Genetic Transplantation: Salmonella enterica Serovar Typhimurium as a Host To Study Sigma Factor and Anti-Sigma Factor Interactions in GeneticallyIntractable Systems

Bacterial strains, plasmids, and primers used in this study are listed in Table 1.

Media, growth conditions, phage P22 transduction methods, and motility assays were as described previously (10, 13). Tetracycline-sensitive (Tc^s) plates were made as described by Maloy (28). Antibiotics in rich media were used at the following concentrations (inμ g/ml): ampicillin (Ap; 100), kanamycin (50), tetracycline (15), and spectinomycin (100). PCR and DNA cloning were performed as described previously (35). DNA primers (Integrated DNA Technologies, Coralville, IA) used in this study are listed in Table 1. PCRs were performed with Taq (Promega, Madison, WI), Accuzyme (Bioline, Randolph, MA), or ThermalAce (Invitrogen, Carlsbad, CA). All enzymes used for DNA cloning were from New England Biolabs (Beverly, MA). DNA sequencing was performed using BigDye v3.1 (Applied Biosystems, Foster City, CA), and the reactions were analyzed using a 3730XL sequencer at the DNA sequencing facility of the Department of Biochemistry, University of Washington, Seattle, WA.

The fliA coding region in S. enterica serovar Typhimurium was replaced by allelic exchange with a tetRA element using λ-Red-mediated recombination (8, 32, 48). In addition, a tetRA element was inserted just before the ATG start codon of flgM. This tetRA element includes the coding sequences of tetR and tetA from transposon Tn10dTc and confers tetracycline resistance (47). The PCR primers for amplification of the tetRA element were flanked by 40-bp sequences of homology for recombination on the chromosome. Primers FliAdeltetR, FliAdeltetA, FlgMAUGtetR, and FlgMAUGtetAwere used for the construction ofΔ fliA5805::tetRA and flgM6085::tetRA, respectively. The PCRs containing genomic DNA of TH3467, DNA primers, dNTPs, and Taq polymerase were amplified as follows; 3 min at 95°C for 1 cycle, 30 s at 95°C, 30 s at 49°C, 2 min at 72°C for 30 cycles, and 10 min at 72°C for 1 cycle. A 1,990-bp product was purified using a QIAquick PCR purification kit (QIAGEN, Valencia, CA). S. enterica serovar Typhimurium strain TH4702 containing aλ -Red plasmid pKD46 that has a temperature-sensitive replicon (8) was grown under inducing conditions in LB plus Ap (100 μg/ml) plus arabinose at 0.2% at 30°C until the optical density at 600 nm reached 0.6 to 0.8. The cells were washed twice in equal volumes of cold sterile water and concentrated 250-fold. Fifty μl of cells was electroporated with 100 to 200 ng of purified PCR fragment using 0.1-cm cuvettes at 200 Ω, 1.6 kV, and 25 μF. Subsequently, one ml of LB was added, and cells were incubated for 1 hour at 37°C. Approximately 0.5 ml of cells was plated on LB plates containing tetracycline and incubated at 37°C. Tc^r colonies were purified on LB plates without antibiotics and incubated at 42°C. Tc^r colonies were screened for Ap^s to confirm the loss of the pKD46 plasmid. To confirm that the tetRA integrated into the correct region on the chromosome, Tc^r and Ap^s colonies were screened by PCR using primers within the tetRA element and in sequences flanking fliA or flgM, respectively.

