RESEARCH ARTICLE

crossm The dev Operon Regulates the Timing of Sporulation during Myxococcus xanthus Development Ramya Rajagopalan, Lee Kroos Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, Michigan, USA

ABSTRACT Myxococcus xanthus undergoes multicellular development when starved.

Thousands of rod-shaped cells coordinate their movements and aggregate into mounds in which cells differentiate into spores. Mutations in the dev operon impair development. The dev operon encompasses a clustered regularly interspaced short palindromic repeat-associated (CRISPR-Cas) system. Null mutations in devI, a small gene at the beginning of the dev operon, suppress the developmental defects caused by null mutations in the downstream devR and devS genes but failed to suppress defects caused by a small in-frame deletion in devT. We provide evidence that the original mutant has a second-site mutation. We show that devT null mutants exhibit developmental defects indistinguishable from devR and devS null mutants, and a null mutation in devI suppresses the defects of a devT null mutation. The similarity of DevTRS proteins to components of the CRISPR-associated complex for antiviral defense (Cascade), together with our molecular characterization of dev mutants, support a model in which DevTRS form a Cascade-like subcomplex that negatively autoregulates dev transcript accumulation and prevents DevI overproduction that would strongly inhibit sporulation. Our results also suggest that DevI transiently inhibits sporulation when regulated normally. The mechanism of transient inhibition may involve MrpC, a key transcription factor, whose translation appears to be weakly inhibited by DevI. Finally, our characterization of a devI devS mutant indicates that very little exo transcript is required for sporulation, which is surprising since Exo proteins help form the polysaccharide spore coat.

Received 9 November 2016 Accepted 28 February 2017 Accepted manuscript posted online 6 March 2017 Citation Rajagopalan R, Kroos L. 2017. The dev operon regulates the timing of sporulation during Myxococcus xanthus development. J Bacteriol 199:e00788-16. https://doi.org/ 10.1128/JB.00788-16. Editor George O'Toole, Geisel School of Medicine at Dartmouth Copyright © 2017 American Society for Microbiology. All Rights Reserved. Address correspondence to Lee Kroos, [email protected].

IMPORTANCE CRISPR-Cas systems typically function as adaptive immune systems in

bacteria. The dev CRISPR-Cas system of M. xanthus has been proposed to prevent bacteriophage infection during development, but how dev controls sporulation has been elusive. Recent evidence supported a model in which DevR and DevS prevent overproduction of DevI, a predicted 40-residue inhibitor of sporulation. We provide genetic evidence that DevT functions together with DevR and DevS to prevent DevI overproduction. We also show that spores form about 6 h earlier in mutants lacking devI than in the wild type. Only a minority of natural isolates appear to have a functional dev promoter and devI, suggesting that a functional dev CRISPR-Cas system evolved recently in niches where delayed sporulation and/or protection from bacteriophage infection proved advantageous. KEYWORDS CRISPR-Cas, Myxococcus xanthus, bacterial development, dev operon, gene regulation, signaling, sporulation

M

yxococcus xanthus is a Gram-negative soil bacterium that exhibits social behaviors (1). In nature, cells of M. xanthus move in groups, feeding on prey bacteria and organic detritus. When starved, thousands of cells aggregate to form mounds that mature into fruiting bodies as rod-shaped cells differentiate into dormant ovoid spores.

May 2017 Volume 199 Issue 10 e00788-16

Journal of Bacteriology

jb.asm.org 1

Rajagopalan and Kroos

Journal of Bacteriology

When favorable conditions return, the spores can germinate to produce a group of rod-shaped cells ready to feed. Interestingly, only a small percentage of the cells in a starved population of M. xanthus become spores. Some cells persist outside fruiting bodies as peripheral rods (2), while the majority of the initial population undergoes cellular lysis (3, 4). The molecular events that determine cell fate during the developmental process are largely unknown, but recently, the timing of commitment to the spore fate was found to be at 24 to 30 h poststarvation under particular laboratory conditions (5). Prior to that period, the addition of nutrients to developing cells prevented spore formation. However, from 24 to 30 h poststarvation, an increasing number of cells became committed to subsequent spore formation despite the addition of nutrients. Although the molecular events that constitute commitment to sporulation are unknown, the addition of nutrients to developing cells caused rapid proteolysis of the key transcription factors MrpC and FruA and halted expression of the dev operon (5), which was known to contain at least three genes (devTRS) important for sporulation (6–8) (Fig. 1). Since the dev operon is a direct target of transcriptional activation by MrpC and FruA (9), a simple model proposes that dev expression governs commitment to sporulation (5). However, recent work showed that deletion of the dev promoter or the first gene of the operon, devI, suppresses the sporulation defects of devR and devS mutants (10). Since devR and devS mutants accumulate more dev transcripts than the wild type, it was proposed that overexpression of devI in the mutants inhibits sporulation. Additional experiments suggested that inhibition is mediated by the product of devI, a small 40-residue protein, in a cell-autonomous fashion. If DevI inhibits sporulation when overproduced, does the loss of DevI advance the timing of sporulation? Here, we carefully address this question and we investigate molecular changes associated with the loss of DevI during development. Another open question from the previous study is the role of DevT during development (10). The developmental defect of a devT mutant was more severe than those of devR and devS mutants, and the devT defect was not overcome by deletion of the dev promoter or devI. This suggested that DevT plays a different role in development than DevR and DevS. The devT mutant studied previously has a small in-frame deletion (6, 10), so it is possible that a DevT variant with altered activity is produced. Alternatively, it is possible that the devT mutant acquired a second-site mutation that affects development (10). Here, we clarify the role of DevT in development. The dev operon encompasses a clustered regularly interspaced short palindromic repeat-associated (CRISPR-Cas) system (8). CRISPR-Cas systems typically function as RNA interference (RNAi)-based adaptive immune systems in bacteria, and components of these systems are being used widely as tools for genetic engineering of many organisms (11). An alternative role of CRISPR-Cas systems in bacteria is in endogenous gene regulation (12). Negative autoregulation by devRS provided one of the first examples of this role (7, 8, 10) (Fig. 1). Whether devT and/or other genes in the dev operon participate in negative autoregulation is unknown, but this would not be surprising given their predicted functions. DevT is similar to Cas8a1 (10), which is the large subunit of the CRISPR-associated complex for antiviral defense (Cascade) (13). DevR and DevS are orthologs of Cas7 and Cas5, respectively, which are repeat-associated mysterious proteins (RAMPs) that are typically part of the Cascade complex. Cas6 is predicted to be an endoribonuclease that cleaves pre-CRISPR RNA (pre-crRNA) to produce mature crRNA, which assembles with Cas6 and other proteins to form the Cascade complex (14). The Cascade complex typically recruits Cas3 to cleave target DNA. In the dev operon, cas6 and cas3 precede the devTRS genes (Fig. 1). Here, we report the phenotypes of cas6 and cas3 mutants (kindly provided by A. Garza) with respect to sporulation and negative autoregulation of the dev operon. RESULTS Deletions of devT cause developmental defects similar to those of deletions of devR or devS. Previously, a devT mutant with an in-frame deletion of codons 408 to 502 May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 2

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

FIG 1 Map of the dev operon. The operon includes eight genes and the downstream CRISPR (8). The devR and devS products negatively autoregulate the dev promoter (7, 8, 10).

(ΔdevT408 –502 mutant) failed to form mounds in submerged culture, and its developmental defect was not suppressed by a deletion of the dev promoter (ΔPdev) or devI (10). Since ΔdevR and ΔdevS mutants form mounds but very few spores, and this defect was suppressed by ΔPdev and ΔdevI deletions, we questioned whether the ΔdevT408 –502 mutant had acquired a second-site mutation that affects development. Therefore, we reconstructed the ΔdevT408 –502 mutant. We also considered the possibility that the ΔdevT408 –502 mutation does not cause a complete loss of DevT function, so we constructed a devT mutant with an in-frame deletion of codons 13 to 541 (ΔdevT13–541 mutant) (i.e., nearly the entire 570-codon gene). Both the reconstructed ΔdevT408 –502 mutant and the ΔdevT13–541 mutant formed mounds that failed to darken even at 72 h poststarvation in submerged culture (Fig. 2A). This phenotype is unlike that of the original ΔdevT408 –502 mutant, which failed to form mounds, but similar to that of ΔdevR and ΔdevS mutants reported previously (10). Failure of mounds to darken

FIG 2 Development of devT mutants under submerged culture conditions. (A) Fruiting body formation by wild-type DK1622 and development of the indicated devT mutants. DK1622 formed dark fruiting bodies by 72 h poststarvation (an arrow points to one), as did the ΔdevT13–541 ΔPdev and ΔdevT13–541 ΔdevI mutants, but the original ΔdevT408 –502 mutant failed to form mounds, and the mounds formed by the reconstructed ΔdevT408 –502 mutant and the ΔdevT13–541 mutant failed to darken. Scale bar, 100 ␮m. Similar results were observed in at least two biological replicates. (B) Cellular shape change by wild-type DK1622 and the indicated devT mutants. DK1622 formed ovoid spores by 48 h poststarvation (an arrow points to one), but shape change was delayed in the devT mutants (arrows point to thickened rods at 48 h and ovoid spores at 72 h). Scale bar, 5 ␮m. Similar results were observed in at least two biological replicates. May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 3

