Characterization of Streptococcus gordonii SecA2 as a Paralogue of SecA^{▿ †}

Bacterial strains, plasmids, and reagents.

The bacterial strains and plasmids used in this study are listed in Table 1. S. gordonii strains were grown in Todd-Hewitt broth (THB; Difco Laboratories) at 37°C in a 5% CO₂ environment. An M99-derived strain that expresses GspB736flag was constructed as previously described (5) but using chloramphenicol as a selectable marker. This enabled the simultaneous use of the nisin-inducible complementation vector pMSP3535, which encodes resistance to erythromycin. A segment of plasmid pB736flag, which includes codons 604 to 736 of GspB fused in-frame to a 3X FLAG (Sigma) coding sequence, was removed by digestion with HindIII and BamHI and then cloned in the suicide vector pEVP3, which had been digested similarly. The resulting plasmid, p736BflagC, was used to transform the secA2 mutant strain PS516 by natural transformation as described previously (4). Transformants were initially selected on sheep blood agar containing 5 μg/ml chloramphenicol. Subsequent retention of the chromosomally integrated plasmid did not require antibiotic selection.

Oligonucleotide primers (Table 2) were synthesized by Sigma-Genosys. Peroxidase-conjugated antibodies and anti-FLAG monoclonal antibodies were purchased from Sigma. Antibiotics and other chemicals were obtained from Sigma or Fisher.

Targeted substitutions within SecA2.

The plasmid pM99secA2 contains the entire secA2 gene, along with an optimized ribosome binding site, in the nisin-inducible expression vector pMSP3535 (2). Mutations in the secA2 sequence were generated by a two-stage PCR procedure. In the first stage, primers 5′A2 and a reverse mutant primer (Table 2) or the complementary forward mutant primer along with primer 3′A2 were used to amplify the 5′- or 3′-end segment of secA2, respectively. The two PCR products were combined for the second stage and then amplified using primers 5′A2 and 3′A2. The 2.5-kb PCR product was then cloned in pMSP3535 or, when possible, restriction fragments thereof were used to replace the corresponding segment of pM99secA2. In some cases, silent mutations that added or removed a restriction site were included in the primer sequence in order to facilitate tracking of the mutation (Table 2). The presence of only the intended alteration in any segment generated by PCR was confirmed by DNA sequence analysis (Sequetech). Plasmids were introduced to PS1226 by natural transformation, and transformants were selected and maintained using 60 μg/ml erythromycin.

Construction of SecA-SecA2 chimeras.

Chimeric secA-secA2 constructs were generated by a two-stage PCR procedure. For the SecA-SecA2 chimera designated SecA2^1-534A^535-838, the ribosome binding site and codons 1 to 534 of secA2 were amplified using primer pairs 5′A2 and IRA2R and codons 535 to 838 of secA were amplified using primer pairs IRA2F and 3′A. The two PCR products were combined and then amplified by overlap extension using primers 5′A2 and 3′A. The SecA^1-534A2^535-793, SecA^1-222A2^220-793, and SecA^1-121A2^114-793 chimeric sequences were generated by a similar process, using the following primer pairs, respectively, for the first-stage reactions: 5′A and IRA2R or IRA2F and 3′A2, 5′A and A1NBDR or A2SSDF and 3′A2, or 5′A and NBD3R or NBD3F and 3′A2. Primers 5′A and 3′A2 were then used for the second-stage PCRs. The second-stage PCR products were cloned in pMSP3535 and used to transform PS1226 as described above.

Analysis of secreted and protoplast proteins.

