Cysteine Scanning Mutagenesis and Topological Mapping of the Escherichia coli Twin-Arginine Translocase TatC Component

Bacterial strains, plasmids, and growth conditions.Trimethylamine-N-oxide (TMAO) reductase activity and cysteine labeling and cross-linking experiments were performed in E. coli strain DADE-P [same as strain MC4100 (9) but with ΔtatABCD ΔtatE pcnB1 zad-981::Tn10d (Km^r); 23]. For assessment of in vivo disulfide bond formation, E. coli strain LM107 was used [F⁻ araD139 Δ(araABC-leu)7679 galU galK Δ(lac)X74 rpsL thi phoR zih-12::Tn10 dsbA::Cm (28)]. Plasmids pUNITATCC4 and pUNITATCC4H have been described previously (23). Each plasmid harbors tatA, tatB, and the cysteine codonless derivative of tatC in pQE60 (QIAGEN). Convenient restriction sites have been engineered between tatB and tatC, allowing easy replacement of each individual gene. In addition, plasmid pUNITATCC4H supplies TatC with a C-terminal hexahistidine tag (TatC_His).

A plasmid, pUNITATC, that contains the tatC gene in pBluescript was constructed following amplification of the tatC gene with primers UNIC1 (5′-GCGCGAATTCCTTAAGCATGTCTGTAGAAGATACT-3′) and UNIC2 (5′-GCGCGGATCCTTATTCTTCAGTTTTTTCGCTTTCT-3′; restriction sites underlined), digestion with EcoRI and BamHI, and cloning into pBluescript that had been previously digested with the same enzymes. Site-specific cysteine mutations were initially introduced into the tatC gene in plasmid pUNITATC using the QuikChange method. Each mutant tatC allele was subsequently excised from pUNITATC by digestion with AflII and BamHI and cloned into pUNITATCC4 that had been similarly digested. For cysteine labeling experiments, tatC alleles were cloned into pUNITATCC4H. This was achieved by amplification of the mutated tatC genes in pUNITATC using primers UNIC1 and TATCH2 (13), digestion with AflII and BglII, and cloning into pUNITATCC4H that had been similarly digested. The individual cysteine substitutions used in this work are shown in Table 1.

All clones obtained from PCR-amplified DNA were sequenced to ensure that no undesired mutations had been introduced. The primer sequences for introduction of site-specific mutations are available on request.

During all genetic manipulations and for cysteine cross-linking experiments, E. coli strains were grown aerobically in Luria-Bertani (LB) medium (36). Strains were normally grown at 37°C unless otherwise stated. Concentrations of antibiotics were as described previously (37). The growth phenotypes of mutants with TMAO as the sole respiratory electron acceptor were determined on M9 minimal medium agar plates (36) supplemented with 0.5% glycerol and 0.4% TMAO and incubated in a gas jar under a hydrogen/carbon dioxide atmosphere. The sodium dodecyl sulfate (SDS) resistance phenotype of mutants was tested on LB agar plates containing 2% SDS (8). For TMAO reductase assays, cells were cultured in modified Cohen and Rickenberg medium (11) supplemented with 0.2% (wt/vol) glucose and 0.4% (wt/vol) TMAO. Subcellular fractions for TMAO reductase assays were prepared from small (30-ml) cultures using a cold osmotic shock protocol (41), and TMAO:benzyl viologen oxidoreductase activity was measured as described previously (40). Spheroplasts for sulfhydryl labeling experiments were prepared from 500-ml cultures using the lysozyme-EDTA method (37). For strains harboring plasmids producing the TatC G144C or S148C variant, 5 mM dithiothreitol (DTT) was included in the incubation buffer during spheroplast preparation.

Sulfhydryl labeling of cysteine residues.Labeling experiments were carried out on spheroplasts prepared from strains expressing tatA, tatB, and single cysteine variants of tatC _His. Three independent reactions were carried out for each sample, one where spheroplasts were treated with the membrane-permeable, sulfhydryl-reactive compound, N^α-(3-maleimidylproprionyl)biocytin (MPB; Molecular Probes) the second where spheroplasts were pretreated with the membrane-impermeable blocking reagent 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid (AMS; Molecular Probes) prior to treatment with MPB, and the third, which was an untreated control sample.