Rajagopalan and Kroos

Journal of Bacteriology

TABLE 1 Sporulation levels Strain Wild-type DK1622 Original ΔdevT408–502 mutant Reconstructed ΔdevT408–502 mutant ΔdevT13–541 mutant Original ΔdevT408–502 attB::pRR031 mutant Reconstructed ΔdevT408–502 attB::pRR031 mutant ΔdevT13–541 attB::pRR031 mutant ΔdevT13–541 ΔPdev mutant ΔdevT13–541 ΔdevI mutant Δcas6 mutant Δcas3 mutant ΔPdev-cas2 mutant ΔdevI-cas2 mutant ΔdevS mutant ΔdevI mutant ΔdevI ΔdevS mutant exoA mutant exoA ΔdevS mutant exoA ΔdevI mutant exoA ΔdevI ΔdevS mutant aNumbers

No. of spores (CFU/ml)a 8 ⫻ 107 9 ⫻ 102 2 ⫻ 102 5 ⫻ 102 5 ⫻ 102 7 ⫻ 106 4 ⫻ 106 4 ⫻ 107 4 ⫻ 107 1 ⫻ 107 5 ⫻ 107 6 ⫻ 107 5 ⫻ 107 6 ⫻ 102 6 ⫻ 107 6 ⫻ 107 0 0 0 0

are the average of the results from at least two biological replicates.

typically correlates with reduced spore formation. To determine whether spores were forming, samples were collected at 24, 48, and 72 h poststarvation, gently dispersed, and observed microscopically. Neither the reconstructed ΔdevT408 –502 mutant nor the ΔdevT13–541 mutant exhibited a cellular shape change at 24 h poststarvation, but a few thickened rods were observed at 48 h, and a few ovoid spores could be seen by 72 h (Fig. 2B). In contrast, the parent strain DK1622 from which the mutants were derived exhibited a shape change as early as 24 h poststarvation, and many cells formed ovoid spores by 48 h. Quantification of mature spores (i.e., those that were heat resistant, sonication resistant, and germination competent) at 72 h poststarvation revealed that the reconstructed ΔdevT408 –502 mutant and the ΔdevT13–541 mutant form far fewer mature spores than DK1622 during submerged culture development (Table 1). Both mutants are indistinguishable from ΔdevR and ΔdevS mutants in terms of their sporulation phenotype (10). These results suggest that the developmental phenotype of the original ΔdevT408 –502 mutant was most likely due to a second-site mutation elsewhere in the chromosome rather than a change in DevT function due to the ΔdevT408 –502 mutation. To further investigate the difference in the developmental phenotype of the original devT mutant compared with those of the newly constructed mutants, we performed a complementation test of each mutant. The devT gene was transcriptionally fused to the dev promoter region (⫺114 to ⫹71 relative to the transcriptional start site) in a plasmid that was transformed into each of the devT mutants. The plasmid integrates ectopically at a phage attachment site in the chromosome. Complementation was observed for both the reconstructed ΔdevT408 –502 mutant and the ΔdevT13–541 mutant but not the original ΔdevT408 –502 mutant. Both of the newly constructed mutants formed darkened mounds (see Fig. S1A in the supplemental material) and many more mature spores when complemented (Table 1), although the number of spores was less than for DK1622, perhaps owing to lower expression of devT from a nonnative context. In contrast, the original ΔdevT408 –502 mutant still failed to form mounds (Fig. S1A) or many mature spores (Table 1), despite the presence of the ectopically integrated plasmid. These results provide strong evidence that the original ΔdevT408 –502 mutant has a second-site mutation that blocks development at an early stage prior to mound formation. It was reported previously that accumulation of the transcription factor FruA was reduced in the original ΔdevT408 –502 mutant (6). We confirmed this result; however, the FruA level was elevated about 2-fold in the reconstructed ΔdevT408 –502 mutant May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 4

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

FIG 3 FruA (A) and MrpC (B) levels during development of devT mutants. The indicated strains were starved under submerged culture conditions. Samples were collected at the indicated times (in hours), and equal amounts of protein (1 ␮g) were analyzed by immunoblotting using anti-FruA or anti-MrpC antibodies. Representative immunoblots are shown from two or three biological replicates. The graph below the immunoblots shows quantification of signal intensities relative to a 15-h sample of wild-type DK1622 on the same immunoblot. Values are the average of the results from the replicates, and error bars indicate one standard deviation from the mean.

compared with wild-type DK1622, and it was elevated slightly less in the ΔdevT13–541 mutant (Fig. 3A). The results are consistent with the evidence presented above that the original ΔdevT408 –502 mutant has an additional mutation that causes earlier blockage of the developmental program. Further, the results suggest that DevT exerts a weak negative regulatory effect on FruA accumulation rather than a positive regulatory effect. Since the transcription of fruA depends on MrpC (15), and the two transcription factors bind DNA promoter regions cooperatively to regulate the transcription of many genes (16–20), including the dev operon (9), we also measured the MrpC level during development of the ΔdevT mutants. Consistent with an earlier blockage of the developmental program in the original ΔdevT408 –502 mutant, MrpC accumulation was reduced relative to wild-type DK1622 (Fig. 3B). In contrast, the MrpC level was elevated about 2-fold in the reconstructed ΔdevT408 –502 mutant and the ΔdevT13–541 mutant. The elevated MrpC level might explain the elevated FruA level (Fig. 3A) since MrpC activates fruA transcription (15). It appears that DevT exerts a weak negative regulatory effect on MrpC accumulation, which in turn reduces the FruA level, overall comprising a negative feedback loop, since MrpC and FruA cooperatively activate dev transcription (9). DevT is not unique in this respect. In a subsequent section, we present results that suggest DevS and DevI also exert weak negative regulatory effects that normally reduce the MrpC and FruA levels during development. Since the reconstructed ΔdevT408 –502 mutant and the ΔdevT13–541 mutant exhibited similar developmental defects, we focused on the ΔdevT13–541 mutant for further May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 5

Rajagopalan and Kroos

Journal of Bacteriology

FIG 4 Levels of dev transcripts in ΔdevT13–541 and Δcas mutants. The indicated strains were starved under submerged culture conditions. At 24 h poststarvation, cultures were harvested, RNA was isolated, and the RNA was subjected to RT-qPCR analysis using primers PdevF-2 and PdevR-3. Values are the averages of the results from three biological replicates relative to wild-type DK1622, and error bars indicate one standard deviation from the mean.

comparison with the ΔdevR and ΔdevS mutants. As noted above, the sporulation defect of ΔdevR and ΔdevS mutants can be suppressed by ΔPdev and ΔdevI mutations (10). To test whether the ΔdevT13–541 mutant sporulation defect can be suppressed similarly, ΔdevT13–541 ΔPdev and ΔdevT13–541 ΔdevI double mutants were constructed and subjected to submerged culture starvation. The double mutants appeared to develop normally (Fig. 2A and Table 1), indicative of suppression, as observed previously for the ΔdevR and ΔdevS mutants (10). The ΔdevR and ΔdevS mutants accumulate about 10-fold more dev transcripts than wild-type DK1622, suggesting that DevR and DevS negatively autoregulate their operon (7, 8, 10). To test whether the ΔdevT13–541 mutant shares this characteristic, reverse transcription-quantitative PCR (RT-qPCR) was performed on RNA extracted at 24 h poststarvation using primers that amplify positions ⫹237 to ⫹440 relative to the dev transcriptional start site. The dev transcripts were about 10-fold more abundant in the ΔdevT13–541 mutant than in wild-type DK1622 (Fig. 4), similar to the results observed for the ΔdevR and ΔdevS mutants (10). We conclude that devT plays a role in development similar to that of devR and devS. Since the three genes are predicted to code for components of a Cascade-like complex, we propose that DevT, DevR, and DevS form a complex that negatively autoregulates transcription from Pdev and prevents overexpression of devI that would inhibit sporulation. This model would explain why the ΔPdev and ΔdevI mutations suppress the developmental defects of the ΔdevT13–541, ΔdevR, and ΔdevS mutants (10) (Fig. 2A and Table 1). Deletions of cas6 and cas3 do not cause severe developmental defects or increased accumulation of dev transcripts. The cas6 and cas3 genes lie between devI and devTRS in the dev operon (Fig. 1). Cas6 is predicted to be an endoribonuclease that cleaves pre-CRISPR RNA to produce short mature crRNA (10). The crRNA and Cas6 are predicted to assemble with the DevTRS proteins, forming a Cascade-like complex. Typically, a crRNA targets a Cascade complex to viral or plasmid DNA, and Cas3 is recruited and activated to cleave a target DNA (14). Therefore, we questioned whether mutations in cas6 and cas3 of the dev operon would cause developmental defects similar to those with mutations in devTRS. We obtained mutants with an in-frame deletion in cas6 or cas3 from Anthony Garza (Syracuse University) and induced development under submerged culture conditions. The Δcas6 and Δcas3 mutants appeared to develop normally, forming mounds that darkened, similar to wild-type DK1622 (Fig. S1B). At 72 h poststarvation, the number of mature spores for the Δcas6 mutant was 8-fold less than that of the wild type, while the spore number for the Δcas3 mutant was similar to that of the wild type (Table 1). RT-qPCR performed on RNA extracted from the Δcas6 and Δcas3 mutants at 24 h poststarvation showed dev transcript levels similar to May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 6

