SUPPLEMENTARY INFORMATION

Transcription

1 SUPPLEMENTARY INFORMATION Temporal competition between differentiation programs determines cell fate choice Anna Kuchina 1,2, Lorena Espinar 3, Tolga Çağatay 1,2, Alejandro O. Balbin 1,2, Fang Zhang 1,2, Alma Alvarado 1,2, Jordi Garcia-Ojalvo 3 and Gürol M. Süel 1,2 1 Green Center for Systems Biology, University of Texas Southwestern Medical Center, Dallas, TX Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX Departament de Física i Enginyeria Nuclear, Universitat Politècnica de Catalunya, Edif. GAIA, Rambla Sant Nebridi s/n, E-8222 Terrassa, Spain

2 TABLE OF CONTENTS S1. SUPPLEMENTARY FIGURES... 2 Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Figure S Supplementary Table S S2. MATERIALS AND METHODS S2.1 Strain construction Supplementary Table S S2.2 Promoter definitions S2.3 NOE strain construction S2.4 Imaging and growth conditions S2.4.1 Preparation for microscopy S2.4.2 Time-lapse microscopy and image analysis S2.4.3 Calculation of competence probability S2.4.4 Prediction of DA cells frequency 25 Supplementary Figure S Supplemental references

3 S1. SUPPLEMENTARY FIGURES A P spoa P comg x% >3% B C 2 Total cell distribution 15 Competence distribution Cell count Cell count Normalized P spoa activity (x) Normalized P spoa activity (x) Supplementary Figure S1. Calculation of competence initiation probability as function of P spoa activity. (A) A frame corresponding to >3% P comg activity of a cell from the time-lapse movie of AcomG strain (strain definitions and genetic background can be found in Supplementary 2

4 Information, S2.1 and Supplementary Table S2) showing how competence and total cell distributions in panels (B) and (C), respectively, were obtained. P comg (red) and P spoa (blue) activities were measured as described in Figure 1B. A competent cell of interest was framed in a rectangular ROI, and the P spoa activities of all cells inside the ROI were measured. n=54 competent cells were measured in this way, giving rise to competence distribution shown in panel (C), while the P spoa activities of all cells inside each corresponding ROI were summed for total cell distribution shown in Panel (B). For more details, see Supplementary Information, S (B) Histogram showing the distribution of P spoa activities for all cells inside the ROI shown in panel (A) (see Supplementary Information, S2.4.3). Only data for P spoa activities between 2% and 8% are shown, because data below 2% is within error and cells above 8% are too close to spore formation (see Figure 1A). Error bars represent counting error. (C) Histogram showing the number of competence initiation events for each specified P spoa activity, again, between 2% and 8% of P spoa activities. The resulting distribution, normalized by the distribution of P spoa activities for total sporulating cells shown in panel (B), is used to calculate the probability of competence shown in Figure 1D. Error bars represent counting error. 3

5 A Normalized P spof activity Normalized P spoa activity B 4 Competence distribution C 5 Total cell distribution 3 4 Cell count 2 1 Cell count Normalized P spof activity Normalized P spof activity D.2 Competence probability Normalized P spof activity Supplementary Figure S2. Competence initiation probability as function of P spof activity. (A) Positive correlation between P spoa and P spof expression measured in the same cell (r=.83). Each dot represents the fluorescence values from one frame of quantitative time traces of sporulating cells simultaneously expressing P spoa- CFP and P spof -YFP, n=2, from strain F-A. 4

6 Panels (B) through (D) show the calculation of competence initiation probability as function of P spof activity, which was performed in the same way as in Figure S1 and Supplementary Information, S2.4.3 (for P comg, the activation threshold of 35% was used). (B) and (C) show the distributions of P spof activities either measured at competence initiation (panel B), or for all cells inside the ROI (panel C), similar to Figure S1B and C. (D) Conditional probability of competence as a function of normalized P spof activity, calculated as competence distribution (panel B) normalized by the distribution of P spof activities for total sporulating cells (panel C). Error bars represent counting error. 5

