Genetic evidence for interaction of sigma E with the spoIIID promoter in Bacillus subtilis

During sporulation in Bacillus subtilis, new RNA polymerase sigma factors are produced. These sigma factors direct the transcription of genes that are required for this cellular differentiation. In order to determine the role of each sigma factor in this process, it is necessary to know which promoters are recognized by each sigma factor. The spoIIID gene product plays an important role in the establishment of mother cell-specific gene expression during sporulation. We found that substitution of an alanine at position 124 of the sporulation-specific sigma factor sigma E suppressed the effect of a single-base-pair transition at position -13 of the spoIIID promoter. This alanine substitution in sigma E did not suppress the effect of a transversion at position -12 of the spoIIID promoter. The allele specificity of the interaction between sigma E and the spoIIID promoter is strong evidence that sigma E directs transcription from the spoIIID promoter during sporulation. Position 124 in sigma E is located within a region that is highly conserved among the regions in other sigma factors that probably interact with the -10 regions of their cognate promoters.

The sigma subunit of bacterial RNA polymerase determines the specificity of promoter utilization. In Bacillus subtilis, new sigma factors are produced during sporulation, resulting in the transcription of genes that are required for this cellular differentiation (reviewed in references 19 and 27). The production of these sigma factors appears to play an important role in regulating gene expression during sporulation. However, since several sigma factors are present simultaneously during sporulation, it has been difficult to determine which sigma factor is responsible for the transcription of each sporulation-essential gene. Therefore, the exact role of each sigma factor remains unclear.
Recently, a genetic approach has been used to provide compelling evidence that CA (11) and (rH (1,32) interact directly with specific promoters during sporulation. These studies and those on &0 in Escherichia coli (5,24,29) support the model that sigma factors govern the specificity of promoter utilization by making sequence-specific contacts at two regions of their cognate promoters, the -10 and -35 regions. In these studies, single-amino-acid substitutions in the sigma factors were found to suppress the effects of specific single-base-pair substitutions in promoters. For example, substitution of threonine at position 100 of aH suppressed the effect of a base pair substitution at position -13 of the spoVG promoter but not the effects of substitutions at other positions in the spoVG promoter (32). Substitution of alanine at position 96 of crH partially suppressed the effect of a substitution at position -12 of the spoVG promoter (1). The position-specific suppression of promoter mutations by the amino acid substitutions at positions 96 and 100 of crH is strong evidence that this region of the sigma factor directly contacts the -10 region of the spoVG promoter. These results and those of similar experiments with u70 from E. coli (5,24) inspired experiments with A in which substitution of arginine at position 196 of CiA was shown to suppress the effects of a base pair substitution at position -11 in the spolIG promoter (11). The allele specificity of this suppres-* Corresponding author. sion was taken as strong evidence that CA interacts with the spoIIG promoter during sporulation. Subsequently, suppression experiments have been used to demonstrate interaction of aA with three additional promoters in B. subtilis (9,31). These results provide important clues concerning the molecular interactions of sigma factors with their cognate promoters; however, the more significant implication for the study of sporulation probably is that these altered sigma factors provide tools that can be used to show which sigma factors interact with which promoters during sporulation.
During endospore development the cell is divided into two compartments in which differential gene expression results in different developmental fates. The mother cell becomes a terminally differentiated cell that provides specialized products to the developing endospore, while the forespore compartment develops into the endospore, which will remain dormant until germination. The spoIIID gene product is a DNA-binding protein that is necessary for mother cellspecific gene expression (12,14). Analysis of spoIIID-lacZ fusions demonstrated that spoIIID expression begins 2 h after the onset of sporulation and that this expression is dependent upon the product of sigE (14,25). sigE encodes the RNA polymerase sigma factor cE, which is produced about 2 h after the onset of sporulation. The temporal pattern of spolIID expression as well as its dependence on sigE and other sporulation genes has been used to suggest that spoIIID transcription is directed by cE (14,25). In an alternative model, the dependency of spoIIID expression on sigE could result because aE directs the expression of an unknown factor that is required for spolIID expression. In this model, cE does not interact directly with the spoIIID promoter. We wished to determine whether genetic suppression experiments similar to those described above for crH and (rA could be used to demonstrate that aE interacts with the spolIID promoter during sporulation. Our strategy required localization of the spoIIID promoter, characterization of single-base-pair substitutions in the spoIIID promoter, and isolation of a mutant allele of sigE which resulted in a single-amino-acid substitution in vE that suppressed the effect of a base pair substitution in the spoIIID promoter.  (23) with vigorous shaking at 37°C. When the culture reached an OD550 of 0.3, it was diluted fivefold in DS medium and incubation at 37°C was continued. Samples (100 ml) were harvested during sporulation, and RNA was isolated as previously described (8). The procedure for primer extension analysis has been described previously (8). The sequence of the oligonucleotide that was used to prime the DNA synthesis was 5'-GCTTCTTCGT TGTATGG-3'. The size standards were generated by using the same oligonucleotide with pBK39 (14) as a template in dideoxy sequencing reactions with modified T7 DNA polymerase (Sequenase).
Cloning and mutagenesis of the spollID promoter. A 130-bp XmnI-DraI fragment from pBK39 (14) containing the spoI-IID promoter region from positions -72 to +63 was cloned into pUC18 (30) at the HincII site, creating pUC18-IIID. The orientation of the spoIlIlD promoter with respect to the HinclI site was determined by restriction endonuclease mapping and confirmed by nucleotide sequencing. The 160-bp HindIII-BamHI fragment from pUC18-IIID was cloned between the HindIll and BamHI sites of both M13mpl8 and Ml3mpl9, generating mpl8-IIID and mpl9-IIID, respectively.
Single-nucleotide substitutions in the -10 region of the spolIID promoter were constructed by oligonucleotide-directed mutagenesis by the procedure of Kunkel (15) as described previously (22). The sequences of the oligonucleotides used to generate these mutants were as follows: 5'-GAATGCTATTACACTGT-3' was used to construct mpl8IIID12T, and 5'-AGAATGCTGATACACTG-3' was used to construct mpl8IIIDI3G. The replicative form (RF) of each mpl8HIID derivative was prepared, and the nucleotide sequence was determined to confirm that the correct mutations were constructed.
Transfer of spoIIID-lacZ transcriptional fusions to SP,I.
Promoter-lacZ fusions were constructed by excising the 160-bp HindIlI-BamHI fragment from the RF of each mpl8-IIID derivative and cloning each between the HindIlI and BamHI sites of pZAB/HNB (9), which contains a promoterless derivative of the E. coli lacZ gene. Nucleotide sequencing of the promoter fusions was used to confirm that the proper constructions were made. After the pZAB derivatives were linearized with SstI, they were used to transform strain ZB307A (20) to chloramphenicol resistance.
Homology between the pZAB derivatives and the prophage SP,Bc2del2::Tn917 pSK10A6 in the B. subtilis chromosome allowed a double crossover event to occur, generating strains in which the spoIIID-lacZ fusion is on the SP,B prophage and is now selectable by resistance to chloramphenicol. The SP,B phage was induced from the chromosome of each strain by heat as described previously (22), generating the following phage lysates: SPPIIID-lacZ, SPPIIID12T-lacZ, and SPPIIIDJ3G-lacZ.
Cloning and mutagenesis of sigE. A 1.1-kb PstI fragment containing sigE was cloned into the PstI site of pJ89 to create pJ8900. pJ89 (6) is an expression plasmid which contains the pUCfl origin for replication in E. coli, the gene encoding bleomycin resistance, and the pUB110 origin for replication in B. subtilis. The lac promoter (Piac) in pJ8900 was changed to the consensus sequence for a sigma A promoter (Pla,BS) (9) by oligonucleotide-directed mutagenesis, generating pJ8903.
To obtain single-stranded phagemid DNA for mutagenesis of the sigE gene, E. coli RY2504 containing pJ8903 was infected with the helper phage M13K07 and grown in the presence of kanamycin at 70 p.g/ml, bleomycin at 7 gg/ml, and uridine at 0.25 ,ug/ml. The mutation in sigE was constructed by oligonucleotide-directed mutagenesis by the procedure of Kunkel (15). The mutation in the sigE allele resulted in changing the methionine (M) to an alanine (A) at position 124 (sigE124MA). The nucleotide sequence of the oligonucleotide used to make the sigEJ24MA allele was 5'-TCTTAAATACGCCAGGATTTCA-3'. To ensure that the proper mutant sigE allele was constructed, the DNA sequence of the mutated gene was determined by using the oligonucleotide 5'-TACGGAATTAATATAG-3', which is complementary to a region in sigE. Construction of B. subtilis strains containing sigE alleles under control of the Pspac promoter. B. subtilis strains containing sigE alleles under Pspac promoter control were constructed in two stages, as described previously for the construction of sigA alleles controlled by Pspac (9,11). First, the sigE and mutant sigE alleles were cloned downstream from the Pspac promoter in pAG58bleo-J. Second, the Pspac-sigE alleles were cloned into pTV21A2 so they could be easily transferred into the chromosome. The plasmids pJ8903 and pJsigE124MA, containing the wild-type and mutant sigE alleles, respectively, were digested with PstI, generating 1.1-kb fragments, which were made blunt-ended with T4 DNA polymerase and cloned into the SmaI site of pIC20H (18), creating plCsigE and pICsigEJ24MA. The sigE alleles were excised from the plasmids as XbaI fragments and ligated into the XbaI site of pAG58bleo-1, generating plasmids pSPIGMAsigE and pSPIGMA124MA, respectively. These plasmids have the wild-type and mutant sigE alleles under the control of the Pspac promoter and are selectable by resistance to bleomycin in E. coli and phleomycin in B. subtilis. To ensure that the wild-type and mutant sigE alleles were cloned in the correct orientation with respect to the Pspac promoter, the plasmids were characterized by restriction endonuclease analysis and by nucleotide sequencing analysis.
The PstI-NcoI fragment from pSPIGMA plasmids, which contains the wild-type and mutant sigE allele under the control of the Pspac promoter, was excised and ligated to the two 5-kb PstI-NcoI fragments from pTV21A2. The ligated mixture was used to transform a B. subtilis strain that contained pTV21A2 to phleomycin resistance at a temperature permissive for plasmid replication, as described previously (9,11). These transformants were sensitive to chloramphenicol. In this transformation, the homology from pug/ml and formed less than 10 heat-resistant spores per ml was chosen. B. subtilis EU100 was transformed with pTVSPIGMAsigE and pTVSPIGMAsigE124MA at the nonpermissive temperature for plasmid replication and selected for growth on phleomycin with sensitivity to chloramphenicol, creating EU101 and EU102, respectively. I-Galactosidase assays. B. subtilis EU101 and EU102 were lysogenized with the SPP derivatives that contained wildtype and mutant spoIIID promoters fused to lacZ and were selected by growth on LB supplemented with 5 ,ug of chloramphenicol per ml. Only those transductants that were able to grow on DS medium containing 5 gg of chloramphenicol, 0.5 ,ug of phleomycin, and 1 gg of erythromycin per ml were used for further analysis. DS medium containing chloramphenicol (5 ,ug/ml) was inoculated with a'fresh transducta,nt and grown until an OD' of 0.6 was attained, at which time the cultures were divided. To one half of the cultures, 1 mM IPTG (isopropylthiogalactopyranoside) was added. At hourly intervals, 1 ml of the cell culture was harvested, and ,-galactosidase assays were done as described previously (10).

RESULTS
Primer extension analysis of the spolIlD transwript. We used primer extension analysis to map the 5' end of the spolIlD transcript. The promoter for spoIlID was thought to be located' less than 230 bp upstream from the spoIIID coding region, since a DNA fragment containing this region and the coding region could complement spoIIID mutations in trans when inserted in the amy locus of the B. subtilis chromosome (14). Two different primers were used in R. Losick's laboratory to map the 5' end of the spoIIID transcript (13,28). We confirmed their results with a third primer. Our primer extension analysis identified a transcript, which was weakly detected 2 h after the onset of sporulation but was abundant 4 h after the onset of sporulation (Fig. 1). Comparison of the electrophoretic mobility of the primer extension product with that of the products of the dideoxy sequencing reactions in Fig. 1 was used to map the 5' end of the transcript on the DNA sequence. The primer extension product comigrated with a dideoxyguanosine product (Fig.  1) that corresponds to C in the nontranscribed strand (Fig.   2), 160 bases upstream from the start of4he spoIIID coding sequence (14). The 5' end of the spoIIrb transcript probably represents the end of the primary transcription product rather than a 5' end that was generated by processing of a longer transcript, since a similar transcript was generated in vitro by RNA polymerase containing iE (data not shown). We assigned the start point of transcription to the adenine residue located 1 bp downstream from the cytosine (Fig. 2) because in in vitro transcription.experiments with crE RNA polymerase, the dinucleotide ApA primed transcription from this promoter more efficiently than did ApC (data not shown). Direct evidence that the s~equence upstream from this putative start point of the spollID transcript acts as the promoter was provided by the mtftagenesis experiments described below.
Analysis of spQIIID promoter mutations. To determine whether the DNA surrounding the putative start of the spoIHD transcript had'promoter activity in vivo and to examine the effects of mutations in this promoter, we cloned a DNA fragment containing this region (from 72 bp upstream from the transcription start point to 63 bp downstream from the transcription start point) upstream from a promoterless derivative of lacZ from E. coli as described in Materials and The resulting cDNAs were subjected to electrophoresis on polyacrylamide sequencing gels next to size standards that were produced by using the same oligonucleotide in dideoxy sequencing reactions with a DNA template containing the spolIID promoter region (lanes d through g). The position of the primer extension product is indicated by the arrows.
Methods. This spoIIID-lacZ fusion was carried into the chromosome of B. subtilis on a SP3 specialized transducing phage. We found that ,B-galactosidase began to accumulate about 2 h after the onset of sporulation in an otherwise wild-type strain that had been lysogenized with the SPfIIID-lacZ transducing phage (data not shown).
To determine whether the region upstream from the start point of transcription was promoting this expression of lacZ, we introduced two different single-base-pa-ir substitutions-a transition and a transversion, 13 and 12 bp, respectively, upstream from the putative start point of transcription (Fig.   2). These specific base substitutions were chosen based on our previous studies of seven promoters that are used in vitro by c.E RNA polymerase (21). In these studies, we found l\ /1\ /\  The circled positions indicate the positions at whici substitutions have been found to specifically suppress mutations in the -10 regions of promoters, C70 (24, 2' cH (1,32), and aE (this paper). The sequences are aligr to Stragier et al. (26) except that the amino acid sequer been corrected (17). The conserved regions are shade that the sequence CATATTNT was conserved region of these promoters. The underlined b sequence are the most highly conserved amonl moters (21). We noted that these most highl) bases are also found in the -10 region of promoter (AATACACT) (Fig. 2). In our previot the effects of mutations in two promoters (21), w a single substitution of a G at the first pos consensus created promoters that were used leb y crE RNA polymerase. Therefore, we made t tion at position -13 in the spoIIID promote substituted T for A at the position in the spoIl (-12) that may be homologous to the second pc consensus sequence, since T was not found at in any of the promoters used by cE RNA polyme (21). We also were aware that regardless of i factor uses the spolIlD promoter in vivo, thes would probably affect promoter activity if transc being initiated from this sequence, because thi sequences at the -10 region of most types of pi important determinants of promoter strength ( these single-base substitutions reduced the proni of this region. These mutant derivatives of th provided us with the opportunity to test wheth acid substitution in cE could suppress the effec tion in this promoter. Suppression of the effect of a promoter mul amino acid substitution in rFE. The positions at acid substitutions in & , JA , and CrH have be suppress the effects of single-base-pair substiti -10 region of their cognate promoters lie withi regions of these sigma factors (Fig. 3). It is like of these sigma factors use a similar motif to inte with the -10 regions of their cognate prom interacts at the -10 region of its cognate prc manner similar to that of CH, then possibly the the methionine at position 124 of (JE, like its hc (the threonine at position 100), would interact -13 of its cognate promoters (Fig. 3). If tI correct, then alteration of the side chain at po substitution of alanine for methionine may prod (FE that is unable to discriminate between two pi differ only by the base pair at position -13. loss-of-specificity mutation has been observedX well as with sequence-specific DNA-binding pi To test this model, we constructed strains that contained the structural gene for cE under the control of an inducible I at the -10 promoter. One set of strains contained the wild-type allele of ases in this sigE. A second set of strains contained a mutant allele of g these pro-sigE (sigEI24MA), which resulted in the substitution of y conserved alanine at position 124 of (rE. Transductants of these strains the spolIlD were isolated that contained prophage carrying either the us studies of wild-type or mutant derivatives of the spoIIID promoter ie found that fused to lacZ (Fig. 4). The accumulation of P-galactosidase ition of the was monitored in cultures of each strain incubated in the ss efficiently presence or absence of the inducer (IPTG) of sigE expreshis substitusion (Fig. 5). In the strain containing the wild-type allele of er. We also sigE and the wild-type spoIIID promoter, P-galactosidase base-pair substitutions at two positions in the -10 region of this promoter reduce its activity. In other work (7), we have shown that this promoter (spoIlID) is used efficiently in vitro by cE RNA polymerase. We had previously noted conserved sequences at the -10 and -35 regions of promoters that are used by CrE RNA polymerase in vitro (21). Recently, Foulger and Errington (4) have compiled additional sequences of promoters that are probably used by aE RNA polymerase. (They included the spoIIID promoter in their list of sequences.) They used these sequences to suggest a modified consensus for the -10 and -35 regions of promoters used by (rE RNA polymerase. Their consensus for the -10 region of CE-dependent promoters retains the most highly conserved features of the sequence found previously (21) but did not include the C that was noted at the first position of the -10 region (4). This position is occupied by an A in the spoIIID promoter, and we found that substitution of a G at this position decreased promoter activity. Evidently, the base pair at this position plays an important role in signaling recognition of the promoter by cE RNA polymerase. In fact, the results of the suppression experiment suggest that this base pair interacts with the amino acid at position 124 of orE. The addition of the new sequences of (uE-dependent promoters by Foulger and Errington (4) and recently by R. Losick and his colleagues (16) probably increases the accuracy with which the consensus sequence represents the sequences that signal recognition of promoters by cE RNA polymerase. However, this assumption must be tested by additional mutagenesis of promoters that are used by urE RNA polymerase.
The correlation of spoIIID promoter activity with the production of aE during sporulation and its dependence on a functional allele of sigE had been used to suggest that (rE directs expression from the spoIIID promoter (14,25). These results, taken together with our results that show allelespecific suppression of a spoIIID promoter mutation by an amino acid substitution in cE, provide a compelling argument that the spoIIlD promoter is used by aE RNA polymerase during sporulation.
During sporulation, spoIIID promoter activity is restricted predominantly to the mother cell compartment (14). Since the activities of several other promoters, thought to be used by cE, are also restricted to the mother cell compartment, it has been suggested that CE activity is restricted to this compartment (2,25). This model predicts that the activities of all promoters used by CE RNA polymerase are restricted to the mother cell compartment. In this regard, one of the important implications of our result is that the 124MA allele of sigE provides a tool with which to test whether specific promoters are used by CE RNA polymerase during sporulation. Specific suppression of a base substitution in a promoter by the 124MA allele of sigE would provide evidence that the promoter is used by cE RNA polymerase during sporulation.
Our results confirm and extend the observations that suggest that most sigma factors use a similar motif to make sequence-specific contacts with the -10 regions of their cognate promoters. Our results suggest that the methionine at position 124 of CE interacts with the base pair at position -13 of the spoIIID promoter, whereas CH uses a threonine for the homologous interaction with its cognate promoters (1, 32) (Fig. 3). Suppression experiments also support the model that the arginine at position 96 in orH contacts position -12 of the spoVG promoter (1) and that the glutamine at  [29]) contacts position -12 in the lacBS promoter (9). Therefore, it seems likely that the asparagine at position 120 in cE has a homologous function (Fig. 3). The role of the other amino acids in this region of each sigma factor remains to be discovered. However, since cE is not essential for growth of B. subtilis and since it appears to interact with promoters in a manner similar to that of other sigma factors, examination of the effects of additional amino acid substitutions in cuE can be expected to provide additional insights into the function of sigma factors.