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

those of wild-type DK1622 (Fig. 4). These results show that cas6 and cas3 do not function in development in the same way as devTRS. We propose that DevTRS form a subcomplex that negatively autoregulates transcript accumulation and prevents devI overexpression, while DevTRS form a larger Cascade-like complex with Cas6 and crRNA that performs a different function during development, which is nonessential for development under laboratory conditions (see Discussion). Timing of sporulation is advanced in mutants that fail to express devI. If DevTRS proteins negatively autoregulate transcription in order to limit expression of devI, as we propose, the loss of devI might advance the timing of sporulation, since none of the DevI sporulation inhibitor would be produced. To test this prediction of our model, we examined the development of the ΔdevI and ΔPdev mutants more carefully than in our previous work (10). The mutants were induced to develop under submerged culture conditions, and phenotypes were observed at 18 to 36 h poststarvation, the critical period during which rod-shaped cells have moved into mounds and begun to undergo sporulation. Cellular shape change was examined by microscopic observation, as described previously (5). In agreement with previous results, thickened rods and ovoid spores were first observed at 24 h poststarvation for wild-type DK1622, with many rounded cells being observed by 30 h (Fig. 5A). For the ΔdevI and ΔPdev mutants, thickened rods were observed as early as 21 h poststarvation, indicating that the timing of cellular shape change was advanced. Likewise, thickened rods were observed by 21 h poststarvation for a ΔdevI ΔdevS double mutant, whereas a ΔdevS single-mutant control did not exhibit cellular shape change even by 30 h poststarvation. We also quantified the formation of sonication-resistant spores, as described previously (5). DK1622 formed spores by 27 h poststarvation, and the number increased 10-fold by 36 h (Fig. 5B). Strikingly, the ΔdevI, ΔPdev, and ΔdevI ΔdevS mutants formed sonication-resistant spores earlier than the wild type. The mutants exhibited spores by 21 h poststarvation, and the number increased 10-fold by 27 h. As expected, the ΔdevS mutant control did not exhibit spores even by 36 h poststarvation. We conclude that the timing of cellular shape change and sonication-resistant spore formation is advanced in mutants that fail to express devI, consistent with our hypothesis that DevI is a sporulation inhibitor. Our findings that the ΔPdev mutant forms a normal number of mature spores at 72 h poststarvation (10), undergoes cellular shape change, and forms sonication-resistant spores earlier than normal (Fig. 5) suggest that products of the dev operon function only to delay sporulation. However, we could not rule out the possibility that expression of one or more dev operon products expressed from another promoter (e.g., readthrough from an upstream promoter or transcription from a suboperonic promoter downstream of Pdev) functions positively in development. To test this possibility, we constructed a mutant with a deletion spanning from ⫺38 upstream of the Pdev transcriptional start site to ⫹8407 near the 3= end of cas2, the last gene before the CRISPR at the dev locus (Fig. 1). In the resulting ΔPdev-cas2 mutant, cellular shape change and sonication-resistant spore formation were advanced indistinguishably from the ΔPdev and devI mutants (Fig. 5). The number of mature spores for the ΔPdev-cas2 mutant was similar to that of wild-type DK1622 (Table 1). Moreover, a ΔdevI-cas2 mutant with a deletion spanning from the second codon of devI (⫹45 downstream of the Pdev transcriptional start site) to ⫹8407 also exhibited a shape change earlier than normal (data not shown) and a number of mature spores similar to that for DK1622 (Table 1). The ΔdevI-cas2 mutant has the dev promoter but not the dev genes. We conclude that expression of the dev genes from the dev promoter normally delays sporulation, but the dev operon is dispensable for spore formation. Taken together with our previous results that suggested that DevI is a sporulation inhibitor (10), our results presented here show that DevI acts as a timer of sporulation under the control of DevTRS-mediated negative autoregulation and that the cas6, cas3, fused cas4 and cas1, and cas2 genes are not required to promote development, at least under submerged culture conditions in the laboratory. May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 7

Rajagopalan and Kroos

Journal of Bacteriology

FIG 5 Timing of sporulation in mutants that fail to express devI. (A) Cellular shape change. The indicated strains were starved under submerged culture conditions and samples were collected, gently dispersed, and examined microscopically at the times indicated. Photos show densely packed cell aggregates presumed to be nascent fruiting bodies. Arrows indicate thickened rods or ovoid spores. Scale bar, 5 ␮m. Similar results were observed in at least two biological replicates. (B) Sonication-resistant spores. The indicated strains were starved under submerged culture conditions and samples were collected at the indicated times (in hours) for the measurement of sonication-resistant spores. Values (log10) are the averages of the results from at least three biological replicates, and error bars represent one standard deviation from the mean.

MrpC is elevated in the absence of devI. The earlier sporulation of mutants that fail to express devI is reminiscent of several mutants that appear to accelerate the developmental program (21). In some cases, accelerated development is correlated with elevated levels of MrpC and FruA early during development (22–24). Therefore, we tested whether the timing of MrpC and FruA accumulation is advanced in the absence of devI. Immunoblots revealed that the MrpC level is about 2-fold higher in the ΔdevI mutant than in wild-type DK1622 at 15 and 18 h poststarvation (Fig. 6A). Since MrpC appears to activate fruA transcription (15), we expected that FruA would be elevated in the ΔdevI mutant. FruA was on average slightly elevated at 15 to 24 h poststarvation in the ΔdevI mutant (Fig. 6B). These results suggest that DevI exerts a weak negative regulatory effect that normally reduces the MrpC and FruA levels during development. May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 8

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

FIG 6 MrpC (A) and FruA (B) levels during development in the absence of devI. The indicated strains were starved under submerged culture conditions. Samples were collected at the indicated times (in hours), and equal amounts of protein (1 ␮g) were analyzed by immunoblotting using anti-MrpC or anti-FruA antibodies. Representative immunoblots are shown from two or three biological replicates. The graph below the immunoblots shows quantification of signal intensities relative to a 15-h sample of wild-type DK1622 on the same immunoblot. Values are the averages of the results from replicates, and error bars indicate one standard deviation from the mean.

We conclude that elevated MrpC and FruA correlate with earlier sporulation of the ΔdevI mutant. How does the absence of devI cause the MrpC and FruA levels to be elevated? First, we compared the levels of mrpC and fruA transcripts in the ΔdevI mutant with those in the wild type using RT-qPCR. No significant difference was observed at 12 to 24 h poststarvation (Fig. S2). In both strains, the transcript levels remained constant. This suggested that the absence of devI upregulates the transcription factors posttranscriptionally. Since the effect on MrpC was larger than that on FruA (Fig. 6), we focused on MrpC and compared the rates of protein turnover in the ΔdevI mutant and wild type. Both strains were treated with chloramphenicol, an inhibitor of protein synthesis, at 18 h poststarvation. Samples were collected at the indicated times after chloramphenicol addition or, as a control, untreated samples were collected at equivalent times. Immunoblot signals were quantified and used to calculate the half-life of MrpC after chloramphenicol treatment. The MrpC half-life was similar in the ΔdevI and in wild-type DK1622 (Fig. S3). This suggests that an equivalent number of mrpC transcripts (Fig. S2A) are translated more efficiently in the ΔdevI mutant, accounting for the higher level of MrpC protein (Fig. 6A). Taken together, our results suggest that one function of DevI is to exert a mild inhibitory effect on the translation of mrpC transcripts, constituting a negative feedback loop in the regulatory network that reduces early accumulation of MrpC and FruA, perhaps accounting for the delayed sporulation of the wild type relative to devI mutants (Fig. 5). We also measured the levels of MrpC and FruA in the ΔdevS and ΔdevI ΔdevS mutants. The MrpC and FruA levels in the ΔdevS single mutant were comparable to May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 9

Rajagopalan and Kroos

Journal of Bacteriology

FIG 7 Levels of exo transcripts during development in the absence of devI. (A) exo transcripts at 30 h poststarvation. The indicated strains were starved under submerged culture conditions. At 30 h poststarvation, cultures were harvested, RNA was isolated, and the RNA was subjected to RT-pPCR analysis using primers Pexo-F-3 and Pexo-R to determine the exo transcript level relative to one of the biological replicates of wild-type DK1622. Values are the averages of the results from three biological replicates, and error bars indicate one standard deviation from the mean. (B) exo transcripts earlier during development of wild-type DK1622 and the ΔdevI mutant. Samples were collected at the indicated times and analyzed as described for panel A, except the exo transcript level was determined relative to the 12-h sample of wild-type DK1622 in the same experiment. Values are the averages of the results from three biological replicates, and error bars indicate one standard deviation from the mean.