The fliA and flgM genes were replaced in S. enterica serovar Typhimurium with homologs from A. aeolicus and C. trachomatis using λ-Red-mediated replacement of the tetRA element as described below. The fliA homologs from A. aeolicus (fliA) and C. trachomatis (rspD) were PCR amplified using genomic DNA from A. aeolicus (kind gift from K. O. Stetter) and pLF28 (39), respectively. Primer sets to amplify the fliA homologs are as follows: for A. aeolicus, AAGTG and AA28R; for C. trachomatis, CTFLIAGTG and CTFLIASTOP. PCR amplification was performed as described above, except the annealing temperature used was 45°C. The PCR products were purified, electroporated into strain TH8350 (ΔfliA::tetRA) as described above except for subsequent plating of the cells onto Tc^s medium (28), and incubated at 42°C overnight. Colonies were purified once on Tc^s medium at 42°C and then on LB plates at 37°C. The constructs were screened for the loss of the tetRA element by PCR, and the region was subsequently PCR amplified for DNA sequencing to confirm the replacement of S. enterica serovar Typhimurium fliA with the fliA homologs of A. aeolicus and C. trachomatis. The coding region of flgM in S. enterica serovar Typhimurium was replaced with the homolog from A. aeolicus as described for fliA using A. aeolicus genomic DNA, primers AAFLGM1 and AAFLGM2, and TH8349.

The coding regions of fliA and flgM from A. aeolicus were mutagenized by error-prone PCR with genomic DNA from TH8125, primers FliA#3 and FliA#4 (for σ²⁸), FlgM5′UP and FlgN reverse (for flgM), dNTPs, and 5 units of Promega Taq per reaction. The PCR amplifications were performed as described above. Electrocompetent cells of strains TH8350 (for fliA) and TH8349 (for flgM) were prepared as described above, and transformants were selected on Tc^s medium and incubated at 42°C. Colonies were further purified on Tc^s medium at 42°C and then on LB medium at 37°C. The colonies were screened for Lac⁺ on MacConkey medium. The fliA or flgM regions from the Lac⁺ isolates were DNA sequenced.

The fliA coding sequences from S. enterica serovar Typhimurium, A. aeolicus, and C. trachomatis were cloned under the arabinose promoter to construct pMC147, pJK558, and pJK629 as follows. A 770-bp EcoRI fragment blunted with Klenow fragment containing S. enterica serovar Typhimurium fliA from pMS531 was ligated into pBAD24 digested with NcoI and blunted with Klenow fragment to generate pMC147. A. aeolicus fliA was PCR amplified using A. aeolicus genomic DNA and primers AQAA1 and AQAA2. The PCR product was digested with BspHI and HindIII and ligated into pBAD24 digested with NcoI and HindIII to generate pJK588. A 515-bp NcoI-PstI fragment from pLF28 containing C. trachomatis rspD was ligated into pBAD24 digested with NcoI-PstI to generate pJK629. A 500-bp NcoI-HindIII fragment from pET-rsbW C. trachomatis L2 containing His-tagged rsbW from C. trachomatis (gift from T. P. Hatch) was ligated into pET15b digested with NcoI-XhoI to generate pJK627.

β-Galactosidase assays were performed as described by Maloy (27). Cells were grown in LB or LB plus Ap (100 μg/ml) plus 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) (in strains containing plasmids) at 37°C until mid-log phase. Assays were performed in triplicate, and the values were normalized as percentages of the wild-type (WT) value.

Cells were grown in LB until mid-log phase, and 1 ml of cells was centrifuged and resuspended in sodium dodecyl sulfate sample buffer. The samples were run on a 10% Tricine-sodium dodecyl sulfate polyacrylamide gel and electrotransferred in CAPS buffer onto polyvinylidene difluoride membranes (29, 38). The membrane was probed with rabbit anti-FliC antibodies (Salmonella H antiserum i rabbit, catalog number 228241; Difco) purified as described previously (15) and developed using an ECL-Plus kit (Amersham Biosciences, Piscataway, NJ).

Amino acid alignments were performed using Clustal W (6). The sequence alignments were edited using Jalview alignment editor (7). Aquifex aeolicus σ²⁸/FlgM cocrystal analysis used coordinates PDB ID 1SC5 from the RCSB protein database (www.rcsb.org/pdb/ ) and RasMol v. 2.7 (2, 36). The analysis of helix interactions was done with PDBsum (23). The analysis of residue contacts was determined on PDB ID 1SC5 using CSU (contact of structural units) software (40).