Proteins from S. gordonii culture supernatants or protoplasted cells were prepared and analyzed by Western blotting as described previously (3), with minor modifications. Strains were grown for 18 h in THB containing erythromycin (60 μg/ml) and then either fractionated immediately or diluted fivefold into fresh THB containing nisin (200 ng/ml) and incubated for 6 h in the absence of erythromycin (the final cell density of all cultures was approximately 8 × 10⁸ CFU per ml). For analysis of secreted products, cells were removed by centrifugation at 16,000 × g for 5 min, and the spent medium was combined directly with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and heated at 70°C for 10 min. For analysis of cytoplasmic components, protoplasts (generated by digestion of the cell wall with mutanolysin) were lysed by suspension in SDS-PAGE sample buffer, followed by boiling for 5 min. When probing for GspB736flag, the proteins present in 25 μl of spent culture medium or in protoplasts recovered from 25 μl of the culture were loaded onto gels and compared by Western blotting for the relative amounts of the secreted and the nonsecreted protein. Anti-FLAG monoclonal antibody was used at 1 μg/ml, along with peroxidase-conjugated anti-mouse antibodies at a 1:10,000 dilution. A 50-kDa cytoplasmic protein that cross-reacts with the anti-FLAG antibody served as a loading control (not shown). For detection of the SecA2 variants and SecA-SecA2 chimeras, a rabbit polyclonal anti-SecA2 serum was used at a 1:2,000 dilution, along with peroxidase-conjugated anti-rabbit antibodies at a 1:25,000 dilution.

Effect of sodium azide on secretion of GspB736flag.

The inhibition of protein transport by sodium azide was performed as described by Jongbloed et al. (24), with modifications. Strains were grown for 18 h in 4 ml of THB containing 60 μg/ml erythromycin. After centrifugation at 3,200 × g for 10 min, cells were suspended in 2 ml fresh medium, and then 500-μl aliquots were combined immediately with 500 μl of THB containing either 0, 3, 9, or 30 mM sodium azide. The cultures were incubated at 37°C for 90 min, and then the medium was prepared and analyzed for GspB736flag content as described above.

Overexpression and purification of SecA and SecA2.

For expression of N-terminally His₆-tagged SecA2_Sg (hereinafter referred to as 6His-SecA2), the secA2 gene was removed from pM99secA2 by digestion of the plasmid with BamHI and SalI and was cloned into pET28c (Novagen) that had been similarly digested, forming pET28c.secA2. This resulted in an in-frame fusion of secA2 with the His₆ coding sequence, along with a short spacer sequence. Plasmids used to express the secA-secA2 chimeras were constructed in the same manner. A SecA2^1-707 expression plasmid was generated by deleting a HindIII fragment spanning the 3′ end of secA2 from pET28c.secA2. For 6His-SecA, the entire secA gene was amplified from strain M99 chromosomal DNA using primers 5′A and 3′A. The 2.5-kb PCR product was digested with BamHI and XhoI and ligated to pET28c that had been digested similarly, forming pET28c.secA. Plasmids were propagated initially in E. coli TOP10 cells (Invitrogen), sequenced to ensure the absence of any mutations in secA or secA2, and then introduced to BL21(DE3) cells (Novagen). For expression of maltose binding protein (MalE) fusion proteins, secA or secA2 was amplified from pET28c.secA or pET28c.secA2 using primers 5′MBPA and T7t or 5′MBPA2 and T7t, respectively. The PCR products were digested with BamHI and XhoI (secA) or BamHI and SalI (secA2) and then ligated to pMAL-c2X (New England Biolabs) that had been digested with BamHI and SalI. The sequences were verified as described above, and the plasmids were then used to transform the E. coli strain BL21 (Novagen).