For each reaction, intact spheroplasts (400 μg of protein) were resuspended in 500 μl of 50 mM Tris-HCl (pH 7.6) and 0.5 M sucrose. For treatment with AMS, 0.5 mM AMS (in water) was added to the sample, which was incubated for 20 min at room temperature with gentle shaking. Excess AMS was removed by gentle centrifugation, and the spheroplasts were resuspended in 500 μl of 50 mM Tris-HCl (pH 7.6) and 0.5 M sucrose. For reaction with MPB, 1 mM MPB (in dimethyl formamide) was added to the spheroplast suspension, which was incubated for 20 min at room temperature with gentle shaking. The control sample (spheroplast suspension alone) was also incubated for 20 min at room temperature with shaking. After the incubation, all samples were quenched by the addition of 500 μl of 50 mM Tris-HCl (pH 7.6), 0.5 M sucrose, and 100 mM β-mercaptoethanol and incubated for 15 min at room temperature with gentle shaking. Spheroplasts were then disrupted by sonication on ice for five 15-s bursts (10-micron amplitude) with 15-s intervals. Cell debris was removed by centrifugation at 10,000 × g at 4°C for 5 min, and membrane proteins were recovered by ultracentrifugation at 278,000 × g for 30 min. The pellet was resuspended in 100 μl of 50 mM Tris-HCl, pH 6.8, and samples were denatured by the addition of 1% SDS and incubated at 37°C for 15 min prior to immunoprecipitation.

Immunoprecipitation of TatC protein.Protein A Sepharose CL-4B (Amersham) was resuspended to 10 mg/ml in buffer A [25 mM Tris-HCl, pH 6.8, 0.9% (wt/vol) NaCl, 1% Triton X-100, 5 mM EDTA, pH 8.0]. Twenty microliters of anti-TatC antibody was added per ml of protein A Sepharose, and the samples were incubated for 3 h at 4°C with gentle shaking to allow binding of the antibody to protein A. After incubation, unbound antibodies were removed by centrifugation at room temperature at 10,000 × g, and the Sepharose beads were washed twice with buffer A. The pelleted beads were resuspended to 10 mg/ml in buffer A, and 10 mg of antibody-loaded protein A Sepharose was added to each denatured membrane protein fraction. Samples were incubated at 4°C overnight with gentle shaking to allow binding of the antigen to the immunoglobulins. The Sepharose beads were then pelleted by centrifugation at 10,000 × g for 2 min to remove unbound proteins. The pellet was washed twice with 1 ml of buffer A, and the beads were resuspended in 1 ml of 25 mM Tris-HCl, pH 6.8, and again pelleted by centrifugation at 10,000 × g for 2 min. The pellet was finally resuspended in 25 μl standard SDS loading buffer (50 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.1% bromophenol blue, 1% β-mercaptoethanol [final concentrations given]) and denatured by incubation at 56°C for 20 min prior to SDS-PAGE.

Other protein methods.Membrane preparation and disulfide cross-linking were carried out essentially by the methods of Lee et al. (23) with the exception that oxidant-induced cross-linking was carried out for 1 hour at either 20°C or on ice as indicated below. The following modification was included for the reduction of spontaneous cross-linked samples. Membrane fractions in buffer 1 (50 mM Tris-HCl, pH 7.5, 5 mM MgCl₂, 10% (vol/vol) glycerol; 50-μl final volume) were treated with 1 μl of 0.5 M DTT on ice for 1 h. Membranes were washed twice with buffer 1 to remove excess DTT and resuspended in buffer 1. Helical wheels were drawn using the “helical wheel” program of the Wisconsin Package, version 10.2 using the default setting of 3.6 amino acid residues per turn.

Polyclonal anti-TatC antibodies were raised in rabbits using the peptide NH₂-CGKGRNREEENDAEAESEKTEE-COOH, by Affiniti (BIOMOL International, Exeter, United Kingdom). Anti-tetra-His tag antibody (mouse) was obtained from QIAGEN, SDS-PAGE and immunoblotting were carried out as described previously (22, 43), and immunoreactive bands were visualized with the ECL detection system (Amersham Biosciences). MPB-labeled protein was detected using streptavidin-horseradish peroxidase (HRP) (Amersham Biosciences). Protein concentrations were estimated by the method of Lowry et al. (25). Spheroplast intactness was assayed by monitoring the activity of the cytoplasmic marker enzyme glucose-6-phosphate dehydrogenase (3).