7 A B Sporulation probability Sporulation Competence Normalized P spoa activity Competence probability Sporulation probability Sporulation Competence Normalized P spof activity Competence probability Supplementary Figure S3. Probability of sporulation as function of P spoa and P spof activity. Panels (A) and (B) show the conditional probability of sporulation as function of P spoa and P spof activity, respectively. The probability of sporulation was obtained by measuring fluorescence from single cells of a sporulating microcolony of strains A-IIR, n=29 (panel A) and F-IIR, n=1836 (panel B). To obtain the probability of sporulation, the distributions of either P spoa values (panel A) or P spof values (panel B) in cells that have formed a forespore (measured by P spoiir activation), was normalized by the distributions of corresponding values in total observed cells. Error bars represent counting error. The corresponding conditional probability of competence as function of either P spoa (panel A) or P spof (panel B) expression measured in Figures 1D, S1 and S2 is plotted with the second y axis (red) for reference. 6

8 Spore Fluorescence (a.u.) 1 P spoa.8 P comg.6.4 Division.2 Spore Competent Supplementary Figure S4. Sister cells choose cell fates independently. Sample quantitative traces of sister cells, one of which is sporulating (dashed traces) while the second becomes competent (solid traces). The division after which the cells split from the common ancestor is indicated as a vertical line. P spoa -CFP fluorescence is shown in blue and P comg -mcherry fluorescence is in red. Microscopy images with the measured cell outlined are shown on top at corresponding time points for illustration. 7

9 A Fluorescence (a.u.) P spoa P comg B Sporulation C Competence Normalized PabrB f luorescence (a.u.) P abrb P abrb D E Normalized PsinI f luorescence (a.u.) P sini P sini Normalized P spoa f luorescence (a.u.) Normalized P comg f luorescence (a.u.) Supplementary Figure S5. The pattern of cross-regulatory genes expression in single sporulating and competent cells. (A) Average fluorescence time traces of competent cells from A-comG strain showing P spoa -YFP (blue) and P comg -CFP (red). Individual traces (n=67) are normalized in 8

10 amplitude and aligned in time with respect to P comg initiation (3% of peak activity) prior to averaging. The blue area and light red curves correspond to standard deviation. Panels (B) through (E) depict normalized single cell fluorescence of either P abrb -YFP (panels B and C, strain A-comG-abrB) or P sini -YFP (panels D and E, strain A-comGsinI) plotted against fluorescence of either P spoa -CFP (B and D) or P comg -mcherry (C and E) measured in the same cell. Panels B and D show sporulating cells (B, n=4; D, n=25) and panels C and E show competent cells (C, n=1; E, n=1). The traces are colored according to time with darker color corresponding to the beginning of the trace. 9

11 A B Fluorescence (a.u.)25 Competence P comg Fluorescence (a.u.) DA P comg C DA Sporulating 16.3h 16.7h 18.h 19.3h 21.h 22.h 24.3h 28.3h 35.7h 38.3h Supplementary Figure S6. Single-cell dynamics of competence and Dual-Activity (DA) state. Panels (A) and (B) show quantitative time traces of (A) the cell undergoing the competence event shown in Figure 3A filmstrip; (B) the DA cell shown in the Figure 3B filmstrip. P comg activity measured by CFP fluorescence intensity is shown in red. The point of forespore appearance is indicated by an arrow and cell cartoon. (C) Filmstrip of a typical DA cell (upper panel) demonstrating that cells in the DA state form spores with similar dynamics and appearance as normal sporulating cells (lower panel). The DA cell is visualized by competence reporter P comg -CFP (red). Time in hours is indicated below each frame. 1

12 A P abrb B P sini Pre-decision 15 Execution Pre-decision 6 Execution Fluorescence (a.u.) 1 5 Fluorescence (a.u.) 4 2 C DA D DA PabrB f luorescence (a.u.) P abrb P spoa fluorescence (a.u.) PsinI f luorescence (a.u.) P sini P spoa fluorescence (a.u.) Supplementary Figure S7. Cross-regulatory genes expression in DA cells. Panels (A) and (B) show mean P abrb -YFP (A) and P sini -YFP (B) fluorescence measured in single cells of strains A-comG-abrB (n=17 DA and 4 spores), and A-comG-sinI (n=16 DA and 25 spores), respectively. The fluorescence measurements were taken from the time point either 4 minutes before initiation of (9% of maximum value), labeled pre-decision, or 4 minutes after the maximum expression of P spoa, labeled execution. Error bars represent standard error of the mean (SEM). Panels (C) and (D) depict normalized single cell fluorescence of either P abrb -YFP (panel C, strain A-comG-abrB, n=17) or P sini -YFP (panel D, strain A-comG-sinI, n=16) plotted 11