To express the fusion proteins, strains were grown in Luria broth containing 50 μg/ml kanamycin (pET constructs) or carbenicillin (pMAL constructs) at 25°C to an optical density at 600 nm of 0.7. Expression was then induced by the addition of isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 0.1 mM, and incubation was continued for an additional 4 h. Cells were harvested by centrifugation at 3,200 × g, and the pellets were stored frozen at −80°C. For purification of His₆-tagged proteins, cell pellets were suspended in 5 mM imidazole, 500 mM NaCl, 20 mM Tris, pH 7.5, and cells were disrupted by sonication. After removal of cellular debris by centrifugation for 15 min at 3,200 × g, the cell lysate was filtered through a 0.45-μm filter and applied to Ni-nitrilotriacetic acid agarose (Qiagen). The resin was washed extensively with 30 mM imidazole, 500 mM NaCl, 20 mM Tris, pH 7.5, and the bound protein was then eluted into 100 mM EDTA, 500 mM NaCl, 20 mM Tris, pH 6.5. Where indicated, 0.1% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS) was included in all buffers used for lysis and purification.

For purification of MalE fusion proteins, cell pellets were suspended in TBS (150 mM NaCl, 50 mM Tris, pH 7.5). Cells were disrupted by sonication, and the insoluble debris was removed as described above. The lysate was applied to amylose resin (New England Biolabs), which was then washed extensively with TBS. The bound protein was eluted into TBS containing 10 mM maltose.

The eluted 6His-SecA, MalE-SecA, and MalE-SecA2 proteins were concentrated and exchanged into TK buffer (100 mM KCl, 50 mM Tris, pH 7) in a 100-kDa-cutoff centrifugal filter device (Amicon). 6His-SecA, 6His-SecA2, and His₆-tagged chimeras that had been purified in the presence of CHAPS were exchanged into TK buffer containing 400 mM NaCl and 0.1% CHAPS. Protein concentrations were determined by using Quick Start Bradford dye reagent (Bio-Rad), with bovine serum albumin as a standard.

Production of anti-SecA2.

A rabbit polyclonal antiserum was generated against polyacrylamide gel slices containing 6His-SecA2 (Covance).

ATPase assays.

ATP hydrolysis rates were measured by a modified malachite green assay (30), using PiColorLock Gold reagent as described by the manufacturer (Innova Biosciences). In brief, enzymes were diluted to 0.2 μM in TK buffer, and 100-μl (20 pmol) amounts were added to wells of a microtiter plate. Control wells contained buffer only. For measurement of the effect of azide on ATPase activity, the enzymes were diluted into TK buffer containing 3, 9, or 30 mM sodium azide. Reactions were initiated by the addition of 100 μl of a 2× substrate solution (100 mM Tris, pH 7.5, 0.5 mM ATP) that included 0, 0.2, 1, or 5 mM MgCl₂, yielding final Mg²⁺ concentrations of 0, 0.1, 0.5, and 2.5 mM, as indicated in Fig. 4. Plates were incubated at 30°C for 30 min, and the amount of inorganic phosphate (P_i) released was determined by adding the PiColorLock Gold reagent, measuring the absorbance at 635 nm, and extrapolating from a P_i standard curve. Enzymatic rates were calculated after subtraction of amounts of P_i in control wells that had no enzyme.

SecA2_Sg is highly similar to SecA_Ec.

SecA_Ec is comprised of four major structural domains (Fig. 1) which together facilitate interactions with ATP, preproteins, the SecB chaperone, and the SecYEG translocon (recently reviewed in references 14 and 37). Two discontinuous regions of SecA_Ec, the nucleotide binding folds NBD and IRA2 (also referred to as NBF1 and NBF2), form two faces of a nucleotide binding cleft and are responsible for the binding and hydrolysis of ATP. A preprotein cross-linking domain (PPXD) extends out of NBD. The C domain includes a helical scaffold domain (HSD), a helical wing domain (HWD), an intramolecular regulator of ATPase (IRA1), and the extreme carboxy-terminal region. The scaffold domain serves as a platform to which the two nucleotide binding folds are anchored. The IRA1 region has been shown to coordinate preprotein binding and ATP hydrolysis (54). The same region has recently been referred to as a “two-helix finger” domain which may push the preprotein through the SecYEG channel (15, 56). The extreme carboxy-terminal region of SecA_Ec interacts with SecB and may regulate or increase the fidelity of preprotein binding (16, 18). Residues within nearly every domain of SecA_Ec have been implicated in SecY interactions during the process of translocation (23, 31, 54, 56).

