Engineering Bacillus subtilis for the formation of a durable living biocomposite material

Materials and chemicals

Reagent grade tetraethyl orthosilicate (TEOS), ammonium molybdate tetrahydrate, 4-methylaminophenol sulfate, anhydrous sodium sulfite, all acids and bases, and other chemicals were purchased from Sigma-Aldrich (Sigma-Aldrich Corp., MO, USA).

Remel RYU flagella stain, polysine microscope slides, plain precleaned glass microscope slides, Alexa Fluor^TM 488 C₅ maleimide, Syto16, and Trump’s Fixative were purchased from ThermoFisher Scientific (Waltham, MA, USA). 22 mm 1.5 cover glasses were purchased from Azer Scientific (Morgantown, PA, USA). 200 mesh copper grids with 10 nm formvar and 1 nm carbon, aqueous 2% uranyl acetate, and EM grade: 10% glutaraldehyde, 0.2 M sodium cacodylate trihydrate pH 7.4, 16% paraformaldehyde, and 4% osmium tetraoxide were purchased from VWR International, an Electron Microscopy Sciences reseller (Radnor, Pennsylvania, USA). SynaptoRed^TM C2 was purchased from Avantor delivered by VWR a Biotium reseller (Radnor, PA, USA). FluoroShield was used in fluorescent microscopy experiments and was purchased from AbCam (Cambridge, MA, USA). Silicon wafers were used for SEM and obtained from Ted Pella (Reeding, CA, USA). Schaeffer and Fulton Spore stain kit (containing malachite green 50 g/L and safranin O 5 g/L) was obtained from MilliporeSigma (Burlington, MA, USA).

Spectra/Por Dialysis Tubing (MWCO 6-8 kDa) used for dialysis of proteins was purchased from Spectrum LifeSciences (Rancho Dominguez, CA, USA). HisTrap™ FF columns (GE Healthcare, IL, USA) were used for all protein purifications except for proteins used for TEM that were purified from E. coli and secreted proteins from Bacillus that were purified using TALON metal affinity resin from TaKaRa Bio USA, Inc. (Mountain View, CA, USA). Two mL of TALON metal affinity resin was used in a Batch/Gravity flow protein purification protocol following manufacturer’s directions. Pierce™ BCA Protein Assay Kit used for protein concentration measurements was purchased from ThermoFisher Scientific (Waltham, MA, USA). Ultrapure water was prepared by filtering deionized water through a Milli-Q water purification system (Millipore, Billerica, MA, USA) to reach a final electrical resistance higher than 18.2 MΩ cm^-1.

All supplies were purchased from New England Biolabs (NEB) (Ipswhich, MA, USA). Q5® High-Fidelity DNA polymerase was used for PCR amplifications. The Q5® Site-Directed Mutagenesis kit was used for site-directed mutagenesis and the NEBuilder® HiFi DNA Assembly Master Mix was used for Gibson assembly of DNA fragments, and restriction enzymes (BglII, BamHI, and PstI) and T4 DNA Ligase were used for general cloning. Primers were ordered from Integrated DNA Technologies (Coralville, IA, USA).

40% acrylamide and bis-acrylamide solution (37.5:1), TEMED and Precision plus all blue prestained protein standard (catalog# 161-0373) were purchased from Bio-Rad (Hercules, CA, USA). Other chemicals for SDS-PAGE analysis and Coomassie brilliant blue staining were purchased from Sigma-Aldrich.

Bacterial strains, media, and general molecular biology methods

Escherichia coli TOP10 (Invitrogen, Carlsbad, CA, USA) was used for cloning and plasmid propagation. E. coli T7 Express (C2566) (NEB) and used for protein expression and purification. Bacillus subtilis 168 (ATCC 23857) was used as a host strain (referred to as WT strain). All plasmids and strains used in this study are listed in Supplementary Data 1.

LB (Luria broth; Tryptone 10 g/L, NaCl 5 g/L, Yeast extract 5 g/L) medium was used for the growth of E. coli and B. subtilis strains.

B. subtilis strains were also grown in modified Spizizen’s minimal medium (SMM)^73,74, containing, per liter: K₂HPO₄–17.5 g, KH₂PO₄–7.5 g, Na₃-Citrate 2 H₂O–1.25 g, MgSO₄ 7 H₂O – 0.25 g, N-source (L-glutamine–10 g, (NH₄) ₂ SO₄–0.5 g), trace elements (CaCl₂ 2 H₂O – 7.35 mg, MnCl₂ 4 H₂O–1 mg, ZnCl₂–1.7 mg, CuCl₂ 6 H₂O–0.43 mg, CoCl₂ 6 H₂O–0.6 mg, Na₂MoO₄ 2 H₂O–0.6 mg), Fe-Citrate (FeCl₃ 6 H₂O–1.35 mg, Na₃-Citrate 3 H₂O–10 mg), glucose–5 g, tryptophan–50 mg.