13 against fluorescence of P spoa -CFP measured in the same DA cell. The traces are colored according to time with darker color corresponding to the beginning of the trace. 12

14 A B C Fluorescence (a.u.) Fluorescence (a.u.) Fluorescence (a.u.) Native P comg P spoa -comk P spoiig -comk P comg P spoa SpoIIE P spoiig P spoiir P comg P comg P spoiir P spoiig P spoa P comg SpoIIE Supplementary Figure S8. Dynamics and robustness of sporulation component activities in DA cells. In panels (A) through (C) time traces of single DA cells from strains expressing pair-wise combinations of one of the sporulation reporters P spoa (blue), SpoIIE (orange), P spoiig (magenta) or P spoiir (green) and the competence reporter P comg (red) fused to fluorescent proteins (YFP and CFP respectively) are shown in shaded colors, with mean trace shown on top in bright color. All traces have been normalized in amplitude and aligned in time with respect to P comg activation (>3% of peak activity) which is set as zero time point. The data from the first four figures (from the left) in panels (A) through (C) was combined to produce the fifth figure. Strain definitions and genetic background can be found in Supplementary Table S2. 13

15 (A) Strains with no ectopic ComK expression, left to right: A-comG, n=15; IIE-comG, n=7; IIG-comG, n=13; IIR-comG, n=4. (B) Strains in which ComK is expressed from the P spoa promoter, from left to right: [AcomG] A-K, n=36; [IIE-comG] A-K, n=27; [IIG-comG] A-K, n=29; [IIR-comG] A-K, n=39. (C) Strains in which ComK is expressed from the P spoiig promoter, from left to right: [AcomG] IIG-K, n=33. [IIE-comG] IIG-K, n =21. [IIG-comG] IIG-K, n=48. [IIR-comG] IIG-K, n=38). 14

16 A PspoIIG-comK NoE Sporulation C NoE+Stable protease P comg P spoa B Competence NoE P comg Supplementary Figure S9. NoE perturbation does not affect normal sporulation and competence. (A) Filmstrips of a sporulating microcolony from B. subtilis strain expressing P spoiig -comk ([A-comG] IIG-K ) (top panel) and NoE strain (lower panel). Time in hours is indicated on each panel. (B) Filmstrip showing a typical competence event visualized by CFP expressed from P comg promoter (in red) followed by sporulation in NoE strain. Time in hours is indicated on each panel. (C) Filmstrip showing a representative failed sporulation attempt in presence of simultaneous ComK and SpoA expression in cells of the NoE strain expressing stable E. coli SspB protease factor (see Supplementary Information, S2.3). P comg CFP is shown in red and P spoa -YFP is in blue. Time in hours is indicated on each panel. 15

17 % comga localization % 1 % Supplementary Figure S1. ComGA localization is observed in competent, but not DA cells. ComGA cell pole localization events visualized by comga-yfp in [comg-comga] IIG-K strain. Competent cells, n = 17. DA, n =

18 1 P spoiig (Native) P spoiig (P spoiig -ynea) Fluorescence (a.u.) Supplementary Figure S11. Ectopic ynea expression delays sporulation in single cells. Averaged P spoiig reporter time traces from single cells of A-IIG strain (black, n=5) and strain [IIG] IIG-yneA in which ynea is ectopically expressed from the P spoiig promoter (magenta, n=5). For [IIG] IIG-yneA strain, representative cells that express P spoiig for extended period of time have been selected. For A-IIG strain, sample sporulating cells are shown for comparison. The time traces are normalized with respect to maximum and aligned at P spoiig initiation (>3% of maximum activity). Dashed traces represent standard deviation. 17