A comparison of the primary amino acid sequences of SecA_Ec, SecA_Sg, and SecA2_Sg, together with structural predictions, indicates that the overall domain configuration of the two S. gordonii paralogues is nearly identical to that of SecA_Ec, having two nucleotide binding folds, a preprotein binding domain located within NBD, and HSD and HWD domains (Fig. 1; see also Fig. S1 in the supplemental material). Perhaps the most significant difference between SecA2_Sg and the SecA orthologues is the absence of a SecB binding site at the extreme carboxy-terminal end of SecA2_Sg. This is not entirely surprising, since homologues of SecB have not been detected in gram-positive bacteria.

A comparison of the primary amino acid sequence alignment also indicates that SecA2_Sg has most of the conserved residues known to be essential for SecA_Ec function (Table 3; see also Fig. S1 in the supplemental material). Of note, SecA2_Sg has the nine conserved motifs (Q, I, Ia, Ib, and II through VI) known to line the nucleotide binding cleft of SecA_Ec and other helicases (35, 44), although the IRA2 motifs IV and VI have nonconserved residues. The alignment also shows that SecA2_Sg has a tyrosine residue in the PPXD that is essential for preprotein binding and release by SecA_Ec (27) and a tryptophan residue in IRA1 that coordinates preprotein binding with ATP hydrolysis (54). This suggests that SecA2_Sg is likely to coordinate the activities of preprotein binding and ATP hydrolysis in a manner similar to that of SecA_Ec.

Selected conserved residues are essential for the function of SecA2_Sg.

The conservation of key amino acids suggested that SecA2_Sg functions analogously to SecA_Ec and, if so, that these residues would also be essential for the proper functioning of SecA2_Sg. To test this directly, we made substitutions for conserved residues in NBD, PPXD, IRA2, or IRA1 (Fig. 2A) whose functions have been characterized in SecA_Ec and examined the effects on transport of GspB. We have previously used a carboxy-terminal FLAG-tagged truncated version of GspB (GspB736flag) as a model substrate for the accessory Sec system (3, 5). GspB736flag lacks the C-terminal cell wall anchoring motif and is freely secreted into the culture medium. As for native GspB, export of this truncated protein is dependent on a functional accessory Sec system. Mutations were generated in the secA2 sequence of the nisin-inducible expression plasmid pM99secA2, and the variant alleles were used to complement a secA2 mutation in the GspB736flag-expressing strain PS1226.

In a PS1226-derived strain carrying the vector only (Fig. 2B, lane 1), GspB736flag was not exported into the culture medium, and the preprotein was retained within the protoplasted cells. The nonexported, glycosylated preprotein appeared as a single band at 150 kDa. When complemented with wild-type secA2 in trans, GspB736flag was secreted into the culture medium, and very little of the preprotein was retained in the protoplasts (Fig. 2B, lane 2). As described previously (5), secreted GspB736flag appeared as a broad band of glycosylated forms that migrated under SDS-PAGE with an apparent molecular mass of 130 to 150 kDa, along with a minor amount of a nonglycosylated form having an apparent molecular mass equivalent to the predicted mass of 80 kDa. The SecA2^E205A (Fig. 2B, lane 3) and SecA2^Y303A (lane 4) variants, which have alterations in NBD and PPXD, respectively, failed to restore transport. A conservative arginine-to-lysine substitution for a residue in IRA2 of SecA_Ec (R509) has been shown to abolish protein transport (32). However, the analogous mutation (R490K) in SecA2_Sg had no effect on transport (Fig. 2B, lane 5). In contrast, a nonconservative substitution for the same residue (R490A) abolished the ability of SecA2 to support translocation (Fig. 2B, lane 6). Another residue in IRA2 of SecA_Ec (D512) has been noted to be important for nucleotide binding (53). A nonconservative substitution for the analogous residue in SecA2_Sg (D493A), however, resulted in only a minor decrease in GspB736flag export (Fig. 2B, lane 7). Lastly, a substitution for the conserved tryptophan residue in IRA1 (Fig. 2B, lane 8) resulted in a modest reduction in transport, which parallels the effect of mutating the analogous residue in SecA_Ec (54). None of the transport defects could be attributed simply to differences in SecA2 expression levels, as equivalent amounts of the wild-type and mutant SecA2 proteins were detected upon Western blotting of the protoplasts (Fig. 2C).