Given the similarities in the critical domains of σ²⁸ (see alignments in Fig. 1), we tested whether σ²⁸ homologs of A. aeolicus and C. trachomatis would complement an S. enterica serovar Typhimurium σ²⁸ mutant. The genes encoding σ²⁸ of S. enterica serovar Typhimurium (fliA_SeT), A. aeolicus (fliA_Aa), and C. trachomatis (rspD_Ct) were cloned under the arabinose-inducible promoter (P_araB) and introduced into the S. enterica serovar Typhimurium fliA mutant strain TH4387. The plasmid-containing strains were assessed for motility on plates containing 0.2% arabinose (Fig. 2A). Levels of complementation for motility by P_araB-A. aeolicus fliA⁺ andP_araB-C. trachomatis rspD⁺ were 39% and 58%, respectively, of the level for P_araB-S. enterica serovar Typhimurium fliA⁺ (Fig. 2A). Similar results were previously observed with A. aeolicus fliA⁺ from E. coli. (44). Interestingly, a P_araB-C. trachomatis rspD⁺ plasmid did not restore motility in E. coli with a fliA deletion (39). Although E. coli and S. enterica serovar Typhimurium possess nearly identical flagellar systems (25), the dramatic differences in complementation for motility with P_araB-C. trachomatis rspD⁺ demonstrates that differences in the flagellar regulons between the two species can be uncovered.

The coding region of chromosomal S. enterica serovar Typhimurium fliA was replaced with either the fliA from A. aeolicus or the rspD from C. trachomatis by targeted replacement of a tetRA element using a λ-Red recombination system (Fig. 3A; also see reference 8 and Materials and Methods). Therefore, σ²⁸-mediated gene expression in these hybrid strains remained under the regulation of the S. enterica serovar Typhimurium flagellar regulon. Theseσ ²⁸ hybrid strains were weakly motile (data not shown). Because motility was weak, we analyzed whether the hybrid strains expressed the σ²⁸-dependent flagellar filament protein FliC. Western blot analysis using anti-FliC serum on cellular lysates showed the presence of FliC protein in S. enterica serovar Typhimurium hybrid strains with A. aeolicus σ²⁸ or C. trachomatisσ ²⁸ (Fig. 2B, lanes 3 and 4, respectively). However, the levels of FliC protein were reduced in theσ ²⁸ hybrid strains by fourfold relative to the level for WT S. enterica serovar Typhimurium (Fig. 2B). Although motility was reduced in the σ²⁸ hybrid strains, the presence of FliC protein and the ability to transcribe class 3 flagellar promoters (see below) indicate the conservation of σ²⁸ function. Together, these results indicate that σ²⁸ from diverse species can interact with the heterologous host transcriptional machinery for flagellar biogenesis.

To investigate the interactions of cognate sigma factor and anti-sigma factor activities between species, hybrid S. enterica serovar Typhimurium strains were constructed by use of λ-Red (Fig. 3A; also see Materials and Methods). The fliA⁺ gene (σ²⁸) of S. enterica serovar Typhimurium was replaced with the σ²⁸ homologs from A. aeolicus and C. trachomatis, fliA⁺ and rspD⁺, respectively. Further, the negative regulatory gene flgM⁺ of S. enterica serovar Typhimurium was replaced with A. aeolicus flgM⁺. We assayedσ ²⁸-mediated gene expression in the hybrid strains using lac gene transcriptional fusion reporters to twoσ ²⁸-dependent promoters, motA and fliC. The motA and fliC genes encode vital components of the flagellum; MotA is required for flagellar rotation, and FliC is the main structural component of the flagellar filament (25).