Topological analysis of the TatC protein.Topological fusion studies using compartment-specific marker proteins have given conflicting results about the number of transmembrane segments in the E. coli TatC protein. In silico predictions suggest that in general, TatC proteins (with the exception of the much larger fused dimeric TatC proteins of some members of the domain Archaea) possess six transmembrane spans. An initial fusion study suggested that TatC had only four transmembrane domains (18), while two later studies using similar fusion technology suggested that the protein had six transmembrane domains (4, 21). A general criticism of fusion studies is that they can influence the topology of the protein to which they are fused. Moreover, they usually inactivate the protein under study. We therefore sought to investigate the presence or absence of the controversial transmembrane domains four and five (Fig. 1) by assessing the cellular location of the intervening loop region using the noninvasive technique of cysteine substitution and topological labeling.

The cysteine substitutions we constructed to probe the topology of TatC are shown in Fig. 1. We initially constructed control cysteine substitutions at residues L9 (the N terminus of TatC has an unequivocal cytoplasmic location [7]) and G144 (periplasmic location, directly preceding transmembrane domain four). To probe the topological organization of predicted transmembrane helices four and five, we constructed individual cysteine substitutions in the loop region between these two predicted transmembrane domains at positions I183, P186, and K191 (Fig. 1). Each of these substitutions was introduced into a plasmid-encoded tatC _His gene in which all of the native cysteine codons had been replaced by alanine and that was also coexpressed with tatA and tatB. Plasmids were subsequently transformed into a strain lacking all Tat components and that carried the pcnB1 mutation to limit plasmid copy number (24), and the activity of the plasmid-encoded Tat system was assessed by measurement of the periplasmic activity of the Tat substrate protein TMAO reductase.

As shown in Fig. 2A, a Tat system harboring a cysteineless derivative of TatC_His shows more than 90% of the periplasmic TMAO reductase activity of a strain producing the His-tagged but otherwise native TatC. This is in agreement with our previous observations that none of the four native cysteine residues in TatC are essential for function (8, 23). A strain producing a TatC harboring the control substitution at L9C had periplasmic TMAO reductase activity that was about 60 to 70% of that of the wild type, indicating that the Tat system retained significant activity. By contrast, however, a strain producing TatC with the control G144C substitution showed much lower Tat activity (around 20% of that of the wild type). Since we wanted to ensure that our topological labeling studies were carried out on TatC proteins that were functional, we constructed an additional substitution of a residue in periplasmic loop 1, S148C. A strain producing a Tat system harboring this variant of TatC gave periplasmic TMAO reductase levels that were more than 70% of that of the wild type and therefore significantly more active than the G144C variant of TatC. The cysteine substitutions at I183, P186, and K191 designed to test the topology of TatC all gave TatC proteins that supported periplasmic TMAO reductase levels close to that of the wild type, indicating that none of our test substitutions significantly affected the activity and hence the topology of TatC.

• Open in new tab
• Download powerpoint

FIG. 2.

Periplasmic TMAO reductase activities of E. coli tat mutant strains producing TatA, TatB and cysteine-substituted TatC proteins that were used for sulfhydryl labeling (A) or disulfide cross-linking (B and C). Periplasmic TMAO reductase activities were measured from the ΔtatABCD ΔtatE pcnB derivative, DADE-P (Δ), or DADE-P carrying either pUNITAT2 (29) (wild type [WT]) and (A) pUNITATCC4H (Cys⁻) or pUNITATCC4H encoding cysteine-substituted TatC proteins or (B and C) pUNITATCC4 (Cys⁻) or pUNITATCC4 encoding cysteine-substituted TatC proteins. In each case, the amino acid position of the substitution is shown under each column. One hundred percent activity was taken to be that determined for E. coli strain DADE-P carrying pUNITAT1 and corresponds to an activity of typically 0.85 μmol benzyl viologen oxidized/min/mg protein. In panel B, the substitutions in transmembrane helix one (TMH1) to transmembrane helix six (TMH6) are shown. The error bars represent the standard errors of the means (n = 3 or 4).