SMM was prepared by mixing stock solutions which included, 5X SMM base (K₂HPO₄–87.5 g/L, KH₂PO₄–37.5 g/L, Na₃-Citrate∙2H₂O–6.25 g/L, MgSO₄ ∙ 7H₂O–1.25 g/L), N-source (sterilized using a 0.2 µm filter) (glutamine–40 g/L, (NH₄)₂SO₄–2 g/L), 100X trace elements (CaCl₂ ∙ 2H₂O–0.735 g/L, MnCl₂ ∙ 4H₂O–0.1 g/L, ZnCl₂–0.17 g/L, CuCl₂ ∙ 6H₂O–0.043 g/L, CoCl₂ ∙ 6H₂O–0.06 g/L, Na₂MoO₄ ∙ 2H₂O–0.06 g/L), 100X Fe-citrate (sterilized using a 0.2 µm filter) (FeCl₃ 6 H₂O–0.135 g/L, Na₃-Citrate.3H₂O–1 g/L), 50% glucose (500 g/L) and tryptophan (sterilized using a 0.2 µm filter) (5 g/L).

Fresh solution of sterile N-source and glucose were prepared every time while stocks of other solutions were stored at 4 °C. The working solution of SMM was prepared by mixing 200 mL of 5X SMM, 10 mL of 100X trace element, 10 mL of 100X Fe-Citrate, 10 mL of 50% glucose, 10 mL from 5 mg/mL stock of Tryptophan, and 250 mL of N-source into 510 mL of sterile water.

SM1 and SM2 media as described previously by Bennallack et al.,⁷⁵ were used for the preparation of B. subtilis competent cells.

For plasmid maintenance, LB was supplemented with 100 μg/mL ampicillin for E. coli strains while LB and SMM were supplemented with 10 μg/mL tetracycline or 20 μg/mL erythromycin for Bacillus strains.

Transformation of E. coli with plasmids followed standard molecular biology techniques. Transformation of B. subtilis strains was done using a method described by Bennallack et al.⁷⁵. Transformants were screened by colony PCR using a method modified from Jamal et al.⁷⁶. To prepare samples for colony PCR, a single colony of B. subtilis was mixed in 300 µL sterile water. Cells were spun down (21,130 x g) and 200 µL of the supernatant was removed. Cells were resuspended in the remaining 100 µL of sterile water followed by 20 mins of sonication in a coldwater bath sonicator (Bransonic 3510R-MTH, CT, USA). After sonication, cells were spun down as above and 2 µL of the supernatant was used as a template for a 10 µL PCR reaction. Plasmids from positive B. subtilis transformants were isolated and their sequences verified by Sanger sequencing. Recombinant strains with verified plasmid sequences were stored as glycerol stocks and used in this work.

Construction of plasmids and in-frame deletion in B. subtilis

All plasmids and strains used in this study are described in Supplementary Data 1. Plasmid sequences were verified by Sanger sequencing (ACGT, Inc. Wheeling, IL, USA). Amino acid sequences and nucleotide sequences for all proteins and genes reported in this work are found in Supplementary Data 2 and 3. Primer pairs for Bacillus genomic manipulations and flagellin engineering are listed in Supplementary Data 4.

Plasmids were constructed using a combination of Gibson Assembly (HiFi® DNA assembly kit, NEB) for fragment assembly, site-directed mutagenesis (Q5® mutagenesis kit, NEB) for shorter insertions, deletions, and insertions, or restriction enzyme cloning and ligation for larger fragment assembly following manufacturers’ instructions and as described previously^33,41.

Briefly, for the C-terminal fusion of biomineralization peptides to His-EutM in pCT5-His-EutM^33,41, primers were designed to insert the corresponding into the plasmid using the Q5® mutagenesis kit according to the manufacturers’ instructions.

To construct the pCT-His-dtTomato-SpyCatcher plasmid, His-EutM was combined with the fluorescent protein in a previously constructed pCT5-His-EutMSpyC plasmid³⁵ by HiFi Assembly. His-tdTomato-SpyCatcher was amplified from tdTomato-pBAD⁷⁷.

To create the EutM Bacillus expression vectors, His-EutM-SpyCatcher was amplified from pCT5-His-EutMSpyC and inserted into our in-house, cumate inducible Bacillus expression vector pCT5-bac2.0⁴¹ by HiFi Assembly. For a control plasmid, pCT-empty was created by deletion of the sfgfp gene from pCT5-bac2.0 using the Q5® mutagenesis kit according to the manufacturers’ instructions. All signal peptides (Supplementary Table 1) were fused to His-EutM-SpyCatcher by site-directed mutagenesis (Q5® mutagenesis kit, NEB) with appropriate primers. The pCT-EutMCotB plasmid for Bacillus expression was made by replacing the SpyCatcher sequence in pCT-EutMSpyC with the CotB sequence by site-directed mutagenesis. To create pCT-EutMCotB-EutMSpyC, the complete SacB-His-EutM-SpyCatcher expression cassette (including cumate inducible promoter) from pCT-EutMSpyC was amplified and subcloned into pCT-EutMCotB. The hag^T209C::SpyTag⁵⁸⁸ expression cassette (including its native promoter) was added to pCT-EutMCotB-EutMSpyC by HiFi assembly with the hag expression module amplified from the corresponding pBBRm34 plasmid (see below). Finally, to construct pCT-Purple-Hag^T209C::SpyT⁵⁸⁸, the chromoprotein was obtained from pSB1C3-tsPurple⁷⁰ and assembled into pCT-EutMCotB-EutMSpyC-hag^T209C::SpyT⁵⁸⁸ by replacing the EutM expression modules.