those in the ΔdevI mutant (Fig. 6) and similar to those in the reconstructed ΔdevT408 – 502 mutant and the ΔdevT13–541 mutant (Fig. 3). MrpC was elevated about 2-fold, and FruA was, on average, slightly elevated. These results suggest that like DevI and DevT, DevS exerts weak negative regulatory effects that normally reduce the MrpC and FruA levels during development. We propose that DevTRS form a complex that inhibits MrpC accumulation, and since MrpC appears to activate fruA transcription (15), FruA accumulation is also inhibited. Interestingly, the MrpC level was even more elevated in the ΔdevI ΔdevS double mutant (about 3-fold at 15 to 24 h) than in either single mutant (Fig. 6A), suggesting that DevI and DevTRS inhibit MrpC accumulation by more than one mechanism. Unexpectedly, the ΔdevI ΔdevS double mutant had a slightly lower level of FruA than the wild type at 18 to 24 h poststarvation (Fig. 6B), despite its elevated MrpC level (Fig. 6A). This result suggests posttranscriptional control of FruA production in the double mutant (see Discussion). We conclude that elevated MrpC, but not elevated FruA, correlates with earlier sporulation of the ΔdevI ΔdevS mutant. Expression of the exo operon is impaired in the absence of devI, but exoA and csgA are still required for sporulation. Mutations in devRS impair expression of the exo operon (25), whose products are required for the maintenance of cell shape change and spore maturation (26, 27). Since the timing of sporulation is advanced in devI mutants (Fig. 5), we hypothesized that developmental expression of the exo operon would also be advanced. Initially, we measured exo transcripts at 30 h poststarvation. We found that exo transcripts were about 3-fold lower in the ΔdevI mutant than in wild-type DK1622 and 15 to 20-fold lower in the ΔdevS and ΔdevI ΔdevS mutants than in the wild type (Fig. 7A). The low level of exo transcripts in the ΔdevS mutant was expected (25), but the low level in the ΔdevI and ΔdevI ΔdevS mutants was surprising, since these mutants form mature spores (10). To test whether exo transcripts are higher in the ΔdevI mutant than in the wild type earlier during development, RT-qPCR was performed on RNA extracted at 12 to 24 h poststarvation. The exo transcripts in the mutant were as low as or lower than those in the wild type (Fig. 7B). As the level of exo transcripts rose at 21 to 30 h poststarvation, it was highly variable between biological replicates (Fig. 7, error bars). Despite the low induction of exo in some replicates of May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 10

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

wild-type DK1622 and especially the ΔdevI mutant, the formation of spores was not impaired for these replicates, suggesting that high induction of exo expression is not required for normal development, including in the absence of devI. To investigate whether the normal requirement of exo for sporulation is bypassed in the absence of devI, an insertion mutation in exoA, the first gene of the exo operon, was made in the ΔdevI and ΔdevI ΔdevS mutants, and in wild-type DK1622 and the ΔdevS mutant as controls. During submerged culture development, mounds of the wild type started to darken by 30 h poststarvation and continued to darken by 48 h (Fig. S4A). Mounds of the exoA mutant did not darken as much as those of the wild type, as reported previously for the development of exoC (also called fdgA) mutants on starvation agar (26, 27). As expected, mounds of the ΔdevS exoA double mutant failed to darken by 48 h poststarvation (Fig. S4A). Mounds of the ΔdevI exoA and ΔdevI ΔdevS exoA mutants were indistinguishable from those of the exoA mutant, failing to darken as much as those of the wild type. These three strains formed fewer sonication-resistant spores than the wild type at 27 and 30 h poststarvation, whereas no sonicationresistant spores were seen for the ΔdevS exoA mutant (Fig. S4B). None of the exoA mutants produced mature spores (Table 1). These results confirm the importance of the exo operon for sporulation under submerged culture conditions, as shown previously on starvation agar (26, 27). Our results also indicate that the requirement for exo is not bypassed in the absence of devI and/or devS. Since the full induction of exo transcription depends on csgA (25, 26), which codes for the C-signal that is normally required for sporulation (28, 29), yet devI mutants form spores (10) without fully inducing exo transcription (Fig. 7), we tested whether the requirement for csgA is bypassed in the absence of devI. An insertion mutation in csgA was introduced into strains, as described above for exoA. All the strains failed to form compact mounds by 48 h poststarvation (Fig. S4C), and no sonication-resistant spores were observed microscopically, demonstrating that the need for csgA is not bypassed in the absence of devI and/or devS. Taken together, our results show that in the absence of devI and/or devS, transcription of the exo operon is weakly induced during sporulation (Fig. 7); yet, a normal number of mature spores form in the absence of devI, and this sporulation depends on exoA and csgA. DISCUSSION Our results clarify the role of devT during M. xanthus development and demonstrate that devI normally delays sporulation. We found that devT null mutants are indistinguishable from devR and devS null mutants in terms of their developmental defects. This suggests that DevT acts in concert with DevR and DevS to regulate development. Since DevTRS are similar to components of the Cascade complex (10), we propose that DevTRS form a Cascade-like subcomplex that negatively autoregulates dev transcript accumulation and prevents overexpression of devI. This regulation is crucial, since DevI appears to inhibit sporulation if overproduced (10). DevI also appears to inhibit sporulation transiently, since our results show that wild-type DK1622 forms spores about 6 h later than do devI null mutants. We propose that the primary effect of DevI is to inhibit sporulation and that DevTRS relieve the inhibition by downregulating dev transcript accumulation and, hence, DevI production. These effects are highlighted in red in Fig. 8, which shows a simple model of the gene regulatory network (GRN) governing sporulation. Based on our analysis of other molecular markers in the GRN, we infer that both DevI and DevTRS exert weak negative regulation on MrpC accumulation (Fig. 8, blue) and weak (DevI) to strong (DevTRS) positive regulation on exo transcript accumulation (Fig. 8, green). Below, we describe the GRN governing M. xanthus sporulation briefly, and then we focus on the implications of our results for regulation by proteins encoded by the dev operon. GRN governing sporulation. MrpC and FruA are transcription factors that cooperatively bind to the dev promoter region and appear to activate transcription (9). Also, MrpC appears to activate transcription of fruA (15). Hence, MrpC, FruA, and the dev May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 11

Rajagopalan and Kroos

Journal of Bacteriology

FIG 8 Model of the gene regulatory network governing M. xanthus sporulation. Starvation and C-signal (the product of csgA) enhance the activity of transcription factors MrpC and FruA, respectively, which exert positive regulation, indicated by black arrows, including combinatorial control of dev transcription. DevTRS are proposed to negatively autoregulate dev transcript accumulation, and DevI is proposed to strongly inhibit spore formation when overproduced in devT, devR, or devS null mutants (red lines). DevI is proposed to transiently inhibit spore formation when produced normally (as in wild-type strain DK1622) by the same mechanism as when overproduced and/or by weak negative regulation of MrpC (solid blue line) at the level of translation. DevTRS also exert weak negative regulation of MrpC, which has no apparent impact on sporulation (dashed blue line). Also dispensable for sporulation are strong and weak positive regulation of exo that appears to be mediated by DevTRS and DevI, respectively (thick and thin green arrows, respectively).

operon form the core of the GRN governing sporulation (Fig. 8). The regulatory inputs and outputs of MrpC and FruA are complex and have been reviewed recently (30). Briefly, starvation and C-signal enhance MrpC and FruA activity, respectively, and MrpC positively autoregulates and stimulates C-signal production (Fig. 8). FruA appears to activate transcription of the exo operon, whose products are required for sporulation, and FruA and MrpC regulate many other genes important for sporulation that are not shown in Fig. 8 (30). Regulation by DevTRS. We propose that DevTRS form a Cascade-like subcomplex with at least one crucial regulatory function: negative autoregulation of dev transcript accumulation to prevent DevI overproduction, and two dispensable regulatory functions, weak negative regulation of MrpC and strong positive regulation of exo (Fig. 8). Both bioinformatics and characterization of mutants provide evidence that DevTRS form a Cascade-like subcomplex. The cotranscription of devTRS (8) and similarity to components of the Cascade complex (10) suggest that DevTRS might form a Cascadelike subcomplex. Mutant characterization initially suggested that DevT plays a distinct role in development (6, 10), but here, we showed that devT null mutants (Fig. 2 and 4 and Table 1) are indistinguishable from devR and devS null mutants (10) in exhibiting normal mound formation, defective cellular shape change, about 105-fold reduced formation of mature spores, and about 10-fold elevated accumulation of dev transcripts. Also, devT and devS null mutants both show about 2-fold elevated accumulation of MrpC and FruA (Fig. 3 and 6). Taken together, the bioinformatics and the characterization of mutants strongly suggest that DevTRS function together as a Cascade-like subcomplex. We use the term “Cascade-like subcomplex” because Cascade complexes include Cas6 and crRNA, and they recruit Cas3 to cleave target DNA (14), but our characterization of cas6 and cas3 null mutants suggests that Cas6 and Cas3 do not function together with DevTRS in negative autoregulation (Fig. 4) or sporulation (Table 1). We propose that negative autoregulation by DevTRS is crucial to prevent DevI overproduction (Fig. 8). The mechanism of negative autoregulation is unknown, but previous work established that a transcriptional dev-lacZ fusion with as little as the first 71 bases of the dev transcript appears to be subject to negative autoregulation (i.e., about 6-fold lower LacZ expression is observed in the wild type than in a devS null mutant) (8). This implies either that negative autoregulation by DevTRS targets the first 71 bases of the dev transcript or that negative autoregulation involves a mechanism that does not rely on the dev transcript. A variety of CRISPR-Cas-related gene regulatory mechanisms have been described (12). Some of these mechanisms rely on crRNA, but May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 12