Western blot analysis (Fig. 3) showed that each of the TatC variants was present in the membrane at a level similar to that of the wild-type protein. Interestingly, in the absence of a reducing agent, the TatC protein harboring the G144C substitution was found almost 100% in a disulfide-bonded dimer form. A small amount of the S148C TatC variant was also present as a dimer. These results suggest that the first periplasmic loop of TatC is found in close proximity (3 to 4 Å) to the same loop from a neighboring TatC and strongly suggests that TatC is arranged as at least a homodimer. It is also consistent with the fact that both of these residues are exposed at the periplasmic face of the membrane, where they would be accessible to the disulfide bond insertion machinery. However, we note that the level of disulfide-bonded dimer of the TatC G144C variant in whole cells of a dsbA strain was similar to that in the wild-type (dsbA ⁺) strain (data not shown).

• Open in new tab
• Download powerpoint

FIG. 3.

TatC variant G144C forms a strong cross-linked dimer in untreated membranes. Western blot analysis of membranes isolated from the ΔtatABCD ΔtatE pcnB strain, E. coli strain DADE-P harboring plasmid producing TatA, TatB, and either wild-type TatC_His (pUNITAT2), the TatC_His cysteine-less derivative (pUNITATCC4H), or single cysteine substitutions as indicated. Samples were analyzed as isolated or were reduced by treatment with 10 mM DTT for 1 h (indicated by the symbols − and +, respectively, above each lane), and all samples were analyzed by SDS-PAGE under nonreducing conditions.

In order to probe the location of the cysteine residue in each of the substituted TatC proteins, we used the following procedure. MPB is a membrane-permeable sulfhydryl-reactive compound that will react with cysteine residues regardless of their topological location (42). Thus, for intact spheroplasts (right-side-out membrane vesicles) it will react with cysteine residues that are located on the inside (cytoplasmic) or the outside (periplasmic location). Since it labels cysteine residues with a biotin moiety, labeling can be detected by probing with streptavidin-conjugated HRP during Western blotting. By contrast, AMS is a membrane-impermeable sulfhydryl-reactive compound that should react only with cysteine residues on the outer face of the cytoplasmic membrane (19). Thus, if spheroplasts are preincubated with an excess of AMS, cysteine residues that are exposed to the outside will react with this compound. Subsequent exposure to MPB will result in labeling of any remaining free cysteine residues. Therefore, cysteine residues that are inaccessible to AMS (on the inside of spheroplasts) will label with MPB, whereas those that are AMS accessible (i.e., on the outside) will be blocked from MPB labeling.

Patterns of MPB labeling of spheroplasts prepared from cells expressing a cysteineless variant of TatC along with single TatC variants L9C, G144C, and S148C are shown in Fig. 4A. For a TatC protein lacking cysteine residues, there is no labeling of the protein, showing that MPB labeling is dependent on the presence of cysteine. For TatC variant L9C, this residue was clearly labeled by MPB, regardless of whether spheroplasts containing this TatC derivative had been previously exposed to AMS. This pattern of labeling is consistent with a cytoplasmic location for this residue. It should be noted, however, that there was some reduction in the intensity of labeling with MPB after preexposure to the AMS block. This might be explained in part at least by the fact that there was some low-level breakage of spheroplasts during their preparation (data not shown). For TatC variant S148C, the labeling of the cysteine residue with MPB was totally blocked by prior exposure to AMS, confirming that this residue is exposed at the periplasmic side of the membrane. For TatC variant G144C, where we reduced the disulfide bond with DTT prior to the labeling experiments, MPB labeling of the cysteine was also blocked by preincubation with AMS, consistent with the exposure of this residue to the periplasmic face of the cytoplasmic membrane.

• Open in new tab
• Download powerpoint

FIG. 4.

