Mutagenesis of Region 4 of Sigma 28 from Chlamydia trachomatis Defines Determinants for Protein-Protein and Protein-DNA Interactions

Abstract

Transcription factor σ²⁸ in Chlamydia trachomatis (σ²⁸_Ct) plays a role in the regulation of genes that are important for late-stage morphological differentiation. In vitro mutational and genetic screening in Salmonella enterica serovar Typhimurium was performed in order to identify mutants with mutations in region 4 of σ²⁸_Ct that were defective in σ²⁸-specific transcription. Specially, the previously undefined but important interactions between σ²⁸_Ct region 4 and the flap domain of the RNA polymerase β subunit (β-flap) or the −35 element of the chlamydial hctB promoter were examined. Our results indicate that amino acid residues E206, Y214, and E222 of σ²⁸_Ct contribute to an interaction with the β-flap when σ²⁸_Ct associates with the core RNA polymerase. These residues function in contacts with the β-flap similarly to their counterpart residues in Escherichia coli σ⁷⁰. Conversely, residue Q236 of σ²⁸_Ct directly binds the chlamydial hctB −35 element. The conserved counterpart residue in E. coli σ⁷⁰ has not been reported to interact with the −35 element of the σ⁷⁰ promoter. Observed functional disparity between σ²⁸_Ct and σ⁷⁰ region 4 is consistent with their divergent properties in promoter recognition. This work provides new insight into understanding the molecular basis of gene regulation controlled by σ²⁸_Ct in C. trachomatis.

Chlamydia trachomatis is a leading causative pathogen of bacterial sexually transmitted diseases and ocular infections (trachoma) in humans (36). The obligate intracellular parasitic feature of Chlamydia has hindered genetic and biochemical studies. As a result, little is known about how gene expression is regulated in Chlamydia. Recent studies indicate that gene expression takes place coordinately at the level of transcription throughout the developmental cycle of Chlamydia (2, 33, 37). Bacterial RNA polymerase (RNAP) is a key enzyme that controls transcription (15). Chlamydial RNAP is identical to the well-studied Escherichia coli RNAP in containing subunits α₂ββ′ and a σ factor, despite the lack of a recognizable ω subunit (43). The σ factors associate with the RNAP core to form an RNAP holoenzyme that initiates promoter-specific transcription (15, 19, 46). Three known σ factors, σ⁶⁶, σ⁵⁴, and σ²⁸, and some transcription regulators have been revealed in the C. trachomatis genome (43). Both chlamydial σ⁶⁶ and C. trachomatis σ²⁸ (σ²⁸_Ct), but not σ⁵⁴, belong to members of the E. coli σ⁷⁰ family that share up to four conserved amino acid regions (regions 1 to 4) (26). Chlamydial σ⁶⁶ serves as the primary σ factor that directs transcription of housekeeping genes or most constitutively expressed genes (2, 7, 23, 27, 33), whereas σ²⁸_Ct directs transcription of several late-stage genes that are required for differentiation from the noninfectious but metabolically active reticulate bodies to the infectious but metabolically inactive elementary bodies (3, 39, 48). σ²⁸_Ct also performs its task of transcription by responding to stressful conditions (39, 50). It is not known how the σ²⁸_Ct function is regulated in Chlamydia. C. trachomatis does not appear to encode an identifiable anti-σ²⁸ homologue, which is known as FlgM in Salmonella (18, 22). A putative chlamydial RsbW, encoded by the rsbW gene, was shown to bind σ²⁸_Ct in a glutathione S-transferase pull-down assay (A. L. Douglas and T. P. Hatch, personnel communication). However, there is no direct evidence to show an inhibitory effect of RsbW on chlamydial σ²⁸_Ct-specific transcription (17, 22). Even less is known about the function of chlamydial σ⁵⁴, and only two putative σ⁵⁴-like promoters have been reported (28). Defining the molecular basis of σ factor action is central for understanding the mechanism by which Chlamydia completes its unique intracellular developmental cycle (16) and adapts to environmental cues.

Several lines of evidence demonstrate that the DNA-σ and σ-core interactions are essential for the process of transcription initiation and elongation (5, 11, 30, 31, 35). Such molecular interactions are well characterized in E. coli σ⁷⁰-mediated transcription regulation (5, 11, 25). Regions 2 and 4 of E. coli σ⁷⁰ recognize the promoter −10 and −35 elements, respectively (4, 11, 34). Region 2 of σ⁷⁰ also interacts with the β′ coiled coil in core RNAP. This interplay is essential for holoenzyme formation (1, 5, 12, 34). Also, the σ⁷⁰ region 4 interacts with the flexible flap structure in the β subunit (β-flap). This interaction is required to properly position σ⁷⁰ regions 2 and 4, allowing for simultaneous contact with the −35/−10 promoter elements. Furthermore, the σ⁷⁰ region 4/β-flap interaction can be a target of transcription factors, such as the anti-σ⁷⁰ factor T4 AsiA (13, 21). By sequence analogy with E. coli σ⁷⁰, conserved region 4 in an alternative σ factor is likely to interact with the β-flap for recognition of the −35/−10 promoters. Indeed, direct interactions of region 4 of Helicobacter pylori σ²⁸ with the β-flap (6), as well as those of region 4 of E. coli σ³⁸ with the β-flap (27, 36) have been identified. An apparent difference in the strengths of the E. coli σ³⁸/β-flap and E. coli σ⁷⁰/β-flap interactions was found, although E. coli σ³⁸ and σ⁷⁰ recognize similar promoter consensus sequences (19). The question is raised of whether the variability of σ region 4 and the β-flap interaction contributes to promoter recognition in a σ-specific manner.

Previously, we have characterized the range of promoter elements that are recognized by chlamydial σ²⁸ (39, 40). This study was performed in order to explore the details of σ²⁸_Ct action in the process of transcription. Specifically, we report that certain σ²⁸_Ct determinants are required for contact with the chlamydial β-flap of RNAP or the −35 promoter element. Because σ²⁸_Ct plays a role in regulation of genes important for late-stage morphological differentiation and the stress response, our data can shed light on molecular mechanisms underlying these processes. Importantly, given the uncertainty about whether a transcription regulator would affect the σ²⁸_Ct activity, our ability to define the molecular interactions of the transcription machinery provides us with a useful genetic tool and information to facilitate the discovery and characterization of potential regulatory factors for σ²⁸_Ct in Chlamydia, a genetically intractable pathogen.

