Regulation of dev, an Operon That Includes Genes Essential for Myxococcus xanthus Development and CRISPR-Associated Genes and Repeats

Bacterial strains, plasmids, and primers.Strains and plasmids used in this work are listed in Table 1. The sequences of oligonucleotide primers used for PCR and reverse transcription-PCR (RT-PCR) are available upon request.

Growth and development. E. coli DH5α strains containing plasmids were grown at 37°C in Luria-Bertani medium (76) containing 50 μg/ml of either ampicillin or kanamycin (Km). M. xanthus strains were grown at 32°C in CTT broth (1% Casitone, 10 mM Tris-HCl, 1 mM KH₂PO₄, 8 mM MgSO₄) or on agar (1.5%) plates (27) unless specified otherwise. When necessary, 40 μg/ml Km was used for selection. TPM (10 mM Tris-HCl, 1 mM KH₂PO₄, 8 mM MgSO₄) agar (1.5%) plates were used for fruiting body development as described previously (50). For isolation of RNA, development was carried out in submerged culture (53).

Construction of plasmids.The plasmid pBJ131, containing an in-frame deletion in devS, was constructed in two steps. First, the 2.8-kb EcoRI-HindIII dev fragment from pLT5 (85) was inserted into EcoRI-HindIII-digested pBluescript II KS(+). The resulting plasmid was digested with Tth111I, the ends were made blunt using the fill-in reaction of the Klenow fragment of DNA polymerase I, and ligation was carried out in the presence of a 10-bp XhoI linker (New England Biolabs), yielding pBJ131. The in-frame devS deletion (ΔdevS) was removed from pBJ131 as an EcoRI-HindIII fragment and ligated into EcoRI-HindIII-digested pBJ113 (34) to create pBJ113ΔdevS. The ΔdevS mutation deletes codons 118 to 164 of the predicted 214-codon devS open reading frame.

The plasmid pBJ139 contains an 8.3-kb PstI-BamHI segment spanning from a PstI site in the MXAN_7267 gene upstream of the dev operon (Fig. 1) to the BamHI site near one end of Tn5 lac Ω4414, which inserted in devR in the proper orientation to fuse transcription of lacZ to the dev promoter (50, 85). Using pBJ139 as a template, an upstream primer containing a HindIII site, and a downstream primer containing a BamHI site, a 1,768-bp HindIII-BamHI fragment was generated by PCR and restriction digestion and inserted into pUC19 to construct pPV1767. The insert in this plasmid was sequenced at the Michigan State University Genomics Technology Support Facility, and the sequence was identical to that obtained in the genomic sequencing project (20). This plasmid served as the template for PCR amplification of M. xanthus DNA segments using upstream primers containing a HindIII site and downstream primers containing a BamHI site. Each PCR product was cloned using pCR2.1-TOPO as described by the manufacturer. Each DNA insert was sequenced at the Michigan State University Genomics Technology Support Facility to ensure that the correct sequence was obtained. Each pCR2.1-TOPO derivative was digested with HindIII and BamHI, and the DNA insert was purified by agarose gel extraction and subcloned into pREG1727 that had been digested with HindIII and BamHI. The 1,005-bp HindIII-BamHI fragment from pPV1004 was also subcloned into pMC1403KmattPTT (a gift from S. Inouye), fusing the predicted 10th codon of the MXAN_7266 gene in frame with the 8th codon of lacZ in pPV01004TF.

The QuikChange site-directed mutagenesis kit (Stratagene) was used to create mutations in the putative −35 and −10 regions of the dev promoter. The plasmid pPV1515 containing dev upstream DNA from bp −934 to 581 served as the DNA template for the mutagenesis with various combinations of primers. Candidate mutant plasmids were sequenced at the Michigan State University Genomics Technology Support Facility to identify plasmids with only the desired mutations. Each mutant plasmid was digested with HindIII and BamHI, and the mutant DNA insert was purified by agarose gel extraction and subcloned into pREG1727, which had been digested with HindIII and BamHI.

Construction of M. xanthus strains and determination of lacZ expression during development. M. xanthus DK11209 has an in-frame deletion in devS, designated ΔdevS. It was constructed by electroporation of pBJ113ΔdevS into wild-type DK1622. The plasmid contains the positive selection marker for Km resistance and a negative selection marker, galK. A transformant of DK1622 with pBJ113ΔdevS was initially selected on a CYE (1% Casitone, 0.5% yeast extract, 0.1% MgSO₄ · 7H₂O) agar (1.5%) plate (9) containing 40 μg/ml Km and named DK11206. To obtain recombinants that had replaced the wild-type devS gene with the ΔdevS allele, DK11206 was grown in CYE broth to stationary phase and plated on CYE agar containing 2% galactose. The galactose-resistant colonies were screened for the presence of the ΔdevS mutation by Southern blot hybridization, and one containing the ΔdevS allele was named DK11209.

Each pREG1727 derivative and pPV01004TF were transformed by electroporation (40) into M. xanthus ΔdevS mutant DK11209 and/or wild-type DK1622. Transformants were selected on CTT-Km plates. Based on previous experience in our laboratory (8, 15, 16, 26, 64), the majority of transformants have a single copy of the plasmid integrated at the Mx8 phage attachment site (designated attB in Table 1). To eliminate colonies with unusual developmental lacZ expression, we screened at least 10 transformants on TPM agar plates containing 40 μg of 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside per ml. Any colonies with unusual lacZ expression were discarded. Of the remaining transformants, three independent isolates of each construct were chosen for development. In all cases, the three transformants gave similar results when developmental β-galactosidase activity was measured as described previously (50).