In summary, substitutions for residues in NBD, PPXD, and IRA1 generally had effects on transport comparable to those that have been reported for SecA_Ec. However, replacement of residues in IRA2 had a less dramatic effect on export than did replacement of the analogous residues of SecA_Ec. These data indicate that SecA2_Sg functions in a manner similar to SecA_Ec but also suggest that the IRA2 domain might interact with NBD or nucleotides somewhat differently than SecA_Ec.

SecA2-mediated transport is sensitive to azide.

Azide is a potent inhibitor of the ATPase activity of SecA. This small molecule mimics the γ phosphate of ATP and stabilizes the ADP-bound state of the enzyme (6). Whereas azide levels of 2 mM are sufficient to suppress the general Sec system of E. coli, streptococci can readily grow in medium containing this concentration of azide, and 15 mM azide or more has been used to inhibit the general Sec system of Bacillus subtilis and streptococci (5, 10, 24). It is unknown, however, whether these differences in the amount of azide required to inhibit protein transport are due to an intrinsic difference in the sensitivity of the respective SecA proteins to azide. To better assess the impact of azide on SecA2_Sg function, we first examined the effect of azide on SecA-dependent export in S. gordonii. We chose to examine the secretion of two GspB736flag variants with modified N-terminal signal sequences that we have previously characterized. The first variant consists of GspB736flag fused to a canonical signal peptide (5). The second GspB736flag variant has undergone replacement of the three glycine residues in the hydrophobic core of the signal sequence, along with a deletion of the extended N region. Alteration of the hydrophobic core redirects the preprotein to the general Sec pathway (3), and deletion of the extended N region renders the signal peptide similar to a canonical signal sequence (5). This second variant was expressed in a strain containing a deletion of secA2, to preclude export via the accessory Sec pathway, as well as a deletion of gtfA to avoid possible hindrance of export by carbohydrate linkage to the serine-rich GspB domains (50). When examined in S. gordonii cells, the presence of 1.5 mM azide led to a slight reduction in the secretion of both GspB736flag variants. Higher concentrations of azide (4.5 and 15 mM) resulted in progressively greater levels of suppression (Fig. 3A).

To assess whether azide has a comparable effect on SecA2_Sg activity, we examined the export of GspB736flag (with its native signal sequence) in the presence of azide. As was seen with SecA_Sg, the transport activity of SecA2_Sg was highly sensitive to azide. As little as 1.5 mM azide resulted in a substantial decrease in the amount of GspB736flag exported, and transport was strongly suppressed by 15 mM azide (Fig. 3B, upper panel).

Substitutions for certain residues in NBD or IRA2 of SecA_Ec have been shown to confer altered sensitivity to azide. For example, the replacement of Y134 in SecA_Ec with cysteine, serine, or asparagine results in a decreased sensitivity to azide, whereas the replacement of A507 with valine produces a hypersensitive phenotype (21, 31). We introduced substitutions in the analogous residues of SecA2_Sg (Y124 in motif Ia of NBD and A488 in motif V of IRA2, respectively) and then compared the effects of azide on export of GspB736flag. Replacement of Y124 in SecA2_Sg led to a slightly increased resistance to azide. That is, the SecA2^Y124S variant strain exported more GspB736flag than did the wild-type SecA2-expressing strain at high azide concentrations (Fig. 3B, middle panel). In contrast, replacement of A488 had no discernible effect on the sensitivity to azide (Fig. 3B, lower panel). These results demonstrate that, like SecA_Ec, SecA2_Sg is highly sensitive to azide and that replacement of a residue in motif Ia of NBD can reduce the sensitivity to azide. However, alteration of a residue in motif V of IRA2 had no apparent effect on azide sensitivity. This is consistent with the possibility that the action of IRA2 of SecA2_Sg during the process of translocation may be subtly different from that of SecA_Ec.

