THE BACTERIAL ATPASE SECA FUNCTIONS AS A MONOMER IN PROTEIN TRANSLOCATION *

The ATPase SecA drives the post-translational translocation of proteins through the SecY channel in the bacterial inner membrane. SecA is a dimer which dissociate into monomers under certain conditions. To address the functional importance of the monomeric state, we generated an E. coli SecA mutant that is almost completely monomeric (>99%), consistent with predictions from the crystal structure of B. subtilis SecA. In vitro, the monomeric derivative retained significant activity in various assays and in vivo, it sustained 85% of the growth rate of wild type cells and reduced the accumulation of precursor proteins in the cytoplasm. Disulfide cross-linking in intact cells showed that mutant SecA is monomeric and that even its parental dimeric form is dissociated. Our results suggest that SecA functions as a monomer during protein translocation in vivo.

Many bacterial proteins are transported posttranslationally across the inner membrane by the Sec machinery, which consists of two essential components (1)(2)(3)(4). One is the SecY complex, which forms a conserved heterotrimeric proteinconducting channel in the inner membrane (5,6). The other is SecA, a cytoplasmic ATPase, which "pushes" substrate polypeptide chains through the SecY channel (7). SecA interacts not only with the SecY channel (8), but also with acidic phospholipids (9)(10)(11) and with both the signal sequence and the mature part of a substrate protein (12). It also binds the chaperone SecB, which ushers some precursor proteins to SecA (8,13,14). When associated with the SecY complex, SecA undergoes repeated cycles of ATP-dependent conformational changes, which are linked to the movement of successive segments of a polypeptide chain through the channel (15,16). However the mechanism employed by SecA to translocate substrates polypeptide chains through the SecY channel remains largely unknown.
An important issue concerning the function of SecA is its oligomeric state during translocation. SecA is a dimer in solution (17,18) and previous work argued that this is its functional state (19). An X-ray structure of B. subtilis SecA also indicates the existence of a dimer (7). However, recent evidence raises the possibility that SecA might actually function as a monomer: In solution, SecA dimers are in rapid equilibrium with monomers (20,21). Although the equilibrium favors dimers, it is shifted almost completely towards monomers in the presence of membranes containing acidic phospholipids or upon binding to the SecY complex (21). A synthetic signal peptide had a similar effect, although this result is controversial (22). A monomeric derivative of SecA containing six point mutations retained some in vitro translocation activity (21), but the low level of translocation precluded any firm conclusion. In addition, the previous results do not exclude models in which SecA cycles between monomeric and oligomeric states during the translocation of a polypeptide chain (22,23). Most importantly, the functional oligomeric state of SecA in vivo remains to be established.
In this report, we have tested whether SecA can function as a monomer in vivo. Guided by the X-ray structure of B. subtilis SecA (7), we generated a monomeric derivative of E. coli SecA that lacks the first eleven amino acids. The monomeric derivative retains significant activity in various in vitro assays and can substitute for 2 endogenous SecA in vivo. Cross-linking experiments in intact cells indicate that even the parental dimeric derivative is mostly present as a monomer. Our results therefore suggest that SecA functions as a monomer. Table 1 lists the bacterial strains and plasmids used in this work. The segment containing the arabinose promoter and the AraC gene on pBAD22 was amplified by PCR using the primers BclIBAD5pr (gtacgtgatc atgcataatgtgcctgtcaaatggac) and SecABAD3pr (gtgcgatcgttacgactaccgaaaactttagttaacaattt gattagcatggtgaattcctcctgctagccc).

