Fatty Acid Synthesis Knockdown Promotes Biofilm Wrinkling and Inhibits Sporulation in Bacillus subtilis

Abstract

Many bacterial species typically live in complex three-dimensional biofilms, yet much remains unknown about differences in essential processes between nonbiofilm and biofilm lifestyles. Here, we created a CRISPR interference (CRISPRi) library of knockdown strains covering all known essential genes in the biofilm-forming Bacillus subtilis strain NCIB 3610 and investigated growth, biofilm colony wrinkling, and sporulation phenotypes of the knockdown library. First, we showed that gene essentiality is largely conserved between liquid and surface growth and between two media. Second, we quantified biofilm colony wrinkling using a custom image analysis algorithm and found that fatty acid synthesis and DNA gyrase knockdown strains exhibited increased wrinkling independent of biofilm matrix gene expression. Third, we designed a high-throughput screen to quantify sporulation efficiency after essential gene knockdown; we found that partial knockdowns of essential genes remained competent for sporulation in a sporulation-inducing medium, but knockdown of essential genes involved in fatty acid synthesis exhibited reduced sporulation efficiency in LB, a medium with generally lower levels of sporulation. We conclude that a subset of essential genes are particularly important for biofilm structure and sporulation/germination and suggest a previously unappreciated and multifaceted role for fatty acid synthesis in bacterial lifestyles and developmental processes. IMPORTANCE For many bacteria, life typically involves growth in dense, three-dimensional communities called biofilms that contain cells with differentiated roles held together by extracellular matrix. To examine how essential gene function varies between vegetative growth and the developmental states of biofilm formation and sporulation, we created and screened a comprehensive library of strains using CRISPRi to knockdown expression of each essential gene in the biofilm-capable Bacillus subtilis strain 3610. High-throughput assays and computational algorithms identified a subset of essential genes involved in biofilm wrinkling and sporulation and indicated that fatty acid synthesis plays important and multifaceted roles in bacterial development.

Keywords: CRISPR interference; biofilms; essential genes; high-throughput screening; spatial organization; sporulation.

FIG 1

CRISPRi effectively inhibits gene expression during colony growth on LB and MSgg agar. (A) In this study, a CRISPRi library was created to individually inhibit the expression of each essential gene and select nonessential genes. (Left) The nuclease-deactivated Cas9 (dcas9) gene was inserted under a xylose-inducible promoter (P_xyl) at the lacA locus. For the control RFP depletion strain (top), the rfp gene was inserted under the control of a vegetative promoter (P_veg) at the amyE locus and a P_veg-rfp-targeting sgRNA construct was inserted at the thrC locus. For the library of 302 strains, a P_veg-gene targeting sgRNA construct was inserted at the amyE locus. (Right) dCas9 binds to the sgRNA and inhibits transcription. (B) CRISPRi effectively inhibits expression in colonies grown on LB or MSgg and is titratable. Colonies were spotted onto plates with various concentrations of xylose at t = 0 h, and RFP fluorescence at 24 h is shown. All images were contrast adjusted identically. Scale bar, 5 mm. (C) CRISPRi knockdown remains effective for at least 48 h in colonies on LB and MSgg and has greater dynamic range on MSgg. RFP fluorescence was quantified from colonies grown with various concentrations of xylose for 12, 18, 24, 36, 42, and 48 h. RFP levels decreased over time and with xylose concentration.

FIG 2

Systematic quantification reveals essential gene knockdowns with medium-specific growth phenotypes. (A) Full knockdown of many essential genes in liquid media (left, LB+1% xylose; right, MSgg+1% xylose) led to growth impairment. Black lines show the parent control, and blue (left) or orange (right) lines show strain 3610 knockdown strains with poor growth (OD₆₀₀ <0.075) at 5 h (left) or 8 h (right). (B) Four gene knockdowns exhibited dramatically different growth patterns in LB+1% xylose and MSgg+1% xylose. patA and asnB full knockdowns grew better in MSgg than in LB. glyA and folD full knockdowns grew better in LB than in MSgg. Dashed lines show the average of the parent control in each medium: the asnB full knockdown and the corresponding control were grown at 30°C, while all other cultures were grown at 37°C. Blue and orange lines represent growth in LB+1% xylose and MSgg+1% xylose, respectively. Solid lines are the average growth curve over n = 6 biological replicates, and shading represents 1 standard deviation. (C) Full knockdown of many genes resulted in growth impairment on solid medium. Colony sizes at 24 h are plotted. Parent controls (n = 44) are shown in gray.