The biochemical properties of SecA2_Sg differ from those of SecA_Sg.

Although SecA2_Sg is predicted to be an ATPase, the ability of this putative enzyme to hydrolyze ATP has not been demonstrated directly. For SecA_Ec, three rates of ATP hydrolysis have typically been assessed in vitro. The rate of ATP hydrolysis by the isolated enzyme in an unactivated state (the intrinsic or basal rate) is relatively low and is highly sensitive to the level of Mg²⁺ (19). The ATPase activity is increased approximately twofold by the addition of anionic phospholipids or lipid vesicles containing SecYEG (the membrane-associated rate), and it is then maximally stimulated upon the addition of a preprotein substrate (the translocation rate).

As a first step toward characterizing the differences between SecA_Sg and SecA2_Sg in vitro, we compared the basal rates of ATP hydrolysis by these proteins. Both proteins were first expressed with N-terminal His₆ tags in E. coli. Although 6His-SecA was easily purified to homogeneity by using nickel affinity chromatography, purification of 6His-SecA2 was hampered by the apparent insolubility of this protein. The expression of 6His-SecA2 under a variety of conditions typically used to enhance solubility upon overexpression in E. coli (expression at low temperature, addition of ethanol to enhance the production of chaperones, addition of low levels of the inducing agent, etc.) yielded only minimal improvements in solubility which, at best, was found to be less than 0.1 mg per ml in aqueous solutions. Purification of 6His-SecA2 in the presence of the nondenaturing zwitterionic detergent CHAPS improved the yields of soluble protein, although solubility was still fairly low (approximately 0.5 mg per ml).

The basal rate of ATP hydrolysis by 6His-SecA2 was measured and compared with that of 6His-SecA purified in the presence or absence of CHAPS. Since the basal ATPase activity of SecA_Ec is highly sensitive to Mg²⁺ levels, due in part to the effect of this cation on the conformation of SecA_Ec (19), we assessed the activity of these enzymes over a range of Mg²⁺ concentrations. As has been reported for SecA_Ec, 6His-SecA showed maximal activity in the presence of very low levels of Mg²⁺ (0.1 mM) and was inhibited by higher levels (Fig. 4A). In contrast, 6His-SecA2 activity was highest in 2.5 mM Mg²⁺, with a maximal rate of hydrolysis that was approximately 30% of that of 6His-SecA. The low rate of ATP hydrolysis by 6His-SecA2 is due in part to the IRA1 domain, as removal of this region resulted in a twofold-higher ATPase activity (see Fig. S2 in the supplemental material). Somewhat surprisingly, the basal hydrolysis rates of 6His-SecA and 6His-SecA2 were unaffected by even high levels of azide (Fig. 4B).

We next examined whether fusion of MalE to SecA2 could enhance protein solubility and recovery. The MalE-SecA2 fusion protein was indeed more amenable to purification and was soluble to at least 1 mg per ml in aqueous solutions. ATP hydrolysis by the MalE-tagged protein was then assessed and compared with that of MalE-SecA. The N-terminal MalE fusion to SecA_Sg resulted in approximately twofold higher rates of hydrolysis in the presence of Mg²⁺ than were found for 6His-SecA. Activity was still highest, however, when the Mg²⁺ concentration was low (Fig. 4C). The MalE-SecA2 fusion protein also exhibited elevated rates of ATP hydrolysis, but it differed from MalE-SecA in two aspects. As seen with 6His-SecA2, ATP hydrolysis was undetectable in the absence of Mg²⁺. Second, the rate was maximal at 0.5 mM Mg²⁺ and was less strongly suppressed by higher levels.