Primer extension.Primer extension analysis was carried out as described previously (17, 41). The hot-phenol-chloroform method (76) was used to isolate total RNA from M. xanthus DK1622 cells that had developed in submerged culture for 12 h. The primer (OAG477) used for this analysis was 5′-AACCTCCAGTCGTTCCAGCA-3′. DNA was sequenced by the dideoxynucleotide chain termination method (77) using the Thermo Sequenase radiolabeled terminator cycle sequencing kit (U.S. Biochemical Corp.) and primer OAG477.

RT-PCR.One-step RT-PCRs were performed according to the manufacturer's instructions (QIAGEN) on RNA prepared as described above except that cells had developed for 18 h. As a negative control for each reaction, the 30-min incubation at 50°C, during which RT normally synthesizes cDNA, was omitted prior to the PCR.

Gene organization and repeat sequences at the dev locus.DNA sequence analysis of the dev locus suggested the presence of an operon that includes seven or eight genes, based on their orientation and proximity (Fig. 1). The translational stop and start codons are predicted to overlap in three cases (cas6-cas3, cas3-devT, and devR-devS) and are closely spaced in three other cases (30, 0, and 3 bp separate devT-devR, devS-cas4-cas1, and cas4-cas1-cas2, respectively) (20). The cas designation stands for CRISPR-associated, and CRISPR is an acronym for clustered regularly interspaced short palindromic repeats (31). CRISPR are a class of DNA repeats found in nearly one-half of all sequenced bacterial and archaeal genomes (23). Three CRISPR are present in the M. xanthus genome, and one of these lies downstream of the dev operon (Fig. 1). This CRISPR includes 28 copies of a 36-bp repeat (GTGCTCAACGCCTTTCGGCATCACGGCGAGCGGGAC) and 4 copies that match this sequence at 31 to 35 positions. Of these 32 repeats, 23 span from 253 to 1,861 bp downstream of cas2. This cluster of repeats is followed by the MXAN_7258 gene, which has the same orientation as dev and cas genes and is predicted to encode a protein that is not similar to others in the database. The remaining nine repeats are clustered from 114 to 690 bp downstream of the MXAN_7258 gene. Comparative genomic analysis has led to the hypothesis that CRISPR-Cas systems are defense mechanisms against invading phages and plasmids, functioning in a manner analogous to eukaryotic RNA interference systems (65). It is intriguing that cas and dev genes might be cotranscribed with downstream CRISPR at the dev locus.

Localization of the dev promoter and demonstration of negative autoregulation by devS.Upstream of cas6 is an open reading frame encoding 40 amino acids that is in the same orientation as other genes at the dev locus (Fig. 1). This small predicted MXAN_7266 gene is 202 bp upstream of cas6 and 135 bp downstream of the MXAN_7267 gene, which is also in the same orientation (Fig. 1). We hypothesized that one of these intergenic regions harbors the dev promoter. To test this hypothesis, we PCR amplified a 1,768-bp DNA fragment that included the 3′ end of the MXAN_7267 gene, the MXAN_7266 gene, and the 5′ end of cas6 (note that in Fig. 1 the 1,768-bp fragment is shown to span from −934 to +834 relative to the dev transcriptional start site [TSS], based on RNA mapping described below). This fragment was inserted into pREG1727, creating a cas6-lacZ transcriptional fusion. After transformation into M. xanthus, the plasmid integrates at the Mx8 phage attachment site in the chromosome (16). The plasmid was transformed into wild-type DK1622 and into the ΔdevS mutant DK11209. The ΔdevS mutation is an in-frame deletion in devS, and a preliminary study suggested that this mutation increases dev expression (A. G. Garza and B. Julien, unpublished data), consistent with a previous study that indicated that dev expression is subject to negative autoregulation (85). We reasoned that elevated expression in the ΔdevS mutant might facilitate our effort to locate the dev promoter. Indeed, developmental expression from the cas6-lacZ fusion was low in the wild type, reaching a maximum of about 50 U at 24 to 30 h into development, while in the ΔdevS mutant expression rose markedly at 24 h and continued to increase at least until 48 h, reaching over 1,100 U (Fig. 2). Expression in the ΔdevS mutant was similar to that in M. xanthus bearing Tn5 lac Ω4414 (50), an insertion near the 3′ end of devR that disrupts negative autoregulation (85). Expression from Tn5 lac Ω4414 appeared to begin about 6 h earlier during development than expression from the cas6-lacZ fusion (Fig. 2). Except for this difference in timing, expression from the cas6-lacZ fusion in the ΔdevS mutant showed the pattern expected for the dev promoter not subject to negative autoregulation. We infer that the dev promoter is located on the 1,768-bp fragment used to generate the cas6-lacZ fusion and that cas and dev genes are indeed cotranscribed at the dev locus (see below). The results in Fig. 2 also demonstrate that activity of the dev promoter is subject to strong (over 20-fold) negative autoregulation mediated by DevS, even when the cas6-lacZ fusion is integrated into the M. xanthus chromosome at the Mx8 phage attachment site, over 2 Mbp from the native dev locus.

• Open in new tab
• Download powerpoint

FIG. 2.