FIG 3

Knockdown of fatty acid synthesis leads to increased wrinkling density. (A) Our library of 302 gene knockdowns in strain 3610 covered a variety of wrinkling phenotypes when grown on MSgg agar plates for 48 h. Wild-type controls are outlined in white dotted rectangles. The distance between the centers of adjacent colonies is 9 mm. (B) The image analysis platform (Materials and Methods) automatically quantified the degree of wrinkling for each strain. The total wrinkling level (white pixels postprocessing) within the colony boundary (pink circles) was extracted for each colony. (C) Many strains exhibited significantly lower wrinkling than parent controls, while fatty acid and gyrase knockdowns exhibited higher wrinkling density (wrinkling intensity normalized to colony area). The dark gray dashed histogram represents the parent control, and the light gray histogram represents the 302 strains from the gene knockdown library. Insets show representative images of strains with higher wrinkling density. Scale bar, 5 mm. (D) Matrix gene (eps and tapA) expression is uncorrelated with wrinkling density. Reporter levels are shown, normalized so that the average parent control was 1. All P_eps-mNeongreen gene knockdown strains were significantly different from the parent control (P < 0.016). Fatty acid mutants are in yellow, DNA replication-related mutants are in red, ribosome-related mutants are in blue, cell-wall-related mutants are in gray, biosynthesis mutants are in green, and a secretion mutant is in purple. For P_tapA-mNeongreen, all gene knockdowns except for murG were statistically significant (P < 0.045).

FIG 4

A high-throughput sporulation screen reveals reduced sporulation efficiency due to knockdown of fatty acid synthesis. (A) A growth-based, high-throughput assay was used to identify mutants defective in sporulation. Strains are grown to saturation, vegetative cells are heat killed, and cultures are inoculated into fresh medium. Any strains without viable spores will not survive the heat kill, and hence those wells will remain clear during outgrowth. (B) Sporulation mutants exhibit distinct patterns of growth in DSM (left) and are unable to outgrow after heat killing (right) for both strain 3610 (top) and strain 168 (bottom). Average growth curves over n = 6 biological replicates are shown as solid lines, and shading represents 1 standard deviation. (C) There are fewer spores in a 24-h culture of strain 168 compared to strain 3610. Cultures were grown in DSM or LB, heat killed, and plated to measure CFU. *, P < 0.02. (D) The sporulation efficiency (ratio of CFU post- versus pre-heat killing) was lower for strain 168 than for strain 3610. (E) The number of spores in the inoculum correlates with growth lag. Cultures post-heat kill were serially diluted 10-fold, and growth curves were measured. The time to reach an OD₆₀₀ of 0.2 was negatively correlated with the number of spores (calculated from CFU). Dark blue circles represent strain 3610 grown in DSM, light blue triangles represent strain 168 grown in DSM, and orange squares represent strain 3610 grown in LB. (F) All essential gene basal knockdowns are competent for sporulation in DSM, and basal knockdown of fatty acid synthesis reduces sporulation in LB. The spore outgrowth curves following 24 h of growth in DSM (left) or LB (right) and heat killing are shown. Solid black lines show the average of parent controls (n = 40 biological replicates), and light gray lines show growth curves of the CRISPRi library. For cultures pregrown in LB (right), several knockdowns exhibited a delay in growth or did not grow at all (colored lines). Fatty acid mutants are in yellow, DNA replication-related mutants are in red, ribosome-related mutants are in blue, cell-wall-related mutants are in gray, biosynthesis mutants are in green, and a secretion mutant is in purple. (G) Manganese and calcium nitrate promote sporulation in LB. The sporulation efficiency of strain 3610 (WT) in DSM and in LB with various DSM components added is shown. Note that the Δspo0A mutant did not sporulate at all in DSM and hence is not plotted. Sporulation efficiencies in DSM and LB plus all additives were statistically indistinguishable (P = 0.84). Sporulation efficiencies in LB+MnCl₂ and LB+Ca(NO₃)₂ were significantly higher than in LB without any additives (P < 0.015). (H) Fatty acid knockdowns exhibit reduced sporulation efficiency in LB. Colors are the same as in panel F. The Δspo0A mutant did not sporulate at all in LB and hence is not plotted. All knockdowns exhibited significantly different sporulation efficiency than the parent control (P < 0.04).