Sulfhydryl labeling of spheroplasts overproducing TatA, TatB, and cysteine variants of TatC. Spheroplasts were incubated with buffer alone (−), with MPB (lab), or with AMS followed by MPB (block), as described in Materials and Methods. Spheroplasts were subsequently lysed, and membranes were isolated, dispersed with SDS, and immunoprecipitated with anti-TatC antisera. Samples were separated by SDS-PAGE (12.5% acrylamide) and electroblotted, and MPB-labeled protein was detected with a 1:5,000 dilution of streptavidin-HRP (Strep). The same blot was then stripped and reprobed with a 1:5000 dilution of anti-tetra-His (α-His) antibody (QIAGEN) to detect TatC protein. (A) Labeling reactions with control samples expressing a cysteine-less TatC variant and the single cysteine substitutions L9C (cytoplasmic location) and G144C and S148C (both periplasmic [peri] location). (B) Labeling reactions with TatC cysteine variants designed to probe the cellular location of the loop region between transmembrane helices four and five: I183C, P186C, and I191C. The positions of these residues on a cartoon of TatC are shown at the right-hand side of each panel (cytoplasmic [cyto] or periplasmic [peri] location).

In order to test the location of the loop region between predicted transmembrane helices four and five, we next incubated spheroplasts harboring either the I183C, P186C, or K191C variants of TatC with one or both of the cysteine-reactive compounds. As shown in Fig. 4B, each of these cysteine residues was labeled when spheroplasts were incubated with MPB. Moreover, since labeling with MPB was not prevented by prior incubation of spheroplasts with the membrane-impermeable AMS reagent, these observations strongly suggest that each of these three residues are located at the cytoplasmic side of the membrane. Taken together, these results support in silico predictions (6, 15) and the more recent fusion studies (4, 21) which suggest that the E. coli TatC protein has six membrane-spanning domains.

Cysteine scanning mutagenesis of TatC.In order to probe the functions of residues within the transmembrane domains of TatC and as a prerequisite for disulfide mapping experiments, we constructed single cysteine substitutions in each of the six transmembrane domains of TatC. We initially selected six consecutive residues predicted to be towards the center of each transmembrane domain, as judged following alignment of multiple TatC sequences and then manual assignment of the positions of each of the transmembrane α-helices. These six residues were designed to cover almost two turns of a putative α-helix and were individually replaced by cysteine. The position of each of these substitutions is shown in Fig. 1. In transmembrane helix one, we constructed individual substitutions of I28, V29, I30, F31, L32, and C33 (which was originally alanine in our cysteineless derivative of TatC), in transmembrane helix two, we constructed substitutions of I81, L82, S83, A84, P85, and V86, in transmembrane helix three F124, A125, Y126, F127, V128, and V129, in transmembrane helix four G166, V167, S168, F169, E170, and V171, in transmembrane helix five A200, F201, V202, V203, G204, and M205, and in transmembrane helix six E227, I228, G229, V230, F231, and F232. Each of these substitutions was introduced into a plasmid-encoded tatC gene where all of the native cysteine codons had been replaced by alanine and that was also coexpressed with tatA and tatB. As above, in order to assess the functionality of Tat systems harboring TatC variants with each of these amino acid substitutions, plasmids were transformed into a strain lacking all Tat components and harboring the pcnB1 allele, and the periplasmic activity of TMAO reductase was assessed.

As shown in Fig. 2B, strains with Tat systems harboring almost all of the cysteine substitutions showed periplasmic TMAO reductase activities that were greater than 50% of the activity of the wild type. The few exceptions to this were TatC V128C and TatC V203C substitutions (40 to 50% of wild-type activity) and TatC Y126C, which had very low activity. It should be noted that a Y126A substitution of TatC also similarly resulted in very low Tat activity (8), indicating that the presence of a tyrosine at this position, which is highly conserved among TatC proteins, is important for TatC function. We conclude that in general, cysteine substitution of transmembrane residues of TatC is well tolerated.