MATERIALS AND METHODS

Bacterial strains and growth.

The bacterial strains used are listed in Table 1. C. trachomatis was propagated and purified as previously described (38). DH5α or XL1-Blue was used as the host for cloning. Salmonella enterica serovar Typhimurium fliA, encoding Salmonella σ²⁸, mutant strains TH5504 and TH7034 were kindly provided by Kelly T. Hughes (University of Utah). Reporter strains BN317 (34) and KS1 (8) were kindly provided by Ann Hochschild (Harvard University). Reporter strains ZY101 containing the test promoter Plac_hctB-35 linked to lacZ on an F′ episome were constructed as described previously (45) (see Fig. 6). Briefly, the EcoRI-HindIII-digested DNA fragment carrying the synthetic promoter Plac_hctB- 35 was cloned into pFW11 via EcoRI and HindIII sites. The resultant plasmid was introduced into strain CSH100, allowing homologous recombination of Plac_hctB- 35 and lacZ onto an F′ episome. Further mating with strain FW102 was performed in order to finalize construction of strain ZY101. In this test promoter, an ectopic chlamydial hctB promoter −35 element with the sequence TAAAGTTT was centered at bp −45.5 relative to the transcription initiation site of lac. Using a similar strategy, the reporter strain ZY102 was constructed. ZY102 is identical to ZY101, except that it contains a mutated ectopic chlamydial hctB promoter −35 element (gAAAGTTT). E. coli and S. enterica serovar Typhimurium strains were grown in Luria-Bertani (LB) medium at 37°C. MacConkey agar supplemented with 1% lactose (Difco) plus 0.02% arabinose (MacConkey-lactose-arabinose) was used in indicator plates for β-galactosidase activity. When required, medium was supplemented with 25 μg/ml chloramphenicol, 50 μg/ml kanamycin, 10 μg/ml tetracycline, and/or 100 μg/ml carbencillin.

TABLE 1.

Strains and plasmids used in this study

Species and strain or plasmid	Description	Reference or source
C. trachomatis ATCC VR-346	Serovar F/IC-Cal-13	ATCC
E. coli
KS1	MC1000 F′ lacIq integrated with a single copy of promoter plac OR2-62 controlled by lacZ in the chromosome	8
ZY101	MC1000 F′ lacIq integrated with a single copy of Plac_hctB-35(TAAAGTTT)::lacZ on an F′episome	This study
ZY102	MC1000 F′ lacIq integrated with a single copy of Plac_hctB-35(gAAAGTTT)::lacZ on an F′ episome	This study
BN317	MC1000 F′ lacIq integrated with a single copy of placCons-35C::lacZ on an F′ episome	34
S. enterica serovar Typhimurium
TH7034	Isogenic with wild-type strain LT2; fliA5886 [R91C L207P] fliC5050::MudJ	K. T. Hughes
TH5504	Isogenic with wild-type strain LT2; fliA5647::FRT	K. T. Hughes
Plasmids
pFW11		45
pS28H	ParaB-His6 chlamydial fliA	39
pES28H	ParaB-His6E. coli fliA	39
pHσ28Ct-Ec4	ParaB-His6 hybrid fliA; encoding σ28Ct residues 1-178 fused to region 4 of σ28Ec residues 166 to 239	This study
pHσ28Ec-Ct4	ParaB-His6 hybrid fliA; encoding σ28Ec (residues 1-165) fused to region 4 of σ28Ct (residues 179-253)	This study
pRV5	Promoter probe vector containing less lacZ; ligation of the XmnI-SphI fragment of pQF50KgroE and the MfeI (blunted)-SphI fragment of pTara; origin of replication of p15A; Cmr	This study
pRVPhctB	pRV5 with promoter of chlamydial hctB (−125 to +45)	This study
pQF50KgroE	Promoter reporter vector containing PgroE::lacZ	24
pTara	Expression vector carrying ParaB-T7RNAP	47
pACλcI	PlacUV5-directed synthesis of the λCI protein	8
pACλcI-β-flap	PlacUV5-directed synthesis of the λCI protein fused via three alanines to residues 858-946 of the β subunit of E. coli RNAP	25
pACλcI-β-flapCt	PlacUV5-directed synthesis of the λCI protein fused via three alanines to residues 798-887 of the β subunit of C. trachomatis RNAP	This study
pBRα	PlacUV5/Plpp-directed synthesis of the α subunit of E. coli RNAP	8
pBRαLN	PlacUV5/Plpp-directed synthesis of N terminus (residues 1-248) of the α subunit of E. coli RNAP	8, 10
pBRα-σ66	PlacUV5/Plpp-directed synthesis of the αNTD fused to C. trachomatis σ66 region 4 (residues 468-571 of σ66)	This work
pBRα-σ70	PlacUV5/Plpp-directed synthesis of the αNTD fused to E. coli σ70 region 4 (residues 528-613 of σ70)	10
pBRα-σ70D581G	PlacUV5/Plpp-directed synthesis of the αNTD fused to E. coli σ70 region 4 carrying the D581G substitution.	25
pBRα-σ28	PlacUV5/Plpp-directed synthesis of the αNTD fused to chlamydial σ28Ct region 4 (residues 148-253 of σ28)	This study
pBRα-σ28E206G	pBRα-σ28 encoding the E206G substitution in the β moiety of the fusion	This study
pBRα-σ28Y214C	pBRα-σ28 encoding the Y214C substitution in the β moiety of the fusion	This study
pBRα-σ28E222G	pBRα-σ28 encoding the E222G substitution in the β moiety of the fusion	This study
pBRα-σ28Q236L	pBRα-σ28 encoding the Q236L substitution in the β moiety of the fusion	This study
pBRα-σ28L243K	pBRα-σ28 encoding the L243K substitution in the β moiety of the fusion	This study
pGPhctB	Contains a chlamydial-hctB-controlled guanine-less cassette and an E. coli fliC promoter-controlled guanine-less cassette; Apr	40
pGPhctBΔup	Derivative of pGPhctB with the AT-rich sequence upstream of the promoter −35 element deleted	40
pGS9	Derivative of pGPhctB containing a 9-bp spacer between the −35/−10 elements	40
pGS14	Derivative of pGPhctB containing a 14-bp spacer between the −35/−10 elements	40
pGPtar	Contains an E. coli tar-controlled guanine-less cassette and an E. coli fliC promoter-controlled guanine-less cassette; Apr	40
pGPtar+	Derivative of pGPtar fusing a DNA sequence upstream of PhctB to the core promoter of Ptar	40
pGHR	Contains a PhctB-controlled guanine-less cassette and a chlamydial rRNA P1 promoter-controlled guanine-less cassette	This study
pLF28	Low-copy-number expression vector carrying chlamydial fliA; Tcr	39
pLF28a	Derivative of pLF28 with a silent mutation at nucleotide position 723 to remove the natural HindIII site	This study

FIG. 6.