Bacterial strains and plasmids -
The kanamycin resistance gene was amplified from p B A D 1 8 u s i n g B g l I I K a n 3 p r (gcttagatctgcctcgtgaagaaggtgttgctgac) and SecMKan5pr (gtgcgaacgctgttttcttaagcacttttccgcacaacttatcttcatt caagccacgttgtgtctcaaaatctc). The PCR-amplified fragments were digested with either BclI (araC-P BAD ) or BglII (kan) and ligated to each other. The ligation mixture was digested with BclI and BglII and the product fragment was amplified by PCR using SecMKan5pr and SecABAD3pr. The amplified linear cassette [SecM(-50--1)-kan-araC-P BAD -SecA ] was used to replace the chromosomal region between SecM(+1) to SecA(-1) by homologous recombination using strain DY378 as described (24). The resulting strain, EO527, requires arabinose for growth. Strain EO528 was generated by P1 transduction of the p r l A 4 allele into strain EO527. The presence of the prlA4 mutation was confirmed by sequencing. Strain EO529 was generated by P1 transduction of kan-araC-P BAD -SecA from EO527 into strain SMG96. D11/N95 was constructed from pT7N95-SecA (21) by PCRbased deletion of residues Leu2 to Gly11. N95 and D11/N95 were amplified by PCR and cloned between the NcoI and BamHI sites of pDSW204 (25) generating pDSW204N95 and pDSW204D11/N95, respectively. D11/N95 was also cloned between the BamHI and NotI sites of pET21 (Novagene). The mutations S636C and Q801C were introduced into pT7N95-SecA(C98S) using PCR-based site directed mutagenesis (Invitrogen) and confirmed by sequencing. The double-cycteine construct, N95(CC), was used to generate D11/N95(CC) by PCR-based deletion. Both constructs were then amplified by PCR and cloned between the NcoI and BamHI site of pDSW204 yielding plasmids pDSW204N95CC and pDSW204D11/N95CC, respectively. Over-expression and purification of proteins -Expression and purification of SecYE His6 G, SecY(prlA4)E His6 G, SecA and its derivatives and were done as described (21,26). D11/N95 was additionally purified by Superdex200 GF. ProOmpA was expressed and purified from inclusion bodies as described (27). Cross-linking and sucrose gradients -SecA and derivatives in buffer (50 mM K-HEPES pH 7.5, 100 mM KCl, 4 mM MgCl 2 , 1mM DTT) were cross-linked with 20 mM EDC 1 or analyzed by sucrose gradient centrifugation as described (21). In vitro assays -SecA and D 11/N95 were labeled with 125 Iodine using Iodogen (Pierce) as described (15). Liposomes containing reconstituted wild type SecY or SecY(prlA4) complexes were prepared as described (26). Proteoliposomes binding assays were done according to (8) as detailed by (21). ATPase and translocation assays were carried out at 30ºC and 37ºC, respectively, as described (21). To follow the kinetics of translocation 40 ml aliquots were withdrawn from a master mixture (380 ml) at different times, and mixed with 160 ml ice-cold 75 mM KCl, 50 mM K-HEPES pH 7.5. Samples were then processed as usual. Disulfide cross-linking of SecA derivatives -Strains EO529 and EO527 expressing N95(CC) from pDSW204N95(CC) were grown in LB containing ampicillin (100 mg/ml) at 30ºC for five hours. Where indicated, 150 mM IPTG was added after 3.5 hours. EO529 expressing D11/N95(CC) from pDSW204D11/N95(CC) was grown for five hours at 30ºC with or without 150 m M IPTG. Where indicated, cells were grown for additional 20 minutes in the presence of 1mM diamide (Sigma). Aliquots of 1.5ml of culture were withdrawn, mixed with 150 ml 100% TCA, incubated on ice for 20 minutes and centrifuged at 14K rpm for 10 minutes. The pellet was washed with ice-cold acetone, incubated on ice for 20 minutes and centrifuged again. Pellets were air-dried and dissolved with 150 ml 20 mM iodoacetamide, 1% SDS, 0.1 M Tris-Cl pH 8.0. After one hour at 22ºC 40 ml 6% SDS, 50% glycerol, 0.1% bromophenol-blue were added. In vitro crosslinking of SecA derivatives in buffer (50 mM K-HEPES pH 7.5, 50 mM NaCl, 4 mM MgCl 2 ) was done with 40 mM diamide for 15 minutes at 22ºC. The reaction was stopped with 30 mM iodoacetamide, 1% SDS, 50 mM Tris-Cl pH 8.0 and after one hour 12 ml 6% SDS, 50% glycerol, 0.1% bromophenol-blue were added. In vivo pulse-chase labeling of proOmpA -Cells were grown at 30ºC in minimal glycerol medium (M63 salts with 0.5% glycerol, vitamin B 1 mg/ml, vitamin B 5 mg/ml, 1 mM MgSO 4 , 18 amino acids, each at 50 µg/ml, and 0.005% yeast extract). Early log phase cells (7.5 ml) were pulsed with 0.1 mCi of [ 35 S]methionine for 20 seconds and chase was initiated by adding 750 µl of 1% methionine, chloramphenicol 1 mg/ml. Samples (1 ml) were removed at the indicated times and precipitated with TCA as described (28). Acetone washed pellets were solubilized with 25 mM Tris-Cl, 1 mM EDTA, 1% SDS and OmpA was immunoprecipitated using OmpA antibodies and protein A.