Disulfide mapping studies of TatC using single cysteine variants.Once we had constructed a bank of cysteine substitutions, we used them to map contact sites between TatC proteins. Disulfide mapping studies with TatB gave patterns of cross-linking that were consistent with the likelihood that TatB formed a homo-oligomeric ring located in the center of the TatBC complex (23). Given the probability that TatC is localized on the periphery, this in principle could give rise to two different types of cross-links. Cross-links could arise because the cysteines are located at contact sites between TatC monomers within the TatBC complex. Alternatively, cross-links could occur between TatC proteins that are present in different TatBC complexes. One way that these may be distinguished is to examine the effect of protein mobility on disulfide bridge formation. Cross-links that occur between different complexes are more susceptible to conditions that slow down the motion of complexes in the membrane, whereas cross-links within a complex are less so (32). One way that membrane fluidity can be altered is by lowering the temperature, which results in the reversible change of state from a disordered to an ordered array of fatty acyl chains (14, 26). Therefore, we undertook cross-linking experiments at room temperature and on ice in order to distinguish cross-links likely to be formed between complexes from those formed within a single complex.

For cross-linking analysis, a plasmid encoding tatA, tatB and single cysteine-codon-substituted tatC was expressed in an otherwise tat mutant strain, membranes were prepared, and the ability of the single cysteine residue to form a disulfide bond with the identical residue in a neighboring TatC protein was assessed after exposure to oxidizing conditions. As shown in Fig. 3, we observed an almost 100% disulfide cross-linked dimeric form of TatC variant G144C in untreated membranes, demonstrating that this part of TatC is in close proximity with a neighbor. To test whether this cross-link was likely to be between TatC proteins in the same TatBC complex, we reduced the disulfide bond and then treated the sample with the oxidant copper phenanthroline at either room temperature (20°C) or on ice. As shown in Fig. 5A, disulfide-bonded dimer was formed to similar extents at both ambient and low temperatures, strongly suggesting that the cross-link was between TatC protomers within a TatBC complex.

• Open in new tab
• Download powerpoint

FIG. 5.

The TatC G144C, A26C, and Y36C variants form disulfide cross-links when oxidized at room temperature or on ice. Membrane samples were prepared from the ΔtatABCD ΔtatE pcnB strain, DADE-P, coexpressing single cysteine-substituted TatC variants G144C (A), A26C (B), and Y36C (C) together with TatA and TatB from pUNITATCC4. For membranes harboring the TatC G144C or A26C substitution, a sample as prepared was retained (lanes marked C), and the remainder of the membranes were treated with DTT to reduce preformed dimer as described in Materials and Methods. Samples (100 μg membrane protein) were then subjected to oxidizing conditions (by the addition of copper phenanthroline) at either room temperature (ORT) or on ice (O0°C). A sample of the DTT-treated membranes are also shown (lanes marked R). For membranes harboring the TatC Y36C variant (which shows no disulfide-bonded dimer in the membrane as isolated), samples (100 μg membrane protein) were incubated for one hour with either buffer alone at room temperature (C) or oxidant at either room temperature (ORT) or on ice (O0°C). Samples (10 μg membrane protein) were resolved by SDS-PAGE (12.5% acrylamide), and TatC proteins were visualized by immunoblotting with anti-TatC antisera.

Figure 6 summarizes the results we obtained for disulfide cross-linking, performed at room temperature, of each of the 36 cysteine-containing TatC variants with substitutions in transmembrane regions. It is clear (Fig. 6A) that none of the cysteine substitutions in transmembrane helix one were in close enough proximity to similar residues in other TatC molecules to form a disulfide cross-link. By contrast, cross-linking analysis of cysteine residues in TatC transmembrane helix two (Fig. 6B) showed two positions where the introduction of cysteine residues and exposure to oxidizing agent resulted in the formation of a disulfide-bonded dimer. A TatC variant harboring a cysteine substitution at residue 82 produced a low level of cross-linked dimer, while a substitution at position 86 gave an almost complete cross-link to dimer form. Plotting the positions of those residues on a helical wheel projection of this part of the protein, depicted in Fig. 7A, shows that both of these residues would fall on the same helical face, indicating that there is facial specificity for the self-interaction of this portion of transmembrane helix two. None of these cross-links were observed when oxidation was carried out on ice (not shown), which suggests that interactions are probably between TatC proteins located in separate TatBC complexes.

• Open in new tab
• Download powerpoint

FIG. 6.