Developmental lacZ expression from the 1,768-bp fragment in wild-type M. xanthus and a devS mutant. The fragment spanning from −934 to +834 was inserted into pREG1727 in the correct orientation to fuse transcription of the dev promoter to lacZ. The resulting plasmid was transformed into M. xanthus wild-type DK1622 (○) and ΔdevS mutant DK11209 (□), and developmental lacZ expression was determined for three independent transformants. As a control, developmental lacZ expression was determined for three colonies of M. xanthus DK5279 (•), which harbors Tn5 lac Ω4414 inserted in devR. The graph shows the average β-galactosidase specific activity expressed in nanomoles of o-nitrophenyl phosphate per minute per milligram of protein, and error bars show 1 standard deviation of the data. The error bars are too small to be seen for some data points.

The ΔdevS mutant formed mounds at the normal time on starvation (TPM) agar, but the mounds failed to darken, and formation of heat- and sonication-resistant spores capable of germination was reduced about 100-fold (data not shown). The developmental defects of the ΔdevS mutant are similar to those reported previously for M. xanthus bearing the Tn5 lac Ω4414 insertion in devR (34, 49, 85).

cas and dev genes are cotranscribed.To test directly whether cas and dev genes are cotranscribed at the dev locus, RNA was prepared from M. xanthus wild-type strain DK1622 after 18 h of development and subjected to RT-PCR analysis. The expected PCR products were observed with primers designed to detect transcription across each junction between genes in the predicted dev operon and across the region downstream of cas2 including the first two repeats of the CRISPR (Fig. 3). Omission of the RT step resulted in no PCR products, demonstrating dependence on RNA (not contaminating DNA) in the samples. The finding that the dev transcript spans the intergenic region between the MXAN_7266 gene and cas6, together with evidence that the dev promoter is located on the 1,768-bp fragment used to generate the cas6-lacZ fusion (Fig. 2), suggested that the dev promoter is located between the MXAN_7267 and MXAN_7266 genes. We conclude that the dev operon includes at least eight genes. The implications of the finding that dev transcription continues into the CRISPR will be addressed in Discussion.

• Open in new tab
• Download powerpoint

FIG. 3.

RT-PCR analysis of the dev transcript. RNA prepared from M. xanthus wild-type DK1622 at 18 h into development (lanes 3 to 20) or from the ΔdevS mutant DK11209 at 23 h (lanes 1 and 2) was subjected to RT-PCR (odd-numbered lanes) or, as a control, only to PCR (even-numbered lanes). The primers were designed to detect transcription across the following junctions (lanes and predicted PCR product size in parenthesis): the MXAN_7266 gene and cas6 (lanes 1 to 4; 510 bp), cas6 and cas3 (lanes 5 and 6; 260 bp), cas3 and devT (lanes 7 and 8; 420 bp), devT and devR (lanes 9 and 10; 280 bp), devR and devS (lanes 11 and 12; 180 bp), devS and cas4-cas1 (lanes 13 and 14; 220 bp), cas4-cas1 and cas2 (lanes 15 and 16; 200 bp), cas2 and the first repeat of CRISPR plus the first unique insert (lanes 17 and 18, 320 bp), and cas2 and the second repeat of CRISPR plus the second unique insert (lanes 19 and 20; 400 bp). Lane M contains the 100-bp ladder (New England Biolabs).

Mapping the 5′ end of dev mRNA.RNA isolated from M. xanthus wild-type DK1622 that had undergone development for 12 h in submerged culture was subjected to primer extension analysis (Fig. 4). The primer was designed to detect mRNA 5′ ends upstream of the MXAN_7266 gene. A single extension product was detected, mapping the dev mRNA 5′ end to the position indicated in the sequence shown in Fig. 4. We designated this position as +1, the putative TSS of the dev promoter. A primer extension product of the same size and greater abundance was observed with RNA from the ΔdevS mutant DK11209 (data not shown). This suggests that higher expression in the ΔdevS mutant (Fig. 2) results, at least in part, from greater accumulation of the dev transcript with the same 5′ end.

• Open in new tab
• Download powerpoint

FIG. 4.

Mapping the 5′ end of the dev developmental transcript by primer extension analysis. RNA was isolated from M. xanthus wild-type DK1622 cells that had undergone 12 h of development in submerged culture. The primer OAG477 hybridized to +71 to +90 relative to the mRNA 5′ end mapped in this experiment. The extension product in lane PE is indicated by an arrow. The other lanes show DNA sequencing products generated with primer OAG477. A portion of the DNA sequence is shown to the left, with the inferred TSS indicated by a right-angle arrow.

5′ deletions cause a graded loss of dev promoter activity.To define the 5′ region required for dev promoter activity, we tested a series of 5′ deletions with a common 3′ end at +834 (Fig. 1). Each fragment was inserted into pREG1727 to generate a transcriptional fusion to lacZ, the plasmid was then transformed into the M. xanthus ΔdevS mutant (to avoid negative autoregulation), and plasmid integrants were measured for developmental lacZ expression. A 5′ deletion to −535 altered the pattern of expression (Fig. 5). Relative to the fragment with its 5′ end at −934, the fragment ending at −535 exhibited earlier expression but expression did not continue to increase after 36 h, resulting in only one-half as much activity at 48 h. A 5′ deletion to −114 markedly decreased expression (Fig. 1 and 5). A deletion to +14 abolished expression, as expected for deletion of the dev promoter. This supports the +1 assignment of the dev TSS based on primer extension analysis (Fig. 4). The graded loss of expression observed at 48 h for the 5′ deletions to −535 and −114 suggests that at least two upstream DNA elements positively regulate dev promoter activity.