The flagellin gene (hag gene) was amplified from B. subtilis genomic DNA using primer pair (NP-hag_fwd and NP-hag_rev) and cloned into the pRBBm34 backbone using HiFi assembly. The flagellin T209C mutation was introduced by Q5®-mutagenesis with primer pairs (NP-hag-T209C_fwd and NP-hag-T209C_rev). A SpyTag was inserted into nine different sites between D1b-N and D1-C domains using Q5®-mutagenesis with primer pairs listed in Supplementary Data 4.

To generate the ΔlytC marker-less deletion construct, the pHBintE vector carrying a temperature-sensitive origin of replication, erythromycin and ampicillin resistance was used as the backbone. The lytC upstream region was amplified from B. subtilis genomic DNA using primer pair dlytC_upstream_fwd and dlytC_upstream_rev, while the lytC downstream region was amplified using primer pair dlytC_downstream_fwd and dlytC_downstream_rev (Supplementary Data 4) and the deletion cassette was cloned into pHBintE vector using HiFi assembly. The plasmid for ΔflhG marker-less deletion was created in a similar manner: The flhG upstream region was amplified using primer pair dflhG_upstream_fwd and dflhG_upstream_rev while the flhG downstream region was amplified using primer pair dflhG _downstream_fwd and dflhG_downstream_rev (Supplementary Data 4). The flhG deletion cassette was cloned into pHBintE vector using HiFi assembly.

The lytC deletion plasmid (temperature-sensitive pHBintE-dlytC) was transformed into B. subtilis using the method described by Bennallack et al.⁷⁵. After transformation, plates were incubated at a temperature permissive for plasmid replication (30 °C) and for selection of erythromycin-resistant transformants. Colonies were then transferred into 4 mL LB medium without antibiotic added followed by several rounds of dilution and regrowth at a temperature nonpermissive for plasmid replication (37 °C). For screening of the deletion mutant appropriate dilutions of the cells were plated onto LB agar. Colonies were streaked on LB plates with and without erythromycin to identify erythromycin sensitive colonies that have successfully evicted the plasmid. Screening the erythromycin sensitive colonies for desired double homologous recombinants was performed using colony PCR with primer pair dlytC_confirmation_fwd and dlytC_confirmation_rev as shown in Supplementary Data 4. Transformation of flhG deletion construct was done into B. subtilis ΔlytC host using the pHBintE-dflhG plasmid. In-frame deletion of flhG followed by screening was performed with the same method as above for lytC but the primer pair used for screening of colonies for flhG deletion mutants was dflhG_confirmation_fwd and dflhG_confirmation_rev.

Cultivation of recombinant B. subtilis strains

Glycerol stocks of recombinant Bacillus strains (WT and ΔlytC ΔflhG deletion strain) transformed with pCT expression plasmids (pCT-empty, pCT-EutMSpyC, pCT-EutMCotB, pCT-EutMCotB-EutMSpyC, pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸) were streaked onto LB plates supplemented with tetracycline (10 μg/mL) and grown overnight at 30 °C. A single colony was used to inoculate 4 mL of selective LB. After 19 h of growth at 30 °C, 220 rpm strains were diluted 3% into 50 mL of LB or SMM supplemented with tetracycline (10 μg/mL) in a 250 mL flask. The flasks were cultured at 37 °C, 180 rpm until OD₆₀₀ reached 0.4–0.7 and protein expression was induced with 10 μM cumate. The induced cultures were then grown at 20 °C or 30 °C, 100 rpm for 48 to 96 h as described in this study. Final scaffold building block expression and secretion was done in SMM with 10 µM cumate for induction and subsequent growth at 20 °C, 100 rpm for 48 h (referred to as ‘standard conditions’).

For co-cultivation experiments, glycerol stocks of three B. subtilis ΔlytC ΔflhG strains transformed with pCT-empty, pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ or pCT-Purple- Hag^T209C::SpyT⁵⁸⁸ were first spread onto selective LB plates, and single colonies were used to inoculate 4 mL of selective LB. After 19 h of growth at 30 °C, cultures were diluted 3% into 50 mL of selective SMM in a 250 mL flask. Individual cultures were then grown at 37 °C until OD₆₀₀ 0.4–0.7. 25 mL of each culture was then mixed in a new 250 mL flask as follows: Co-culture 1: B. subtilis ΔlytC ΔflhG pCT-empty + B. subtilis ΔlytC ΔflhG pCT-Purple-Hag^T209C::SpyT⁵⁸⁸, Co-culture 2: B. subtilis ΔlytC ΔflhG pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ + B. subtilis ΔlytC ΔflhG pCT-Purple-Hag^T209C::SpyT⁵⁸⁸. Protein expression was then induced with 10 μM cumate and co-cultures grown at 20 °C, 100 rpm for 48 h.

All expression and cultivation experiments were performed with three biological replicates (i.e. three independent cultures).

Analysis of scaffold building block expression and secretion by Bacillus strains

Extracellular secretion and intracellular accumulation of scaffold building blocks by recombinant Bacillus strains was analyzed by preparing four different fractions from cultures for SDS-PAGE analysis: (i) culture supernatant, (ii) urea supernatant containing urea solubilized scaffolds that co-precipitated with cells, (iii) lysed cell pellets after solubilization of co-precipitated scaffolds, and (iv) total protein (secreted-precipitated and intracellular) from completely lysed cells with sonication followed by lysozyme treatment. The preparation of all four protein fractions is shown as a flow chart in Supplementary Fig. 5a. Fractions i-iii correspond to samples A, B, and C in Figs. 2, 3, and 8. Comparison of protein bands in fraction ii (i.e., absence of bands except for solubilized EutM protein) and iv indicates that urea does not break open cells Supplementary Fig. 5b.