Bacterial one-hybrid assay: characterization of the σ²⁸_Ct substitutions that affect interaction with the hctB −35 DNA element. (A) Cartoon illustrating the test promoter with wild-type or mutant ectopic hctB −35. The ectopic hctB −35 element serves as a binding site for the tethered σ²⁸_Ct region 4 moiety. When they interact with each other, transcription from the test promoter is activated, resulting in an increase of reporter lacZ (β-galactosidase). (B) Effects of substitution in the σ moiety of the α-σ²⁸_Ct chimera on transcription from the test promoter. Strain ZY101 harboring plasmids directing the synthesis of the α-σ²⁸_Ct or derivatives was grown in the presence of different concentrations of IPTG as indicated. β-Galactosidase activity was measured in duplicate on at least three independent occasions. Values shown are the averages from one experiment. (C) Change (n-fold) in the β-galactosidase activity before and after IPTG (100 μM) induction. Shown are changes of transcription from the test promoter with wild-type hctB −35 (TAAAGTTT) (in ZY101, black bars) or mutant hctB −35 (gAAAGTTT) (in ZY102, light gray bars). ZY101 and ZY102 harboring plasmids that directed the synthesis of α-σ²⁸_Ct or derivatives were grown in the absence or presence of IPTG and assayed for β-galactosidase activity. β-Galactosidase activity was measured in duplicate on at least three independent occasions.

Plasmid construction.

The plasmids used in this study are listed in Table 1. The hybrid σ²⁸ genes (Fig. 1) were generated by overlapping PCR amplification and recloned into expression vectors pS28H and pES28H (39), creating pHσ²⁸_Ct-Ec4 and pHσ²⁸_Ec-Ct4. The hybrid σ²⁸ genes are under the control of an arabinose-inducible P_ara promoter. Plasmids pACλcI, pBRα, pBRαLN and pACλcI-β-flap (8-10, 25) were kindly provided by Ann Hochschild (Harvard University). pACλcI-β-flap_Ct directs synthesis of λcI fused to chlamydial β-flap under the control of a lacUV5 promoter. pBRα-σ²⁸ encodes the E. coli N-terminal domain of the α subunit of RNAP (αNTD) fused to the chlamydial σ²⁸ region 4 under the control of tandem lpp and IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible lacUV5 promoters. pBRα-σ⁶⁶ encodes αNTD fused to region 4 of chlamydial σ⁶⁶. Plasmid pGHR, carrying promoters of chlamydial hctB and the rRNA, was created by inserting the hctB promoter region into XbaI-EcoRV-digested pMT504 (44). pGHR together with pGS₉, pGP_hctB, pGS₁₄, pGP_hctBΔ_up, pGP_tar, or pGP_tar+ (40) was used as a supercoiled template in the in vitro transcription assay. Inserts in all constructs were confirmed by restriction mapping and DNA sequencing. Details of primers and procedures for these constructs are available upon request.

FIG. 1.

In vitro promoter selectivity of holoenzymes containing hybrid σ²⁸ proteins. (A) Construction of hybrid σ²⁸, showing the organization of σ²⁸_Ct, σ²⁸_Ec, σ²⁸_Ec-Ct4 (σ²⁸_Ec residues 1 to 165 fused to region 4 of the σ²⁸_Ct [residues 179-253]), and σ²⁸_Ct-Ec4 (σ²⁸_Ct residues 1 to 178 fused to region 4 of σ²⁸_Ec [residues 166 to 239]). (B) Gel analysis of in vitro transcription products using the RNAP holoenzyme containing σ²⁸_Ct or hybrid σ²⁸_Ec-Ct4 protein. (C) Gel analysis of in vitro transcription products using the RNAP holoenzyme containing σ²⁸_Ec or hybrid σ²⁸_Ct-Ec4 protein. RNAP is designated by the σ²⁸ proteins as indicated on the left. Above each lane is shown the test promoter used in the transcription assay. On the right side of each panel are shown transcripts from different test promoters. P_fliC serves as an internal control in the same reaction.

Mutant σ²⁸_Ct library and genetic screening.

For convenience in cloning, a silent mutation (T to A) was introduced at position 723 of the σ²⁸_Ct coding region using PCR-based site-directed mutagenesis in order to remove a natural HindIII site in pLF28 (40). The resultant plasmid was named pLF28a. Error-prone PCR was performed with Taq DNA polymerase (New England Biolabs) for introducing random mutations of the σ²⁸_Ct gene into appropriate fragments of pLF28a. The mutagenized PCR fragments were digested with HindIII-SphI, and a 187-bp DNA fragment containing region 4 of σ²⁸_Ct was ligated with the HindIII-SphI-digested fragment of pLF28a. DH5α was transformed with the ligation mixture in order to generate a mutant library of σ²⁸_Ct region 4. This mutant library was then subjected to a genetic screen in S. enterica serovar Typhimurium fliA mutant strain TH7034. The selection strains formed white or pink colonies on MacConkey-lactose-arabinose indicator plates when transformed with mutant σ²⁸_Ct (the wild type was red). The plasmids carrying potential mutations were analyzed by restriction enzyme digestion, and the full-length σ²⁸_Ct gene was sequenced in order to confirm the presence of mutations.

Expression and purification of wild-type σ²⁸_Ct and derivatives.