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

the CRISPR spacers at the dev locus do not match the first 71 bases of the dev transcript or the dev promoter region (8, 10). However, we have discovered an antisense transcript originating from about 300 bases downstream of the dev transcriptional start site (⫹1) (D. Srivastava, R. Rajagopalan, and L. Kroos, unpublished data). If the antisense transcript extends through ⫹1 and into the promoter region, it would provide complementarity to RNA and DNA that could target DevTRS to the dev transcript and/or to DNA in the vicinity of the dev promoter. Targeting of DevTRS to the dev transcript could lead to the recruitment of ribonucleases that degrade the transcript (even a dev-lacZ fusion transcript), analogous to a well-established mechanism observed for type IIIB CRISPR-Cas systems (31, 32). Targeting of DevTRS to DNA in the vicinity of the dev promoter could inhibit the initiation or elongation of transcription, analogous to the use of defective Cas9 to engineer inhibition of transcription (33). Although further work will be needed to elucidate the mechanism of negative autoregulation by DevTRS, our results with a devT null mutant are consistent with results reported previously for devR and devS null mutants (10), and the results strongly support the model that negative autoregulation by DevTRS is crucial to preventing DevI overproduction. Specifically, devT, devR, and devS null mutants each accumulate about 10-fold more dev transcripts than the wild type (Fig. 4) (10), and the sporulation defect of each mutant is rescued by introduction of a devI null mutation (Table 1) (10). In addition to the apparently crucial role of DevTRS in preventing overproduction of DevI, our results also implicate DevTRS in at least two other regulatory functions that appear to be dispensable for development: weak negative regulation of MrpC and strong positive regulation of exo (Fig. 8). We observed that MrpC accumulation is elevated about 2-fold in devT (Fig. 3B) and devS (Fig. 6A) null mutants. The mechanism of regulation is unknown, and there are many possibilities, since the regulatory inputs of MrpC are complex (30). The elevated level of MrpC likely accounts for the elevated level of FruA that we observed in devT (Fig. 3A) and devS (Fig. 6B) null mutants, since MrpC appears to activate the transcription of fruA (15). Elevated levels of MrpC and FruA have been correlated with early aggregation in some studies (22–24), but we did not observe early aggregation of devT or devS null mutants in our experiments. Perhaps MrpC and/or FruA were not elevated enough to cause early aggregation. We also did not observe early aggregation of a ΔdevI ΔdevS double mutant, which exhibited a higher level of MrpC than ΔdevI or ΔdevS single mutants (Fig. 6A). However, the FruA level was not elevated in the ΔdevI ΔdevS mutant (Fig. 6B). This result was surprising, since MrpC appears to activate the transcription of fruA (15). The result suggests posttranscriptional control of FruA production in the double mutant (e.g., reduced stability or translation of fruA mRNA, or reduced stability of FruA). Perhaps the unexpectedly low level of FruA in the ΔdevI ΔdevS mutant accounts for its failure to aggregate earlier and its failure to form more sonication-resistant spores at 21 h poststarvation than the devI mutant (Fig. 5). In any case, our results provide no indication that the apparent weak negative regulation of MrpC by DevTRS further delays aggregation or sporulation in the context of the delay brought about by DevI in the wild type, so the negative feedback loop from DevTRS to MrpC appears to be dispensable for normal development (Fig. 8). Positive regulation of exo by DevTRS also appears to be dispensable for development (Fig. 8). It was surprising to find that the ΔdevI ΔdevS double mutant accumulated about 15-fold fewer exo transcripts than the wild type at 30 h poststarvation, because the ΔdevI ΔdevS mutant sporulates normally (10), yet exo mutants fail to form mature spores (25–27). In agreement with the previous reports, we found that an exoA mutant fails to form mature spores (Table 1). The exoA mutant did form sonication-resistant spores at 27 and 30 h poststarvation, albeit in reduced numbers compared with the wild type (see Fig. S4B in the supplemental material), but the spores failed to mature (i.e., become heat resistant and germination competent), consistent with the model that Exo proteins are necessary to form the polysaccharide spore coat (27). The ΔdevI ΔdevS exoA triple mutant exhibited a sporulation defect indistinguishable from the exoA single mutant (Table 1 and Fig. S4), showing that the requirement for exoA is not May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 13

Rajagopalan and Kroos

Journal of Bacteriology

bypassed in the absence of devI and devS. The requirement for csgA was also not bypassed in the absence of devI and devS (Fig. S4C), which was not surprising, since C-signal activates FruA, and this affects the expression of many genes in addition to exo (25, 30, 34, 35). What is surprising is that the low level of exo transcripts in the ΔdevI ΔdevS mutant is sufficient for sporulation, although it is possible that the spore coat of the ΔdevI ΔdevS mutant differs from that of the wild type in a way that does not affect heat and sonication resistance or germination competence. Since the ΔdevS single mutant exhibited a low level of exo transcripts similar to that of the ΔdevI ΔdevS double mutant (Fig. 7A), we infer that DevI overproduction in the ΔdevS mutant is not responsible for low exo transcript accumulation. Rather, we propose that DevTRS positively regulate exo by increasing transcript synthesis and/or stability. Interestingly, DevTRS may also regulate other genes positively. Expression of the nfs promoter fused to an mCherry reporter gene was reduced in a devRS mutant (36). Nfs proteins appear to work together with Exo proteins to form the polysaccharide spore coat (27, 37, 38). Positive regulation of exo, nfs, and perhaps other genes by DevTRS suggests a general mechanism of control. The mechanism probably does not involve the level of FruA, which is lower in the wild type than in devT (Fig. 3) and devS (Fig. 6) null mutants, but the mechanism could involve activation of FruA by C-signal, if DevTRS somehow feed back positively on C-signal. Regulation by DevI. Our results suggest three regulatory functions of DevI: inhibition of sporulation, weak negative regulation of MrpC, and weak positive regulation of exo (Fig. 8). We propose that the primary effect of DevI is to inhibit sporulation (Fig. 8). As discussed above, negative autoregulation by DevTRS appears to be crucial to relieve the inhibition of sporulation by DevI. The mechanism of sporulation inhibition by overproduced DevI in devT, devR, and devS null mutants is unknown, but the mechanism does not appear to involve an extracellular signal (10). Rather, DevI appears to be a small 40-residue protein that acts in a cell-autonomous manner. Small proteins have been shown to affect sporulation of Bacillus subtilis and various processes (signal transduction, enzyme activity, transport of ions and macromolecules across membranes, and cell division) in many bacteria (39). Recently, the 37-residue CmpA protein of B. subtilis has been shown to function as an adapter that targets SpoIVFA for proteolysis by ClpXP (40). SpoIVFA plays a key role in assembly of the spore coat. CmpA appears to detect misassembled SpoIVFA and target it for degradation, thus blocking sporulation and ensuring that defective cells eventually lyse. DevI could likewise function as an adapter that targets a protein important for rod-shaped cells to become oval spores. One such protein is MreB (27), but there are likely many others. We are pursuing a genetic approach to discover the mechanism by which overproduced DevI inhibits sporulation of devT, devR, and devS null mutants. An important question is whether the strong inhibition of sporulation apparently caused by overproduced DevI in devT, devR, and devS null mutants involves the same mechanism as the transient inhibition of sporulation apparently caused by DevI in the wild type. Our careful examination of mutants expected to fail to make DevI (ΔdevI, ΔPdev, ΔdevI ΔdevS, and ΔPdev-cas2) revealed that cellular shape change and sonicationresistant spore formation of the mutants are advanced by about 6 h compared with wild-type DK1622 (Fig. 5). None of the mutants formed aggregates earlier than normal during development. In the ΔdevI mutant, MrpC and FruA levels are elevated about 2-fold relative to the wild type (Fig. 6), and these differences are not due to altered transcript levels (Fig. S2) or altered MrpC stability (Fig. S3), suggesting that increased translation of mrpC transcripts in the ΔdevI mutant accounts for the elevated MrpC. Since MrpC appears to activate fruA transcription (15), we expected a higher level of fruA transcripts in the ΔdevI mutant than in the wild type, but this was not observed (Fig. S2B). Perhaps, fruA transcripts are synthesized more abundantly but are less stable in the ΔdevI mutant and are translated more efficiently, or FruA is more stable (to explain the elevated FruA level). In any case, our results suggest weak negative regulation May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 14