The following workflow was used for sample preparation: 5 mL of culture broth was centrifuged at 3220 x g, at 4 °C for 10 mins to separate culture supernatant from pelleted cells and co-precipitated scaffolds.

For SDS-PAGE analysis of soluble scaffolds in the culture supernatant (fraction i, or A in Fig. 2), proteins from 1 mL of supernatant were precipitated by adding 200 μL of 100% Trichloroacetic acid (TCA) and 1 h incubation on ice. Precipitated proteins were centrifuged at 21,130 x g for 10 mins at 4 °C and the supernatant was removed. The precipitate was washed with 1 mL of pre-chilled acetone and the tube was centrifuged again at 21,130 x g for 10 mins at 4 °C. The acetone was removed and the washing step was repeated twice. The dried precipitate was resuspended in 100 μL of 1X SDS-loading buffer (2% w/v SDS, 0.1 M DTT) (6X standard SDS-loading buffer diluted with 50 mM Tris-HCl, 4 M urea, pH 7.5) for a 10-fold concentrated sample.

For SDS-PAGE analysis of secreted scaffolds that co-precipitate with cells after centrifugation of 5 mL cultures (fraction ii), the collected pellet was carefully resuspended in 1 mL of urea buffer (4 M urea, 50 mM Tris-HCl, pH 7.5). The resuspended solution was centrifuged at 3,220 x g at 4 °C for 10 mins and the supernatant was analyzed by SDS-PAGE.

For SDS-PAGE of the remaining scaffolds in the urea washed pellet (fraction iii), cells were lysed with 200 μg/mL final concentration of lysozyme in 500 μL of 50 mM Tris-HCl, pH 7.5 with incubation at 37 °C for 1 h. The clarified supernatant after centrifugation (3,220 x g, 10 mins) was analyzed by SDS-PAGE.

For SDS-PAGE of total protein (fraction iv), pelleted cells and co-precipitated scaffolds from 5 mL of culture was resuspended into 500 μL of 50 mM Tris-HCl pH 7.5 in an eppendorf tube. The suspension was sonicated in a coldwater bath sonicator (Bransonic 3510R-MTH, CT, USA) followed by lysis with lysozyme as above. The resulting suspension was mixed with SDS-loading buffer.

All protein samples were denatured by boiling for 10 mins at 100 °C and 8 μL of each sample was loaded onto a 15% SDS-PAGE gel. A PowerPAC 300 power supply (Bio-Rad) was used for SDS-PAGE analysis. Coomassie brilliant blue staining was then used for visualization of protein bands.

De novo sequencing of secreted EutM-SpyCatcher proteins

For de novo sequencing of secreted EutM-SpyCatcher, protein bands separated by SDS-PAGE were cut from gels and submitted to the Center for Mass Spectrometry & Proteomics, University of Minnesota for analysis by LC-MS using a Thermo Scientific LTQ Orbitrap Velos mass spectrometer. Data were analyzed and viewed using the PEAKS Xpro Studio 10.6 software (Bioinformatics Solutions Inc., Waterloo, ON, Canada).

Analysis of EutM-SpyCatcher attachment to Hag^T209C::SpyTag⁵⁸⁸ displaying flagella

Covalent isopeptide bond formation between secreted EutM-SpyCatcher and Hag^T209C::SpyTag⁵⁸⁸ displayed on flagella of B. subtilis ΔlytC ΔflhG was analyzed by SDS-PAGE of sheared flagella isolated from co-cultures. Glycerol stocks of three B. subtilis ΔlytC ΔflhG strains (pCT-empty, pCT-EutMSpyC, pRBBm34-Hag^T209C::SpyT⁵⁸⁸) were first spread onto selective LB plates, and single colonies were used to inoculate 4 mL of selective LB liquid cultures. After 19 h of growth at 30 °C, cultures were diluted 3% into 50 mL of fresh, selective SMM and grown at 37 °C until OD₆₀₀ of 0.4–0.7 was reached. 25 mL of each culture was then mixed in a new 250 mL flask as follows: Co-culture 1: B. subtilis ΔlytC ΔflhG pCT-empty + B. subtilis ΔlytC ΔflhG pRBBm34-Hag^T209C::SpyT⁵⁸⁸, Co-culture 2: B. subtilis ΔlytC ΔflhG pCT-EutMSpyC + B. subtilis ΔlytC ΔflhG pRBBm34-Hag^T209C::SpyT⁵⁸⁸. Protein expression was induced with 10 μM cumate and co-cultures grown at 20 °C, 100 rpm for 96 h.

Cell pellets were collected from 50 mL cultures by centrifugation (3220 x g, 30 mins) and resuspended in 20 mL PBS (pH 7.5). The suspension was then sonicated at low power (30% power, 5 sec on and 5 sec off, three times) (Branson 450 Digital Sonifier, CT, USA) to shear off flagella. Turbo^TM DNase (Invitrogen, Waltham, MA, USA) was then added at a 0.5 U/mL final concentration and incubated on ice for 10 mins. Cells were then removed by centrifugation at 3220 x g for 10 mins at 4 °C, followed by an ultra-centrifugation step (39,800 x g for 2 h at 4 °C) to collect sheared flagella which were then resuspended in 100 µL 50 mM Tris-HCl (pH 7.5) for SDS-PAGE analysis.

