Figure 2. Growth on biofilm-promoting medium, increased knockdown, and competition against parent-RFP all broaden the distribution of fitnesses across the library.

(A) Basal knockdown (light blue) of essential genes on LB agar (which does not promote biofilms) resulted in similar colony sizes as parent-GFP controls (gray); only 13 of 302 colonies had a colony size two standard deviations below the mean of the controls. By contrast, on biofilm-promoting MSgg agar the distribution of colony sizes spread to smaller values, with 80 colonies more than two standard deviations below the mean of the controls. Full knockdown (dark blue) inhibited growth of most strains. Data are from measurements at 16 hr using the sacA::gfp library. Vertical solid lines are means of the control distributions, and dashed lines show two standard deviations below the mean. (B) 17 (LB) and 11 (MSgg) knockdown strains in the library competed poorly against the parent-RFP strain at basal knockdown (light blue), while 41 (LB) and 46 (MSgg) had competitive fitness similar to parent-GFP+parent-RFP controls (gray) even at full knockdown (dark blue). Data are from competition ratios at 16 hr using the sacA::gfp library. Low-fitness strains were defined as having fitness at least two standard deviations below the mean of the data, and neutral-fitness strains were defined as having fitness at least one standard deviation above the mean. Vertical solid lines are means of the distribution from basal knockdown, and dashed lines show two standard deviations below and one standard deviation above the mean. (C) Strains with low competitive fitness for basal knockdown generally had lower growth rates in liquid monoculture than parent control strains. Colonies were inoculated into liquid LB, and OD₆₀₀ was monitored over time. Ribosome-related genes are shown in shades of purple, and a cell envelope-related gene (tagG) is shown in yellow. Curves are means, and shaded regions represent one standard deviation (n = 3). Inset: maximum growth rates. *: p<0.03, Student’s unpaired t-test with a Benjamini–Hochberg multiple-hypothesis correction. (D) On both LB and MSgg agar, basal knockdown of ygsA, which is involved in gene regulation, exhibited high competitive fitness (left) and tagG and ribosomal-gene knockdowns exhibited low competitive fitness (right). GFP (knockdown strain) is false-colored green, and RFP (parent) is false-colored magenta. Images are from 16 hr using the sacA::gfp library. Scale bar: 5 mm. (E) Competitive fitness of the strains with the highest and lowest values was approximately constant after 16 hr. Curves are means, and shaded regions represent one standard error of the mean (n = 3 independent measurements). Inset and dashed black line: a reaction-diffusion model of co-culture colony growth with physically realistic parameters indicates that knockdowns (magenta) with maximum growth rate 20% lower than the parent (green) reproduce the colony size (bottom, inset) and competitive fitness (bottom, dashed black line) of ribosomal protein knockdowns after 16 hr (Materials and methods, Figure 2—figure supplement 1C).

Figure 2—figure supplement 1. Monoculture colony size screen, and analysis of knockdowns with low and high competitive fitness.

(A) Monoculture colonies on MSgg agar (basal knockdown) exhibited more variation in size than monoculture colonies on LB agar after 16 hr of growth. With full knockdown, there were many monoculture colonies with small size relative to controls on LB and on MSgg after 16 hr. The CRISPRi library is within the white boxes, and colonies outside of the boxes are parent-GFP controls. The distance between the centers of neighboring colonies is 9 mm. Data for colony areas is in Supplementary file 2. (B) A few knockdowns exhibited high competitive fitness in co-culture despite having reduced colony sizes in monoculture in basal knockdown conditions (orange), and one non-essential gene knockdown in the library had reduced competitive fitness (blue). The rest of the strains are in gray, and the controls are in black. (C) A reaction-diffusion model of colony growth (Materials and methods) recapitulates competitive fitness when D_C = 0.003 mm² hr⁻¹; competitive fitness is not strongly perturbed by 10-fold changes in K/n₀ and D_n/D_C. Parameters M₁ = 0.0175 min⁻¹, M₁/M₂ = 0.8, r₀ = 1 mm, and C₀/(n₀/b)=0.001 are estimates from experimental data. The black line shows the simulated competitive fitness corresponding to the colony in the inset of Figure 2E in which both competitive fitness and colony size data were recapitulated with K/n₀ = 0.05 and D_n/D_C = 100. Competitive fitness was largely unchanged if D_n/D_C = 1000 (dark pink, partly beneath black line) or D_n/D_C = 10 (light pink, partly beneath black line), or if K/n₀ = 1.0 (dark blue) or K/n₀ = 0.005 (light blue, partly beneath dark pink line). (D) Strains that competed well at full knockdown included genes related to metabolism, gene regulation, prophage, and cell envelope (divIB), along with yloU, a gene of unknown function. Merged images show the parent-RFP false-colored in magenta and the knockdown strain false-colored in green. The 100% parent-RFP, 50% parent-RFP+50% parent-GFP, and 100% parent-GFP controls are shown to the right of the knockdowns. Scale bar: 5 mm. (E) Competitive fitnesses of the knockdowns with the highest fitness values at 16 hr were mostly stable over time with full CRISPRi induction. Curves are means, and shaded regions represent one standard error of the mean (n = 3 biological replicates).