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

of MrpC by DevI at the level of translation (Fig. 8) and weak negative regulation of FruA by DevI posttranscriptionally. This negative feedback may explain the observed delay of wild-type sporulation compared with mutants unable to make DevI (Fig. 5). Alternatively, the delay might be due to a relatively low level of DevI in the wild type (due to negative autoregulation by DevTRS) transiently inhibiting sporulation by the same (unknown) mechanism that at a higher level of DevI in devT, devR, and devS null mutants strongly inhibits sporulation. It is important to note that the presumed high level of DevI in devT and devS null mutants does not decrease the levels of MrpC and FruA (Fig. 3 and 6). Rather, the levels of the two transcription factors are increased about 2-fold, as in the ΔdevI mutant. We infer that in the absence of DevTRS, overproduced DevI is either unable to negatively regulate translation of mrpC transcripts or, if such negative regulation does occur, the absence of DevTRS elevates the level of MrpC by another mechanism. As noted above, the regulatory inputs of MrpC are complex (30), so there are many potential mechanisms. Our finding that the MrpC level is higher in the ΔdevI ΔdevS double mutant than in either single mutant (Fig. 6A) suggests that DevTRS and DevI regulate the MrpC level by more than one mechanism. Clearly, though, the strong inhibition of sporulation apparently caused by overproduced DevI in devT and devS null mutants is not due to low levels of MrpC and FruA. Positive regulation of exo by DevI appears to be weaker than that by DevTRS and likewise dispensable for development (Fig. 8). The level of exo transcripts was about 3-fold lower in the ΔdevI mutant than in the wild type at 30 h poststarvation, and the levels were on average lower in the ΔdevI mutant than in the wild type at 21 and 24 h poststarvation, although the levels varied considerably between experiments (Fig. 7), and we do not understand this variability. Despite the lower levels of exo transcripts, the ΔdevI mutant appeared to sporulate normally (10), suggesting that positive regulation of exo by DevI is dispensable. We propose that DevI positively regulates exo by increasing transcript synthesis and/or stability. Whether DevI positively regulates other genes is unknown. The mechanism of regulation is also unknown but probably does not involve the level of FruA, which is lower in the wild type than in the ΔdevI mutant (Fig. 6), although the mechanism could involve activation of FruA by C-signal if DevI feeds back positively on C-signal. Functions of cas genes in the dev operon. We have proposed that DevTRS function together with Cas6, crRNA, and Cas3 in a Cascade-like complex to protect developing M. xanthus from lysogenization by bacteriophage Mx8, because the first spacer of the cotranscribed CRISPR matches the Mx8 integrase gene (8, 10). Our results suggest that neither Cas6 nor Cas3 functions together with DevTRS in negative autoregulation (Fig. 4) or sporulation (Table 1), although Cas6 appears to play some role in sporulation based on the 8-fold defect of the Δcas6 mutant. Cas4/Cas1 and Cas2 are predicted to be involved in acquiring new spacers in the dev CRISPR and possibly in another CRISPR elsewhere in the M. xanthus chromosome (10), so Cas4/Cas1 and Cas2 are not expected to be involved in development. Evolution of the dev operon. Our findings that the dev operon is dispensable for sporulation and that spore formation occurs about 6 h earlier in mutants lacking devI, Pdev, or the entire dev operon shed light on the evolution of the dev operon. As discussed previously (10), the DevTRS proteins are not conserved in most myxobacteria for which a genome sequence is available. This lack of conservation is not surprising in light of the dispensability of the dev operon for M. xanthus sporulation. Further, most natural isolates of M. xanthus appear to lack a functional dev promoter and definitely lack devI, yet these strains form a normal number of spores (10), which suggests that the dev operon is dispensable for sporulation (unless readthrough transcription from an upstream promoter or transcription from a suboperonic promoter was expressing dev genes, a possibility ruled out by our results with the ΔPdev-cas2 mutant). Based on our results, natural isolates lacking devI might form spores 6 h earlier than our laboratory wild-type strain DK1622 or natural isolates with a functional dev operon. Only 5 out of May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 15

Rajagopalan and Kroos

Journal of Bacteriology

26 natural isolates appeared to have a functional dev promoter and devI (10). We speculate that natural isolates with a functional dev operon evolved recently in niches where delayed sporulation and/or protection from phage Mx8 infection proved advantageous. MATERIALS AND METHODS Bacterial strains, plasmids, and primers. The strains, plasmids, and primers used in this study are listed in Table S1 in the supplemental material. Deletions of devT in the M. xanthus chromosome were obtained by allelic exchange using positive selection for kanamycin resistance, followed by negative selection for galactose resistance (41). We used a strategy similar to that described previously (10) to construct the plasmids used to make the deletions and to perform the allelic exchange. The details of plasmid construction and a brief description of the allelic exchange procedure follow. To reconstruct the ΔdevT408 –502 mutant with an in-frame deletion of codons 408 to 502, DNA fragments were generated by PCR using primers devT-old-Eco-F1 and devT-old-Bam-R1. Chromosomal DNA from M. xanthus strain DK11207 (ΔdevT408 –502) was used as the template. The DNA fragment generated by PCR was cloned into pCR2.1-TOPO (Invitrogen), as described by the manufacturer, to generate plasmid pRR024, the DNA sequence was verified, and the fragment was subcloned into pBJ113 using restriction sites BamHI and EcoRI to generate plasmid pRR025, which was used for allelic exchange. To construct the ΔdevT13–541 mutant with an in-frame deletion of codons 13 to 541, DNA fragments were generated by PCR using primers dev-Eco-F and dev-Int-R, and dev-Int-F and dev-Bam-R. Chromosomal DNA from M. xanthus strain DK1622 was used as the template. The DNA fragments generated by PCR served as the template for sequence overlap extension (SOE) PCR using primers dev-Eco-F and dev-Bam-R. The desired SOE PCR fragments were cloned into pCR2.1-TOPO to generate plasmid pRR022, the DNA sequence was verified, and the fragment was subcloned into pBJ113 using restriction sites BamHI and EcoRI to generate plasmid pRR023, which was used for allelic exchange. To construct the ΔPdev-cas2 mutant with a deletion spanning from ⫺38 upstream of the Pdev transcriptional start site to ⫹8407 near the 3= end of cas2, DNA fragments were generated by PCR using primers PD-Eco-Fwd and DO-OL-Rev, and DO-OL-Fwd1 and DO-Bam-Rev1. Chromosomal DNA from M. xanthus strain DK1622 was used as the template. The DNA fragments generated by PCR served as the template for SOE PCR using primers PD-Eco-Fwd and DO-Bam-Rev1. The desired SOE PCR fragments were cloned into pCR2.1-TOPO to generate plasmid pRR026, the DNA sequence was verified, and the fragment was subcloned into pBJ113 using restriction sites BamHI and EcoRI to generate plasmid pRR027, which was used for allelic exchange. To construct the ΔdevI-cas2 mutant with a deletion of devI through cas2, DNA fragments were generated by PCR using primers PD-Eco-Fwd and DO-OL-Rev2, and DO-OL-Fwd2 and DO-Bam-Rev1. Chromosomal DNA from M. xanthus strain DK1622 was used as the template. The DNA fragments generated by PCR served as the template for SOE PCR using primers PD-Eco-Fwd and DO-Bam-Rev1. The desired SOE PCR fragments were cloned into pCR2.1-TOPO to generate plasmid pRR032, the DNA sequence was verified, and the fragment was subcloned into pBJ113 using restriction sites BamHI and EcoRI to generate plasmid pRR033, which was used for allelic exchange. To perform allelic exchange, the plasmids described above were transformed into the appropriate M. xanthus strains using electroporation (42), with outgrowth in Casitone-Tris (CTT) liquid (see below for description of medium) prior to plating on CTT agar supplemented with 40 ␮g/ml kanamycin (Km) sulfate. Transformants with a single crossover were grown in CTT-yeast extract (CTTYE) liquid without selection to allow a second crossover event and then plated on CTT agar supplemented with 2.5% galactose to select for M. xanthus isolates that had lost the plasmid. Isolates that had successfully completed the allelic exchange were identified by colony PCR. To generate pRR031, which was used to complement the devT deletion strains, primers C-devT-Fwd1727 and C-devT-Int-Rev and primers C-devT-Int-Fwd and C-devT-Rev- 1727 were used to generate PCR products using chromosomal DNA from M. xanthus strain DK1622 as the template. The products were combined with BamHI-digested pPV0184 in a Gibson assembly reaction to enzymatically join the overlapping DNA fragments (43). The DNA sequence of the joined fragments was verified and the plasmid was transformed into the appropriate M. xanthus strains using electroporation (42), with outgrowth in CTT liquid prior to plating on CTT agar supplemented with 40 ␮g/ml Km sulfate to select transformants. To generate pRR028, which was used to create an insertion mutation in csgA, an internal fragment of csgA was amplified using primers csgA Fwd and csgA Rev with chromosomal DNA from M. xanthus strain DK1622 as the template. The fragment was cloned into pCR2.1-TOPO, and the DNA sequence was verified. The plasmid was transformed into the appropriate M. xanthus strains using electroporation (42), with outgrowth in CTT liquid prior to plating on CTT agar supplemented with 40 ␮g/ml Km sulfate to select transformants. Likewise, pRR029 was generated by amplifying an internal fragment of exoA using primers exoA-F2 and exoA-R1 with chromosomal DNA from M. xanthus strain DK1622 as the template, cloning the fragment into pCR2.1-TOPO and verifying the sequence, and transforming the plasmid into the appropriate M. xanthus strains to create an insertion mutation in exoA. Growth and development. Escherichia coli strains containing plasmids were grown at 37°C in Luria-Bertani (LB) medium supplemented with 50 ␮g/ml kanamycin sulfate, as required. M. xanthus strains were grown at 32°C in CTTYE liquid medium (1% Casitone, 0.2% yeast extract, 10 mM Tris-HCl [pH May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 16