• Open in new tab
• Download powerpoint

FIG. 5.

Developmental lacZ expression from 5′ deletions. Fragments spanning from −535 (▪), −114 (▴), or +14 (⋄) to +834 were inserted into pREG1727, the resulting plasmids were transformed into M. xanthus ΔdevS mutant DK11209, and developmental lacZ expression was determined for three independent transformants. The data for the 1,768-bp fragment from −934 to +834 (same as in Fig. 2) are shown for comparison (□). The meaning of points and error bars is the same as described in the Fig. 2 legend. The error bars are too small to be seen for some data points.

Mutational analysis of the putative dev promoter.Upstream of +1 are sequences that match the E. coli σ⁷⁰ consensus promoter −35 and −10 regions (62) at five of six and three of six positions, respectively, although the spacing between these sequences (20 bp) is not optimal (Fig. 6). To test whether these sequences are important for dev expression, multiple-base-pair mutations were made in each in the context of a fragment spanning from −934 to +581. This fragment, bearing the wild-type sequence, exhibited slightly higher expression at 36 and 48 h than the fragment from −934 to +834 in the ΔdevS mutant (Fig. 1 and 6; see below). The multiple-base-pair changes in the putative dev promoter −35 and −10 regions completely abolished expression (Fig. 6). These results, together with our 5′ deletion analysis (Fig. 5), provide strong evidence that the mRNA 5′ end mapped by primer extension (Fig. 4) reflects the start site of dev transcription.

• Open in new tab
• Download powerpoint

FIG. 6.

Mutational analysis identifies the dev promoter. The top part shows the sequence upstream of the inferred TSS, which is indicated by the right-angle arrow. Sequences that match the E. coli σ⁷⁰ consensus promoter −35 and −10 regions (62) are in boldface. The putative dev promoter −35 and −10 regions are underlined, and the 6-bp mutation made in each region is indicated. Below is shown developmental lacZ expression from fragments that span from −934 to +581, after insertion into pREG1727 and transformation into M. xanthus ΔdevS mutant DK11209. Expression from three independent transformants bearing the wild-type promoter (▪), the −35 region mutation (▴), or the −10 region mutation (○) was measured. Likewise, developmental lacZ expression was measured for a fragment spanning from −934 to +833 (•). The data for the 1,768-bp fragment from −934 to +834 (same as in Fig. 2) are shown for comparison (□). The meaning of points and error bars is the same as described in the Fig. 2 legend. The error bars are too small to be seen for some data points.

The higher expression observed for the fragment spanning from −934 to +581 might be due to the fact that this fragment creates an in-frame fusion between cas6 and trpA, which precedes lacZ in the plasmid vector we used (pREG1727), whereas the fragment ending at +834 creates an out-of-frame fusion, possibly causing a polar effect on expression of the downstream lacZ reporter due to a premature translation stop in trpA (61, 64). To address this possibility, a fragment from −934 to +833, which creates an in-frame fusion between cas6 and trpA, was tested. Higher expression was observed than for fragments ending at +834 or +581 in the ΔdevS mutant (Fig. 1 and 6), suggesting that the out-of-frame fusion at +834 does cause a polar effect on lacZ expression and that, comparing the two in-frame fusions, the segment between +581 and +833 contains a positive regulatory element. It is worth noting that a downstream regulatory element might affect a postinitiation event (e.g., transcription elongation or termination, mRNA stability, or translation), instead of or in addition to affecting transcription initiation.

3′ deletions imply the presence of additional downstream regulatory elements.To further investigate the role of downstream DNA in dev promoter activity, fusions at different downstream end points were constructed and tested for developmental lacZ expression as described above. The 3′ deletions have a common 5′ end at −934 (Fig. 1). A 3′ deletion to +280 markedly decreased developmental lacZ expression in the ΔdevS mutant (Fig. 1 and 7). The 3′ end of this fragment is between the end of the MXAN_7266 gene and the predicted translational start of cas6 at +367 (Fig. 1). We made two other fusions with 3′ end points in predicted untranslated regions. Compared with the fusion at +280, a fusion at +219 showed considerably higher developmental lacZ expression and a fusion at +32 showed considerably lower expression (Fig. 1 and 7). This comparison indicates that a strong positive regulatory element lies between +32 and +219 and that a negative regulatory element lies between +219 and +280.

• Open in new tab
• Download powerpoint

FIG. 7.

Developmental lacZ expression from 3′ deletions. Fragments spanning from −934 to +280 (▴), +219 (•), or +32 (▵) were inserted into pREG1727, the resulting plasmids were transformed into M. xanthus ΔdevS mutant DK11209, and developmental lacZ expression was determined for three independent transformants. The meaning of points and error bars is the same as described in the Fig. 2 legend. The error bars are too small to be seen for some of the points.