Relativeβ -galactosidase levels of motA-lac and fliC-lac expression in WT S. enterica serovar Typhimurium increased 3- and 1.7-fold, respectively, in the absence of the S. enterica serovar Typhimurium FlgM (Fig. 4, compare lanes 2 and 3 in both panels). Likewise, in S. enterica serovar Typhimurium strains with A. aeolicusσ ²⁸, in the absence of A. aeolicus FlgM, motA-lac and fliC-lac expression was increased around 2.5-fold (Fig. 4, compare lanes 6 and 7). This is in support of structure studies on A. aeolicus σ²⁸/FlgM interaction which suggest that FlgM binding to σ²⁸ inhibits its interaction with core RNAP (41). Interestingly, in the presence of S. enterica serovar Typhimurium FlgM, A. aeolicus σ²⁸-mediated levels of motA-lac and fliC-lac expression were similar to that of an S. enterica serovar Typhimurium flgM deletion strain, indicating the absence of negative regulation on A. aeolicus σ²⁸ activity (Fig. 4, compare lanes 7 and 8). In S. enterica serovar Typhimurium σ²⁸ strains, the relative levels of β-galactosidase activity in the S. enterica serovar Typhimurium flgM deletion strain were similar to those of strains containing A. aeolicus FlgM (Fig. 4, compare lanes 3 and 4), suggesting that A. aeolicus FlgM has no effect on S. enterica serovar Typhimurium σ²⁸ activity. These findings show inhibition of A. aeolicusσ ²⁸ activity by its cognate negative regulator FlgM but not by S. enterica serovar Typhimurium FlgM. Similarly, S. enterica serovar Typhimurium σ²⁸ activity is inhibited by its cognate negative regulator FlgM but not by A. aeolicus FlgM. Collectively, these studies indicate that the anti-σ²⁸ FlgM factors interactionswith σ²⁸, and their negative regulation onσ ²⁸ activities are species specific.

Hybrid strains of S. enterica serovar Typhimurium containing C. trachomatis rspD⁺ (C. trachomatis σ²⁸) were constructed in motA-lac and fliC-lac backgrounds. C. trachomatis RsbW, the putative negative regulator of RspD (39), was encoded in trans by plasmid pJK627. However, there was no difference in the relative β-galactosidase levels in the presence of RsbW or in its absence (Fig. 4, lanes 10 and 11, respectively). S. enterica serovar Typhimurium FlgM also did not negatively regulate C. trachomatis σ²⁸ activity (Fig. 4, lanes 12). C. trachomatis RsbW is 36% identical and 58% similar to homolog RsbW in Bacillus subtilis. In B. subtilis, anti-sigma factor RsbW and anti-anti-sigma factor RsbV, as well as other regulatory proteins, regulate σ^B activity but differ mechanistically from anti-σ²⁸ FlgM homologs (16). The lack of anti-σ²⁸ activity by C. trachomatis RsbW in S. enterica serovar Typhimurium suggests that additional factors for C. trachomatisσ ²⁸ inhibition may be required or thatσ ²⁸ inhibition may be independent of RsbW.

We sought to isolate mutants defective in A. aeolicusσ ²⁸/FlgM interactions using a genetic screen. S. enterica serovar Typhimurium strain TH8125 contains A. aeolicus fliA⁺ and A. aeolicus flgM⁺ genes as well as aσ ²⁸-dependent fliC-lac reporter construct. TH8125 is white (Lac⁻) on MacConkey lactose (Mac-lac) medium, and the same strain deleted for A. aeolicus flgM is red (Lac⁺). Therefore, a phenotypic screen was used to identify mutants of TH8125 defective in A. aeolicus σ²⁸/FlgM interactions (red on Mac-lac). Error-prone PCR was used to mutagenize A. aeolicus fliA (σ²⁸) and A. aeolicus flgM genes. The mutagenized gene products were electroporated into S. enterica serovar Typhimurium strain TH8125 that had a tetRA insertion in either A. aeolicus fliA or A. aeolicus flgM. Subsequent recombination of fliA and flgM was achieved by use of λ-Red (Fig. 3B). The loss of tetracycline resistance, indicative of replacement with the mutagenized gene products, was selected on tetracycline-sensitive medium (28). The transformants were subsequently screened on Mac-lac medium, and red (Lac⁺) colonies were sequenced for putative mutations in either A. aeolicus fliA or A. aeolicus flgM.