RESULTS
Generation of a monomeric derivative of SecA. We previously generated a monomeric SecA derivative by mutagenizing into alanines six residues that we suspected to be important for dimerization (21). The X-ray structure of B. subtilis SecA (7) later showed that four of the six residues were indeed close to the interface between the monomers. However, the derivative exhibited low translocation activity in vitro, probably because the specific mutations had additional effects. The X-ray structure of B. subtilis SecA now offers a more rational design of a monomeric derivative: In the crystal, SecA is a dimer with the monomers arranged head-totail (7). Most of the intersubunit contacts are contributed by the first nine residues of each subunit (Met1-Phe9), which contact side chains in the C-terminal domain of the other subunit ( Fig. 1). To test whether these residues are crucial for dimer stability, we deleted the corresponding sequence (L 2 IKLLTKVFG 11 ) from N95, an E. coli SecA derivative that lacks the last 70 residues. We used N95 as a starting point, because it is shorter than SecA and yet is fully dimeric and functional (21, 29,30). Also, the last 70 residues contribute slightly to dimer formation 2 . After sucrose gradient centrifugation the parental N95 dimer (189kDa) migrated in fractions 15-17, close to full-length SecA (fraction 18, 204kDa; Fig. 2A). In contrast, the derivative D11/N95, lacking residues Leu2-Gly11, migrated in fractions 11-13 close to the position of BSA (fraction 9, 70kDa) and consistent with it being monomeric. Quantification of the experiment showed that at most 1% of D 11/N95 exists as dimers 3 . Crosslinking experiments using EDC (17,21) supported the conclusion that D11/N95 is monomeric (Fig. 3E).
The monomeric derivative D11/N95 is active in vitro. Next we compared the activity of D11/N95 with that of wild type, full-length SecA in several in vitro assays 4 . We first measured their binding affinity to the SecY channel. Increasing amounts of SecA or D11/N95 were mixed with a constant amount of 125 I-SecA and incubated with proteoliposomes containing the SecY complex ( Fig. 3A). As the amount of unlabeled competitor was increased, the level of 125 I-SecA bound to reconstituted SecY decreased. A binding constant of 224±19 nM was calculated for wild type SecA (Fig.3A, circles), in close agreement with previous data (31). The monomeric derivative, D11/N95, had an only two-fold lower affinity of 447±32 nM (triangles). About 28% of the bound 125 I-SecA could not be competed away by D11/N95, perhaps because the N-terminal residues of SecA have a secondary binding site at SecY.
Previous experiments had shown that the prlA4 signal sequence suppressor mutation in SecY (32) enhances the activity of a monomeric derivative of SecA approximately five-fold (21). We therefore tested proteoliposomes containing the mutant SecY complex for its ability to bind D11N95. The affinities for SecA and D11/N95 were a little higher than with wild type SecY (173±12 nM and 305±23 nM, respectively), and again, the monomeric SecA derivative, D11/N95, had only a slightly reduced affinity for the SecY complex.
We next compared the translocation ATPase activity of SecA and D11/N95 under conditions where the ATPase activity is proportional to the protein concentration (Fig. 3B). SecA and D11/N95 exhibited the same low basal ATPase activity (~6.5 mol ATP/mol protein/ min; Fig.3B, closed and open triangles, respectively). Addition of both SecY complex-containing proteoliposomes and translocation substrate (proOmpA), stimulated the ATPase activity of SecA and of D11/N95 by a factor of 16 and 10, respectively (closed and open circles, r e s p e c t i v e l y ) .
W i t h S e c Y(prlA4) proteoliposomes the factor of stimulation was 13 and 12, respectively (closed and open squares). Thus, the monomeric derivative has almost the same level of translocation ATPase activity as wild type SecA.