Protein purification

His-EutM, His-EutM-silica binding peptide fusion proteins and His-EutM-SpyCatcher were expressed in E. coli and purified by metal affinity chromatography as described in detail in:^33,34,35,40. Briefly, proteins were isolated from 500 mL recombinant E. coli C3566 LB cultures (37 °C, 220 rpm) transformed with pCT5-plasmids encoding EutM proteins (Supplementary Data 1). Gene expression was induced with 50 μM cumate at OD₆₀₀ of 0.4–0.7 and cultures continued to be incubated overnight. Cells were collected by centrifugation (4000 x g, 30 min, 4 °C) and resuspended into either 30 mL EutM purification buffer (20 mM Tris-HCl, 250 mM NaCl, 5 mM imidazole, 4 M Urea, pH 7.5) for His-EutM and His-EutM-silica binding peptide fusion proteins, our EutM-SpyCatcher purification buffer (50 mM Tris-HCl, 250 mM NaCl, 20 mM imidazole, pH 8.0)³⁵. Resuspended cells were lysed by sonication on ice (30 min, power 50%, pulse on 10 s, pulse off 20 s with a Branson Sonifier) followed by centrifugation (12,000 x g, 40 min, 4 °C) and passing of the supernatant through 0.2 μm ultrafilter. The soluble protein was then loaded onto a 5-mL HisTrap^TM HP column (GE Healthcare Life Sciences, Pittsburgh, PA) using an ÄKTA FPLC system. Bound protein was eluted with 250 mM imidazole added to the purification buffer. His-tdTomato-SpyCatcher protein was expressed and purified using the EutM-SpyCatcher protocol³⁵.

For silica binding assays and SEM analysis of purified His-EutM and His-EutM silica binding fusion protein, purified proteins were dialyzed against deionized water.

For TEM analysis of EutM-SpyCatcher scaffold formation in SMM, proteins expressed in E. coli or secreted by Bacillus cultures were purified with a 2 mL TALON metal affinity resin using a Batch/Gravity flow protein purification protocol following manufacturer’s directions (TaKaRa Bio USA, Inc., see above). EutM-SpyCatcher protein from Bacillus was eluted with the same elution buffer (50 mM Tris-HCl, 250 mM NaCl, 250 mM imidazole pH 8.0) used for His-EutM-SpyCatcher as previously described^33,34,35,40. EutM-SpyCatcher from E. coli was isolated from 125 mL cultures using the EutM-SpyCatcher protocol³⁵ but with a TALON gravity column instead of a HisTrap™ FF column (GE Healthcare, IL, USA). For the purification of His-EutM-SpyCatcher proteins secreted by B. subtilis ΔlytC ΔflhG, recombinant strains harboring pCT-EutMSpyC were grown and induced under standard conditions described above. Cells and scaffolds from twelve 50 mL cultures (total volume: 600 mL) were pelleted at 3,220 x g, at 4 °C for 10 mins and scaffolds solubilized into 30 mL of 50 mM Tris-HCl, pH 7.5 containing 4 M urea as described above. The solubilized scaffolds were then purified by TALON metal affinity chromatography as described above for E. coli. Purified protein from E. coli or Bacillus cultures were then dialyzed against SMM after elution from the TALON metal affinity resin.

For GFP labeling of scaffolds in B. subtilis ΔlytC ΔflhG transformed with EutMCotB–EutMSpyC–Hag^T209C::SpyT⁵⁸⁸ plasmid, His-tagged eGFP and SpyTag-eGFP were expressed in E. coli from pET28 expression plasmids and then purified by metal affinity chromatography as described in³⁵.

Preparation of hydrolyzed TEOS as silica source

Silicic acid solutions (i.e. silica) were prepared by hydrolyzing TEOS. TEOS was hydrolyzed by adding water and 1 M hydrochloric acid at a 1:5.3:0.0013 molar ratio of TEOS:water:HCl and stirring vigorously. Under these conditions, hydrolysis is complete after 80 mins⁷⁸ for use as silica source in silica binding and biomineralization experiments.

Silica binding assays on EutM scaffolds

Silica precipitation was measured using the blue silicomolybdate method^66,79. Silica was added to purified protein scaffolds, at a concentration of 1 mg/mL in deionized water, to a concentration of 100 mM and the suspension was mixed for 2 h at room temperature (RT). The precipitate was then collected by centrifugation (3 mins at 18,800 x g), and the resulting pellet was washed 3 times with deionized water. Precipitated silica in the scaffold pellet was depolymerized by adding 1 M NaOH for 10 mins. The samples were then diluted to 1.6 mL and 0.15 mL of solution A (containing 20 g/L ammonium molybdate tetrahydrate and 60 mL/L concentrated hydrochloric acid in deionized water) was added to the sample. After 10 mins, 0.75 mL of solution B (containing 20 g/L oxalic acid, 6.67 g/L 4-methylaminophenol sulfate, 4 g/L anhydrous sodium sulfite, and 100 mL/L concentrated sulfuric acid in deionized water) was added. The blue color was allowed to develop for 2 h at RT, and the absorbance was measured at 810 nm. All samples were run in triplicate.