The MXAN_7266 gene is translated.The predicted 40-amino-acid product of the MXAN_7266 gene does not exhibit significant similarity to any protein in the database. The gene's putative GTG start codon is preceded 6 bp upstream by the sequence AGGAGCG, which could be a ribosome binding site. To determine whether this short open reading frame is translated, an in-frame fusion was constructed between its predicted 10th codon and the 8th codon of lacZ in the translational fusion vector pMC1403KmattPTT. For comparison, the predicted 10th codon of the MXAN_7266 gene (end point at +71 relative to the dev TSS) was also fused in-frame with trpA, which precedes lacZ, in pREG1727, as explained above. The upstream end point was −934 in both constructs, and both integrate at the Mx8 phage attachment site in the M. xanthus chromosome. The plasmids were transformed into the ΔdevS mutant DK11209, and developmental lacZ expression was measured (Fig. 8). For both fusions, expression was similar to that observed for the fragment from −934 to +833, which creates an in-frame fusion between cas6 and trpA in pREG1727, as explained above. We conclude that dev mRNA is translated to produce the 40-amino-acid MXAN_7266 gene product.

• Open in new tab
• Download powerpoint

FIG. 8.

Developmental lacZ expression from a MXAN_7266 gene-lacZ translational fusion and a short fragment bearing the dev promoter. The fragment spanning from −934 to +71 was inserted into pMC1403KmattPTT (▵) and pREG1727 (▪), creating translational and transcriptional fusions to lacZ, respectively. The fragment spanning from −114 to +71 was also inserted into pREG1727 (○). The resulting plasmids were transformed into M. xanthus ΔdevS mutant DK11209, and developmental lacZ expression was determined for three independent transformants. The data for the fragment from −934 to +833 (same as in Fig. 6) are shown for comparison (•). The meaning of points and error bars is the same as described in the Fig. 2 legend.

A short DNA fragment exhibits unexpectedly high dev promoter activity.The DNA fragment spanning from −934 to +834 caused lacZ to be expressed at a sevenfold-higher level than the fragment from −114 to +834, at 48 h into development (Fig. 1 and 5). Based on this finding, we expected lacZ expression from the fragment spanning from −114 to +71 to reach about 390 U, one-seventh of that observed for the fragment from −934 to +71 (Fig. 1 and 8). Surprisingly, expression from the −114 to +71 fragment rose to 1,060 U on average (Fig. 1 and 8). In the absence of DNA downstream of +71, the positive effect of upstream DNA between −934 and −114 is only 2.6-fold (i.e., 2,720 U/1,060 U), not the 7-fold (i.e., 1,120 U/160 U) observed for fusions at +834 (Fig. 1). This suggests that upstream and downstream regulatory elements interact functionally.

Negative autoregulation by DevS involves DNA upstream and downstream of the dev promoter.Developmental lacZ expression from the fragment from −934 to +834 reached a much higher level in the ΔdevS mutant than in the wild type (Fig. 2), indicating that DevS is required for negative autoregulation of the dev operon. To investigate the DNA requirements for DevS-mediated negative autoregulation, expression from shorter fragments containing the dev promoter was measured in wild-type cells. Interestingly, the short fragment from −114 to +71 exhibited considerably higher expression than the fragment from −934 to +834 in the wild type (Fig. 9). In the ΔdevS mutant, these two fragments caused lacZ expression to rise to similar levels (Fig. 1). Hence, DevS-mediated negative autoregulation appeared to require DNA upstream of −114 and/or downstream of +71.

• Open in new tab
• Download powerpoint

FIG. 9.

Developmental lacZ expression from dev promoter fragments in wild-type M. xanthus. A fragment spanning from −114 to +71 (▪) was inserted into pREG1727, the resulting plasmid was transformed into M. xanthus wild-type DK1622, and developmental lacZ expression was determined for three independent transformants. The data for the 1,768-bp fragment from −934 to +834 (same as in Fig. 2) are shown for comparison (○). The meaning of points and error bars is the same as described in the Fig. 2 legend. The error bars are too small to be seen for some of the points.

The pattern of developmental lacZ expression in the wild type was different from that in the ΔdevS mutant background. Expression reached a maximum at 24 to 30 h into development in the wild type (Fig. 9), but activity continued to rise until 48 h in most of the ΔdevS mutant strains (Fig. 2 and 5 to 8). Normal DevS function seems to prevent lacZ expression from continuing to rise between 24 and 48 h into development. In the wild type during this period, cells differentiate into spores, sequestering β-galactosidase activity, and others may lyse, causing β-galactosidase to be lost due to degradation or diffusion. By blocking developmental progression, the ΔdevS mutation may permit β-galactosidase activity to continue accumulating in sonication-sensitive cells, allowing its detection in our assay. Therefore, comparison of levels of lacZ expression at 24 h in the ΔdevS mutant and wild type provides a better measure of DevS-mediated negative autoregulation than comparison of the maximum expression during a 48-h time course of development. Applying this measure, the ratio (ΔdevS mutant/wild type) is 6.4 for the fragment from −934 to +834, 5.3 for the fragment from −934 to +833, and 2.0 for the fragment from −114 to +71 (Table 2). We conclude that DevS-mediated negative autoregulation depends at least in part on DNA upstream of −114 and/or downstream of +71.

A fragment with more upstream DNA (−934 to +71) restored the ΔdevS mutant/wild type ratio to 5.6 (Table 2). Hence, negative autoregulation by DevS can be mediated by DNA upstream of −114.

A fragment with more downstream DNA (−114 to +834) also restored the ΔdevS mutant/wild type ratio to a similar value, 5.8 (Table 2), indicating that DevS-mediated negative autoregulation can occur through DNA downstream of +71.

