UV light in the presence of either unmodified or modified diatoms was monitored. A decrease in the absorbance of the methylene blue solution at 656 nm was observed only in the presence of TiBALDH-modified diatoms following exposure to UV light (Supplementary Figure 3).

Discussion

The preservation of the diatom growth pattern upon the chemical modification of the culture medium is an aspect of paramount importance. Exposure of diatoms to sub-lethal doses of alternative precursors results in alterations to the architecture of the diatom, explained in part by interruptions to the processes within the silica deposition vesicle (SDV) and modification of cell organelles34,35. Accordingly, the threshold concentration of chemical entities that can be incorporated into the culture medium to modify the diatom morphology and/or composition must fulfil a delicate balance: sufficiently high that any modification is detectable, yet sufficiently low to avoid cytotoxic concentrations that alter diatom growth patterns. In this study, TiBALDH added to the culture at a final concentration of 200 μM met these criteria.

The increase in cell density observed in the presence of TiBALDH not only confirms that the presence of Ti does not adversely affect the growth pattern of T. weissflogii, but also suggests that Ti is capable of being incorporated into the diatom frustule by a metabolically associated process. Further, exploration of the exact mechanism is required but is hampered by the lack of Ti specific inhibitors.

This Ti content achieved using TiBALDH is enhanced compared to that seen with TiOSO4. However; the TiBALDH treatment was insufficient to create a fully SiO2–depleted valve. Figure 4b illustrates the formation of new valves within the silica deposition vesicle in the parent diatom. Incorporation of the precursor requires uptake of the precursor into the parent diatom, deposition of the precursor within the SDV, followed by division of the parent diatom into two daughter diatoms. The daughter diatom consists of an epitheca composed of original material from the parent diatom and a hypotheca composed of material from the precursor. Incorporation of the precursor into both valves of the frustule requires a minimum of two cell cycles (Figure 4c). As additional Na2SiO3 was not added to TiBALDH treated cultures, one would expect the formation of six fully SiO2–depleted frustules after three cell cycles from one parent SiO2–diatom (Figure 4c). However, this was not observed experimentally.

Titanium dioxide particles have been shown to have bactericidal properties when irradiated with UV light36,37,38, however this mechanism is a matter for discussion and possible suggestions including oxidation of intracellular enzymes leading to decreases in bacterial cell respiration and death36, or disruption of the bacterial cell membrane leading to cell death39. Nonetheless, the generation of free radicals is the root cause of the bactericidal properties of TiO2 particles. Irradiation of TiBALDH-modified T. weissflogii in co-culture with E. coli leads to a decrease in bacteria abundance. This behaviour is not observed in co-cultures of T. weissflogii and E. coli and is expected to be due to photocatalysis of TiO2. As an alternative to photocatalysis, it can be postulated that the difference between E. coli colony numbers in the presence of either unmodified or TiBALDH-modified diatoms under illumination is due to metabolic differences. However, the lack of an observed difference between the co-culture systems under dark conditions (Figure 9) does not support this idea and indicates a photocatalytic mechanism underlying the differences observed under UV irradiation.

In summary, the use of TiBALDH as a precursor in diatom cultures allowed for a metabolic insertion of up to 14.2 ± 5.1 pg Ti per diatom valve. TiBALDH was chosen because of its stability in the culture medium, which is one of the critical issues that favours precursor uptake. Enhanced metabolic insertions were found with experimental conditions (e.g. multiple precursor additions) that prolong the exponential growth phase of the culture. The resemblance (in terms of both the pore size and distribution across the valve) between TiO2-modified diatoms and regular diatoms suggests that, in vivo, silaffins and polyamines are determinants of TiBALDH precipitation into the patterned structure that characterizes diatom frustules. It is worth noting that the modification of the chemical composition of living diatoms have several implications. For instance, results generated in this work revealed how the suppression or the preservation of T. weissflogii in co-cultures with E. coli depended upon chemical composition. Moreover, co-cultures of TiBALDH-modified T. weissflogii diatoms exhibited an unprecedented decay of E. coli under illumination but not in dark conditions indicative of a certain photocatalytic activity originating from TiO2 incorporated into the diatom frustules. While further investigation is clearly warranted to fully explore the potential of these features, the photocatalytic activity of living cultures of TiBALDH-modified T. weissflogii opens exciting possibilities in a number of applications. For instance, TiBALDH-modified diatoms can exert bactericidal effects. In this context it is worth noting that the TiO2 concentration in the diatoms studied herein is nearly 1000-fold below that used in previous reports on photocatalytic TiO2 particles exhibiting bactericidal effects39.

Methods

Axenic Thalassiosira weissflogii cultures were grown in artificial seawater (ASW) prepared according to Berges et al.40 enriched with Guillard's f/2 marine enrichment media without silicates (Sigma Aldrich) according to manufacturer recommendations. Cultures were silica depleted according to the procedure detailed by Hildebrand et al.41 for a minimum of 24 hours before inoculation with the precursor. Following silica depeletion cultures were inoculated at 1 × 104 cells/mL or 5 × 104 cells/mL in a final volume of 200 ml. Cultures were grown in polystyrene tissue culture flasks.

Sodium metasilicate nonhydrate (Na2SiO3) or titanium bis(ammoniumlactacodihydroxide) (TiBALDH) were added to cultures at a final concentration of 200 μM and grown at a 14 hour:10 hour light:dark cycle at a light intensity of 3000 lux and temperature range of 16–22°C. Multiple addition cultures received further additions of Na2SO3 or TiBALDH at a final concentration of 200 μM at 48, 96 and 144 hours. Cultures were collected at 192 hours post inoculation and diatoms cleaned according to procedure detailed in Supplementary Information. Cell density was monitored using a haemocytometer.