Optimization of silica concentration for EutM-CotB mediated biomineralization

To test and optimize conditions for silica gel formation by secreted EutM-CotB scaffolds, recombinant Bacillus cultures were incubated with six different concentrations of silica. B. subtilis WT strain harboring pCT-EutMCotB or pCT-empty (control) were first cultured in 50 mL SMM under optimal scaffold secretion conditions described above (20 °C, 100 rpm, for 48 h with 10 μM cumate added for induction of protein expression). Cultures were grown as three independent biological replicates. OD₆₀₀ and pH measured after 48 h were for B. subtilis pCT-empty: OD₆₀₀ = 1.73 ± 0.09, pH = 6.94 ± 0.01, and for B. subtilis pCT-EutMCotB: OD₆₀₀ = 1.26 ± 0.16, pH = 6.90 ± 0.02. Then, 2.5 mL culture aliquots were transferred to fresh culture glass tubes, incubated with 50–500 mM silica at 20 °C for 24 h and finally inspected for gel formation.

Six-well plate biomineralization experiments

Five mL of Bacillus cultures were transferred into sterile 6-well plates. Silica was added to cultures to a concentration of 100 mM. Samples were allowed to gel at 20 °C, 100 rpm for 1 h. The biomineralization experiments were performed with three biological replicates (i.e., three independent cultures for each experiment).

Silica block formation and mechanical testing

Silica gel plugs were prepared in 10 mL syringes (ID = 15 mm) with the tops removed and then tops were covered instead with sterilized caps to maintain sterile conditions. Silica was added to 1 mL cultures to a concentration of 200 mM. Samples were incubated at 25 °C for 5 h, and the gel plugs were pushed out of the syringe. Mechanical properties of the gel plugs were measured using an extensional DMA Rheometer (TA Instruments RSA-G2) with 15 mm compression disks. A frequency sweep was performed with a gap of 4.5 mm and an oscillation strain of 1%.

Larger silica gel blocks from ‘purple’ co-cultures of B. subtilis ΔlytC ΔflhG pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ + B. subtilis ΔlytC ΔflhG pCT-Purple-Hag^T209C::SpyT⁵⁸⁸ were prepared in the same manner, except that 3 mL of culture was used and the material was cured for 24 h.

ELM regeneration

Small piece (~5 mm³) from silica plugs fabricated from B. subtilis ΔlytC ΔflhG pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ cultures (biomineralized with 200 mM silica and cured for 24 h at 25 °C) was used to inoculate fresh cultures for ELM fabrication under standard conditions (see above). Scaffold building block expression, secretion, and biomineralization in six-well plates by the regrown cultures was confirmed as described above.

Protein analysis of silica gels

Silica gel plugs (15 mm × 5 mm) from 1 mL cultures of B. subtilis ΔlytC ΔflhG pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ or pCT-empty as a control were fabricated in the same manner as described above. After 24 h curing at 25 °C, the entire 1 mL solidified silica gels were dissolved in 200 µL 6x SDS-loading buffer at 100 °C for 1 h and 8 µL of spun down sample supernatant was loaded onto 15% SDS-PAGE gel.

Flagella light microscopy

One mL of Bacillus culture was centrifuged at 2,000 x g for 5 mins and gently resuspended in 1 mL of PBS, pH 7.5. The sample was pelleted as above, and the pellet was resuspended to an OD₆₀₀ of 10. Five µL of sample was loaded onto a polysine microscope slide, covered with a coverslip, and 10 µL Remel RYU flagella stain was applied to the edge of the coverslip.

Slides were examined using a Nikon Eclipse 90i microscope with a 100 × 1.3 numerical aperture oil-immersion objective (University Imaging Center (UIC), University of Minnesota (UMN)). Images were captured using a Nikon D2-Fi2 color camera running on Nikon’s NIS-Elements software ver 5.2.1.00.

Spore staining

The Schaeffer and Fulton Spore stain kit was used for spore staining with a modified protocol. Briefly, 500 µL of Bacillus culture was placed into a glass test tube. One drop (≈150 µL) of malachite green was added to the tube and immersed into boiling water for 10 min. If samples were dehydrating before 10 min, deionized water was added dropwise.

Following boiling, 50 µL of the sample was spread by pipette onto a plain, precleaned glass slide. The slide was allowed to air dry and then heat fixed with flame. The slide was decolorized with distilled water, counterstained with safranin for ≈20 s and again decolorized.

Slides were examined using a Nikon Eclipse 90i microscope with a 100 × 1.3 numerical aperture oil-immersion objective (UIC UMN). Images were captured using a Nikon D2-Fi2 color camera running on Nikon’s NIS-Elements software ver 5.2.1.00.

SEM and TEM imaging of silica precipitation

For SEM imaging, EutM scaffolds (1 mg/mL purified proteins in water) were attached to silicon wafers by incubating the protein suspensions on a silicon wafer for 1 h. In samples with silica, the wafers with attached protein were incubated in 100 mM silica for 2 h. Samples were washed with water, fixed with Trump’s Fixative for 2 min, and washed with water again. Samples were washed with increasing ethanol concentrations of 25, 50, 75, and 100%, and then supercritically dried using a Tousimis Critical-Point Dryer (Model 780 A). Samples were sputter-coated with 8 nm of iridium and viewed with a Hitachi SU8230 field emission gun scanning electron microscope (University of Minnesota Characterization Facility).