To compare the translocation activities of SecA and D11/N95, they were incubated with SecY-complex containing proteoliposomes, ATP, and [ 35 S-Met]-proOmpA. Translocated proOmpA was detected by its resistance to protease treatment (Fig. 3C). We found that D11/N95 translocated 16% of the amount of proOmpA compared to wild type SecA (Fig. 3C, 'SecY'). With SecY(prlA4) complex-containing proteoliposomes, the translocation efficiency of D11/N95 increased to 74% (Fig. 3C, 'prlA4'). The kinetics of translocation was almost the same for SecA and D11/N95 ( Fig. 3D; see inset for raw data). As expected, no proteaseprotected proOmpA was seen in the absence of ATP, or when Triton X-100 was added during proteolysis (Fig. 3D, inset). Together, these results show that the D11/N95 monomer retains significant activity in several in vitro assays.
To exclude the possibility that D11/N95 dimerizes when bound to the SecY(prlA4) complex, we performed cross-linking experiments with the carbodiimide EDC. Wild type SecA on its own showed strong dimer crosslinks, and these were reduced upon addition of proteoliposomes containing SecY(prlA4) complex (Fig. 3E, lane 1 . D 11/N95 did not give rise to dimer crosslinks in the absence or presence of proteoliposomes (Fig. 3E, lanes 3,4), indicating that it remains monomeric when bound to SecY(prlA4).
The monomeric derivative D11/N95 supports cell growth. To test the activity of D11/N95 in vivo, we constructed E. coli strain EO527, in which the 5' regulatory region of the chromosomal SecA gene (~300 bp) was replaced with the tightly regulated arabinose promoter. This promoter is active in the presence of arabinose, but is turned off completely in the presence of glucose. Indeed strain EO527 exhibited robust growth in arabinose (doubling time of 45 min), but upon a switch to glucose, the cells stopped growing after four hours (Fig. 4A). This was paralleled by the expression of SecA: while in the presence of arabinose SecA was expressed at similar levels throughout the experiment (Fig. 4B, '+Arabinose'), its expression dropped to negligible levels four hours after the switch to glucose ('+Glucose'). These data are consistent with SecA being essential for viability of E. coli (33).
Next we tested whether D 11/N95 can replace SecA and support growth of EO527 cells, when expression of chromosomal SecA is shut down. N95 and D11/N95 were cloned into the plasmid pDSW204 under the control of an attenuated IPTG-driven promoter. The two plasmids and, as a control, the empty vector were each introduced into EO527 cells. It should be noted that the deletion of the sequence coding for Leu2-Gly11 also removes a segment, which plays a key role in translation initiation (34), resulting in much lower expression levels of the D11/N95 construct. As expected, all strains grew on plates containing arabinose owing to the expression of chromosomal SecA (Fig. 4C, left plate, lower half). On glucose plates lacking IPTG, only the N95-expressing construct gave viable cells (middle plate, lower half). As expected, the incomplete repression of the promoter by the Lac repressor allowed N95 to be made at levels sufficient for growth. Immunoblotting showed that even in the absence of IPTG, N95 was still made in significant amounts ( Supplementary Fig. S1). The D11/N95 construct did not support growth in the absence of IPTG, consistent with its negligible basal expression level (see Fig. 5D). In the presence of IPTG, both N95 and D11/N95 supported growth. Immunoblots showed that, in the presence of arabinose, the full-length SecA protein was synthesized from the chromosomal gene, while in the presence of IPTG and absence of arabinose, only the plasmid encoded D11/N95 protein was made (Fig. 4D, lane 2 versus 1). These data show that D11/N95 is responsible for cell growth under these conditions.
We performed similar experiments using strain EO528, which differs from EO527 by harboring the prlA4 mutation in its chromosomal SecY gene (Fig. 4C, upper halves of the plates). Again, D11/N95 supported cell growth only in the presence of IPTG (compare right and middle plates). Interestingly, N95 supported growth in the absence, but not in the presence of IPTG. This is explained by the massive overexpression of N95 in the presence of IPTG, which inhibits growth of strain EO528 (Supplementary Fig.  S1).