Small Gene devI Delays M. xanthus Sporulation

Journal of Bacteriology

8.0], 1 mM KH2PO4·K2HPO4, 8 mM MgSO4, [final pH 7.6]) with shaking, except that outgrowth after electroporation was in CTT liquid medium (the same as CTTYE except with no yeast extract). CTT agar (1.5%) agar was used for growth on solid medium. Submerged culture development was done in 6-well plates or petri dishes with MC7 (10 mM morpholinepropanesulfonic acid [MOPS] [pH 7.0], 1 mM CaCl2) as the starvation buffer (44), as described previously (5). Briefly, log-phase cultures of M. xanthus growing in CTTYE liquid medium were collected by centrifugation, resuspended in MC7, and added to either 6-well plates or petri dishes. Upon incubation at 32°C, cells form a biofilm on the bottom of the plates and undergo development. At the indicated times, samples were collected as described previously (5). To observe cellular shape change, samples were left undisturbed for 5 to 10 min to allow cell aggregates to settle to the bottom of the tube, and cell aggregates were removed for microscopic observation, as described previously (5). For measurement of sonication resistance, samples were sonicated and refractile spores were counted microscopically using a Neubauer counting chamber, as described previously (5). Mature spores that were heat and sonication resistant and capable of germination were measured, as described previously (5). Microscopy. Images of fruiting bodies at low resolution were obtained with a Leica Wild M8 microscope equipped with an Olympus E-620 digital camera. High-resolution images of cell aggregates from nascent fruiting bodies were obtained with an Olympus BX51 microscope equipped with an Olympus DP30BW digital camera and using a differential interference contrast (DIC) filter and a 100⫻ oil-immersion objective. Immunoblot analysis. Cells developing in 6-well plates were collected and thoroughly mixed, as described previously. One sample (50 ␮l) was immediately mixed with an equal volume of 2⫻ sample buffer (0.125 M Tris-HCl [pH 6.8], 20% glycerol, 4% sodium dodecyl sulfate [SDS], 0.2% bromophenol blue, 0.2 M dithiothreitol), boiled for 5 min, and then stored at ⫺20°C. A second sample (400 ␮l) was stored at ⫺20°C, thawed, sonicated, and centrifuged, and the total protein concentration of the supernatant was determined, as described previously (5), using a Bradford assay kit (45) (Bio-Rad Laboratories), according to the manufacturer’s instructions. Equal amounts of total protein (typically 1 ␮g) in the first samples were subjected to SDS-PAGE and immunoblotting, as described previously (46). Anti-MrpC antibodies were used at a 1:10,000 dilution, and anti-FruA antibodies were used at a 1:1,000 dilution. Protein-antibody complexes were detected using horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (Bio-Rad) at a 1:5,000 dilution and chemiluminescence (Western Lightning; PerkinElmer), according to the manufacturer’s instructions. Signals were detected using a ChemiDoc MP imaging system (Bio-Rad), with exposure times short enough to ensure that the signals were not saturated. Signal intensities were quantified using Image Lab 5.1 (Bio-Rad) software. RNA extraction and analysis. The MC7 overlay of cells developing in a petri dish was replaced with 0.5 ml of RNase stop solution (5% phenol [pH ⬍ 7] in ethanol) and 4.5 ml of MC7 (47). Cells were scraped from the bottom of the plate, and the sample was aspirated into a 15-ml centrifuge tube, flash frozen in liquid nitrogen, and stored at ⫺80°C. After thawing and centrifugation at 8,700 ⫻ g and 4°C for 10 min, RNA was extracted using the hot-phenol method and digested with DNase I (Roche), as described previously (22). Total RNA (1 ␮g) was used to synthesize cDNA using SuperScript III reverse transcriptase (Life Technologies) and random primers (Promega), according to the manufacturers’ instructions. The control reactions omitted reverse transcriptase during the cDNA synthesis step. Quantitative PCR (qPCR) was done in 20-␮l reaction mixtures using SYBR green PCR master mix (Applied Biosystems), 1 ␮l of cDNA product, and 20 pmol each primer in a model 7300 real-time PCR system (Applied Biosystems). A standard curve was generated for each set of qPCRs using DK1622 chromosomal DNA, and gene expression was quantified using the relative standard curve method (user bulletin 2; Applied Biosystems). 16S rRNA was used as the internal standard. The primer sequences used for qPCR are listed in Table S1 in the supplemental material. Protein stability. Protein half-lives were calculated assuming a first-order kinetic degradation reaction, as described previously (48). Signal intensities, after subtracting background, were normalized to the relevant intensity at t ⫽ 0, and the natural log of the resulting values was plotted versus minutes after chloramphenicol (200 ␮g/ml) treatment, as described previously (5). The slope of a linear fit of the data was used to calculate the protein half-life.

SUPPLEMENTAL MATERIAL Supplemental material for this article may be found at https://doi.org/10.1128/ JB.00788-16. SUPPLEMENTAL FILE 1, PDF file, 0.4 MB. ACKNOWLEDGMENTS We thank Anthony Garza, Bryan Julien, Dale Kaiser, Lotte Sogaard-Andersen, and Mitch Singer for providing bacterial strains and Penelope Higgs for suggesting primers for fruA RT-qPCR. This research was supported by NSF grant MCB-1411272 and by salary support for L.K. from Michigan State University AgBioResearch. May 2017 Volume 199 Issue 10 e00788-16

jb.asm.org 17

Rajagopalan and Kroos

Journal of Bacteriology

REFERENCES 1. Yang Z, Higgs P. 2014. Myxobacteria: genomics, cellular and molecular biology. Caister Academic Press, Norfolk, United Kingdom. 2. O’Connor KA, Zusman DR. 1991. Development in Myxococcus xanthus involves differentiation into two cell types, peripheral rods and spores. J Bacteriol 173:3318 –3333. https://doi.org/10.1128/jb.173.11 .3318-3333.1991. 3. Wireman JW, Dworkin M. 1977. Developmentally induced autolysis during fruiting body formation by Myxococcus xanthus. J Bacteriol 129: 796 – 802. 4. Lee B, Holkenbrink C, Treuner-Lange A, Higgs PI. 2012. Myxococcus xanthus developmental cell fate production: heterogeneous accumulation of developmental regulatory proteins and reexamination of the role of MazF in developmental lysis. J Bacteriol 194:3058 –3068. https://doi .org/10.1128/JB.06756-11. 5. Rajagopalan R, Kroos L. 2014. Nutrient-regulated proteolysis of MrpC halts expression of genes important for commitment to sporulation during Myxococcus xanthus development. J Bacteriol 196:2736 –2747. https://doi.org/10.1128/JB.01692-14. 6. Boysen A, Ellehauge E, Julien B, Sogaard-Andersen L. 2002. The DevT protein stimulates synthesis of FruA, a signal transduction protein required for fruiting body morphogenesis in Myxococcus xanthus. J Bacteriol 184:1540 –1546. https://doi.org/10.1128/JB.184.6.1540-1546.2002. 7. Thöny-Meyer L, Kaiser D. 1993. devRS, an autoregulated and essential genetic locus for fruiting body development in Myxococcus xanthus. J Bacteriol 175:7450–7462. https://doi.org/10.1128/jb.175.22.7450-7462.1993. 8. Viswanathan P, Murphy K, Julien B, Garza AG, Kroos L. 2007. Regulation of dev, an operon that includes genes essential for Myxococcus xanthus development and CRISPR-associated genes and repeats. J Bacteriol 189: 3738 –3750. https://doi.org/10.1128/JB.00187-07. 9. Campbell A, Viswanathan P, Barrett T, Son B, Saha S, Kroos L. 2015. Combinatorial regulation of the dev operon by MrpC2 and FruA during Myxococcus xanthus development. J Bacteriol 197:240 –251. https://doi .org/10.1128/JB.02310-14. 10. Rajagopalan R, Wielgoss S, Lippert G, Velicer GJ, Kroos L. 2015. devI is an evolutionarily young negative regulator of Myxococcus xanthus development. J Bacteriol 197:1249 –1262. https://doi.org/10.1128/JB .02542-14. 11. Wiedenheft B, Sternberg SH, Doudna JA. 2012. RNA-guided genetic silencing systems in bacteria and archaea. Nature 482:331–338. https:// doi.org/10.1038/nature10886. 12. Westra ER, Buckling A, Fineran PC. 2014. CRISPR-Cas systems: beyond adaptive immunity. Nat Rev Microbiol 12:317–326. https://doi.org/10 .1038/nrmicro3241. 13. Makarova KS, Haft DH, Barrangou R, Brouns SJ, Charpentier E, Horvath P, Moineau S, Mojica FJ, Wolf YI, Yakunin AF, van der Oost J, Koonin EV. 2011. Evolution and classification of the CRISPR-Cas systems. Nat Rev Microbiol 9:467– 477. https://doi.org/10.1038/nrmicro2577. 14. van der Oost J, Westra ER, Jackson RN, Wiedenheft B. 2014. Unravelling the structural and mechanistic basis of CRISPR-Cas systems. Nat Rev Microbiol 12:479 – 492. https://doi.org/10.1038/nrmicro3279. 15. Ueki T, Inouye S. 2003. Identification of an activator protein required for the induction of fruA, a gene essential for fruiting body development in Myxococcus xanthus. Proc Natl Acad Sci U S A 100:8782– 8787. https:// doi.org/10.1073/pnas.1533026100. 16. Mittal S, Kroos L. 2009. A combination of unusual transcription factors binds cooperatively to control Myxococcus xanthus developmental gene expression. Proc Natl Acad Sci U S A 106:1965–1970. https://doi.org/10 .1073/pnas.0808516106. 17. Mittal S, Kroos L. 2009. Combinatorial regulation by a novel arrangement of FruA and MrpC2 transcription factors during Myxococcus xanthus development. J Bacteriol 191:2753–2763. https://doi.org/10 .1128/JB.01818-08. 18. Son B, Liu Y, Kroos L. 2011. Combinatorial regulation by MrpC2 and FruA involves three sites in the fmgE promoter region during Myxococcus xanthus development. J Bacteriol 193:2756 –2766. https://doi.org/10 .1128/JB.00205-11. 19. Lee J, Son B, Viswanathan P, Luethy P, Kroos L. 2011. Combinatorial regulation of fmgD by MrpC2 and FruA during Myxococcus xanthus development. J Bacteriol 193:1681–1689. https://doi.org/10.1128/JB.01541-10. 20. Robinson M, Son B, Kroos L. 2014. Transcription factor MrpC binds to promoter regions of many developmentally-regulated genes in Myxococcus xanthus. BMC Genomics 15:1123. https://doi.org/10.1186/1471 -2164-15-1123. May 2017 Volume 199 Issue 10 e00788-16