For the production of N-terminally His₆-tagged σ²⁸_Ct and derivatives, pS28H and derived plasmids were introduced into S. enterica serovar Typhimurium strain TH7034. Cells were grown in LB containing carbenicillin. The fusion proteins were induced by treatment with 0.02% arabinose for 2 h. Proteins were purified under denaturing conditions on an Ni-nitrilotriacetic column (Qiagen), followed by a HiTrap HP column (GE Healthcare), as described previously (42). Purified proteins were renatured by serial dialysis. The final dialysis buffer contained 50% glycerol, 50 mM Tris-HCl (pH 7.9), 0.01% (vol/vol) Triton X-100, 0.1 mM EDTA, 150 mM NaCl, and 0.1 mM dithiothreitol. Protein was stored at −20°C for future use. Protein concentrations were determined using a protein assay kit (USB Corporation) and confirmed by Western blot analysis.

In vivo assay of chlamydial hctB promoter activity in Salmonella: relative β-galactosidase activity.

To examine the effect of mutant σ²⁸_Ct proteins on transcription from the chlamydial hctB promoter, pS28H or derived plasmids were introduced into Salmonella fliA mutant strain TH5504 cells harboring pRV_hctB, which carries the reporter hctB::lacZ. Cells were grown in LB containing carbencillin and chloramphenicol to ensure selection of both plasmids. The protein was induced by the addition of 0.02% (wt/vol) arabinose starting from an optical density at 600 nm of ≈0.3 for 1 hour. Aliquots of cells were collected. Cell lysates in sodium dodecyl sulfate (SDS) loading buffer were electrophoresed on a 10% glycine-SDS-polyacrylamide gel, followed by Western blot analysis. Levels of cellular σ²⁸_Ct and RNAP β subunit were measured by Western blot analysis with polyclonal anti-σ²⁸_Ct antibody (a gift from Thomas P. Hatch, University of Tennessee) and monoclonal antibody 8RB13 against the β subunit of RNAP (NeoClone) as probes, respectively. The β-subunit band was used as an internal standard for correcting the amount of protein loading. The activity of β-galactosidase was measured as described previously (29) and then normalized to the abundance of cellular σ²⁸_Ct protein. The resultant relative β-galactosidase activity was used to evaluate σ²⁸_Ct function.

Core-binding assays: coimmunoprecipitation.

To test the in vivo core-σ binding, cellular lysates of Salmonella fliA mutant TH7034 expressing σ²⁸_Ct (from pLF28 or its derivatives) in buffer (10 mM Tris [pH 7.5], 150 mM NaCl) were used for coimmunoprecipitation. To test core-σ binding in vitro, a reaction mixture (15 μl) containing purified σ²⁸ proteins (4 pmol) and E. coli core RNAP (1 pmol) (Epicenter) in buffer A (10 mM Tris-Cl [pH 7.5], 1 mM β-mercaptoethanol, 0.3 M NaCl, 10 mM MgCl₂, 0.1 mM EDTA, 0.01% Triton 100, 250 μg/ml bovine serum albumin) was used. Monoclonal antibody B8R13 (1 μl) against the β subunit was incubated with cell lysate or a mixture of σ²⁸ and core at 4°C overnight, and then complexes of core and σ were purified using protein A-agarose (Sigma). After three time washes with buffer A to remove unbound σ factor, the bound proteins were separated on a 10% SDS-polyacrylamide gel, followed by Western blot analysis. Bands of the bound σ²⁸_Ct and β subunit were probed with polyclonal σ²⁸_Ct antibody and β antibody (8RB13), respectively. Protein bands were quantified with Quantity One software (Bio-Rad).

In vitro transcription assays.

The σ²⁸RNAP holoenzyme was formed by incubation of a threefold excess of purified His₆-tagged σ²⁸ protein or derivatives with E. coli core RNAP enzyme (Epicenter) on ice for 15 min. In some cases, the ratio of mutated σ²⁸ to core enzyme was increased in order to make σ²⁸-saturated RNAP. σ⁷⁰RNAP holoenzyme was purchased from Epicenter. The in vitro transcription system contained RNAP holoenzyme, supercoiled plasmids (1 μg), 10 mM Tris-HCl (pH 8.0), 200 mM NaCl, and 5 mM dithiothreitol, and transcription was initiated by adding 400 μM ATP, 400 μM UTP, 1.2 μM CTP, 0.20 μM [α-³²P]CTP (3,000 μCi/mmol), and 100 μM 3′-O-methylguanosine 5′-triphosphate (GE Healthcare). The reaction was performed as described previously (39). mRNA made by the RNAP holoenzyme was separated and detected on a 6% (wt/vol) polyacrylamide-8 M urea polyacrylamide gel electrophoresis. Signal intensities from autoradiographs were determined with Quantity One software (Bio-Rad).

RESULTS

Replacement of σ²⁸_Ct region 4 with σ²⁸_Ec region 4 changes the behavior of RNAP holoenzyme in vitro.

Previously, we found that σ²⁸_CtRNAP is permissive in recognizing promoters with altered length between the promoter −35 and −10 elements, and preferentially activates promoters with upstream AT-rich sequences; in contrast the σ²⁸RNAP of E. coli (σ²⁸_EcRNAP) does not have such a preference (40). In order to test whether differences in region 4 of σ²⁸ were related to the observed disparity, we assessed the complementary functionality of σ²⁸_Ct region 4 by making a reciprocal pair of hybrid σ²⁸ proteins. Specifically, region 4 of σ²⁸_Ec was exchanged with the relevant region 4 of σ²⁸_Ct (Fig. 1A). S. enterica serovar Typhimurium fliA mutant strain TH7034 carrying the constructs directing the synthesis of the hybrid σ²⁸ (σ²⁸_Ec-Ct4 and σ²⁸_Ct-Ec4) appeared as red colonies on MacConkey-lactose-arabinose indicator plates, suggesting that both σ²⁸_Ec-Ct4 and σ²⁸_Ct-Ec4 transcribe from the same fliC promoter that was recognized by wild-type σ²⁸_Ct (23, 40).