To compare the efficiency of D11/N95 with wild type SecA in a quantitative manner, we determined the growth rates of cells expressing similar levels of either protein. In the presence of arabinose, EO527 cells had a growth rate that approached that of the parental strain DY378 (Fig. 5A). In the absence of arabinose and presence of IPTG, plasmid-born D 1 1 / N 9 5 supported a growth rate that reached 85% of that of the parental strain DY378 (Fig. 5B). Immunoblotting for SecA showed that after induction, the levels of SecA and D11/N95 were about equal (Figs. 5C versus 5D). These data therefore indicate that the monomeric derivative is almost as efficient as wild type SecA in supporting cell growth. Similar results were obtained with strain EO528, which has a SecY(prlA4) background (Figs. 5E-H). Again, comparable growth rates corresponded to similar expression levels of SecA and D11/N95.
Finally, we tested whether D 11/N95 can prevent the steady-state accumulation of precursor proteins with uncleaved signal sequences, which is observed under SecA deficiency. When EO527 cells were depleted of SecA by incubation in the absence of arabinose, precursors to DegP, MBP and OmpA were prominently present, as demonstrated by immunobloting with specific antibodies (Fig.  6A, lane 1). In contrast, when wild type SecA was expressed in the presence of arabinose, only the mature forms were seen (lane 2). When D11/N95 was expressed, the accumulation of precursors was significantly diminished though not entirely eliminated (lanes 3 versus 1). With strain EO528, having a SecY(prlA4) background, precursor accumulation was further reduced (lanes 4-6): with MBP and OmpA the monomeric derivative was as effective as wild type SecA, and with DegP it was even more efficient (lanes 6 versus 5). Similar conclusions could be drawn from 'pulse-chase' experiments in which the processing of proOmpA into mature OmpA was followed (Fig. 6B). In EO527 cells expressing wild type SecA in the presence of arabinose, proOmpA was processed into mature OmpA within less than a minute ('SecA'). In the absence of arabinose, even after 10 minutes only 50% of proOmpA was processed ('none'). When D 11/N95 was expressed, 50% of the labeled proOmpA was processed within one minute ('D11/N95'). In conclusion, the monomeric SecA derivative clearly supports translocation in vivo, albeit less efficiently than wild type SecA.
Probing the oligomeric state in vivo by disulfide cross-linking. To test the oligomeric state of the SecA derivatives in vivo, we introduced cysteines at defined positions and tested their ability to form disulfide bridged SecA-dimers. Guided by the X-ray structure of the B. subtilis SecA dimer (7), we identified Gly587 and Arg750 as residues in close proximity across the subunit interface (Fig. 1). The equivalent residues of E. coli (SecA)N95, Ser636 and Gln801, were changed to cysteines generating the single cysteine mutants and the double mutant (N95(CC)). We first tested these mutants in vitro for their ability to form disulfide-bridged dimers. The purified mutant proteins were treated with the oxidizing reagent diamide (35,36), and separated on a non-reducing SDS gel (Fig. 7A). The double mutant N95(CC) showed two main crosslinked products in the molecular weight region where dimers are expected (lane 2). In other studies crosslinked products containing the same components in different linkages also had different mobilities in SDS gels (37,38). Several weaker bands were seen in the low molecular weight region, perhaps caused by crosslinking to contaminating proteolytic N95 fragments. The two dimer crosslinks were also seen when the single cysteine mutant proteins were mixed together before oxidation (lane 3). This is in agreement with previous experiments showing that SecA dimers are in equilibrium with monomers (Or et al., 2002). In addition, this experiment shows that both dimer crosslinks contain a single disulfide bridge. The single-cysteine mutant Ser636Cys did not give crosslinks (lane 4), as expected from the X-ray structure. Surprisingly, however, the single cysteine mutant Gln801Cys gave a strong crosslinked band (lane 5), even though in the X-ray structure residues 801 in the two subunits are quite distant (Fig. 1). This band corresponds to the lower crosslinked band seen with the double mutant (lane 5 versus lane 2). Thus, there seem to be two different conformations of the dimer, one corresponding to the X-ray structure, which gives rise to the upper crosslinked band, and one in which residues Gln801 in the two subunit come close, which gives rise to the lower band. All crosslinked bands disappeared when the samples were reduced (lanes 7, 8), indicating that they are indeed formed by disulfide bridges.