21. Kroos L. 2008. Bacterial development in the fast lane. J Bacteriol 190: 4373– 4376. https://doi.org/10.1128/JB.00580-08. 22. Higgs PI, Jagadeesan S, Mann P, Zusman DR. 2008. EspA, an orphan hybrid histidine protein kinase, regulates the timing of expression of key developmental proteins of Myxococcus xanthus. J Bacteriol 190: 4416 – 4426. https://doi.org/10.1128/JB.00265-08. 23. Nariya H, Inouye S. 2005. Identification of a protein Ser/Thr kinase cascade that regulates essential transcriptional activators in Myxococcus xanthus development. Mol Microbiol 58:367–379. https://doi.org/10 .1111/j.1365-2958.2005.04826.x. 24. Nariya H, Inouye S. 2006. A protein Ser/Thr kinase cascade negatively regulates the DNA-binding activity of MrpC, a smaller form of which may be necessary for the Myxococcus xanthus development. Mol Microbiol 60:1205–1217. https://doi.org/10.1111/j.1365-2958.2006.05178.x. 25. Licking E, Gorski L, Kaiser D. 2000. A common step for changing cell shape in fruiting body and starvation-independent sporulation of Myxococcus xanthus. J Bacteriol 182:3553–3558. https://doi.org/10.1128/JB .182.12.3553-3558.2000. 26. Ueki T, Inouye S. 2005. Identification of a gene involved in polysaccharide export as a transcription target of FruA, an essential factor for Myxococcus xanthus development. J Biol Chem 280:32279 –32284. https://doi.org/10.1074/jbc.M507191200. 27. Müller FD, Schink CW, Hoiczyk E, Cserti E, Higgs PI. 2012. Spore formation in Myxococcus xanthus is tied to cytoskeleton functions and polysaccharide spore coat deposition. Mol Microbiol 83:486 –505. https://doi .org/10.1111/j.1365-2958.2011.07944.x. 28. Hagen TJ, Shimkets LJ. 1990. Nucleotide sequence and transcriptional products of the csg locus of Myxococcus xanthus. J Bacteriol 172:15–23. https://doi.org/10.1128/jb.172.1.15-23.1990. 29. Kim SK, Kaiser D. 1990. C-factor: a cell-cell signaling protein required for fruiting body morphogenesis of M. xanthus. Cell 61:19 –26. https://doi .org/10.1016/0092-8674(90)90211-V. 30. Kroos L. 2016. Highly signal-responsive gene regulatory network governing Myxococcus development. Trends Genet 33:3–15. 31. Hale CR, Zhao P, Olson S, Duff MO, Graveley BR, Wells L, Terns RM, Terns MP. 2009. RNA-guided RNA cleavage by a CRISPR RNA-Cas protein complex. Cell 139:945–956. https://doi.org/10.1016/j.cell.2009.07.040. 32. Hale CR, Majumdar S, Elmore J, Pfister N, Compton M, Olson S, Resch AM, Glover CV, III, Graveley BR, Terns RM, Terns MP. 2012. Essential features and rational design of CRISPR RNAs that function with the Cas RAMP module complex to cleave RNAs. Mol Cell 45:292–302. https://doi.org/ 10.1016/j.molcel.2011.10.023. 33. Peters JM, Silvis MR, Zhao D, Hawkins JS, Gross CA, Qi LS. 2015. Bacterial CRISPR: accomplishments and prospects. Curr Opin Microbiol 27: 121–126. https://doi.org/10.1016/j.mib.2015.08.007. 34. Kroos L, Kaiser D. 1987. Expression of many developmentally regulated genes in Myxococcus depends on a sequence of cell interactions. Genes Dev 1:840 – 854. https://doi.org/10.1101/gad.1.8.840. 35. Ueki T, Inouye S. 2005. Activation of a development-specific gene, dofA, by FruA, an essential transcription factor for development of Myxococcus xanthus. J Bacteriol 187:8504 – 8506. https://doi.org/10.1128/JB.187.24 .8504-8506.2005. 36. Müller FD, Treuner-Lange A, Heider J, Huntley SM, Higgs PI. 2010. Global transcriptome analysis of spore formation in Myxococcus xanthus reveals a locus necessary for cell differentiation. BMC Genomics 11:264. https:// doi.org/10.1186/1471-2164-11-264. 37. Wartel M, Ducret A, Thutupalli S, Czerwinski F, Le Gall AV, Mauriello EM, Bergam P, Brun YV, Shaevitz J, Mignot T. 2013. A versatile class of cell surface directional motors gives rise to gliding motility and sporulation in Myxococcus xanthus. PLoS Biol. 11:e1001728. https://doi.org/10.1371/ journal.pbio.1001728. 38. Holkenbrink C, Hoiczyk E, Kahnt J, Higgs PI. 2014. Synthesis and assembly of a novel glycan layer in Myxococcus xanthus spores. J Biol Chem 289:32364 –32378. https://doi.org/10.1074/jbc.M114.595504. 39. Storz G, Wolf YI, Ramamurthi KS. 2014. Small proteins can no longer be ignored. Annu Rev Biochem 83:753–777. https://doi.org/10.1146/ annurev-biochem-070611-102400. 40. Tan IS, Weiss CA, Popham DL, Ramamurthi KS. 2015. A quality-control mechanism removes unfit cells from a population of sporulating bacteria. Dev Cell 34:682– 693. https://doi.org/10.1016/j.devcel.2015.08.009. 41. Ueki T, Inouye S, Inouye M. 1996. Positive-negative KG cassettes for construction of multi-gene deletions using a single drug marker. Gene 183:153–157. https://doi.org/10.1016/S0378-1119(96)00546-X. jb.asm.org 18

Small Gene devI Delays M. xanthus Sporulation

42. Kashefi K, Hartzell P. 1995. Genetic suppression and phenotypic masking of a Myxococcus xanthux frzF⫺ defect. Mol Microbiol 15:483– 494. https:// doi.org/10.1111/j.1365-2958.1995.tb02262.x. 43. Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, III, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343–345. https://doi.org/10.1038/nmeth.1318. 44. Kuner JM, Kaiser D. 1982. Fruiting body morphogenesis in submerged cultures of Myxococcus xanthus. J Bacteriol 151:458 – 461. 45. Bradford M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248 –254. https://doi.org/10.1016/0003-2697 (76)90527-3.

May 2017 Volume 199 Issue 10 e00788-16

Journal of Bacteriology

46. Yoder-Himes D, Kroos L. 2006. Regulation of the Myxococcus xanthus C-signal-dependent ⍀4400 promoter by the essential developmental protein FruA. J Bacteriol 188:5167–5176. https://doi.org/10.1128/JB .00318-06. 47. Overgaard M, Wegener-Feldbrugge S, Sogaard-Andersen L. 2006. The orphan response regulator DigR is required for synthesis of extracellular matrix fibrils in Myxococcus xanthus. J Bacteriol 188:4384 – 4394. https:// doi.org/10.1128/JB.00189-06. 48. Schramm A, Lee B, Higgs PI. 2012. Intra- and interprotein phosphorylation between two hybrid histidine kinases controls Myxococcus xanthus developmental progression. J Biol Chem 287:25060 –25072. https://doi .org/10.1074/jbc.M112.387241.

jb.asm.org 19

The dev Operon Regulates the Timing of Sporulation during Myxococcus xanthus Development.

Myxococcus xanthus undergoes multicellular development when starved. Thousands of rod-shaped cells coordinate their movements and aggregate into mound...
2MB Sizes 0 Downloads 9 Views