In the A. aeolicus fliA mutagenesis screen, 300 Tc^s colonies were isolated, of which 28 were Lac⁺ on Mac-lac medium. Twelve Lac⁺ mutants were sequenced, and seven contained missense mutations in A. aeolicus fliA (Fig. 5A and Table 2).Since the remaining five had multiple mutations in A. aeolicus fliA and are difficult to interpret, they were excluded from the study. β-Galactosidase assay results for A. aeolicus σ²⁸ mutants compared to those for WT (TH8125) demonstrated an increase in activity in the presence of A. aeolicus FlgM (Fig. 5B). The relative activity of A. aeolicus σ²⁸ mutants ranged from 1.3- to 2.9-fold higher than that for WT (TH8125) (Table 2).

To examine whether the mutants were more active due to increased activity of σ²⁸, β-galactosidase assays were performed in the absence of A. aeolicus FlgM (TH8200) (Fig. 5B). The relative activity of A. aeolicus σ²⁸ mutants compared to WT (TH8200) ranged from 0.8- to 1.6-fold higher (Table 2). Mutants T27S, I29M, and T175M had more σ²⁸ activity than WT (TH8200) in the absence A. aeolicus FlgM, suggesting that their ability to bypass FlgM inhibition was due to increased σ²⁸ activity and that they are not defective for FlgM interaction. Mutants I29M and T175M had changes in codon frequency that may have resulted in increased levels of σ²⁸ expression (Table 2). The codon frequency change of mutant T27S was nominal (Table 2), suggesting that an increased σ²⁸ activity is not due to increased translation. The increased activity could be attributed to increased stability, as was found for the H14D mutants of S. enterica serovar Typhimurium σ²⁸ that were able to bypass FlgM inhibition (4).

Mutation R149G had more σ²⁸ activity than WT (TH8125) in the presence of FlgM but had less σ²⁸ activity in the absence of FlgM (TH8200) (Fig. 5B and Table 2). This class of mutant was also found in the S. enterica serovar Typhimuriumσ ²⁸ defective for FlgM interaction (4). Although they had more σ²⁸ activity in the presence of FlgM, they were implied to have less σ²⁸ activity in the absence of FlgM, due to amino acid substitutions that increased the turnover ofσ ²⁸ (4).

Mutants P4S, E189G, and E199G have σ²⁸ activity levels similar to that of WT (TH8200) in the absence A. aeolicus FlgM and moreσ ²⁸ activity than WT (TH8125) in the presence of A. aeolicus FlgM, suggesting they are defective for FlgM interaction (Fig. 5B and Table 2). Residue E189 is buried within the σ²⁸ structure, where there are extensive contacts with the other interacting domains (41). This mutation may be involved in stabilizing the σ²⁸ structure in a manner that renders it resistant to FlgM inhibition through allosteric effects. In contrast, P4 lies at the N terminus ofσ ²⁸, is not buried within theσ ²⁸, and appears too distant to make contacts with FlgM (41). Nonetheless, the P4S mutation could involve a conformational change to alter inhibition by FlgM or modify σ²⁸ protein stability or interaction with RNAP.