For TEM imaging, TEM grids were coated with EutM scaffolds by incubating grids in a protein solution (0.1 mg/mL of purified His-EutM or His-EutM-CotB in water) for 10 s at room temperature. Then, a small amount of 100 mM silica was added to the grid and incubated for 1 h. Grids were then washed with deionized water, fixed with Trump’s fixative for 2 mins, rewashed and stained with 2% uranyl acetate for 10 s. Grids were imaged as described below for other purified protein scaffolds.

T209C flagella staining

Glycerol cultures were inoculated into 4 mL of selective LB medium for plasmid maintenance and grown overnight at 30 °C and 220 rpm. One mL of culture was centrifuged at 2,000 x g for 10 mins. The supernatant was discarded and pelleted cells were gently resuspended in 1 mL of PBS, pH 7.5. The samples were pelleted as above and pellets resuspended with 50 µL PBS, pH 7.2 with 5 µg/mL Alexa Fluor^TM 488 C₅ maleimide (to stain cysteines in flagella with the Hag^T209C subunit) and incubated in the dark for five mins. 450 µL PBS was added to the samples prior to centrifugation as above. The pellet was resuspended in 50 µL PBS with 5 µg/mL SynaptoRed^TM C2 (FM4-64) and incubated for five mins in the dark. Samples were pelleted as above and then resuspended with 1 mL of PBS. After another centrifugation step, the pellets were resuspended to an OD₆₀₀ of 10. Three µL of sample were mixed with an equal volume of anti-fade fluorescent mounting medium FluoroShield and applied to a polysine microscope slide.

Slides were examined using a Nikon Eclipse 90i microscope with a 100 × 1.3 numerical aperture oil-immersion objective (UIC UMN). Illumination was obtained using a Lumencor Sola Light Engine. SynaptoRed^TM C2 was visualized using a dsRed fluorescence cube (excitation 530–560 nm, emission 590–650 nm) with a 3 second exposure time. Alexa Fluor^TM 488 C₅ maleimide was visualized using a GFP fluorescence cube (excitation 450–490 nm, emission 500–550 nm) with a 5.6 second exposure time. Images were captured using a Hamamatsu Orca Flash 4.0 v2 CMOS monochrome camera in black and white using Nikon’s NIS-Elements software ver 5.2.1.00.

In situ labeling of SpyTagged flagella with His-tdTomato-SpyCatcher

Glycerol cultures were inoculated into 4 mL of selective LB medium containing 0.15 mg of purified His-tdTomato-SpyCatcher protein and grown overnight at 30 °C and 220 rpm. One mL of culture was centrifuged at 2000 x g for 10 mins and gently resuspended in one mL of PBS, pH 7.5. The samples were pelleted as above. The pellets were resuspended with 50 µL PBS pH 7.2 with 5 µg/mL Alexa Fluor^TM 488 C₅ maleimide and incubated in the dark for five mins. 450 µL PBS was added to the samples prior to centrifugation as above. Cellular nucleic acid was stained by resuspending the pellets in 50 µL Tris-HCl, pH 8.0 with 10 µM Syto16 and incubated in the dark for 10 mins. 450 µL Tris-HCl, pH 8.0 was added and samples were centrifuged as above. Unused dye was removed by resuspension with 1 mL PBS and samples were centrifuged again. After resuspension to an OD₆₀₀ of 10, 3 µL of sample was mixed with an equal volume of FluoroShield and loaded onto a polysine microscope slide.

Slides were examined using a Nikon Eclipse 90i microscope with a 100 × 1.3 numerical aperture oil-immersion objective (UIC UMN). Illumination was obtained using a Lumencor Sola Light Engine. tdTomato was visualized using a dsRed fluorescence cube (excitation 530–560 nm, emission 590–650 nm) with a 3 second exposure time. Alexa Fluor^TM 488 C₅ maleimide and Syto^TM16 were visualized using a GFP fluorescence cube (excitation 450–490 nm, emission 500–550 nm) with a 0.3–5.6 s exposure time. Images were captured using a Hamamatsu Orca Flash 4.0 v2 CMOS monochrome camera in black and white using Nikon’s NIS-Elements software ver 5.2.1.00.

TEM of purified scaffolds from E. coli and Bacillus

Purified His-EutM-SpyCatcher protein scaffolds were diluted to 1 mg/mL. Ten µL was placed onto 200 mesh copper grids with 10 nm formvar and 1 nm carbon for 3 mins. The fluid was removed and 10 µL of Trump’s fixative (4% formalin, 1% glutaraldehyde, 0.1 M sodium cacodylate) was applied for 5 mins to fix the proteins. The fluid was removed and the grids were rinsed with 10 µL of deionized water three times. Negative staining was completed using 1% aqueous uranyl acetate which was applied and removed immediately to prevent over-staining. Grids were examined using a JEOL-JEM-1400Plus transmission electron microscope with a LaB6 tungsten filament at 60 kV (UIC UMN). Images were acquired using an Advanced Microscopy Techniques XR16 camera using AMT capture Engine software version 7.0.0.187.