19 Native P spoa -comk P spoiig -comk P spoiir -comk Competent DA Competent DA Competent DA Competent DA Mean STD xP spoiig -comk 2xP spoa -comk P spoiig- ynea P spoiig -comk/p spoiig -ynea Competent DA Competent DA Competent DA Competent DA Mean STD P spoa -ynea P spoa -comk/p spoa -ynea NoE Competent DA Competent DA Competent DA Mean STD Supplementary Table S1. Frequencies of competent and DA cells in strains with engineered cross-regulation between competence and sporulation. Listed are percent fractions of competent and DA cells measured for indicated strains. The frequencies were counted as numbers of ComK initiations per total number of cells observed in a defined time-lapse movie frame (visualized as activity of P comg reporter). A minimum of three measurements was made for each strain. STD, standard deviation. 18

20 S2. MATERIALS AND METHODS S2.1 Strain construction Supplementary Table S2 lists Bacillus subtilis strains isogenic to wild-type B. subtilis PY79 strain. Polymerase Chain Reaction (PCR) was utilized to amplify native P spoa, P spoiig, P spoiie, P spoiir, P abrb, P sini, and P comg promoters as well as spoiie, ynea, comga and comk from the wild type B. subtilis PY79 strain (see Promoter definitions, Supplementary Information, S2.2). Amplified fragments were cloned into B. subtilis chromosomal integration vectors. The vectors used in this study are as follows: psac-cm, integrating into the saca locus (constructed by R. Middleton and obtained from the Bacillus Genetic Stock Center); pld3 designed to integrate into the amye locus (kind gift from Jonathan Dworkin, Columbia University); pglt-kan, designed to integrate into the glta locus(middleton and Hofmeister, 24) (constructed by R. Middleton and obtained from the Bacillus Genetic Stock Center); per449, a generic integration vector constructed for integration into the gene of interest (kind gift from Wade Winkler, UT Southwestern); and the bifunctional cloning plasmid php13 carrying the replication origin of the cryptic B. Subtilis plasmid pta16 (5 copies per genome)(haima et al., 1987). Standard B. subtilis transformation protocols were followed to transform B. subtilis strain PY79 with these constructs. B. subtilis strains isogenic to PY79 Genotype Appearance in Figures and Tables A-IIR AmyE::P spoa -yfp, P comg -mcherry (Sp R ) 1A SacA::P spoiir -cfp (Cm R ) A-F AmyE:: P spof -yfp (Sp R ) S2A SacA:: P spoa -cfp, P comg -mcherry (Cm R ) F-comG AmyE:: P spof -yfp (Sp R ) SacA:: P comg -cfp (Cm R ) S2B-D; S6C; S12; Table S1 A-comG SacA:: P spoa -yfp, P comg -cfp (Cm R ) 1B-D; 3A-E; S1; S3A; S5A; S6A-B; S8A; Table S1 IIG-comG AmyE::P spoiig -yfp, P comg -cfp (Sp R ) 3E; S8A IIR-comG SacA:: P spoiir -yfp, P comg -cfp (Cm R ) 3E; S8A IIE-comG AmyE::P spoiie -spoiie-yfp (Sp R ) SacA::P comg -cfp (Cm R ) 3E; 4B-C; S8A 19