Next we performed crosslinking experiments in vivo. Since the cytoplasm of wild type cells is reducing and thus does not favor disulfide bond formation, we used strain EO529, which contains an oxidizing cytoplasm, owing to the absence of TrxB and Gor, two key enzymes in the thiol reducing pathway (36). N95(CC) was functional, and supported growth of EO529 cells in the absence of arabinose when expressed from a plasmid. To test whether N95(CC) can form dimers when highly over-expressed, cells were induced with IPTG and precipitated with TCA. The proteins were treated with iodoacetamide to block free cysteines, and separated on non-reducing SDS gels. Blotting with SecA antibodies revealed the monomer and, in addition, three high molecular weight bands in the 200kDa region (Fig. 7B, lane 3  'Dimers'). A similar pattern of three high molecular weight bands (21) was seen when purified N95 was treated with the cross-linking agent EDC (lane 12). Two of the bands were at the same position as the crosslinked products generated in vitro (Fig. 7A, lane 1 versus lane  2). The third upper band may be a dimer with two Cys636-Cys801 disulfide bridges. When IPTG-induced cells were pre-treated with diamide, the intensity of all three bands was enhanced (Fig. 7B, lane 4). As expected, the bands disappeared when the samples were reduced with DTT (lane 11). These data therefore suggest that over-expressed N95(CC) forms disulfide-linked dimers in vivo. The appearance of the cross-linked dimers was dependent on an oxidizing environment in the cytoplasm: they were not observed in EO527 cells with a reducing cytoplasm, unless diamide was added (lane 10 versus 9). The N95(CC) protein did not form dimers when expressed at lower levels in the absence of IPTG in either strain EO529 or EO527 (lanes 1 and 7), even though more material was loaded onto the gel to compensate for the lower expression level. Diamide treatment produced weak cross-links in strain EO527 (lane 8) but not in EO529 cells (lane 2). Thus, even the N95(CC) protein with an intact N-terminus is largely monomeric when expressed at basal levels in vivo. As expected, the derivative D11/N95(CC) did not give rise to disulfide-linked dimers in EO529 cells, even when diamide was added (Fig. 7B lanes 5 and  6). These data therefore support our assumption that this protein is also monomeric in vivo. It should be noted that deletion of the N-terminus of N95 abolished the crosslinks corresponding to both dimer conformations. In addition, whenever the dimer cross-links were diminished or absent (lanes 1, 2, 5 -8), other prominent cross-links were seen (some are marked in lane 6 by asterisks); apparently, the monomeric form of SecA can interact with additional proteins in the cell 5 .

DISCUSSION
Our results suggest that SecA functions as a monomer in protein translocation. Using sucrose gradient centrifugation and chemical crosslinking, we show that a SecA derivative, D11/N95, lacking the first 11 residues is fully monomeric. It remains monomeric in intact cells, as demonstrated by in vivo disulfide cross-linking experiments. D11/N95 is functional in vitro, as demonstrated by its high affinity for the SecY complex, its translocation ATPase activity, and its ability to translocate proOmpA into r e c o n s t i t u t e d proteoliposomes. More importantly, we show that the monomeric derivative is active in vivo, supporting growth of the cells at~85% the rate of wild type cells. Our in vivo results extend previous in vitro experiments that showed dissociation of SecA dimers in the presence of membranes (21, 22) and provide strong evidence against the notion that SecA functions as a dimer (19). Interestingly, we found that even the parental N95 derivative, which contains the 11 Nterminal residues, is monomeric in cells unless overexpressed. Thus, the equilibrium between dimers and monomers, which in purified N95 preparations is on the side of dimers, appears to be shifted towards monomers under the conditions in vivo. Given that the total concentration of SecA in E. coli cells is about 5 mM 6 and that the dissociation constant of the dimer is 0.3-0.5 mM, one would have expected SecA to be mostly dimeric. Possible explanations for the discrepancy are that the effective concentration of SecA is lowered by its interaction with other cellular components or membranes (39,40), or that the dissociation constant is higher at the ionic conditions in the cell (20,41). Full-length SecA dimers may dissociate less readily than the N95 dimers because the C-terminal 70 residues seem to contribute slightly to the interaction between subunits 7 , but our results clearly show that the translocation mechanism of SecA does not require the presence of dimers.