Interestingly, negative autoregulation by DevS was greatly reduced for the fragment from −934 to +581 (Table 2). In this case, upstream DNA between −934 and −114 is almost completely incapable of mediating negative autoregulation by DevS. This upstream segment was quite capable of mediating negative autoregulation by DevS when the downstream end point was +71 (Table 2). DNA between +71 and +581 appears to interfere with the ability of upstream DNA between −934 and −114 to mediate negative autoregulation by DevS. DNA between +581 and +833 restores DevS-mediated negative autoregulation (Table 2). Taken together, the results demonstrate that DNA upstream of −114 or downstream of +71 can mediate negative autoregulation by DevS independently but that, in the presence of both, DNA between +581 and +833 becomes necessary, suggesting that interactions between upstream and downstream DNA influence DevS-mediated negative autoregulation.

Similarities between the dev promoter region and those of other M. xanthus genes.Sequences in the −35 and −10 regions are crucial for dev promoter activity (Fig. 6). These sequences resemble those recognized by M. xanthus σ^A RNA polymerase (RNAP) in other promoters, but the dev promoter has a longer spacer (20 bp) between its −35 and −10 regions. M. xanthus σ^A is highly similar to E. coli σ⁷⁰ (30). σ^A RNAP has been purified from growing M. xanthus cells and has been shown to transcribe four genes in vitro (5, 26, 93). The promoters of these genes match the E. coli σ⁷⁰ consensus −35 region (TTGACA) at five or six positions but match the −10 region consensus (TATAAT) at only two or three positions, with spacers ranging from 16 to 18 bp. The dev promoter matches the E. coli σ⁷⁰ consensus −35 and −10 regions at five and three positions, respectively; however, the 20-bp spacer would be expected to pose a significant barrier to utilization by σ^A RNAP alone, based on the 16- to 18-bp spacers observed in M. xanthus promoters studied so far and on extensive studies of E. coli promoters utilized by σ⁷⁰ RNAP, which prefers a 17-bp spacer (68). The long spacer may explain, at least in part, why the dev promoter is not transcribed during growth. The barrier posed by the long spacer could be overcome in at least two ways. First, a transcription factor could bind to and bend spacer DNA (2) or bind to the promoter −35 region and interact with region 4 of σ^A in such a way that its interaction with the −10 region is improved (52). If such a transcription factor was developmentally regulated, this could explain developmental regulation of dev transcription. Second, a σ factor that recognizes promoter −35 and −10 regions similar to those recognized by σ^A but that better tolerates a longer spacer could utilize the dev promoter. In E. coli, σ^S and σ⁷⁰ recognize similar promoter −35 and −10 regions and σ^S better tolerates a longer spacer (90). σ^S RNAP selectively transcribes genes induced during stationary phase and stress. Its ortholog in M. xanthus is σ^D (92), which was shown recently to play a direct or indirect role that is essential for expression of dev (97), so perhaps σ^D RNAP transcribes the dev operon.

Slightly upstream of the dev promoter −35 region are sequences similar to those that have been shown to be important for transcription of other developmentally regulated M. xanthus genes. A 5-bp element (consensus sequence GAACA) located 5 to 8 bp upstream of a C box (consensus sequence CAYYCCY, in which Y means C or T) typically spans from approximately bp −63 to −46, and these sequences have been shown to be important for developmental transcription of five different genes or operons (83, 95, 96, 102, 103). Upstream of the dev promoter −35 region is a sequence centered at bp −60 that matches the 5-bp element consensus in three of five positions and two sequences centered at bp −51 and −47 that match the C box consensus at five of seven positions (Fig. 10). It seems likely that one or more of these sequences is important for dev promoter activity during development.

• Open in new tab
• Download powerpoint

FIG. 10.

Sequences upstream of the dev promoter predicted to be important for its activity. Mirror repeats are underlined in a 17-bp sequence centered at −91. A sequence shown in boldface and centered at −60 is similar to 5-bp elements in other developmentally regulated M. xanthus promoter regions (see Discussion for references). Likewise, two C-box-like sequences are shown in boldface with underlining, centered at −51 and −47.

Farther upstream, centered at bp −91, is a 17-bp sequence that includes a 5-bp mirror repeat (Fig. 10). This sequence is very similar to a 16-bp sequence that includes a 6-bp mirror repeat and that is centered at −74.5 in the Ω4400 promoter region (8). The upstream half of that mirror repeat overlaps a site recognized by the DNA-binding domain of FruA (104), a putative response regulator proposed to activate transcription in response to C-signaling (14, 70, 91). The site recognized by the FruA DNA-binding domain is important for C-signal-dependent regulation of Ω4400 promoter activity (102). By analogy, we speculate that FruA (possibly phosphorylated in response to C-signaling) binds upstream of the dev promoter and activates its transcription. This may explain the partial dependence on C-signaling observed previously for lacZ expression from Tn5 lac Ω4414 (48), an insertion in devR (85).

Upstream regulatory elements.Our 5′ deletion analysis of a ΔdevS mutant revealed at least two upstream elements that positively regulate dev promoter activity (Fig. 1 and 5; summarized in Fig. 11). One of these is located more than 535 bp upstream of the TSS, an unusually long distance for a positive cis-regulatory element in bacteria. However, in M. xanthus, activity of the tps (42), csgA (58), Ω4499 (15), and Ω4514 (26) promoters has been shown to depend on sequences located more than 500 bp upstream of the TSS. These sequences have been inferred to be binding sites for transcriptional activators and are typically located in the coding region of an upstream gene, based on our analysis of the M. xanthus genome sequence (20). In the case of the dev promoter, our 5′ deletion analysis suggests that binding sites for transcriptional activators might be located between bp −934 and −535 and between bp −535 and −114. Both of these segments are in the MXAN_7267 gene coding region (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 11.