21 spoiie::neo (Neo R ) A-comG-abrB SacA:: P spoa -cfp, P comg -mcherry (Cm R ) AmyE::P spoiig -comk/p abrb -yfp (Sp R ) A-comG-sinI SacA:: P spoa -cfp, P comg -mcherry (Cm R ) AmyE::P spoiig -comk/p sini -yfp (Sp R ) [A-comG] A-K AmyE::P spoa -yfp, P comg -cfp (Sp R ) SacA:: P spoa -comk (Cm R ) [IIG-comG] A-K AmyE::P spoiig -yfp, P comg -cfp (Sp R ) SacA:: P spoa -comk (Cm R ) [IIR-comG] A-K AmyE::P spoa -comk (Sp R ) SacA::P spoiir -yfp, P comg -cfp (Cm R ) [IIE-comG] A-K AmyE::P spoiie -spoiie-yfp (Sp R ) SacA:: P spoa -comk (Cm R ) spoiie:: P comg -cfp (Neo R ) [A-comG] IIG-K AmyE::P spoiig -comk (Sp R ) SacA:: P spoa -yfp, P comg -cfp (Cm R ) [IIG-comG] IIG-K AmyE::P spoiig -yfp, P comg -cfp (Sp R ) SacA:: P spoiig -comk (Cm R ) [IIR-comG] IIG-K AmyE::P spoiig -comk, P comg -cfp (Sp R ) SacA:: P spoiir -yfp (Cm R ) [IIE-comG] IIG-K AmyE::P spoiie -spoiie-yfp (Sp R ) SacA:: P spoiig -comk, P comg -cfp (Cm R ) spoiie::neo (Neo R ) [A-comG] IIR-K AmyE:: P spoa -yfp, P comg -cfp (Sp R ) SacA:: P spoiir -comk (Cm R ) [comg-comga] IIG-K AmyE:: P spoiig -comk, P comg -cfp (Sp R ) SacA:: P comg -comga-yfp (Cm R ) [comg] 2xIIG-K AmyE::P spoiig -comk (Sp R ) SacA:: P spoiig -comk, P comg -cfp (Cm R ) [comg] 2xA-K AmyE::P spoa -comk (Sp R ) SacA:: P spoa -comk, (Cm R ) GltA:: P comg -mcherry (Neo R ) NoE SacA::P spoa -yfp, P comg -cfp (Cm R ) AmyE::P spoiig -comk, P comg -sspb EC -XP (Sp R ) spoiie::neo (Neo R ) GltA::P spoiie -spoiie-ssra EC (Erm R ) [comg] A-yneA AmyE::P comg -cfp (Sp R ) SacA:: P spoa -ynea, (Cm R ) [comg] A-K/A-yneA AmyE::P spoa -comk (Sp R ) SacA:: P spoa -ynea, P comg -cfp (Cm R ) [comg] IIG-yneA AmyE:: P spoiig -yfp, P comg -cfp (Sp R ) SacA:: P spoiig -ynea (Cm R ) [IIG] IIG-yneA AmyE:: P spoiig -yfp, P spoa -cfp (Sp R ) SacA:: P spoiig -ynea (Cm R ) [comg] IIG-K/IIG-yneA AmyE:: P spoiig -comk, P comg -cfp (Sp R ) SacA:: P spoiig -ynea (Cm R ) 2A;S4; S5B-C; S7A,C 2B; S5D,E; S7B,D 5B; S8B; Table S1 S8B S8B S8B 3F; 5B; S8C; S9A; Table S1 S8C S8C S8C 5B; Table S1 4A,C; S1 5B; Table S1 5B; Table S1 3F; S9; Table S1 Table S1 5B; Table S1 Table S1 S11 5B; Table S1 Supplementary Table S2. Strain definitions and genetic background. 2

22 In the first column, promoters expressing fluorescent proteins are shown as plain text and/or in square brackets, while additional ectopic expression of comk (abbreviated as K ) and/or ynea from various promoters is indicated in subscript. A, P spoa. F, P spof. IIE, spoiie. IIG, P spoiig. IIR, P spoiir. S2.2 Promoter definitions All promoter sequences were Polymerase chain reaction (PCR) amplified from B. subtilis PY79 chromosomal DNA with reverse primer carrying an optimized ribosomal binding sequence (RBS) and spacer sequence (SS) AAGGAGGAA. Promoter sequences were defined as follows: P spoa : chromosomal sequence to (minus-strand); P spof : chromosomal sequence to (minus-strand); P spoiig : chromosomal sequence to (plus-strand); P spoiie : chromosomal sequence 7327 to 7523 (plus-strand); P spoiir : chromosomal sequence to (minus-strand); PcomG: chromosomal sequence to (minus-strand); P AbrB : chromosomal sequence to (minus-strand); PsinI: chromosomal sequence to (plus-strand). S2.3 NoE strain construction Exclusion of competence appears to coincide with the formation of the asymmetric septum and localization of SpoIIE protein (see Figure 3E and Supplementary Figure S8). To disrupt SpoIIE localization, which constitutes a critical early step in forespore formation, we constructed a NoE strain (for detailed genetic composition see Supplementary Table S2). Specifically, we replaced 21