We next tested the in vitro activities of purified σ²⁸_Ec-Ct4 or σ²⁸_Ct-Ec4 in the context of a holoenzyme with several defined σ²⁸ promoters. We chose the chlamydial hctB promoter (P_hctB) (native AT-rich upstream sequence plus core promoter of hctB), chlamydial hctB promoters with spacer lengths of 9 or 14 bp between the −35 and −10 element (P_S9 and P_S14), the core promoter of hctB (P_hctBΔ_up), and the E. coli tar promoter (P_tar) (native GC-rich upstream sequence plus core promoter of tar) and its derivative P_tar+ (AT-rich upstream sequence of hctB plus core promoter of tar) (40). In the reaction mixture containing plasmid templates and reconstituted RNAP holoenzymes, the resultant transcripts were dependent on the addition of RNAP holoenzyme. RNAP holoenzyme made from hybrid σ²⁸_Ec-Ct4, which contains σ²⁸_Ct region 4, exhibited higher activities with P_S9, P_hctB, P_S14, and P_tar+ (P_test/P_fliC transcript ratio of ≥1) but was weakly active with P_hctBΔ_up and P_tar (P_test/P_fliC ration of <1) (Fig. 1B). Such behaviors of σ²⁸_Ec-Ct4 RNAP are comparable to those of wild-type σ²⁸_CtRNAP. In contrast, reconstituted σ²⁸_Ct-Ec4RNAP, which contains σ²⁸_Ec region 4, conferred strong activities for P_hctB, P_hctBΔ_up, P_tar, and P_tar+, but it was weakly active with P_S9 and P_S14 (Fig. 1C). These actions are in agreement with results from using wild-type σ²⁸_EcRNAP. Taken together, the results obtained with the hybrid σ²⁸ proteins indicate that the observed differences in promoter selectivity between σ²⁸_Ct and σ²⁸_Ec are, at least in part, specified by region 4.

Identification of mutants defective in σ²⁸_Ct-dependent transcription in Salmonella.

We next sought to identify residues in region 4 of σ²⁸_Ct that are important for σ²⁸_Ct-dependent transcription. A plasmid library carrying mutagenized σ²⁸_Ct genes driven from an arabinose-inducible P_araB promoter was transformed into Salmonella fliA mutant TH7034, which allowed us to screen mutants defective in transcription from σ²⁸-dependent fliC::lacZ. In this background, only σ²⁸_Ct-dependent transcription can be detected, because no functional Salmonella σ²⁸ exists. Cells were selected for a lower level of σ²⁸-dependent transcription of lacZ (white or pink colonies) than for the wild type (red colony) by plating on MacConkey-lactose-arabinose agar. The candidates were further confirmed by measurement of β-galactosidase activity. Approximately 7,000 clones were screened, and four with single-residue substitutions in σ²⁸_Ct (E206G, Y214C, E222G, and Q236L) were isolated (Fig. 2). These mutants were further characterized as described below.

FIG. 2.

Mutagenesis of σ²⁸_Ct region 4. (A) Alignment of the amino acid sequences of region 4 from E. coli σ⁷⁰, Thermus aquaticus σ^A, E. coli σ³² (σ³²_Ec), Salmonella σ²⁸ (σ²⁸_St), σ²⁸_Ec, Aquifex aeolicus σ²⁸ (σ²⁸_Aa), and σ²⁸_Ct using the CLUSTAL W program. The amino acid sequence of σ²⁸_Sa is identical to that of σ²⁸_Ec. Regions were defined based on previous studies (26, 41). The number for each amino acid position is relative to that from the start of each protein sequence. Identical or similar residues are shown in black or gray shadow, respectively. Asterisks at the bottom indicate residues undergoing mutation. (B) Isolation of σ²⁸_Ct mutants. Mutated nucleotides and the deduced amino acid residues are indicated based on the position of σ²⁸_Ct and corresponding residue positions in different σ factors.

Mutations in region 4 of σ²⁸_Ct decrease transcription from the chlamydial hctB promoter.

We compared the effects of σ²⁸_Ct mutants on transcription from the σ²⁸_Ct-dependent hctB promoter in Salmonella. Plasmid-encoded wild-type σ²⁸_Ct or derivatives were expressed in Salmonella fliA mutant strain TH5504 carrying pRV_hctB, and σ²⁸_Ct-driven transcription from P_hctB::lacZ was assessed by measurement of β-galactosidase activity as described in Materials and Methods. Relative β-galactosidase activities from strains carrying σ²⁸_Ct substitutions Y214C, E222G, and Q236L were significantly lower than that of wild-type σ²⁸_Ct (i.e., 7.9%, 25.1%, and 1.0%, respectively) (Fig. 3A). σ²⁸_Ct E206G also decreased levels of β-galactosidase activity, to about 41.9% relative to wild-type σ²⁸_Ct.

FIG. 3.

Effect of single-residue substitution on σ²⁸-dependent transcription from chlamydial hctB promoter. (A) Activity of mutated σ²⁸_Ct in vivo using a lacZ reporter assay. Cellular β-galactosidase activity was normalized to levels of σ²⁸_Ct determined by Western blot analysis using specific anti-σ²⁸ antibody. We previously found that mutation in region 4 of σ²⁸_Ct did not impair its reactivity with the antibody to σ²⁸_Ct. The β-galactosidase activity from each strain containing the indicated mutated proteins is represented as a percentage relative to the β-galactosidase activity of the strain with wild-type σ²⁸_Ct. Asterisks above the bars indicated significantly reduced activities compared with that of the wild type (P < 0.05). (B) Transcripts from single-round transcription assay in vitro. Plasmid pGHR containing two chlamydial promoters (P_hctB and P1_CtrRNA) was used as a template in the presence of σ⁷⁰RNAP or σ²⁸_CtRNAP as indicated on the top. The transcripts generated from promoters are indicated on the left. The intensity of each transcript band was quantified using Quantity One. The amount of transcript produced by each mutant RNAP is reported as a percentage of the transcription of wild-type σ²⁸_CtRNAP.