Summary of regulatory elements inferred from 5′ and 3′ deletion analyses of the dev promoter region. Numbers indicate the difference (n-fold) in maximum developmental lacZ expression in a ΔdevS mutant (Fig. 1) attributable to the indicated segment. Arrows indicate positive effects, and the dashed line with a barred end indicates a negative effect. Upstream regulatory elements likely affect transcriptional initiation, whereas downstream elements could alternatively or in addition affect postinitiation events.

Downstream regulatory elements.Our 3′ deletion analysis of a ΔdevS mutant revealed at least two positive regulatory elements and one negative element (Fig. 1, 6, and 7; summarized in Fig. 11). These regulatory elements might affect transcription initiation (e.g., as sites for binding of transcriptional activators or repressors, which might exert their effects via DNA looping) or a postinitiation event (e.g., transcription elongation or termination/antitermination or stability or translation of the fusion mRNA). Downstream positive (66, 95) and negative (28) regulatory elements for several M. xanthus genes have been described, but the regulatory mechanisms are not well understood. In the case of fruA, a negative element (xbs) located at +75 to +94 is bound specifically by an unidentified protein (factor X) that is present during growth and decreases early in development (28). When xbs was placed downstream of a vegetative promoter (vegA), it reduced expression of a lacZ reporter located farther downstream, indicating that its mechanism of negative regulation is not specific to transcription initiated at the fruA promoter. Perhaps factor X blocks RNAP elongation. A similar mechanism might explain how the dev negative regulatory element between +219 and +280 functions.

A strong (23-fold) positive regulatory element is located between +32 and +219, based on comparison of lacZ expression from fusions at these two downstream end points (Fig. 1 and 7; summarized in Fig. 11). This element is likely located between +32 and +71 and may exert an even stronger effect, since lacZ expression reaches a 65-fold-higher level for the fusion at +71 than for one at +32 (Fig. 1, 7, and 8); however, this comparison involves an in-frame fusion between the MXAN_7266 gene and trpA, which precedes lacZ in pREG1727, in the case of +71, and a fusion to a predicted untranslated region in the case of +32. Likewise, there may be a positive regulatory element between +280 and +581 that exerts a 7.5-fold effect, based on comparison of the in-frame fusion at +581 with the fusion at +280 to a predicted untranslated region (Fig. 1, 6, and 7).

Interactions between upstream and downstream regulatory elements.Our results strongly suggest functional interaction between upstream and downstream regulatory elements in the presence or absence of DevS. In a ΔdevS mutant, the 7-fold-positive effect of regulatory elements upstream of −114, revealed by our 5′ deletion analysis in the context of +834 as the 3′ end (Fig. 1 and 5; summarized in Fig. 11), was reduced to a 2.6-fold-positive effect when DNA downstream of +71 was removed (Fig. 1 and 8). If the upstream positive regulatory elements are binding sites for transcriptional activators as suggested above, presumably DNA looping would bring the activators close to the promoter (reviewed in references 78 and 94). Downstream DNA may participate in DNA loop formation, influencing the position of activators in a complex analogous to a eukaryotic enhancesome or billboard enhancer (reviewed in references 3 and 67). Testing expression from additional segments with different combinations of the upstream and downstream end points so far employed, as well as new end points, might pinpoint regulatory elements involved in the putative long-range interaction.

Functional interaction between upstream and downstream regulatory elements in the presence of DevS is supported by comparison of levels of lacZ expression from different segments at 24 h into development (Table 2). A segment spanning from −934 to +71 exhibits 5.6-fold-higher expression in a ΔdevS mutant than in the wild type, indicative of DevS-mediated negative autoregulation. The ratio drops to 1.4-fold for the fragment from −934 to +581, indicating almost complete loss of negative autoregulation. Amazingly, DevS-mediated negative autoregulation returns for the fragment from −934 to +833 (ratio, 5.3). Since DevS-mediated negative autoregulation was weak for the fragment from −114 to +71 (ratio, 2.0), it appears that the ability of DNA beyond −114 upstream to mediate negative autoregulation by DevS depends on whether or not DNA between +71 and +581 is present, and if so, DNA between +581 and +833 becomes necessary for negative autoregulation. As proposed above, this apparent long-range interaction between regulatory elements might involve DNA looping, but a better understanding of DevS-mediated negative autoregulation might suggest other mechanisms.

Autoregulation by Dev proteins.The mechanism of dev negative autoregulation is unknown. It involves DevS (Fig. 2 and Table 2), but preliminary data suggest that it does not involve DevR (Garza and Julien, unpublished), so the loss of negative autoregulation in M. xanthus bearing Tn5 lac Ω4414 (inserted in devR) (85) is probably due to a polar effect on expression of devS. Primer extension analysis of developmental RNA isolated from a ΔdevS mutant suggested that the dev transcript, initiating from the same site as in the wild type, accumulates to a higher level (data not shown). This could result from increased transcription initiation or greater stability of the transcript in the absence of DevS.