23 the native SpoIIE gene with a SpoIIE coding sequence that contained a C-terminal Escherichia coli tag (ssra) targeting it for enzymatic degradation in the presence of the E. coli protease factor SspB (Flynn et al., 21; Gottesman et al., 1998; Griffith and Grossman, 28; Levchenko et al., 2). We placed the exogenous SspB protease adaptor protein under the control of the P comg promoter to ensure SpoIIE degradation only during simultaneous expression of ComK (Figure 3F). In addition, we introduced fluorescent proteins (YFP and CFP) under the control of P spoa and P comg promoters, respectively. Finally, for the convenience of observation we increased fraction of DA cells by ectopically inducing ComK expression from P spoiig promoter (see Figure 5 and Supplementary Table S1).This perturbation disrupted sporulation in cells that simultaneously express P spoa and ComK leading to cell death (Supplementary Figure S9C). To allow these prospective DA cells instead to proceed with competence, we added a C-terminal B. subtilis degradation tag (XP) to the exogenous SspB protease to reduce its half-life (Wiegert and Schumann, 21). The resulting NoE re-wiring did not interfere with sporulation (Supplementary Figure S9A) or successful completion of competence events (Supplementary Figure S9B). Compared to the strain expressing ComK from P spoiig promoter ([A-comG] IIG-K ), we observed an almost 5% decrease in the fraction of DA cells and a corresponding increase in competence frequency (Figure 3F and Supplementary Table S1) indicating that our perturbations were able to specifically rescue DA cells to undergo competence without affecting sporulation. 22

24 S2.4 Imaging and growth conditions S2.4.1 Preparation for microscopy For imaging, B. subtilis cells were grown at 37 C in LB with appropriate selection. Antibiotics for selection were added to the following final concentrations: 5 μg/ml chloramphenicol, 5 μg/ml neomycin, 5μg/ml erythromycin and 1 μg/ml spectinomycin. Cells grown to OD 1.8 were resuspended in.5 volume of Resuspension Media (RM; composition [per 1 liter]:.46 mg FeCl 2, 4.8 g MgSO 4, 12.6 mg MnCl 2, 535 mg NH 4 Cl, 16 mg Na 2 SO 4, 68 mg KH 2 PO 4, 96.5 mg NH 4 NO 3, 219 mg CaCl 2, 2 g L-glutamic acid)(sterlini and Mandelstam, 1969) supplemented with.2% glucose. The cells were incubated at 37 C for 1 hour, then diluted 1-fold in RM and applied onto a 1.5% low-melting agarose pad placed into a coverslip-bottom Willco dish for imaging. This protocol is optimized for time-lapse microscopy. RM reduces the growth rate of microcolonies on agarose and leads to sporulation. The imaging was performed as described previously(suel et al., 26). S2.4.2 Time-lapse microscopy and image analysis Growth of microcolonies was observed with fluorescence time-lapse microscopy at 37 C with an Olympus IX-81 inverted microscope with a motorized stage (ASI) and an incubation chamber. Image sets were acquired either every 4 min, or every 2 min with a Hamamatsu ORCA-ER camera. The imaging time has been optimized in order to prevent phototoxicity(suel et al., 26). Custom Visual Basic software in combination with the Image Pro Plus (Media Cybernetics) was used to automate image acquisition and microscope control. 23

25 Time-lapse movie data analysis was performed by custom software developed with MATLAB image processing and statistics toolboxes (TheMathworks)(Suel et al., 26). S2.4.3 Calculation of competence probability The probability of competence initiation (referred to as probability of competence) in a cell expressing a specific level of P spoa activity (denoted as x) can be calculated from the Bayes theorem: P(competence P spoa = x) = P(P spoa =x competence)p(competence)/p(p spoa = x); where: P(competence P spoa = x) is the probability of competence given that the cell has a specific level of P spoa activity (denoted as x), P(P spoa =x competence) is the probability of observing a specific value of P spoa =x during competence, P(competence) is the probability of competence which can be expressed as the ratio of competence initiations number to the number of total observed cells, P(P spoa = x) is the probability of the cell to express a specific value of P spoa =x. The straightforward way to experimentally measure competence probability as a function of P spoa activity would be therefore to obtain P spoa values at comk initiation for each competent 24