The deletion of the first 11 residues from the parental SecA derivative N95 was not entirely without effect on its translocation activity. In vitro, it was significantly less efficient in the translocation of proOmpA into proteoliposomes (16% compared with wild type SecA), perhaps because it dissociates more readily from the SecY channel 8 . However, when tested with mutant SecY(prlA4), which binds SecA with higher affinity (42), its efficiency reached 74%. In vivo, the monomeric derivative supported a growth rate close to that of wild type SecA, both in wild type and in SecY(prlA4) cells, although some precursor accumulation was observed. Thus, the N-terminal residues appear to be less important in vivo, perhaps due to the presence of other factors.
Our conclusion that the monomeric state of SecA is the active species is consistent with a new X-ray structure of B. subtilis SecA in which the protein crystallized as a monomer (43). Compared with the structure of the SecA dimer (7) the monomeric structure shows a drastic conformational change, likely corresponding to the opening of the peptide binding groove (43)). Chemical modification and cross-linking experiments show that conditions that lead to the dissociation of the dimer also result in opening of the groove. An active SecA monomer is also consistent with the demonstration that a detergent-solubilized translocation intermediate contains only one copy of SecA (44).
The ATPase domain of SecA is structurallyrelated to super-family I and II helicases (7). One of them, PcrA, functions like SecA as a monomer and is proposed to use an "inchworm" mechanism to move along a single-stranded nucleic acid strand (45); the two RecA-like nucleotide binding folds move relative to each other during the nucleotide hydrolysis cycle. SecA might undergo similar conformational changes with its nucleotide binding folds while bound to the SecY complex, thereby moving the polypeptide chain through the channel.
The reason why SecA can form dimers and even higher oligomers (17,22,46,47) is unclear. Our results make it unlikely that dimerization is an obligatory step during the translocation cycle. It is even possible is that the dimer would never occur under physiological conditions. Alternatively, however, the dimer may be generated in certain situations, for example, when secretion is blocked and SecA is upregulated (48,49). Dimerization could prevent SecA from interacting with unfolded cytoplasmic proteins. In this scenario, the dimer would only dissociate upon binding to the membrane in preparation for protein translocation. 3 Data not shown. 4 The activity of N95 in these assays was the same as that of full-length SecA; data not shown. 5 These crosslinks do not, however, react with anti-SecY antibodies; data not shown. 6 Data not shown. 7 Data not shown. 8    the presence of proteoliposomes containing either wild type SecY or SecY(prlA4) complex. The numbers below are the fraction of total proOmpA that was translocated and protected from protease. D, proOmpA was translocated by SecA or D11/N95 into proteoliposomes containing SecY(prlA4) complex. Raw data (inset) were quantified by phosphor-imaging and normalized to the value obtained by SecA after 12 minutes. '-ATP'-15 minutes incubation without ATP. 'TX'-Triton X-100 present during proteinase K treatment. E, D11/N95 or SecA were treated with EDC in the presence or absence of 2 ml proteoliposomes containing reconstituted SecY(prlA4). Samples containing 0.5 mg were resolved on a 5% gel and immunoblotted with antibodies against SecA.    The following purified proteins were treated with diamide and samples (0.5 mg) were resolved on a 5% gel and immunoblotted with SecA antibodies (lanes 2-5): N95(CC), N95(C636) plus N95(C801), N95(C636) and N95(C801), respectively; lane1-in vivo cross-linked N95(CC). Lanes 6-8: Same samples as in lanes 1-3, treated with 50 mM DTT. B, Strains EO529 and EO527, containing an oxidizing and reducing cytoplasm, respectively, and expressing N95(CC) from a plasmid, were grown in the absence or presence of IPTG, as indicated. The monomeric derivative D11/N95(CC) was expressed by EO529 cells in the presence of IPTG. Where indicated, the cells were treated with diamide. The samples were resolved on a 5% SDS gel under non-reducing conditions and blotted with SecA antibodies. The sample volumes were adjusted so that equal amounts of the SecA derivatives were loaded: lanes 1, 2, 5, 6 -40 ml; lanes 3, 4, 11 -6 ml; lanes 7, 8 -20 ml; lanes 9, 10 -3 ml. The sample in lane 11 is identical to that in lane 4, except that 50mM DTT was added. The sample in lane 12 contained purified N95/SecA (0.5 mg) treated with the cross-linker EDC. 'Dimers' indicates the positions of the crosslinked bands. '*'-Cross-links with unidentified proteins.