Next, the direct effect of mutant σ²⁸_CtRNAP holoenzyme on transcription from P_hctB was examined in vitro. In the presence of both σ²⁸_CtRNAP and σ^70_CtRNAP holoenzymes, plasmid template pGHR produced two divergent transcripts: a 152-bp transcript from the σ²⁸-dependent P_hctB and a 130-bp transcript from the σ⁶⁶-dependent rRNA P1 (P1_CtrRNA) (data not shown). The RNAP containing σ²⁸_Ct Y214C or Q236L was severely defective in generating transcripts from hctB, showing a yield of less than 10% of the wild-type level (Fig. 3B). The RNAP containing σ²⁸_Ct E206G or E222G reduced transcription activity to 79.1% and 49.0% of that of wild-type σ²⁸_CtRANP, respectively (Fig. 3B). None of the mutant σ²⁸_CtRNAPs was able to transcribe from the P1_CtrRNA. We also noted that the RNAP saturated with mutated σ²⁸_Ct did not mount transcription activity in vitro. Moreover, mutated σ²⁸_CtRNAPs did not change their behavior for transcription from P_S9, P_hctB, P_S14, P_hctBΔ_UP, P_tar, and P_tar+ in vitro (data not shown).

Taken together, these transcription studies indicate that σ²⁸_Ct substitutions Y214C and Q236L severely decreased transcription from hctB, whereas substitutions E206G and E222G moderately affected hctB transcription.

Effect of σ²⁸_Ct substitutions on core binding.

The abilities of wild-type and mutant σ²⁸_Ct to bind core RNAP were compared using coimmunoprecipitations. The levels of mutant proteins that bound to Salmonella core RNAP were similar to or slightly higher than the wild-type level. (Fig. 4). As validation, we also examined potential core-binding differences among these proteins using the E. coli core and purified recombinant σ²⁸_Ct or derivatives in vitro. We found that each of the σ²⁸_Ct mutants effectively bound E. coli core RNAP by coimmunopreciptation relative to wild-type σ²⁸_Ct (data not shown). This confirms that none of the σ²⁸_Ct substitutions severely reduced σ affinity for the core RNAP under our test conditions.

FIG. 4.

Binding of σ²⁸ and derivatives to core RNAP in Salmonella. (A) Immunoblot showing the results of a coimmunoprecipitation assay from Salmonella fliA mutant TH7034, which expresses σ²⁸ or derivatives as indicated above each lane. A strain expressing E. coli σ²⁸, which is identical to Salmonella σ²⁸, was the negative control. Cell lysates were subjected to immunoprecipitation using anti-β monoclonal antibody as described in Materials and Methods. Precipitated proteins were separated by PAGE and immunoblotted with anti-σ²⁸ antibody or anti-β monoclonal antibody. (B) Relative binding affinity of σ²⁸ or derivatives for core. The amount of precipitated σ²⁸ protein is reported as a percentage of the level of β in core RNAP.

Influence of σ²⁸_Ct substitutions on the interaction of σ²⁸_Ct region 4 with the β-flap.

We examined whether substitutions in the σ²⁸_Ct region 4 would disrupt protein-protein interaction of σ²⁸_Ct region 4 with the β-flap in a bacterial two-hybrid assay (8, 10). This assay involves the use of (i) a reporter strain, KS1, which carries a chromosomal copy of the test promoter plac O_R2-62 linked to lacZ; (ii) a plasmid encoding the λcI fused to the β-flap from Chlamydia (λcI-β-flap_Ct) or E. coli (λcI-β-flap); and (iii) a second plasmid encoding the N-terminal domain of αNTD fused to σ²⁸_Ct region 4 or its derivatives. Interaction of the β-flap with σ²⁸_Ct region 4 stabilizes RNAP binding to the test promoter, thus, mediating transcription activation from the test promoter plac O_R2-62::lacZ. This can be monitored by measuring β-galactosidase activity (Fig. 5A).

FIG. 5.

Bacterial two-hybrid assay: characterization of σ²⁸_Ct substitution that affects its interaction with the β-flap. (A) Diagram showing how interaction of the σ²⁸_Ct region 4 (in pBRα-σ²⁸_Ct) with the β-flap (in pACλcI-β-flap_Ct or pACλcI-β-flap) activates transcription from the test promoter plac O_R2-62 located on the chromosome in KS1. (B) Effects of substitution in the σ moiety of the α-σ²⁸_Ct chimera on transcription from plac O_R2-62 in the presence of the chimera λcI-β-flap_Ct. (C) Effects of substitutions in the σ moiety of the α-σ²⁸_Ct chimera on transcription from plac O_R2-62 in the presence of the chimera λcI-β-flap. KS1 cells harboring two compatible plasmids, which direct the synthesis of α-σ²⁸_Ct or derivatives and λcI-β-flap_Ct or pACλcI-β-flap as indicated, were grown in the presence of carbencillin, chloramphenicol, and kanamycin. Protein expression was induced by treatment with different concentrations of IPTG for 1 hour when cultures reached an optical density at 600 nm of 0.3. β-Galactosidase activity was measured in duplicate on at least three independent occasions. Values shown are the averages from one experiment.

As shown in Fig. 5B, in the presence of the chlamydial β-flap fusion (cI-β-flap_Ct), the chimera α-σ²⁸_Ct activates transcription from plac O_R2-62 up to ∼14.0-fold relative to that of the negative control α expressed from pBR-α. Such an increased magnitude compared to no obvious increase in the paired α-σ⁷⁰/cI-β-flap did not surprise us, because σ²⁸_Ct contains a residue at G228 corresponding to σ⁷⁰D581 (25, 34). Substitution at σ⁷⁰D581G stabilizes the folded structure of the tethered σ⁷⁰ region 4 moiety and enhances interaction of σ⁷⁰ region 4 with the DNA −35 element (25, 34). In strains coexpressing chimera λcI-β-flap_Ct and chimera α-σ²⁸E206G, α-σ²⁸Y214C, or α-σ²⁸E222G, the magnitude of transcription activation from plac O_R2-62 decreased about ∼8.3-, ∼4.5-, and ∼6.9-fold, respectively. The decrease of transcription activation observed is not due to instability of chimera proteins, as the chimeras were able to interact with the hctB −35 element in a one-hybrid assay (Fig. 6). In contrast, chimera α-σ²⁸Q236L strengthened the interaction between the β-flap and σ²⁸ in the presence of λcI-β-flap_Ct; transcription from plac O_R2-62 was increased up to ∼23.4-fold (Fig. 5B) compared to the wild-type σ²⁸_Ct (∼14.0-fold increase).