DevT is involved in a positive autoregulatory loop involving the putative response regulator and transcriptional activator FruA (7). Because FruA is activated by C-signaling (14), it has been proposed that the FruA/DevT positive-feedback loop ensures a burst of fruA and dev transcription once a certain level of C-signaling is reached (7). That level of C-signaling may be reached as cells become aligned in fruiting bodies, possibly explaining how dev expression is spatially restricted (34).

In addition to autoregulation, do Dev proteins have other functions? As noted above, an in-frame deletion in devT impairs aggregation and sporulation (7), and an in-frame deletion in devS reduces sporulation about 100-fold (data not shown). A devRS mutant fails to express lacZ inserted at the Ω7536 locus, and the products of this locus are required for sporulation (60). Whether these phenotypes result directly from the absence of DevTRS proteins or indirectly from improper autoregulation of other dev operon products is unknown.

Implications of cotranscription of dev and cas genes with CRISPR.Our RNA analysis indicates that the MXAN_7266 gene (encoding a 40-amino-acid product), cas genes, and at least two repeats of the CRISPR are cotranscribed with devTRS (Fig. 3). As noted above, the predicted MXAN_7266 amino acid sequence does not share significant similarity with any known protein. The cas genes typically are arranged in the order cas3-cas4-cas1-cas2, and these four genes are found only in CRISPR-containing species, invariably adjacent to a CRISPR (31). At the dev locus, cas genes are arranged in the typical order but devTRS lies between cas3 and cas4 (Fig. 1). A similar arrangement is found in Desulfovibrio vulgaris Hildenborough and several other species, and this has been called the three-gene Dvulg CRISPR/Cas subtype (23). At the M. xanthus dev locus, the normally separate cas4 and cas1 genes are fused, creating cas4-cas1 (Fig. 1), and a cas6 gene is present, which is atypical for the Dvulg subtype. Certain Cas proteins exhibit similarity to helicases and nucleases (31, 65), and some of the unique inserts between repeats in CRISPR are similar to segments of bacteriophage and plasmid genes (69). These observations led Makarova et al. (65) to hypothesize that CRISPR-Cas systems are RNA interference-based immune systems in which transcribed unique inserts function as small interfering RNAs (siRNAs) to promote degradation or inhibit translation of target bacteriophage and plasmid mRNAs.

We performed a BLAST (1) search with the unique insert between the first two repeats in the CRISPR downstream of cas2 in the dev operon. This insert is transcribed as part of the dev operon (Fig. 3). The entire 37-bp insert matches perfectly a sequence near the 3′ end of the bacteriophage Mx8 intP gene (Fig. 12A), which encodes a form of the integrase needed for site-specific recombination between the circular phage genome and the M. xanthus chromosome during lysogenization (89). If the hypothesis of Makarova et al. (65) is correct, a function of the dev operon may be to protect developing M. xanthus from lysogenization by Mx8. Without this putative siRNA defense mechanism, starving M. xanthus cells might be particularly vulnerable to Mx8 lysogenization, based on analogy with phage λ infection of E. coli, during which poor growth of the host favors lysogenization rather than lytic phage growth (reviewed in reference 71). Although integrase is also necessary for Mx8 excision (88), the putative siRNA mechanism would not work on the integrase gene of the lysogen because its 3′ end is altered by the site-specific recombination event that occurs during lysogenization (i.e., the attP site in Mx8 lies in the intP gene upstream of the 37-bp sequence shown in Fig. 12A) (89). Hence, the putative siRNA mechanism could not guard against prophage induction during development. A lysogen is stable upon passage through development and spore germination under laboratory conditions (72), so it is not immediately obvious why, evolutionarily, it would be important to protect against Mx8 lysogenization specifically during M. xanthus development. Perhaps in nature the conditions that favor germination are often accompanied by conditions that favor prophage induction, resulting in lysis of the host cell.

• Open in new tab
• Download powerpoint

FIG. 12.

Matches between unique inserts in the CRISPR at the dev locus and other sequences. (A) The antisense strand (top) of the insert between the first two repeats downstream of cas2 (20) matches the sense strand (bottom) of the bacteriophage Mx8 intP gene (89). (B) Part of the insert between the 23rd and 24th repeats downstream of cas2 matches the MXAN_7283 gene (20). The insert between the 23rd and 24th repeats downstream of cas2 contains the MXAN_7258 gene in the same orientation as dev and cas genes. The top sequence is from upstream of the MXAN_7258 gene and would be the antisense strand if transcription from either the dev promoter or another promoter in the same orientation proceeds across this region. The bottom sequence is from the sense strand of the MXAN_7283 gene. Numbers are from the cited references.

In the context of a developmentally regulated operon like dev, known to autoregulate as well as to regulate other developmental genes, the hypothesis of Makarova et al. (65) has an intriguing extension: that transcribed unique inserts function as microRNAs (miRNAs) to inhibit expression of M. xanthus developmental genes. Therefore, we performed a BLAST (1) search with each unique insert in the CRISPR at the dev locus. We found three matches to other sequences in M. xanthus, but only one of these is likely antisense to a transcript (assuming the unique insert is transcribed from the dev promoter) and therefore could function as miRNA (Fig. 12B). This match is to the coding region of a putative Cas protein (MXAN_7283) that is near a different CRISPR, raising the possibility of “cross talk” regulation between different CRISPR-Cas systems.