Immunoblot analysis of cross-linked E. coli TatC proteins containing cysteine substitutions in each transmembrane helix. Membrane samples were prepared from the ΔtatABCD ΔtatE pcnB strain, DADE-P, coexpressing single cysteine-substituted TatC variants together with TatA and TatB from pUNITATCC4. Samples (100 μg membrane protein) were subjected to either oxidizing conditions (O; shown in the leftmost lane for each sample) or reducing conditions (R; middle lane for each sample) or incubated with buffer alone (C; rightmost lane for each sample) each at room temperature as described in Materials and Methods. Samples (10 μg membrane protein) were resolved by SDS-PAGE (12.5% acrylamide), and TatC proteins were visualized by immunoblotting with anti-TatC antisera. The position of the cysteine substitution is shown above each panel. The positions of TatC monomer (*) and dimer (**) forms are indicated. (A) Substitutions in transmembrane helix one; (B) substitutions in transmembrane helix two; (C) substitutions in transmembrane helix three; (D) substitutions in transmembrane helix four; (E) substitutions in transmembrane helix five; (F) substitutions in transmembrane helix six.

• Open in new tab
• Download powerpoint

FIG. 7.

Helical wheel projections of portions of TatC transmembrane helix two (TMH2) to transmembrane helix six (TMH6) that were analyzed by disulfide cross-linking. The percentage next to each residue is the percentage of cross-linked dimer (expressed as a percentage of monomer plus dimer) formed after oxidation at room temperature for 1 h.

The results of disulfide cross-linking experiments of TatC variants expressing single cysteine substitutions in transmembrane helix three are shown in Fig. 6C. The first three residues that we tested (positions 124 to 126) did not show any cross-linking, while the latter three residues (127 to 129) all showed some level of dimer formation in the presence of oxidant. Reasonably strong dimer bands were observed for TatC F127C and V129C, and a weaker cross-link was observed for V128C. When these residues are plotted onto a helical wheel representation, shown in Fig. 7B, there is no obvious facial specificity to the interaction. Moreover, the fact that the first three residues that we tested failed to self-cross-link suggests that these helices in the TatC interface are tilted relative to each other, assuming that this portion of the protein is helical. Again, all cross-links were strongly influenced by mobility because they were lost when oxidation was carried out at lower temperatures (not shown), and we therefore conclude that they are probably formed between different complexes.

A surprisingly large number of self-cross-links were also observed when TatC cysteine variants in transmembrane helix four were exposed to oxidant (Fig. 6D). A rather strong cross-link was seen for the cysteine substitution at residue 169, and a slightly weaker one was seen at G166, both of which would cluster on one face of a α-helix (Fig. 7C). Cross-links were also observed for cysteine substitutions at residues 167 and 171, which would also cluster together on one face of a putative α-helix (Fig. 7C). No cross-links were detected for substitutions at positions 168 and 170, and all observed cross-links were again lost when samples were oxidized on ice (not shown), which indicates that they are formed between complexes. The positions of the cross-linking residues on a helical wheel representation (Fig. 7C) are consistent with two, almost completely opposite, faces of self-interaction for transmembrane helix four.

Analysis of patterns of disulfide cross-linking for TatC cysteine variants in transmembrane helix five are shown in Fig. 6E. Again, an unexpectedly large number of cross-links was detected when samples were oxidized at room temperature, and again, none of these were observed when samples were oxidized at 0°C, suggesting that they are primarily between different TatBC complexes. TatC F201C and G204C gave strong cross-linked dimer bands under oxidizing conditions, whereas TatC V203C and M205C gave much weaker cross-linked dimer bands. TatC A200C and V202C did not form cross-linked dimers. A plot of these residues on a helical wheel projection (Fig. 7D) again shows at least two faces of self-interaction—a strongly interacting face comprising F201 and G204, possibly extending as far as M205, and a weaker face comprising V203.

Figures 6F and 7E summarize the cross-linking results we observed when TatC variants with cysteine substitutions in transmembrane helix six were incubated at room temperature in the presence of oxidant. Surprisingly, every one of the six consecutive cysteine substitutions that we tested showed formation of a significant level of self-cross-linked dimer under oxidizing conditions. These observations suggest that this part of the protein in particular must be highly mobile to allow each of these residues into close proximity with their counterparts in other TatC proteins. Again no cross-links were seen after oxidation on ice (not shown), emphasizing that mobility is essential for cross-link formation and suggesting that the interactions we detect are likely to be between different TatC-containing complexes.