In the presence of an E. coli β-flap fusion (cI-β-flap), the α-σ²⁸_Ct chimera increased transcription from plac O_R2-62 up to ∼8.5-fold, compared to ∼14-fold in the presence of the cI-β-flap_Ct chimera. Substitutions E206G, Y214C, and E222G in σ²⁸_Ct reduced transcription from plac O_R2-62 to ∼3.4-, ∼3.4-, and ∼5.4-fold, respectively. However, the α-σ²⁸_Ct Q236L chimera stimulates transcription from plac O_R2-62 ∼14.4-fold, compared to a change of ∼23.4-fold in the presence of the cI-β-flap_Ct chimera (Fig. 5C).

These results indicate that substitution E206G, Y214C, or E222G in σ²⁸_Ct weakens interactions between σ²⁸_Ct and the β-flap from both E. coli and Chlamydia. Thus, these residues are involved in interaction of the cI-β-flap with σ²⁸ region 4, whereas σ²⁸_Ct residue Q236 seems not to be important for β-flap binding. E. coli and chlamydial β-flap share 62.2% amino acid identity and 73.8% similarity. The finding that σ²⁸_Ct can bind the E. coli β-flap could be relevant to its ability to transcribe from both σ²⁸_Ec- and σ²⁸_Ct-dependent promoters when it associates with the E. coli core enzyme in a heterologous genetic system or in vitro.

Effect of σ²⁸_Ct mutants on binding of the −35 element from the hctB promoter.

We next examined whether residue substitutions in σ²⁸_Ct region 4 would affect interaction of σ region 4 and the −35 element of the promoter, using a one-hybrid assay (34). This in vivo DNA-binding assay is designed to use a test promoter, which contains an ectopic hctB −35 element upstream of the lac promoter in strain ZY101, and a plasmid-encoded α-σ²⁸ fusion. The direct interaction of σ²⁸ region 4 with the ectopic hctB −35 element recruits and stabilizes RNAP to the test promoter, causing an increase of reporter gene (lacZ) expression. This can be monitored by measuring β-galactosidase activity (Fig. 6A).

We introduced a plasmid encoding chimera α-σ²⁸ or its derivatives into the reporter strain ZY101. We chose chimera α-σ⁶⁶ as a negative control, as previous report showed that chlamydial σ⁶⁶ did not recognize the σ²⁸-dependent promoter (48). The σ²⁸L243K mutant bears enhanced transcriptional activity from the hctB promoter (Z. Hua et al., unpublished data) and was the positive control. In the presence of α-σ²⁸, transcription activation from the test promoter was observed, and the increase of stimulation occurred in an IPTG dose-dependent fashion (Fig. 6B). Similarly, expression of α-σ²⁸E206G, α-σ²⁸E222G, or α-σ²⁸Y214C was able to activate transcription from the test promoter. Although the stimulations are small (<1.5-fold), these increases are reproducible in all experiments. Chimera α-σ²⁸L243K increased transcription up to 1.8-fold, indicating that this system functioned. In contrast, chimera α-σ²⁸Q236L failed to stimulate transcription from the test promoter (Fig. 6B). Because α-σ²⁸Q236L interacts with chimera λcI-β-flap (Fig. 5), the inability of α-σ²⁸Q236L to interact with the ectopic promoter is unlikely to be related to the poor protein expression. The substitution σ²⁸_CtQ236L disrupted α-σ²⁸/hctB −35 interactions, which may be an explanation for our findings. As expected, chimera α-σ⁶⁶ was unable to activate transcription from test promoter in ZY101 (Fig. 6B and C); however, α-σ⁶⁶ stimulated transcription from placCons-35C in BN317, which contains a σ⁷⁰-specific ectopic −35 element (data not shown).

In ZY102, which bears a mutant ectopic hctB −35 element (gAAAGTTT), none of the wild-type and mutant α-σ²⁸ chimeras stimulated transcription from the test promoter (Fig. 6C). The first base pair T of the hctB −35 element has been shown to be the major determinant for σ²⁸_Ct recognition (49). This result confirmed that the observed stimulatory effects of α-σ²⁸_Ct, α-σ²⁸E206G, α-σ²⁸Y214C, and α-σ²⁸E222G in ZY101 are a result of the specific interaction of the chimeras with the ectopic hctB −35 element. Substitution σ²⁸_CtQ236L interrupts interaction between σ²⁸ and the hctB −35 element, indicating that σ²⁸_Ct Q236 is essential to be in contact with the −35 element.

DISCUSSION

We focused on characterization of several σ²⁸_Ct mutants with mutations in region 4 that are defective in σ²⁸-dependent transcription. Since there is no successful genetic system available to manipulate genes in Chlamydia, our strategy has been to study σ²⁸-dependent transcription using complementary approaches. By switching σ²⁸_Ct region 4 with σ²⁸_Ec region 4, we found that amino acids in σ²⁸_Ct region 4 might specify the function of the σ²⁸_CtRNAP holoenzyme for use of an altered spacer length between promoter −35 and −10 elements, as well as preference for the UP-like sequences in vitro (Fig. 1). We also screened for σ²⁸_Ct mutants defective in transcription from the σ²⁸-dependent fliC promoter in Salmonella (Fig. 2). We then examined the effects of four single-residue substitutions in σ²⁸_Ct region 4 on transcription from the chlamydial hctB promoter both in vitro and in a fliA mutant Salmonella strain. We found that substitutions σ²⁸_Ct, Y214C and Q236L, significantly decreased transcription from the hctB promoter, while σ²⁸_CtE206G and -E222G moderately reduced hctB transcription (Fig. 3); the promoter specificity was unchanged, and the effect was stronger in vivo than in vitro. A possible explanation for this is that the existence of additional host cell regulatory factors in vivo might lead to a bias for a diminished in vitro effect seen with σ²⁸_Ct mutations. It is unlikely that the difference observed is due to the inhibition of FlgM encoded by the surrogate strain of S. enterica serovar Typhimurium, because FlgM specifically negatively regulates Salmonella σ²⁸ but not σ²⁸_Ct (22).