The E199G mutation found in A. aeolicus σ²⁸ corresponds to location E203 of S. enterica serovar Typhimurium (4). An S. enterica serovar Typhimurium σ²⁸ E203D mutant exhibited 10-fold less affinity to S. enterica serovar Typhimurium FlgM and was the most attenuated for the destabilizing of the S. enterica serovar Typhimuriumσ ²⁸-RNAP holoenzyme by S. enterica serovar Typhimurium FlgM (4). By analogy to S. enterica serovar Typhimuriumσ ²⁸, the A. aeolicus σ²⁸ E199 position is an important contact for FlgM in inhibition ofσ ²⁸ interaction, with RNAP likely to be involved in the “stripping” of σ²⁸ from theσ ²⁸-RNAP holoenzyme complex by FlgM (4, 41). The E199G mutation is located in the conserved region 4 of the σ⁷⁰ family of transcription factors (31). Interestingly, many mutations isolated in S. enterica serovar Typhimurium defective in σ²⁸/FlgM interactions mapped toσ ²⁸ region 4 (4, 20). This domain is also a target for transcriptional regulation in many other systems (12). Anti-sigma factors Rsd, SpoIIAB, and FlgM bind to region 4 of the sigma factors to inhibit its interaction with RNAP (12).

Previous structural studies showed that when FlgM is bound toσ ²⁸, it could occlude the interaction of this domain with the β-flap of RNAP that is important for proper recognition of the −35 promoter element (22, 30, 41, 45). Further, studies suggest FlgM helix H3′ interacts with region 4 (where E199 of A. aeolicus σ²⁸ is located) on theσ ²⁸-RNAP holoenzyme complex, and either by a conformational change in the interaction with FlgM or through enzyme“ breathing,” region 4 is released for binding by helix H4′ of FlgM. This inhibits region 4 interaction with the RNAPβ -flap and the subsequent dissociation ofσ ²⁸ from RNAP (4, 41). E199 inσ ²⁸ lies within the face of the helix in s₄, where it interacts with H3′ helix of FlgM (41). Isolation of E199G in A. aeolicus σ²⁸ that is resistant to FlgM inhibition supports this model.

Three hundred Tc^s clones were isolated from the A. aeolicus flgM mutagenesis screen, of which 25 were Lac⁺ on Mac-lac indicator medium. Twelve Lac⁺ colonies were sequenced, and eight had either single base pair changes or deletions in FlgM (Fig. 6 and Table 3). The remaining four had multiple mutations, and they were excluded from the study. All strains with A. aeolicus flgM mutations were assayed for σ²⁸ activity by use ofβ -galactosidase assays to compare the levels of fliC-lac expression with a strain harboring wild-type A. aeolicus flgM (Table 3).

Of the mutations characterized, several were found in the ribosome binding site, one altered in the start codon, and another introduced a rare codon; all mutations presumably lowered the levels of FlgM, thereby enabling increased σ²⁸ activity (Fig. 6 and Table 3). Three mutations resulted in frameshifts in FlgM after codon positions 18, 59, and 67. Frameshift mutation after codon 67 suggests that the last 21 amino acids in FlgM are important for inhibition of σ²⁸ activity. Previous studies of FlgM show the C terminus is important forσ ²⁸ interaction (9, 41). Alternatively, the resulting frameshift mutation could destabilize FlgM protein; therefore, additional biochemical methods are needed to support the in vivo findings.

We found a bias in favor of A. aeolicus flgM null mutant isolation, because the screen was based on the loss of anti-σ²⁸ activity. Previous isolations of FlgM mutants of S. enterica serovar Typhimurium defective in interaction with S. enterica serovar Typhimuriumσ ²⁸ indicated a bias against flgM null alleles. This is because S. enterica serovar Typhimurium strains carrying flgM null alleles and remaining fliA⁺ are sick, and excessσ ²⁸ activity is lethal in S. enterica serovar Typhimurium. The FlgM mutants obtained were predominantly single amino acid substitution mutants that retained some anti-σ²⁸ activity (5, 9). Since it was easy to isolate flgM null alleles of A. aeolicus in a strain expressing A. aeolicus σ²⁸, we conclude that excess A. aeolicus σ²⁸ activity, unlike excess S. enterica serovar Typhimuriumσ ²⁸ activity, does not inhibit cell growth.