Cysteine scanning mutagenesis and disulfide cross-linking analysis of the remaining residues in TatC transmembrane helix one.Our cysteine cross-linking analysis showed that each of the putative transmembrane helices of TatC exhibited self-interaction with the exception of the first transmembrane helix. These observations would be consistent with the possibility that this part of TatC is tightly buried within the TatC monomer, thus preventing self-contact. In order to test this hypothesis, we constructed a further 15 individual cysteine substitutions in transmembrane helix one, covering residues L20 to V27 and L34 to D40, respectively (Fig. 1).

As shown in Fig. 2C, each of these substitutions gave rise to a TatC protein that was competent to support protein transport. Substitutions in the first part of the helix (positions 20 to 23) gave TatC proteins that were almost as active as the wild type. A previous study showed that L20A substitution of TatC was also active, although a deletion of residues 20 through 22 did inactivate the protein (2). Substitutions further along the helix, in particular the alanine residues at positions 26 and 38, had quite a marked effect on Tat-dependent transport, but even the most severely affected mutations still supported around 25% of the wild-type Tat activity.

The results of disulfide cross-linking analysis, undertaken at room temperature, on these additional cysteine-substituted TatC proteins are shown in Fig. 8. Residues V27C and L34C, which are immediately adjacent to the six non-cross-linking residues of TatC that we tested above, similarly show no or very little self-cross-linking. However, many of the substitutions towards the ends of the helix in either direction show significant disulfide bond formation upon oxidation. The ends of helices in general have greater mobility, and similar nonperiodic cross-linking was also seen with cysteine substitutions at either end of the TatB transmembrane helix (23). Residue A26C also formed a disulfide-bonded dimer prior to the addition of catalyst, suggesting that this residue may be uniquely exposed to an oxidizing environment in the membrane. These observations are therefore not consistent with the idea that transmembrane helix one is buried within the TatC monomer.

• Open in new tab
• Download powerpoint

FIG. 8.

Immunoblot analysis of cross-linked E. coli TatC proteins containing further cysteine substitutions in transmembrane helix one. Membrane samples were prepared from the ΔtatABCD ΔtatE pcnB strain, DADE-P, coexpressing single cysteine-substituted TatC variants together with TatA and TatB from pUNITATCC4. Samples (100 μg membrane protein) were subjected to either oxidizing conditions (O; shown in the leftmost lane for each sample) or reducing conditions (R; middle lane for each sample) or incubated with buffer alone (C; rightmost lane for each sample), each at room temperature. Samples (10 μg membrane protein) were resolved by SDS-PAGE (12.5% acrylamide), and TatC proteins were visualized by immunoblotting with anti-TatC antisera. The position of the cysteine substitution is shown above each panel. The positions of TatC monomer (*) and dimer (**) forms are indicated.

When the cross-linking analysis was repeated at 0°C, most of the cross-links were lost with the exception of homodimeric cross-linking of residues A26C and Y36C (shown in Fig. 5B and C) which still gave significant dimer formation at low temperatures. Therefore, we conclude that most of the observed cross-links are probably between TatC proteins located within different TatBC complexes with the exception of the A26C and Y36C variants which are likely to be cross-links within a single complex. A helical wheel projection of the entire transmembrane helix one is shown in Fig. 9, with the percentage of dimer formation (after oxidation at room temperature) indicated for those residues that show disulfide cross-linking. Residues A26 and Y36 (Fig. 9) do not fall on the same face of a regular α-helix, which suggests either that this helix may be twisted or that this helix can self-interact with more than one TatC partner.

• Open in new tab
• Download powerpoint

FIG. 9.

Helical wheel projection for the entire TatC transmembrane helix one. For those residues that show cross-linking, the percentage of cross-linked dimer (expressed as a percentage of monomer plus dimer) formed after oxidation at room temperature for 1 h is indicated. Residues A26 and Y36 are shown boxed.