The apparent transcriptional deficiency observed might result from σ²⁸_Ct mutations that have a reduced affinity of σ with the core RNAP, a disrupted σ/β-flap interaction, and/or an occluded σ/promoter interaction. Our data do not support a direct strong effect of mutant σ²⁸_Ct on core affinity, since RNAP saturated by mutant σ²⁸ did not increase transcription. Moreover, the mutant σ²⁸_Ct effectively bound core RNAP in a coimmunoprecipitation experiment (Fig. 4). However, modest indirect effects cannot be excluded. By taking advantage of the bacterial two-hybrid system, we found that the three substitutions in σ²⁸_Ct, E206G, Y214C, and E222G, impaired interaction of σ²⁸_Ct with the β-flap from Chlamydia or from E. coli (Fig. 5). Of these residues, only the corresponding residue σ²⁸_CtY214 has been mapped and directly contacts the β-flap in the X-ray structure of the Thermus aquaticus σ^A holoenzyme-DNA complex (34) (Fig. 7A and B). The corresponding residues of σ²⁸_CtE206 and -E222 may indirectly involve such σ/β-flap interactions. The X-ray structure of Aquifex aeolicus σ²⁸, free or in complex with FlgM, the anti-σ²⁸ factor, does not explain our observations, as both core and DNA-binding domains are buried in the folded compact structure (Fig. 7C) (41, 42). We speculate that σ²⁸_Ct is mostly in an active conformation when it associates with the core RNAP in Salmonella or in vitro, consistent with their transcription activities. Perhaps substitution E206G, Y214C, or E222G in σ²⁸_Ct caused a deficiency of σ²⁸_Ct in binding the β-flap, resulting in an unstable or mispositioned region 4 in the σ²⁸_CtRNAP holoenzyme and an inability to simultaneously bind promoter −10 and −35 elements, leading to poor transcription activity. Our results further add support to the idea that RNAP remodeling of σ²⁸ region 4 is required to expose the recognition domains of both the −35 binding site and β-flap on σ²⁸ (41, 42). Previous studies indicated that the σ⁷⁰ residues E555, F563, and E575 (corresponding to σ²⁸_Ct E206, Y214, and E222, respectively) (Fig. 3) have been implicated in interaction with the E. coli β-flap (12, 13).

FIG. 7.

Positions of residues in substitutions that affect the interaction of σ²⁸_Ct region 4 with the β-flap. (A) Locations of residues based on the crystal structures of Thermus aquaticus σ^A holoenzyme and DNA (PDB ID no. 1I9Z) (32). Shown are the surfaces of α (green), β-flap (gray), β′ (cyan), ω (orange), and σ factor (blue). The DNA promoter (magenta) is shown as a cartoon. (B) Enlarged illustration of region 4 of T. aquaticus σ^A (labeled with corresponding amino acid numbers in σ²⁸_Ct) (see amino acid alignments in Fig. 2) and the β-flap contact. The β-flap (gray) is shown as the cartoon. The DNA promoter is hidden. (C) Locations of residues based on the crystal structures of the σ²⁸_Aa/FlgM complex (labeled with corresponding amino acid numbers in chlamydial σ²⁸_Ct) (PDB ID no. 1SC5 and 1RP3) (41). Blue, σ²⁸_Aa; yellow-brown, FlgM. Residues corresponding to the T aquaticus σ^A or σ²⁸_Aa were identified and mutated to the naturally occurring residues in chlamydial σ²⁸_Ct as indicated. This figure was generated using Pymol 0.99rc6 (http://pymol.sourceforge.net).

Our data suggest a role of residue Q236 of σ²⁸_Ct in binding the −35 element of the hctB promoter because (i) substitution Q236L in σ²⁸_Ct disrupted the interaction of σ²⁸_Ct with the hctB −35 sequence as determined by the one-hybrid assay (Fig. 6), and the same protein interacted with the β-flap as determined by the two-hybrid assay (Fig. 5); (ii) Q236 is located in the DNA-binding helix-turn-helix motif in of σ²⁸_Ct; and (iii) in T. aquaticus σ^A, the corresponding residue Q414 is structurally positioned on the DNA-binding interface (4). Consistent with our results, Kourennaia et al. (24) reported that the substitution Q269A in E. coli σ³² (corresponding to σ²⁸_Ct Q236) impeded recognition of the −35 element of the E. coli groE promoter but did not reduce its ability to bind to the core enzyme. The corresponding residue in σ⁷⁰ (Q589) has not been implicated in contact with a specific DNA base. Instead, the adjacent residue R588 of σ⁷⁰ can help position residue 585 for direct contact with bacteriophage P_RM DNA and mediate binding to the DNA activator (20). Presumably, the σ²⁸_Ct substitution Q236L, which changes a polar residue to an aliphatic hydrophobic residue, interrupts the side chain contacting the −35 element sequence and/or indirectly affects its ability to bind DNA by distorting or destabilizing the structure of σ²⁸_Ct region 4.

Given the favorable effect of the presence of AT-rich sequences upstream of the promoter on σ²⁸_Ct-dependent transcription (40), we wondered whether a possible positive influence is induced by the C-terminal domain of the α subunit (αCTD). The interaction of σ region 4 with the αCTD can stimulate transcription from a subset of UP element-dependent promoters by tightening RNAP-promoter associations (14). We failed to detect direct interaction of wild-type or mutant σ²⁸_Ct with the αCTD in a two-hybrid assay (Hua et al., unpublished data). We still cannot rule out the possibility that this association, if any, might be transient and/or weak, thus making it difficult to detect in vivo.

These studies have allowed us to begin defining the determinants in σ²⁸_Ct that are essential for contact with the β-flap and with the −35 element; both are required for recognition of the −35/−10 promoter. Because there is no overlapping promoter recognition specificity between σ²⁸_Ct and the primary σ⁶⁶ (σ⁷⁰) (39, 40), it is fair to assume that there might be some differences regarding the σ/core interaction as well as σ/promoter interaction. Supporting this, σ²⁸_Ct Q236 has been shown to be important for the σ²⁸_Ct-specific −35 element binding. This function has not been reported for the counterpart residue of E. coli σ⁷⁰. While this study does not directly address whether σ²⁸ activity is inhibited by a regulator, observations with several σ²⁸_Ct region 4 mutants defective in contact with the β-flap indicates that inactivation may play a role in chlamydial σ²⁸ activity control. In the future, we will use a stabilized or deficient protein-binding mutant of chlamydial σ²⁸ in order to facilitate identification and characterization of potential σ²⁸ regulators in a two-hybrid assay.