Determination of Ti content in cleaned frustules

SEM-EDX analysis was performed using Hitachi S-4700 SEM with INCA software (Oxford Instruments) to determine the Si and Ti content. Cleaned diatoms suspended in methanol were allowed to air dry on a carbon stub and were subsequently gold coated. Diatoms were analysed only if the valve view was clearly visible. Three diatoms per culture and a minimum of three cultures per treatment group were analysed. The ratio of Ti:Si was determined and represented as pg Ti per valve. The gradient of Ti across the diatom was analysed using the girdle view of the diatom. Nine diatoms were analysed with two spectra collected per diatom identified as girdle 1 and girdle 2. t-tests were performed to determine statistical difference between Si and Ti content across the girdle of the frustule. TEM-EDX analysis was performed on a 90 nm thick cross-sections of fixed T. weissflogii and fixed TiBALDH-modified T. weissflogii as detailed in Supplementary Information. ICP-MS analysis was performed using ICP-MS Elan 6000 Perkin Elmer-Sciex. Digestion of the samples was performed by microwave (high pressure microwave, model ETHOS SEL, Milestone) using equal part solution of HNO3 and HF. XPS analysis was performed using the Kratos AXIS 165 spectrometer. Binding energies were referenced to the C 1 s line at 284.5 eV and 284.8 eV. Cleaned diatom samples were dried at 60°C for 48 hours. The dried sample was ground to a fine powder and dusted on to double sided adhesive tape for analysis.

Quantification of architectural parameters

Quantification of architectural parameters was performed on cleaned frustules grown in the presence of either Na2SiO3 or TiBALDH collected at 192 hours post inoculation following multiple addition of Na2SiO3 or TiBALDH at 48 hour intervals. TEM images of frustules were collected using Hitachi H-7500 TEM with AMT image capture software. Cleaned diatoms suspended in methanol were allowed to air dry on a copper grid. Pore parameters (perimeter, width, length and area) were quantified using ImageJ software. Five sections per diatom and five diatoms per culture were analysed. A minimum of three cultures per treatment group were analysed. Pore parameters were calculated as the mean ± sem per treatment (n = 3). Pore distribution across the valve surface was analysed using ImageJ software. The valve surface was divided into four discrete regions as illustrated in the schematic in Figure 6d. A minimum of eight sections measuring 0.5 μm2 per region were analysed. Pore distribution was calculated as the mean ± sem per treatment (n = 3). AFM measurements were performed under ambient conditions in intermittent contact mode using a commercial AFM system (NanoWizard-II, JPK Instruments, Germany) coupled with an inverted optical microscope (Eclipse Ti-E, Nikon, Japan). Silicon cantilevers (spring constant, k ~ 2.8 N m−1 and resonance frequency, f ~ 75 kHz) with high aspect ratio (1:10 aspect ratio, tip radius < 3 nm), high density, diamond like carbon tips were used (MSS-FMR-13, Nanotools, Germany). Analysis of AFM images was performed using WSxM Software42. Sections within regions 2 and 3 as outlined in the schematic in Figure 7c were analysed. The criteria for measurements of valley-to-valley distance, pore depth, rib-to-rib distance and rib width and rib height is illustrated in Figure 6c. t-tests were performed to determine statistical difference between Na2SO3 and TiBALDH treated cultures regarding architectural parameters.

Investigating the relationship between T. weissflogii/TiBALDH-modified T. weissflogii and E.coli

Escherichia coli, T. weissflogii and TiBALDH-modified T. weissflogii were prepared for the co-culture study as detailed in Supplementary Information. Six treatment groups were prepared to a final volume of 200 ml in sterile de-ionised water; (i) Water, (ii) E. coli, (iii) T. weissflogii, (iv) TiBALDH-modified T. weissflogii, (v) T. weissflogii + E. coli, (vi) TiBALDH-modified T. weissflogii + E. coli. Flasks were inoculated at diatom cell density of ca. 3 × 104 cells/mL and E. coli density of ca. 106 cfu/mL. Flasks were placed on a shaker and exposed to light (HALOLINE ECO 64695 from Osram) or covered with black cloth to eliminate exposure to light. 200 μl aliquots were removed at 0, 1.5, 3, 6, 12 and 24 hours post inoculation. Diatom cell counts were performed using a haemocytometer. Serial dilutions in sterile PBS were performed before plating 100 μl on LB agar plates. Plates were incubated at 37°C for 24 hours and colony counts recorded.

Investigating the degradation of methylene blue in the presence of T. weissflogii or TiBALDH-modified T.weissflogii

Cleaned T. weissflogii and TiBALDH-modified T. weissflogii were suspended in de-ionised water. Methylene blue dissolved in de-ionised water was added at a weight ratio of 0.1 mg dye:100 mg diatom. Six samples were prepared for both modified and unmodified diatoms. All samples were placed in the dark for 1 hour afterwhich they were centrifuged at 2500 g for 20 minutes. The absorbance of the supernatent at 656 nm was measured to ensure that there was no difference between samples before the incubation period. Three samples for each unmodified and modified diatom were placed in the dark for 24 hours. The remaining samples were exposed to UV light at 365 nm. The absorbance of the supernatent was measured at 656 nm following the 24 hour incubation period.