In situ labeling of secreted EutM-SpyCatcher scaffolds with His-SpyTag-eGFP

B. subtilis ΔlytC ΔflhG harboring pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ was grown under standard conditions with (induced) and without (uninduced) cumate added for induction of protein expression. 500 µL of culture was mixed with 0.25 mg His-eGFP or His-SpyTag-eGFP (an estimated 2:1 ratio of SpyCatcher: eGFP established in³⁵) and mixed at RT for 30 mins. Ten µL was applied to a glass slide and covered with a slip. Slides were examined using a Nikon Eclipse 90i microscope with a 40 × 0.75 numerical aperture objective (UIC UMN). Illumination was obtained using a Lumencor Sola Light Engine. His-eGFP and His-SpyTag-eGFP were visualized using a GFP fluorescence cube (excitation 450–490 nm, emission 500–550 nm) with a 0.032–0.2 second exposure time. DIC images were captured with a 0.037–0.039 second exposure. Images were captured using a Hamamatsu Orca Flash 4.0 v2 CMOS monochrome camera in black and white using Nikon’s NIS-Elements software ver 5.2.1.00.

Silica material thin sections and TEM

B. subtilis ΔlytC ΔflhG harboring pCT-EutMCotB-EutMSpyC-Hag^T209C::SpyT⁵⁸⁸ was grown under standard conditions with (induced) and without (uninduced) cumate added for induction of protein expression. Cultures with cell densities of OD₆₀₀ ~3.5 (induced) or ~6 (uninduced) were either directly silica biomineralized or were concentrated prior to silica gel formation. For concentration, cultures were centrifuged at 3,220 x g for 10 mins and the supernatant was removed. Pellets were resuspended to OD₆₀₀ 10 or 20 using the media removed from the pellet. Silica gel plugs were prepared as described above for mechanical testing using 200 mM silica. Gels were pushed out of the syringes and 2 mm³ gel portions were cut from the center of both induced and uninduced samples. The gel pieces were fixed at 4 °C and 10 rpm overnight in 100 mM sodium cacodylate, 2% glutaraldehyde, and 2% paraformaldehyde. The gel pieces were then rinsed with 100 mM sodium cacodylate three times for 30 mins each at 4 °C and 10 rpm. An overnight post-fixation using 4% osmium tetraoxide in 100 mM sodium cacodylate was completed at 4 °C and 10 rpm. The samples were extensively rinsed with water and subjected to an ethanol dehydration series (25, 50, 75, 95, and 100%). Gel samples were trimmed into approximately 1 mm³ pieces, infiltrated with EMbed 812 resin (1:2 resin:ethanol, 1:1, 2:1, 100% resin without hardener, 100% resin with benzyl dimethyl amine hardener repeated once (2x total); all times 8 h-overnight), and polymerized in a 60 °C oven for 48 h. Sections were cut using a diamond knife on a Leica Ultracut UCT microtome at a thickness of 70–100 microns and collected on 200-mesh formvar/carbon-coated copper grids. Sections were stained with 3% aqueous uranyl acetate for 20 mins, rinsed in ultrapure water (5 sec, 5x), stained with Sato’s triple-lead stain⁸⁰ for 3 mins, and rinsed in ultrapure water (5 sec, 5x). Samples were examined using a JEOL-JEM-1400Plus transmission electron microscope with a LaB6 tungsten filament at 60 kV (UIC UMN). Images were acquired using an Advanced Microscopy Techniques XR16 camera using AMT capture Engine software version 7.0.0.187.

Image Analysis

Images were cropped, scale bars added, and channels were colorized and merged using Fiji v. 1.53⁸¹. Supplementary Fig. 9 image edits for part A include B. subtilis ΔlytC ΔflhG strain SpyTag 555 in which the red and green channels were multiplied by 1.25 before merging. Supplementary Fig. 12 image edits for part A include the Hag^T209C control in which the red channel was multiplied by 1.5 before merging with the green channel. In addition, for Hag^T209C::SpyTag⁵⁸⁸ the red channel was multiplied by 1.25 before merging. Supplementary Fig. 12 image edits in part B include the Hag^T209C control in which the red channel was multiplied by 1.5 before merging. In addition, for Hag^T209C::SpyTag⁵⁸⁸ the red channel was multiplied by 1.25 before merging.

Flagella RYU staining images were converted to grayscale and had the contrast and brightness increased by 10 using GNU Image Manipulation Program (GIMP) v. 2.8.18.

Statistical analysis

Statistical analysis of OD and pH data was performed using Origin v. 9.1. Statistical analysis of rheology data was performed using Microsoft Excel 2016. Mean and standard deviations from three biological replicates were calculated using Average and STDEV functions. Standard deviations from three biological replicates are shown as error bars in graphs and plots.

Statistical significance (p-values) of rheology data (Supplementary Table 2) was determined by performing T-test (Paired Two Sample for Means) and ANOVA (One-way, Single-factor) analysis with G’ (storage modulus) and G” (loss modulus) angular frequency averages. T-Test analysis was used to determine the statistical significance of rheology data between two groups of angular frequency data (G’ or G”) from silica gels fabricated with B. subtilis strains transformed with three different plasmids (pCT-empty, pCT-EutMCotB-EutMSpyC, pCT-EutMCotB-EutMSpyC-HagT209C::SpyT588). ANOVA analysis was performed on three groups of angular frequency data (G’ or G”) from silica blocks fabricated by the same three B. subtilis strains. Data were considered statistically significant when p < 0.05.

Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.