The effects of sodium azide on the rate of ATP hydrolysis by the MalE fusion proteins were also examined. Both enzymes were strongly inhibited by azide levels as low as 1 mM (Fig. 4D). The ATPase activity of MalE-SecA2 was even more sensitive to azide than that of MalE-SecA (86% reduction compared with 69% reduction, respectively, at 1 mM azide) and was nearly completely suppressed at 15 mM. The combined results indicate that SecA2_Sg is an ATPase that can provide the energy for protein transport by the accessory Sec system but that there may be subtle differences in the way that the activity of this motor protein is regulated.

The SecA_Sg and SecA2_Sg domains are not readily interchangeable.

Since SecA_Sg and SecA2_Sg appeared to be highly similar both structurally and functionally, we hypothesized that these proteins might have a modular organization, such that domains of SecA_Sg and SecA2_Sg with common functions could be interchanged with little impact on export activity. In contrast, the exchange of any uniquely adapted domains could result in a gain of function, such as the ability to interact with the paralogous transport system. Indeed, a similar type of exchange between domains of the SecA proteins of organisms as disparate as rickettsiae and E. coli (39) or even mycobacteria and E. coli (34) has produced functional chimeras. Accordingly, construction of SecA-SecA2 chimeras that transport GspB through either the SecYEG or the SecY2 translocon could help to address whether SecA2 or SecY2 is uniquely adapted for a glycosylated preprotein.

To determine whether such uniquely adapted domains could be identified, we generated chimeric forms of SecA and SecA2 by exchanging various segments of the two S. gordonii paralogues. In all, eight chimeras were generated. For brevity, results for four chimeras will be presented here. One type of chimera involved an exchange within a very highly conserved motif near the end of IRA2, just ahead of the junction with the C domain (see Fig. S3 in the supplemental material). In effect, this involved replacement of the conserved PPXD and nucleotide binding folds (62% similarity and 52% identity) to adjoin the less similar C domain (43% similarity and 26% identity). Although both of the chimeric proteins (SecA^1-534A2^535-793 and SecA2^1-534A^535-838) were expressed at detectable levels (Fig. 5, lanes 3 and 4, respectively), neither was able to support transport of GspB736flag (data not shown).

Another chimera involved simply replacing the highly conserved NBD of SecA2_Sg, by an exchange at the conserved β strand, or “hinge,” that connects the N-terminal portion of NBD to PPXD. Again, this chimera (SecA^1-222A2^220-793) was expressed at detectable levels (Fig. 5, lane 5) but failed to transport GspB736flag (data not shown). A fourth chimera, which has a replacement of just the N-terminal end of SecA2 up to the conserved motif I (Walker A box) of NBD (SecA^1-121A2^114-793), was also examined. Surprisingly, although the segments exchanged are very highly similar (72% similarity and 61% identity) and the chimera was expressed at levels comparable to that of wild-type SecA2 (Fig. 5, lane 6), the SecA^1-121A2^114-793 chimera also failed to complement the SecA2 mutation (data not shown).

The inability of these chimeric proteins to support transport is evidently not due to a complete loss of the ability to hydrolyze ATP, as all retained low levels of basal ATPase activity (see Fig. S2 in the supplemental material). The lack of function may therefore be due either to disruption of more-complex intramolecular interactions (for example, improper coupling of ATP hydrolysis with the motion of protein transport) or to faulty interaction with the transport channel. Regardless, these data indicate that, despite the high degree of similarity, the SecA_Sg and SecA2_Sg domains are not interchangeable.