26 cell observed in a time-lapse movie, and then normalize the obtained distribution with the distribution of P spoa from total sporulating cells. However, technically it is not feasible to trace all competent cells in a time-lapse movie. To bypass this problem, we analyzed n=48 competent cells from different movies, normalized P comg data relative to its peak and selected a frame for each cell at which its P comg expression peaked over 3% threshold (Supplementary Figure S1A). The P spoa activities of competent cells at this time point were used to obtain the competence distribution (Supplementary Figure S1C). Due to presence of other competent cells, we implemented an automatic MATLAB procedure that placed a rectangular ROI around the cell of interest containing the area that corresponded to total frame area divided by number of competent cells in this frame (Supplementary Figure S1A). Assuming the cell density to be more or less even, the cells inside the ROI would represent the microcolony that would give rise to exactly one competence event. We then calculated the P spoa levels for every cell inside the ROI and normalized them assuming that each ROI contains at least one sporulating cell (P spoa =1) and one vegetative cell or spore (P spoa =). The resulting data from ROIs for all measured competent cells is shown in Supplementary Figure S1B. To obtain the probability of competence as a function of P spoa activity, we normalized the number of cells in each bin from competence distribution (Supplementary Figure S1C) by the number of cells in each bin from total cell distribution (Supplementary Figure S1B). The resulting plot is shown in Figure 1D. S2.4.4 Prediction of DA cells frequency Our data presented in Figures 1D and S2D suggests that ComK initiation is independent from progression to spore formation even at high levels of SpoA (and SpoF) expression. We also determined that in DA cells ComK expression is restricted within a narrow time window between the localization of SpoIIE and expression of SpoIIR, which is the commitment point for spore formation (Figure 3E). Therefore, the DA state can be explained by the simultaneous 25

27 occurrence of two low probability events in the same cell: 1) stochastic initiation of ComK expression and 2) close proximity to spore formation. Since our conclusions (Figures 1D and S2D) indicate that these events are independent, the probability of observing DA cells should correspond to the product of the probability of competence initiation (measured to be ~2%, Table S1) and the probability of a cell to reside in the time window between the localization of SpoIIE and expression of SpoIIR. From the data shown in Figure 3E, this time window was measured to be ~1 hour, while the average time for a cell to reach spore formation under our movie conditions was found to be ~17 hours (Figure S12). Thus, the expected frequency of DA cells under our conditions is 2% times 6%, or ~.1%, which is similar to the experimentally observed frequency of.11% ±.4% (Table S1). 4 Spore count 3 2 mean 1 17 h Supplementary Figure S12. Average time to spore formation. Phase-bright spores were counted in each frame of a typical time-lapse movie of sporulating B. subtilis microcolony (strain F-comG). Under our imaging conditions, ~5% of observed cells reach sporulation in 17 hours. 26

28 Supplemental References: Flynn JM, Levchenko I, Seidel M, Wickner SH, Sauer RT, and Baker TA (21) Overlapping recognition determinants within the ssra degradation tag allow modulation of proteolysis. Proc Natl Acad Sci U S A 98: Gottesman S, Roche E, Zhou Y, and Sauer RT (1998) The ClpXP and ClpAP proteases degrade proteins with carboxy terminal peptide tails added by the SsrA tagging system. Genes Dev 12: Griffith KL, and Grossman AD (28) Inducible protein degradation in Bacillus subtilis using heterologous peptide tags and adaptor proteins to target substrates to the protease ClpXP. Mol Microbiol 7: Haima P, Bron S, and Venema G (1987) The effect of restriction on shotgun cloning and plasmid stability in Bacillus subtilis Marburg. Mol Gen Genet 29: Levchenko I, Seidel M, Sauer RT, and Baker TA (2) A specificity enhancing factor for the ClpXP degradation machine. Science 289: Middleton R, and Hofmeister A (24) New shuttle vectors for ectopic insertion of genes into Bacillus subtilis. Plasmid 51: Sterlini JM, and Mandelstam J (1969) Commitment to sporulation in Bacillus subtilis and its relationship to development of actinomycin resistance. Biochem J 113: Suel GM, Garcia Ojalvo J, Liberman LM, and Elowitz MB (26) An excitable gene regulatory circuit induces transient cellular differentiation. Nature 44: Wiegert T, and Schumann W (21) SsrA mediated tagging in Bacillus subtilis. J Bacteriol 183: