The Rhodobacter sphaeroides ECF sigma factor, σE, and the target promoters cycA P3 and rpoE P11

Article No. jmbi.1999.3263 available online at http://www.idealibrary.com on

J. Mol. Biol. (1999) 294, 307±320

The Rhodobacter sphaeroides ECF Sigma Factor, s E, and the Target Promoters cycA P3 and rpoE P1 Jack D. Newman, Mat J. Falkowski, Brenda A. Schilke Larry C. Anthony and Timothy J. Donohue* Department of Bacteriology University of Wisconsin Madison, Madison, WI 53706, USA

Rhodobacter sphaeroides rpoE encodes a 19.2 kDa protein, sE, related to members of the extra-cytoplasmic function subfamily of eubacterial RNA polymerase s factors. We demonstrate that sE directs transcription from rpoE P1, the promoter for the rpoEchrR operon, and from cycA P3, a promoter for the cytochrome c2 structural gene. Comparison of these sE-dependent promoters reveals signi®cant sequence conservation in their ÿ35 and ÿ10 regions; however, rpoE P1 is over 80-fold stronger than cycA P3. Both promoters contain identical ÿ35 hexamers, ÿ36 TGATCCÿ31, that appear to constitute the preferred sequence, since any single base mutation in this region of cycA P3 reduces promoter function. The higher activity of rpoE P1 appears to re¯ect a better ÿ10 region, ÿ13TAAGAÿ9, as it contains four out of ®ve of the nucleotides found to be important to sE-dependent transcription. We also propose that ChrR acts as an inhibitor of sE, since these two proteins can form a complex, and
chrR mutations increase sE-dependent transcription. ChrR is believed to respond to a signal from tetrapyrrole biosynthesis because loss of function mutations in chrR lead to cohemin resistance. Based on our observations, we present a model in which cohemin resistance is conferred by increasing sE activity. # 1999 Academic Press

*Corresponding author

Keywords: sigma factor; extra-cytoplasmic function; anti-sigma factor; transcription; tetrapyrroles

Introduction Eubacteria rely on RNA polymerase containing the primary sigma factor, s70, to direct transcription of a large number of genes commonly expressed in logarithmic growth phase (Record et al., 1996). In recent years it has become clear that eubacteria also utilize alternative sigma factors to direct a pool of RNA polymerase to transcribe suites of genes appropriate to changing environmental conditions (Wosten, 1998). The extra-cytoplasmic function (ECF) sigma factors are one such family of alternative sigma factors that generally direct the transcription of genes whose products function outside the cytoplasm (Missiakas & Raina, 1998; Wosten, 1998). In different bacteria, the target genes recognized by ECF sigma factors are responsible for such diverse functions as periplasmic protein turnover (Erickson & Gross, 1989; Mecsas et al., Abbreviations used: ECF, extra-cytoplasmic function. E-mail address of the corresponding author: [email protected] 0022-2836/99/470307±14 $30.00/0

1993), or the synthesis of cell surface polymers (Hershberger et al., 1995; Schurr et al., 1995), antibiotics (Paget et al., 1998) or membrane carotenoids (Gorham et al., 1996). Recently completed genome sequences suggest that ECF sigma factors play a major role in gene regulation, as some bacteria may have as many as 15 open reading frames assigned as sigma factors, a third of which fall into the ECF sub-family (Cole et al., 1998). Like many alternative sigma factors, the synthesis and activity of ECF family members is often regulated. Commonly, the activity of ECF sigma factors is controlled by an anti-sigma factor that is encoded by the gene immediately downstream of the sigma factor (Brown & Hughes, 1995). For example, the genes for sE homologues in Escherichia coli or Pseudomonas aeruginosa are co-transcribed with rseA, and mucA, respectively, which encode membrane-bound regulators of their cognate sigma factor (De Las Penas et al., 1997; Gorham et al., 1996; Schurr et al., 1996). These antisigma factors function as reversible inhibitors of sigma factor activity, keeping their activity low # 1999 Academic Press

308 until the appropriate signal is received (Missiakas & Raina, 1998). Despite this common mechanism of control, the signals that regulate individual sE homologues may be quite different depending on the physiological role of the target genes (Missiakas & Raina, 1998). In E. coli, where sE controls the extra-cytoplasmic stress response, the signal is believed to be the accumulation of mis-folded proteins in the periplasm (De Las Penas et al., 1997; Missiakas & Raina, 1997). Alternatively, in Myxococcus xanthus, the ECF sigma factor controlling the carotenoid biosynthetic genes, CarQ, is regulated by CarR, which is thought to sense oxidative stress (Gorham et al., 1996). Our work focuses on the regulation and target genes of the ECF sigma factor, sE, from the photosynthetic bacterium Rhodobacter sphaeroides. We show that one such target gene is cycA, which codes for the periplasmic electron carrier, cytochrome c2. The R. sphaeroides cycA gene is expressed from three different promoters, P1-P3, each responding to a distinct signal (Karls et al., 1998; MacGregor et al., 1998). Activity of cycA P1 is increased by heat shock so it is not surprising that this promoter is recognized by sigma factors related to the E. coli heat shock, sigma factor, s32 (Karls et al., 1998). The cycA P2 promoter is recognized by the s70 homologue of this bacterium, and stimulated by changes in the oxidation-reduction state of the electron transport chain that activate the photosynthesis response regulator, PrrA (Karls et al., 1999). Activity of the third promoter, cycA P3, is increased in strains containing chrR point mutations that result in resistance to the toxic heme analog cobaltic hemin (cohemin) (Schilke & Donohue, 1992). Here, we show that R. sphaeroides rpoE encodes the ECF sigma factor sE, which recognizes both cycA P3 and its own promoter rpoE P1. Comparison of the cycA P3 and rpoE P1 promoters in combination with mutational analysis of cycA P3 have identi®ed nucleotide sequences important to transcription from sE-dependent promoters. The chrR gene immediately downstream of rpoE is also shown to encode a protein that inhibits sE activity in vivo. Our data indicate that the previously identi®ed chrR point mutations of cohemin resistant cells increase sE-dependent transcription by inactivating ChrR.

R. sphaeroides E-Dependent Transcription

as cycA P3; Schilke & Donohue, 1992). These chr mutations were mapped to one locus, which was previously thought to de®ne a single gene product, ChrR (Schilke & Donohue, 1995). However, correction of a frameshift in the published DNA sequence predicts that two genes, rpoE and chrR, are present at this locus and that cohemin resistant cells contain point mutations in chrR (Figure 1; accession number U11283). To con®rm that this locus encodes two polypeptides, the RpoE protein (sE) was His-tagged and functional His-sE was puri®ed from E. coli (see below). In addition, puri®cation of ChrR-His from R. sphaeroides yielded a protein with an apparent molecular mass by SDSPAGE (25 kDa) similar to that predicted from the DNA sequence (23 kDa; data not shown). N-terminal sequencing of ChrR-His showed that the ®rst 12 residues were identical with the amino acid sequence deduced from chrR (Figure 1). The following experiments describe how sE and ChrR control gene expression. rpoE codes for a member of the ECF sigma factor family The rpoE gene encodes the 168 amino acid residue sE protein (19.2 kDa; coordinates 430-936 in accession number U11283) that is predicted to be a member of the ECF sigma factor family (Figure 2). R. sphaeroides sE is most similar to the mycobacterial ECF sigma factors, SigK (32 % identical, 49 % similar) and SigE (25 % identical, 39 % similar), but also shares considerable amino acid similarity with E. coli RpoE (35 %) and P. aeruginosa AlgU (38 %). R. sphaeroides sE is related to ECF sigma factors

Results chr mutations are in the rpoEchrR locus While the heme analogue cobaltic protoporphyrin IX (cohemin) is toxic to R. sphaeroides, chr mutations have been described that lead to cohemin resistance (Schilke & Donohue, 1992). One major effect of chr mutations is to increase abundance of a cycA transcript initiating 52 base-pairs upstream of the start of translation (the promoter that produces this transcript is hereafter referred to

Figure 1. The R. sphaeroides rpoEchrR operon. The organization of the rpoEchrR operon in wild-type (top), TF18 (middle), and
ChrR (bottom) strains. The Nterminal amino acid residues of ChrR con®rmed by protein sequencing are underlined. The chr-4 allele is a point mutation that produces a cysteine to arginine change at residue 38 in ChrR (C38R in bold) (Schilke & Donohue, 1995). Strains TF18 and
ChrR have null mutations in the rpoEchrR genes or the chrR gene, respectively.

309

R. sphaeroides E-Dependent Transcription

throughout its length, including domains implicated in binding core RNA polymerase (region 2.1) and ÿ35 promoter recognition (region 4.2) (Lonetto et al., 1994). s E directs transcription of cycA P3 and rpoE P1 in vitro E

To test if s acts as a sigma factor in vitro, transcription assays were performed using a template (pJDN11) that contains cycA P3 sequences from ÿ53 to ‡11 (relative to transcription start) cloned upstream of a SpoT 40 terminator. When reconstituted EsE was tested for activity with this cycA P3 template, a product of the size expected from the in vivo transcription initiation site was synthesized (Figure 3(a)). Synthesis of this transcript was dependent on cycA P3 sequences, since it was not produced from a control plasmid that lacks this promoter. In addition, synthesis of this transcript was dependent on sE since it was only observed if the sigma factor is present in the reactions. From this we conclude that sE recognizes and directs transcription from cycA P3 in vitro. Cohemin-resistant cells also have elevated expression of the cycA P1 promoter that initiates transcription 24 bp downstream of cycA P3 (Schilke & Donohue, 1995). To determine if EsE recognizes cycA P1, we analyzed transcription from a template that included both of these promoters (extending from ÿ53 to ‡43 relative to cycA P3 and ÿ77 to ‡19 relative to cycA P1). It does not appear that EsE recognizes cycA P1, since this transcript is not produced from this template. However, the amount of the cycA P3 transcript produced from this template was  tenfold greater than the shorter construct containing only cycA P3, indicating that sequences downstream of ‡1 can increase cycA P3 activity. Thus the increase in cycA P1 activity in cohemin-resistant cells is probably not a direct consequence of transcription from this promoter by EsE (see Discussion). For simplicity, all subsequent analysis of cycA P3 expression uses promoter fragments lacking cycA P1. Another feature of ECF family members is that the sE gene is often transcribed by EsE (Missiakas & Raina, 1998). It appeared that the rpoEchrR operon might contain a sE-dependent promoter, since a primer extension product that initiates 132 nt upstream of rpoE is present at higher levels in cells with increased sE activity (data not shown). To test if rpoE P1 is transcribed by EsE in vitro, transcription assays were performed with a plasmid template that contains this promoter (from ÿ39 to ‡17 relative to transcription initiation) cloned upstream of a SpoT 40 terminator (pJDN34). When reconstituted EsE was tested for activity with this rpoE P1 template, a product of the size expected from the in vivo transcription initiation site (Schilke & Donohue, 1995) was produced (Figure 3(b)). Thus, we conclude that rpoE P1 is also a sE-dependent promoter. Quanti®cation of the transcripts produced in vitro shows that rpoE P1 (pJDN34)

produces approximately 84-fold more RNA than a cycA P3 template that contains a promoter element of comparable size (pJDN11; Figure 3(c)). From this, we conclude that rpoE P1 is more active than cycA P3 in vitro. Important sequence elements in the ÿ10 region of an R. sphaeroides s E-dependent promoter As members of the s70 family, ECF sigma factors are believed to recognize sequences centered at ÿ10 and ÿ35 (Wosten, 1998). To test if conserved sequences in the ÿ10 region of rpoE P1 and cycA P3 (Figure 3(c)) contribute to promoter function, we constructed a library of mutant cycA P3 promoters in which each nucleotide of this potential hexamer was changed to the other three possible bases (Figure 4(a)). The effect of each mutation was analyzed in vitro by assaying activity with reconstituted EsE, and in vivo by monitoring LacZ levels from a reporter gene in a
ChrR mutant in order to increase sE activity (see below). Control experiments have shown that transcription from this reporter gene initiates from the same position as genomic cycA P3 (data not shown) indicating that it provides an accurate picture of in vivo promoter activity. We also know that in vivo activity of all cycA P3 mutant promoters is abolished by a
rpoE mutation (data not shown), indicating that they are not recognized by other RNA polymerase holoenzymes. We found that the conserved T and A residues at ÿ12 and ÿ11 of cycA P3 are critical to sE-dependent transcription, since any mutation at either T12 or A11 results in a non-functional promoter in vitro and at least a ®vefold reduction in activity in vivo (Figure 4(a)). A G to A mutation at ÿ10 (G10A), making cycA P3 more similar to rpoE P1, increases promoter function  sevenfold in vitro, while G10T or G10C are  eight or fourfold more active, respectively (Figure 4(a)). In vivo analysis of promoters containing these ÿ10 mutations shows the same general trend (T > A > C) however only the G10T promoter is stronger than wild-type cycA P3 (Figure 4(a)). It appears that T (wild-type) or A is preferred over G or C at ÿ9 since T9G and T9C have  two- to ®vefold less activity than either wild-type cycA P3 or T9A in vitro and in vivo (Figure 4(a)). A ÿ8 mutation that makes cycA P3 more similar to rpoE P1 (G8A) increases activity  tenfold in vitro (Figure 4(a)), while G8T or G8C increases cycA P3 function  seven- or threefold, respectively. Surprisingly, all ÿ8 substitutions produced promoters with essentially the same in vivo activity as cycA P3. Changing the A at ÿ7 only affected promoter strength by twofold or less in vitro (T > G > A ˆ C) or in vivo (G > A > C ˆ T; Figure 4(a)), suggesting that any mutation at this position has little effect on cycA P3 function. In summary, the data show that mutations in cycA P3 at ÿ10 or ÿ8 that make this promoter more like rpoE P1, increase promoter strength. The

310

R. sphaeroides E-Dependent Transcription

Figure 2 (legend opposite)

311

R. sphaeroides E-Dependent Transcription

data predict that a sE-dependent promoter should contain TAwww in the ÿ10 region, where w ˆ T or A. Potential reasons why some individual promoter mutations may produce different effects in vitro and in vivo are presented in the Discussion. cycA P3 and rpoE P1 both contain an optimal ÿ35 element The cycA P3 and rpoE P1 promoters also contain six identical base-pairs centered around ÿ35 (Figure 3(c)). It appears that this conserved hexamer represents an optimal ÿ35 region, since any deviation from ÿ36TGATCCÿ31 reduces cycA P3 function both in vitro and in vivo (Figure 4(b)). The negative effects of these changes are most evident in vitro, as all but three mutations (G35C, G35A and C31T) reduce promoter activity to less than 10 % of cycA P3. Mutations in this hexamer often produce less dramatic effects on cycA P3 function in vivo, as half of these mutant promoters retain 40-55 % of wild-type activity. However, the other half of these mutant cycA P3 promoters have between 14 and 34 % of wild-type activity. Thus it appears that an R. sphaeroides sE-dependent promoter should include TGATCC in the ÿ35 region. Inactivation of chrR increases cycA P3 expression The activity of many ECF family members is negatively regulated by anti-sigma factors that are encoded by a downstream gene (Brown & Hughes, 1995). To test if the product of the downstream gene, ChrR, inhibits R. sphaeroides sE, we analyzed how the presence or absence of each of these proteins altered cycA P3 activity in vivo using a LacZ reporter plasmid (pJDN11z). Activity of sE is very low in wild-type cells, as LacZ levels are only slightly higher than that measured in TF18, lacking the sigma factor (Figure 5). It appears that ChrR inhibits sE activity, since LacZ levels from the cycA P3 reporter gene are increased at least tenfold by a
chrR allele (
chrR-1::dfr). To ensure that this increase in cycA P3 transcription is not due to a polar effect of the
chrR allele, we introduced a plasmid containing chrR (pJDN6) behind its native promoter into the
ChrR mutant. As expected, plasmid-encoded ChrR was suf®cient to reduce cycA P3 expression (Figure 5). Consistent with our observations, introducing a plasmid-encoded rpoE gene into cells containing a
rpoEchrR allele is suf®cient to increase

cycA P3 transcription to an extent comparable to that seen in the
ChrR mutant (Figure 5). To test if previously characterized chrR point mutations (Schilke & Donohue, 1992, 1995) prevent ChrR from inhibiting sE, we compared sE activity in cells containing either the
chrR or chr-4 allele (Figure 1). LacZ levels from the cycA P3 reporter gene in cells containing the chr-4 allele were comparable to those observed in the
ChrR mutant (Figure 5). Thus, the C38R substitution, caused by the chr-4 allele, inactivates ChrR and increases sE activity. Loss of ChrR increases rpoE P1 activity Since we have shown that sE directs transcription of rpoE P1 in vitro, then loss of function mutations in chrR should also increase activity of this promoter in vivo. To test this hypothesis, we analyzed expression from a reporter gene in which a 57 bp region encompassing rpoE P1 (ÿ39 to ‡17 relative to transcription start) was fused to a promoter-less lacZYA operon. It appears that rpoE P1 is also recognized by sE in vivo, as the LacZ levels from this reporter gene were reduced to background in cells lacking the rpoE gene (Figure 5). However, the
chrR or the chr-4 allele increase rpoE P1 transcription approximately 200-fold. From this we conclude that ChrR also inhibits sE function at rpoE P1. Of equal signi®cance, both wild-type and chr mutant cells containing this rpoE P1::lacZ fusion have substantially more LacZ activity than comparable strains containing the cycA P3 reporter gene (Figures 5 and 6). Since rpoE P1 and cycA P3 are both transcribed by EsE, it appears that rpoE P1 is a stronger promoter than cycA P3 in vivo as well as in vitro. Moreover, this analysis of rpoE P1 activity reveals that some fraction of sE is active under aerobic conditions in wild-type cells. ChrR forms a complex with His-s sE If ChrR were an anti-sigma factor, it might bind to sE. To test for formation of such a complex, we asked if ChrR was associated with sE when this sigma factor was puri®ed under non-denaturing conditions. This association was assayed by passing soluble extracts from E. coli cells expressing both His-sE and wild-type ChrR over an Ni2‡ agarose column and analyzing fractions for His-sE or ChrR with speci®c antiserum against each protein. Under these conditions, His-sE bound to the

Figure 2. Alignment of ECF sigma family members. Identical/similar amino acid residues are shaded. The broken lines denote biochemically de®ned regions of eubacterial sigma factors (Lonetto et al., 1994). The sequences were translated from the following GenBank entries: RpoEEcoli, sE E. coli (D13169); AlgUPseae, AlgU P. aeruginosa (L02119); HrpLPsesy, HrpL Pseudomonas syringae (U03854); SigEStrco, sE Streptomyces coelicolor (L29636); SigXBacsu, SigX Bacillus subtilis (L09228); CarQMyxxa, CarQ Myxococcus xanthus (X71062); SigKMyctu, SigK M. tuberculosis (AL021932); SigEMyctu, SigE M. tuberculosis (U87242); RpoERhosp, sE R. sphaeroides (U11283). The alignment was generated using ClustalW (Higgins et al., 1996).

312

R. sphaeroides E-Dependent Transcription

Figure 3. sE directs transcription from cycA P3 and rpoE P1. Radio-labelled products of multiple round transcription assays. The transcript appears as a doublet due to two termination sites within the SpoT 40 transcription terminator. (a) and (b) The components of the in vitro transcription reactions and the plasmid templates used. Promoter fragments are labelled to show the extent of DNA contained on the plasmid template relative to the start site of transcription. Arrows denote transcription products of expected length from each promoter. (a) Templates containing (pJDN11 and pRKK139) or lacking cycA P3 (pRKK96) were tested for transcription with reconstituted EsE (lanes 1, 2 and 4) or with R. sphaeroides core RNA polymerase alone (RNAP) (lanes 3 and 5). (b) Comparison of transcription from rpoE P1 (pJDN34) compared to that of cycA P3 (pJDN11) with reconstituted EsE. (c) Comparison of promoter sequence of rpoE P1 (pJDN34) to cycA P3 (pJDN11). Relative transcript abundance from each promoter in vitro was measured by pixel intensity on a phosphorimager (Molecular Dynamics, Sunnyvale CA) and corrected for product length.

column and (after washing) was eluted with buffer containing 400 mM imidazole (Figure 7, lane 1). It appears that these two proteins form a complex since ChrR was also present in the eluate contain-

ing His-sE (Figure 7, lane 1). Control experiments demonstrate that ChrR binding to this column requires the presence of His-sE since ChrR is not retained when a comparable extract from E. coli

313

R. sphaeroides E-Dependent Transcription

Figure 4. Effect of single base mutations in cycA P3 activity. Analysis of cycA P3 function ( & ) in the (a) ÿ10 or (b) ÿ35 region compared to mutant cycA P3 promoters (ÿ53 to ‡11 relative to transcription start) with single base changes ( ). Top of each panel shows in vivo promoter activity as evidenced by LacZ expression from cycA P3::lacZ fusions in a
ChrR mutant. Assays were performed in triplicate and agreed within 15 % of the average. Bottom of each panel shows the transcripts produced from wildtype and mutant promoters in vitro with reconstituted EsE. The middle of each panel shows the level of transcripts produced as measured by pixel intensity of the transcript on a phosphorimager. Transcription assays were performed in duplicate and results agreed within 15 %. Control experiments show that there is <10 % difference in the production of the Es70-dependent RNA1 transcript from these plasmid templates (data not shown). Thus, the differences in EsE-dependent transcription do not re¯ect variation in template quality or loading error. ND (not detected) is reported for in vitro transcription products whose abundance was less than 5 % of the wild-type control. Activity for the wild-type cycA P3 promoter (ÿ36TGATCCÿ31 and ÿ12 TAGTGAÿ7) in vivo or in vitro is de®ned as 100 %. Activity of all mutant promoters is reported as a percentage of wild-type activity.

cells expressing wild-type sE (with no His tag) and ChrR is processed in an identical manner (Figure 7, lane 2). Formation of this complex also appears to require functional ChrR, since the non-functional chr-4 gene product (C38R), which fails to inhibit sE activity in vivo (Figures 5 and 6), is not retained on the column when extracts from E. coli cells expressing both His-sE and C38R are assayed for complex formation (Figure 7, lane 3). The ability of

ChrR to form a complex with His-sE is consistent with the proposed function of ChrR as an inhibitor, or anti-sigma factor, of sE. Increased s E activity or exogenous heme is sufficient for cohemin resistance If the C38R substitution in ChrR produces cohemin resistance by increasing sE activity, then

314

Figure 5. ChrR is an inhibitor of sE activity at cycA P3. b-Galactosidase activities from R. sphaeroides cells containing a cycA P3::lacZ (pJDN11z) transcriptional fusion. The strains shown are wild-type 2.4.1 ( ), TF18 (
rpoEchrR &),
ChrR (
chrR ), and Chr4 (chr-4 ). Also shown are the results obtained when either rpoE or chrR is placed in trans on a low copy plasmid under control of their native promoter (pJDN16 and pJDN8, respectively). All assays were performed at least three times. Vertical bars denote the average standard deviation from the mean.

R. sphaeroides E-Dependent Transcription

Figure 6. ChrR inhibits sE activity at rpoE P1. b-Galactosidase activities from wild-type 2.4.1 ( ), TF18 (
rpoEchrR &),
ChrR (
chrR ), and Chr4 (chr-4 ) carrying an rpoE::LacZ transcriptional fusion (pJDN30). All assays were performed at least three times. Vertical bars denote the average standard deviation from the mean.

this occurred was not known (Schilke & Donohue, 1992, 1995). Here we show that the R. sphaeroides ECF sigma factor sE recognizes cycA P3 and rpoE P1, and that point mutations like chr-4 inactivate ChrR, an inhibitor of sE activity. We also show that an increase in sE activity is suf®cient for cohemin resistance. By analyzing the activity of cycA P3, rpoE P1, and a bank of mutant cycA P3 promoters, we elucidate some of the characteristics of sE-dependent promoters and the regulation of this ECF sigma factor.


ChrR mutants should also exhibit this phenotype. As predicted,
ChrR cells plate with 100 % ef®ciency in the presence of 30 mg/ml cohemin, conditions that reduce the plating ef®ciency of both wild-type cells and those with a
rpoEchrR allele (Table 1). From this we conclude that increased sE activity can produce cohemin resistance. Previously, we proposed that cohemin is a toxic inhibitor of either heme biosynthesis or its function (Schilke & Donohue, 1992). However, tetrapyrroles such as cohemin are also reported to act as photosensitizers, killing cells under aerobic conditions (Nitzan et al., 1994). To distinguish between these two possible mechanisms of toxicity, we asked if the presence of another known photosensitizer like heme (Ladan et al., 1993; Nitzan et al., 1994) in¯uenced the effect of cohemin. When heme was included along with cohemin, cells plated with 100 % ef®ciency (Table 1). Since the addition of heme prevents cohemin toxicity we believe that cohemin is not acting as a photosensitizer.

Amino acid sequence alignments indicate that sE is a member of the ECF sigma factor family. Since sE directs transcription of cycA, encoding an extra-cytoplasmic cytochrome, it also ®ts the functional de®nition of an extra-cytoplasmic function (ECF) sigma factor. Typical of ECF family members (Missiakas & Raina, 1998), sE production is autoregulated in R. sphaeroides, as this sigma factor recognizes its own promoter, rpoE P1. Finally, the sE gene is encoded in an operon with an inhibitor of its activity, ChrR (Schilke & Donohue, 1995), another regulatory strategy common to ECF sigma factors (Missiakas & Raina, 1998).

Discussion

Sequences important to Es sE promoter recognition

Previous work identi®ed chr mutations that led to both cohemin resistance and increased cycA transcription; however, the mechanism by which

We found that R. sphaeroides sE prefers the sequence TGATCC and TAwww in the ÿ35 and ÿ10 regions of cycA P3. There is some sequence

R. sphaeroides s E has features typical of an ECF sigma factor

315

R. sphaeroides E-Dependent Transcription Table 1. Cohemin sensitivity of wild-type and mutant strains Strain 2.4.1 (wild-type) TF18
ChrR Chr4

Relevant genotype

Plating efficiency (30 mg/ml cohemin) (%)

Plating efficiency (30 mg/ml cohemin ‡ 30 mg/ml hemin) (%)

rpoE chrR
rpoE
chrR rpoE
chrR rpoE chr-4

<0.01 <0.01 99 97

103 92 93 114

similarity between the ÿ35 recognition hexamer of a R. sphaeroides sE-dependent promoter, TGATCC, and the consensus generated from other organisms, GAACTT (Missiakas & Raina, 1998). This conservation in ÿ35 hexamers could re¯ect amino acid similarity among ECF sigma factors in region 4.2 (Figure 2), which typically recognizes these nucleotide sequences. In contrast, it has been proposed that the small degree of DNA sequence conservation between ÿ10 hexamers of sE-dependent promoters from different bacteria is due to the low degree of amino acid sequence similarity in region 2.4 of ECF sigma factors (Missiakas & Raina, 1998). We found that the ÿ10 region recognized by R. sphaeroides sE (TAwww) has limited similarity to the consensus proposed for E. coli sE (TCTGA), even though there is little conservation between these two sigma factors in region 2.4. Both in vitro and in vivo assays show that promoter sequences between ÿ53 and ‡17 are suf®cient for sE-dependent transcription. Comparison of rpoE P1 and cycA P3 reveals a six for six, and a three for six match between their putative ÿ35 and ÿ10 hexamers (Figure 3(c)). The fact that rpoE P1 is stronger than cycA P3 in vitro, suggests that these sequence differences alone may be suf®cient to account for this difference in promoter strength seen in vivo. The 80-fold greater activity of rpoE P1 could simply re¯ect the existence of a better ÿ10 hexamer, since mutations that bring cycA P3 closer to rpoE P1 increase promoter activity by

Figure 7. ChrR co-puri®es with His-sE. Western blot analysis of fractions obtained from af®nity puri®cation of His-sE on a Ni2‡ agarose column. Crude extracts expressing His-sE and ChrR (lane 1), sE and ChrR (lane 2) or His-sE and C38R (lane 3) were applied to separate Ni2‡ agarose columns, washed, and bound protein was eluted as described in the Materials and Methods. Elution fractions from each column were separated by SDS-PAGE prior to Western blotting using antibodies speci®c to sE (top) or ChrR (bottom).

eight- to tenfold each (Figure 4(a)). However, sequences outside these hexamers might also contribute to sE-dependent transcription, as we found the addition of sequences downstream of ‡1 increased cycA P3 activity in vitro. It is also possible that an activator may stimulate EsE function in vivo in a manner similar to CpxA activation of E. coli sE-dependent promoters (Danese et al., 1995; Missiakas & Raina, 1997). Our analysis of mutant cycA P3 promoters in vivo generally agreed with data from in vitro experiments. However, there were a few instances where mutations that produced large effects in vitro lead to only modest changes in vivo. As transcription initiation is known to be a multi-step process, the effect of mutations that signi®cantly increase cycA P3 strength in vitro could be masked by ratelimiting steps that exist in vivo (Record et al., 1996). ChrR is a negative regulator of s E activity Many ECF family members are co-transcribed with anti-sigma factors that control their activity (Brown & Hughes, 1995). We have shown that ChrR inhibits sE activity in vivo, and co-puri®es with His-sE when expressed in a heterologous host, suggesting that this protein is an anti-sigma factor. Database searches indicate that ChrR is not homologous to other known anti-sigma factors and shares only limited amino acid similarity to the Nterminal region of an open reading frame downstream of Mycobacterium tuberculosis SigE (Accession no. U87242). Interestingly, ChrR and the putative mycobacterial protein both contain the Nterminal cysteine-cluster region that is altered by point mutations in cohemin resistant strains (Figure 1). We found that two of the three characterized point mutations (Schilke & Donohue, 1992), chr-4 and chr-9, lead to the C38R substitution (Figure 1), while the third, chr-7, produces a serine to proline substitution at residue 32 (S32P). It remains to be seen whether the amino acid sequence similarity in this region of these two potential anti-sigma factors is indicative of a similar signaling pathway or mechanism of regulation. Since the submission of this paper, the anti-sigma factor RsrA in Streptomyces coelicolor has been shown to regulate the activity of sR through the reversible oxidation of a similar cysteine cluster (Kang et al., 1999). Our data show that loss of function mutations in chrR increase sE activity and are suf®cient to account for the increase in transcription from cycA

316 P3 and rpoE P1 that is seen in cohemin-resistant cells. Previous work also reported that cycA P1 transcription was increased in cohemin-resistant cells that have lost ChrR activity (Schilke & Donohue, 1992). Since EsE fails to transcribe cycA P1 in vitro, this increase is probably an indirect effect of inactivating ChrR. For example, we know cycA P1 is a heat-shock promoter (Karls et al., 1998), and thus it is possible that one or both of the R. sphaeroides heat-shock sigma factors contains a sE-dependent promoter much like the case in E. coli (Erickson & Gross, 1989). Alternatively, it is possible that, the accumulation of inactive ChrR proteins like C38R increases cycA P1 activity by triggering a heat-shock response. Of equal signi®cance, cohemin resistance is observed in both
ChrR and Chr4 strains, which have increased sE activity, while wild-type and
rpoEchrR cells are sensitive to cohemin and have little or no sE activity. Thus the increase in sE activity that accompanies loss of ChrR is suf®cient to confer cohemin resistance. Our ®nding that exogenous heme relieves cohemin toxicity supports the conclusion that cohemin either substitutes for heme in hemoproteins, or acts as a feedback inhibitor of tetrapyrrole biosynthesis (Schilke & Donohue, 1992). Since loss of ChrR activity also leads to elevated levels of heme biosynthetic enzymes (Schilke & Donohue, 1992), it appears that mutations increasing sE activity may compensate for heme limitation by increasing tetrapyrrole biosynthesis. Possible roles of R. sphaeroides s E While R. sphaeroides sE may be classi®ed as an ECF sigma factor, its regulon and physiological role remains unknown. One of the most closely related sigma factors, M. tuberculosis SigE, has been implicated in the response to peroxides, heat, and acid (Wu et al., 1997). Considering that the open reading frame downstream of M. tuberculosis SigE has some N-terminal amino acid similarity to ChrR, it is possible that activity of these two sE homologues respond to analogous signals. However, this does not appear to be the case, as we have yet to observe R. sphaeroides sE activity or the growth of
RpoE mutants to be affected by the stress signals that affect M. tuberculosis SigE (data not shown). In summary, we have shown that R. sphaeroides sE recognizes both cycA P3 and rpoE P1 and that activity of this ECF sigma factor is inhibited by ChrR. In addition, mutations that inactivate ChrR are suf®cient to account for the increased cycA P3 activity that was previously found in coheminresistant cells (Schilke & Donohue, 1992, 1995). It has been proposed that ChrR regulates gene expression in response to tetrapyrrole signals (Schilke & Donohue, 1992). This work provides a mechanism for how ChrR could regulate expression of tetrapyrrole biosynthetic enzymes or tetrapyrrole-binding proteins, such as cytochrome

R. sphaeroides E-Dependent Transcription

c2, by controlling sE activity. Experiments are in progress to determine if ChrR acts as an anti-sigma factor, identify the signals that increase sE activity, and determine if members of the R. sphaeroides regulon include other proteins involved in synthesis or binding of tetrapyrroles.

Materials and Methods Bacterial strains, plasmids, and growth conditions E. coli strains (Table 2) were grown at 37  C in LuriaBertani medium (Maniatis et al., 1989). DH5a was used as a plasmid host and S17-1 was used as a donor for conjugal DNA transfer into R. sphaeroides. R. sphaeroides strains (Table 2) were routinely grown aerobically at 30  C in Sistrom's succinate-based minimal medium A (Sistrom, 1960). Plasmids were maintained in host strains using previously described antibiotic concentrations (Schilke & Donohue, 1995).

Construction of a
ChrR mutant To produce an R. sphaeroides
ChrR mutant, an 500 bp fragment internal to chrR was replaced with a trimethoprim resistance (Tpr) gene, dfr, to create the
chrR-1::dfr allele. To interrupt chrR, an XmnI-EcoRI restriction fragment containing dfr from pTF1 was ligated into pJDN13 that was digested with EcoRI and KpnI (®lled in with phage T4 DNA polymerase). The resulting plasmid (pJDN25I) was digested with EcoRI (®lled in as above) and HindIII, and the restriction fragment was ligated into the EcoRV and HindIII sites of pBS15, replacing wild-type chrR with the
chrR-1::dfr allele. This plasmid (pJDN26) was partially digested with PstI and EcoRI, a 2.4 kb fragment containing
chrR-1:: dfr was puri®ed, and ligated into the same restriction sites in the mobilizable suicide vector, pSUP202 (Allen & Hanson, 1985). The resulting plasmid (pJDN27) was used to place a
chrR-1::dfr allele in R. sphaeroides by homologous recombination. Screening Tpr cells (30 mg/ml) for those sensitive to tetracycline (1 mg/ml) identi®ed a strain (
ChrR) where the
chrR1::dfr allele replaced a wild-type gene. The genotype was con®rmed by PCR ampli®cation of the interrupted gene and sequencing of the PCR product. We also con®rmed that ChrR was not being accumulated in the
ChrR strain by Western blot using antibodies against ChrR.

Purification and amino acid sequencing of ChrR-His R. sphaeroides TF18 (pJDN18) was used as a source of C-terminal hexa-histidine tagged ChrR (ChrR-His). Cells were grown to mid-log phase, harvested by centrifugation (6000 g for 15 minutes) and lysed by sonication (Barber & Donohue, 1998). This crude cell extract was solublized in 6 M urea and ChrR-His was enriched by nickel af®nity chromatography under denaturing conditions. The eluted proteins were electrophoresed on a SDS-12 % PAGE gel and transferred to a PVDF membrane. A protein identi®ed by Western blot as ChrR was used for N-terminal sequencing (Medical College of Wisconsin, Milwaukee, WI).

317

R. sphaeroides E-Dependent Transcription Table 2. Bacterial strains and plasmids Strains or plasmid

Description

Source or reference

A. R. sphaeroides 2.4.1 Wild-type
ChrR
chrR mutation (
chrR-1::dfr) in 2.4.1. TF18
rpoEchrR mutation in 2.4.1. Chr4 chr-4 allele that produces a C38R substitution in ChrR

Laboratory strain This work Schilke & Donohue (1995) Schilke & Donohue (1995)

B. E. coli DH5a S17-1 M15

supE44 lacU169 ( 80 lacZ M15) hsdR178 recA1 endA1 gyrA96 thi-1 relA-1 C600::RP-4 2-(Tc::Mu) (Kn::Tn7) thi pro hsdR hsdM‡ recA lacÿ araÿ galÿ mtlÿ Fÿ recA‡ uvr‡

BRL (1986) Simon et al. (1983) Villarejo & Zabin (1974)

lacIq; Knr N-terminal His-tag expression vector; Apr Apr Apr Apr Suicide plasmid; Apr, Tcr, Cmr, Mob‡ pBR325 derivative Shuttle vector for expression in R. sphaeroides. RK2 replicon, Mob‡;Tcr Vector for in vitro transcription. Filled in SpeI fragment containing Spr gene and pUC21 polylinker ligated in filled HindIII site of pUC19spf0 . Spr gene in opposite orientation as the lac promoter; Apr, Spr Vector for in vitro transcription; Apr KpnI-StuI cycA promoter fragment encompassing cycA P3 and cycA P1 cloned into pRKK137; Apr lacZ-fusion reporter vector; Knr, Spr 3.7 kb SalI restriction fragment containing the rpoEchrR operon in pUC19; Apr 1.8 kb SalI-SacI fragment containing the rpoEchrR operon in pUC19; Apr pBS16 with chr-4 allele of chrR; Apr Tpr, Apr Contains malE gene; Apr Filled in 1.3 kb AseI- XmnI malE restriction fragment from pMal c-2 cloned into filled in StuI-SfiI restriction sites of pBS16. Replaces N-terminal 28 amino acid residues of RpoE with in-frame fusion to MalE; inactivates sE and leaves ChrR intact 3.1 kb HindIII-EcoRI restriction fragment from pJDN6 cloned into HindIII-EcoRI restriction sites of pRK415 pBS16 with DNA coding for ChrR-His fusion. Constructed by ligating double-stranded oligonucleotides coding for His into the BsgI restriction site of pBS16 64 bp restriction fragment (made from overlapping oligonucleotides) encompassing cycA P3 from ÿ53 to ‡11 relative to transcription start, cloned into BglII-StuI restriction sites of pRKK96 114 bp KpnI-HindII restriction fragment containing cycA P3 from pJDN11 ligated into KpnI-StuI restriction sites of pRKK200 960 bp PstI-BanI (filled in) restriction fragment from pBS16 containing rpoE cloned into PstI-SmaI restriction sites of pUC19 533 bp HaeII(filled in)-SacI fragment encompassing rpoE from pJDN13 cloned into BamHI (filled in)-SacI restriction sites of pQE31 to create in frame His-sE fusion 1 kb HindIII-EcoRI restriction fragment from pJDN13 containing rpoE cloned into HindIII-EcoRI restriction sites of pRK415 1.8 Kb rpoEchrR-His EcoRI-HindIII restriction fragment of pJDN10 cloned into HindIII-EcoRI restriction sites of pRK415 401 bp HindIII-StuI promoter fragment of pBS16 cloned into HindIII-HincII restriction sites of pUC21 330 bp HindIII-HinfI (filled in) promoter fragment of pJDN20 in HindIII-HincII restriction sites of pUC21 XmnI-EcoRI restriction fragment encompassing the Tpr gene from pTF1 ligated into EcoRI-KpnI (filled in) restriction sites of pJDN13 EcoRI (filled in)-HindIII restriction fragment from pJDN25I containing rpoE and Tpr in the EcoRV-HindIII restriction sites of pBS15. Creates insertion of Tpr gene into a deletion of chrR gene 2.4 kb PstI-EcoRI restriction fragment from pJDN26 cloned into PstI-EcoRI restriction sites of the suicide plasmid, pSUP202 ÿ39 to ‡17 rpoE P1 promoter fragment generated from PCR of pJDN23 using the JG3 and 1212 primers. PCR product was digested with KpnI-EcoRV and cloned into KpnIStuI sites of pRKK200 ÿ39 to ‡17 rpoE P1 promoter fragment generated from PCR of pJDN23 using JG3 and 1212 primers. PCR product was cut with KpnI-EcoRV and cloned into KpnI-StuI restriction sites of pRKK96 pBS16 with the DNA sequence ATGCACCATCACCATCACCAT inserted at the beginning of the sE coding sequence to make a His-sE fusion. Constructed by PCR using the HISRPOE1 and PBS16 406L24 primers; Apr pJDN40 with the chr-4 allele of chrR. 175 bp StuI-SfiI fragment from pJDN40 ligated into  4.3 kb pBS16c4 StuI-SfiI fragment; Apr

Qiagen (1992) Qiagen (1992) Yanisch-Perron et al. (1985) Vieira & Messing (1991) Erickson & Gross (1989) Simon et al. (1983) Ditta et al. (1985) R. Karls; This work

C. Plasmid pREP4 pQE31 pUC19 pUC21 pUC19spf0 pSUP202 pRK415 pRKK96 pRKK137 pRKK139 pRKK200 pBS15 pBS16 pBS16c4 pTF1 pMal c-2 pJDN6 pJDN8 pJDN10 pJDN11 pJDN11z pJDN13 pJDN14 pJDN16 pJDN18 pJDN20 pJDN23 pJDN25I pJDN26 pJDN27 pJDN30 pJDN34 pJDN40 pJDN41

MacGregor et al. (1998) MacGregor et al. (1998) Karls et al. (1999) Schilke & Donohue (1995) Schilke & Donohue (1995) Schilke & Donohue (1995) Schilke & Donohue (1995) New England Biolabs This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work

318 Cohemin sensitivity assay Cells were grown aerobically to mid-log phase and dilutions plated on medium containing 30 mg/ml cohemin (Porphyrin Products, Logan, Utah) with or without 30 mg/ml heme (Sigma, St. Louis, Missouri) to assess viability. Heme and cohemin stock solutions were made in 50 % DMSO at 150 mg/ml and converted to the free base with Tris buffer (pH 9.5) before dilution into growth medium. Construction of the cycA P3 mutant promoter bank To create a bank of 36 mutant cycA P3 promoters with single base mutations in the putative ÿ35 and ÿ10 hexamers, a series of oligonucleotides extending from ÿ53 to ‡11 relative to the transcription initiation site were designed to be complementary over 12 base-pairs at their respective 30 ends and contain a BamHI restriction site at ÿ53. In one set of experiments, six different 43 bp oligonucleotides (Genosys, The Woodlands, TX) each including a different randomized position of the cycA P3 ÿ35 hexamer, were annealed to a complementary 37 bp oligonucleotide containing a wild-type ÿ10 hexamer (sequence available upon request). The annealed product was extended using phage T4 DNA polymerase to yield a 68 bp double-stranded DNA fragment that was phosphorylated, digested with BamHI, ligated to BglII - StuI digested pRKK96, and transformed into E. coli DH5a. This procedure was modi®ed to introduce changes in the putative ÿ10 hexamer, using an oligonucleotide containing a wild-type ÿ35 hexamer and a complementary set of oligonucleotides with randomized nucleotides in the ÿ10 hexamer. For in vitro transcription assays, each pRKK96 derivative was puri®ed using Qiagen midiprep Tip100 DNA isolation kits (Qiagen Inc., Chatsworth, CA) followed by two extractions with phenol:chloroform:isoamyl alcohol (25:24:1) and one with chloroform:isoamyl alcohol (24:1). To create a mutant promoter bank for in vivo analysis, a set of cycA P3 restriction fragments from the pRKK96 derivatives were isolated by digesting with KpnI and HincII and ligated into the KpnI and StuI sites of pRKK200 to orient the promoters upstream of lacZYA. DNA sequence of the cycA P3 region in the pRKK96 and pRKK200 derivatives was con®rmed by ABI dye-terminator dideoxy-sequencing (UW-Madison Biotechnology Center). Core RNA polymerase and His-s sE purification Core RNA polymerase was puri®ed from an R. sphaeroides
RpoH strain (Karls et al., 1998) as described (MacGregor et al., 1998) using af®nity chromatography and the 4RA2 antibody directed against the a-subunit of RNA polymerase (kindly provided by N. Thompson and R. Burgess). Amino-terminal hexa-histidine tagged R. sphaeroides RpoE (His-sE) was obtained by cloning the rpoE gene into pQE31 (Qiagen Inc., Chatsworth, CA). Over-expression of His-sE in E. coli led to localization of the recombinant protein into inclusion bodies that were solublized in 6 M urea, 50 mM Tris-HCl (pH 8.0), prior to puri®cation by nickel af®nity chromatography under denaturing conditions (Qiagen Inc., Chatsworth, CA). His-sE, >95 % pure as measured by densitometry of protein analyzed by SDS-PAGE, was diluted to a concentration of 4 mg/ml and renatured by dialysis against 100 volumes of a series of three buffers each containing decreasing concentrations of urea (3, 1.5, or 0.75 M) in

R. sphaeroides E-Dependent Transcription 50 mM Tris-HCl (pH 7.9) 1 mM EDTA, 1 mM DTT, 20 % (v/v) glycerol. Prior to storage at ÿ20  C, the sample was dialyzed against two, 400 volumes of the above buffer containing 50 % glycerol. In vitro transcription assays To reconstitute EsE, 0.4 pmol core RNA polymerase and 0.8 pmol His-sE were incubated together at 30o C for 20 minutes. EsE was then added to a 20 ml reaction containing 20 nM supercoiled template in transcription buffer (125 mM KCl, 10 mM magnesium acetate, 40 mM Tris-HCl (pH 7.9), 1 mM DTT, 62.5 mg/ml acetylated bovine serum albumin) and incubated for ten minutes at 30  C. To initiate transcription, ribonucleotides were added to a ®nal concentration of 250 mM GTP, ATP, CTP; 125 mM UTP; and 2 mCi [a-32P]UTP (3000 Ci/ mmol). Samples were incubated at 30  C for 30 minutes, and reactions terminated by the addition of Sequenase stop solution (US Biochemical Corp, Cleveland, Ohio). RNA products were electrophoresed with known DNA sequence ladders and quanti®ed as described (MacGregor et al., 1998). b -Galactosidase assays b-Galactosidase activity assays were performed in triplicate as described (Schilke & Donohue, 1992). All replicates were within 15 % of the average standard error. Co-purification of His-s sE and ChrR and Western blot analysis M15 pREP4 (Table 2) was used to express His-sE and ChrR, sE and ChrR, or His-sE and C38R, from pJDN40, pBS16, or pJDN41, respectively. Cells were grown to mid-exponential phase, harvested by centrifugation, and resuspended in 1 ml PBS (140 mM NaCl, 5 mM KCl, 12 mM phosphate, pH 7.4). Cells were lysed by sonication and extracts clari®ed by centrifugation at 4  C (20,000 g for 30 minutes). Extracts were incubated with 100 ml of pre-equilibrated Ni2‡ agarose (Qiagen Inc., Chatsworth, CA) for one hour prior to loading the slurry to a disposable column (I.D. ˆ 0.5 cm). The column was washed with 100 ml of PBS containing 10 mM imidazole and bound protein was eluted with PBS containing 400 mM imidazole. Western blots were performed using speci®c rabbit polyclonal antibodies raised against His-ChrR and HissE isolated from inclusion bodies as outlined above. Control experiments show that antibodies against HisChrR did not to cross-react with His-sE and vice versa (data not shown). Western blots were performed (Maniatis et al., 1989) using primary antibodies diluted 1:2000 and a 1:20,000 dilution of secondary antibodies (goat anti-rabbit immunoglobulin conjugated to horseradish peroxidase; Pierce; Rockford, IL).

Acknowledgments This work was supported by grant GM37509 from the National Institute of Health to T.J.D. Early in this project, J.D.N. was supported by a National Institute of Health Biotechnology Predoctoral Training grant (GM08349) to the University of Wisconsin-Madison. We also thank

R. sphaeroides E-Dependent Transcription Russell Karls, Tamas Gaal, Jeff Gralnick, and Jen Wolf for their help.

References Allen, L. N. & Hanson, R. S. (1985). Construction of broad-host-range cosmid cloning vectors: identi®cation of genes necessary for growth of Methylobacterium organophilum on methanol. J. Bacteriol. 161, 955-962. Barber, R. D. & Donohue, T. J. (1998). Function of a glutathione-dependent formaldehyde dehydrogenase in Rhodobacter sphaeroides formaldehyde oxidation and assimilation. Biochemistry, 37, 530-537. Brown, K. L. & Hughes, K. T. (1995). The role of antisigma factors in gene regulation. Mol. Microbiol. 16, 397-404. Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S. V., Eiglmeier, K., Gas, S., Barry, C. E. 3rd, Tekaia, F., Badcock, K., Basham, D., Brown, D. & Chillingworth, T., et al. (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature, 393, 537-544. Danese, P. N., Snyder, W. B., Cosma, C. L., Davis, L. J. B. & Silhavy, T. J. (1995). The Cpx two-component signal transduction pathway of Escherichia coli regulates transcription of the gene specifying the stressinducible periplasmic protease, DegP. Genes Dev. 9, 387-398. De Las, Penas A., Connolly, A. & Gross, C. A. (1997). The sE-mediated response to extracytoplasmic stress in Escherichia coli is transduced by RseA and RseB, two negative regulators of sE. Mol. Microbiol. 24, 373-385. Ditta, G., Schmidhauser, T., Yakobson, E., Lu, P., Liang, X.-W., Finlay, D., Guiney, D. & Helinski, D. (1985). Plasmids related to the broad host range vector, pRK290, useful for gene cloning and for monitoring gene expression. Plasmid, 13, 149-153. Erickson, J. W. & Gross, C. A. (1989). Identi®cation of the sE subunit of Escherichia coli RNA polymerase: a second alternate sigma factor involved in hightemperature gene expression. Genes Dev. 3, 14621471. Gorham, H. C., McGowan, S. J., Robson, P. R. H. & Hodgson, D. A. (1996). Light-induced carotenogenesis in Myxococcus xanthus: light-dependent membrane sequestration of ECF sigma factor CarQ by anti-sigma factor CarR. Mol. Microbiol. 19, 171-186. Hershberger, C. D., Ye, R. W., Parsek, M. R., Xie, Z. D. & Chakrabarty, A. M. (1995). The algT (algU) gene of Pseudomonas aeruginosa, a key regulator involved in alginate biosynthesis, encodes an alternative sigma factor (sE). Proc. Natl Acad. Sci. USA, 92, 7941-7945. Higgins, D. G., Thompson, J. D. & Gibson, T. J. (1996). Using CLUSTAL for multiple sequence alignments. Methods Enzymol. 266, 383-402. Kang, J. G., Paget, M. S., Seok, Y. J., Hahn, M. Y., Bae, J. B., Hahn, J. S., Kleanthous, C., Buttner, M. J. & Roe, J. H. (1999). RsrA, an anti-sigma factor regulated by redox change. EMBO J. 18, 4242-4248. Karls, R. K., Brooks, J., Rossmeissl, P., Luedke, J. & Donohue, T. J. (1998). Metabolic roles of a Rhodobacter sphaeroides member of the s32 family. J. Bacteriol. 180, 10-19.

319 Karls, R. K., Wolf, J. R. & Donohue, T. J. (1999). Activation of the cycA P2 promoter for the Rhodobacter sphaeroides cytochrome c2 gene by the photosynthesis response regulator. Mol. Microbiol. 34, 822835. Ladan, H., Nitzan, Y. & Malik, Z. (1993). The antibacterial activity of haemin compared with cobalt, zinc and magnesium protoporphyrin and its effect on potassium loss and ultrastructure of Staphylococcus aureus. FEMS Microbiol. Letters, 112, 173-177. Lonetto, M. A., Brown, K. L., Rudd, K. E. & Buttner, M. J. (1994). Analysis of the Streptomyces coelicolor sigE gene reveals the existence of a subfamily of eubacterial RNA polymerase sigma factors involved in the regulation of extracytoplasmic functions. Proc. Natl Acad. Sci. USA, 91, 7573-7577. MacGregor, B. J., Karls, R. K. & Donohue, T. J. (1998). Transcription of the Rhodobacter sphaeroides cycA P1 promoter by alternate RNA polymerase holoenzymes. J. Bacteriol. 180, 1-9. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1989). Molecular Cloning: A Laboratory Manual, 2nd edit., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Mecsas, J., Rouviere, P. E., Erickson, J. W., Donohue, T. J. & Gross, C. A. (1993). The activity of sE, an Escherichia coli heat-inducible sigma factor, is modulated by expression of outer membrane proteins. Genes Dev. 7, 2618-2628. Missiakas, D. & Raina, S. (1997). Signal transduction pathways in response to protein misfolding in the extracytoplasmic compartments of E. coli: role of two new phosphoprotein phosphatases PrpA and PrpB. EMBO J. 16, 1670-1685. Missiakas, D. & Raina, S. (1998). The extracytoplasmic function sigma factors: role and regulation. Mol. Microbiol, 28, 1059-1066. Nitzan, Y., Wexler, H. M. & Finegold, S. M. (1994). Inactivation of anaerobic bacteria by various photosensitized porphyrins or by hemin. Curr. Microbiol. 29, 125-131. Paget, M. S., Kang, J. G., Roe, J. H. & Buttner, M. J. (1998). sR, an RNA polymerase sigma factor that modulates expression of the thioredoxin system in response to oxidative stress in Streptomyces coelicolor A3(2). EMBO J. 17, 5776-5782. Record, M., Reznikoff, W., Craig, M., McQuade, K. & Schlax, P. (1996). Esherichia coli RNA polymerase, promoters, and the kinetics of the steps of transcription initiation. In Esherichia coli and Salmonella: Cellular and Molecular Biology (Neidhardt, F., ed.), 2nd edit., vol. 1, pp. 792-820, ASM Press, Washington, DC. Schilke, B. A. & Donohue, T. J. (1992). d-Aminolevulinate couples cycA transcription to changes in heme availability in Rhodobacter sphaeroides. J. Mol. Biol. 226, 101-115. Schilke, B. A. & Donohue, T. J. (1995). ChrR positively regulates transcription of the Rhodobacter sphaeroides cytochrome c2 gene. J. Bacteriol. 177, 1929-1937. Schurr, M. J., Yu, H., Martinez, Salazar J. M., Hibler, N. S. & Deretic, V. (1995). Biochemical characterization and posttranslational modi®cation of AlgU, a regulator of stress response in Pseudomonas aeruginosa. Biochem. Biophys. Res. Commun. 216, 874-880. Schurr, M. J., Yu, H., Marteniz-Salazar, J. M., Boucher, J. C. & Deretic, V. (1996). Control of AlgU, a member of the sE-like family of stress sigma factors, by the negative regulators MucA and MucB and Pseu-

320

R. sphaeroides E-Dependent Transcription

domonas aeroginosa conversion to mucoidy in cystic ®brosis. J. Bacteriol. 178, 4997-5004. Simon, R., Priefer, U. & PuÈhler, A. (1983). A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in Gram-negative bacteria. Biotechnology, 1, 784-791. Sistrom, W. R. (1960). A requirement for sodium in the growth of Rhodopseudomonas spheroides. J. Gen. Microbiol. 22, 778-785. Vieira, J. & Messing, J. (1991). New pUC-derived cloning vectors with different selectable markers and DNA replication origins. Gene, 100, 189-194.

Villarejo, M. R. & Zabin, I. (1974). b-Galactosidase from termination and deletion mutant strains. J. Bacteriol. 120, 466-474. Wosten, M. (1998). Eubacterial sigma-factors. FEMS Microbiol. Rev. 22, 127-150. Wu, Q. L., Kong, D., Lam, K. & Husson, R. N. (1997). A mycobacterial extracytoplasmic function sigma factor involved in survival following stress. J. Bacteriol. 179, 2922-2929. Yanisch-Perron, C., Vieira, J. & Messing, J. (1985). Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene, 33, 103-119.

Edited by M. Gottesman (Received 2 June 1999; received in revised form 22 September 1999; accepted 30 September 1999)