TatB and TatC Form a Functional and Structural Unit of the Twin-arginine Translocase from Escherichia coli *

In Escherichia coli, a subset of periplasmic proteins is exported via the twin-arginine translocation (Tat) pathway. In the present study, we have purified the Tat complex from E. coli, and we show that it contains only TatA, TatB, and TatC. Within the purified complex, TatB and TatC are present in a strict 1:1 ratio, suggesting a functional association. This has been confirmed by expression of a translational fusion between TatB and TatC. This Tat(BC) chimera supports efficient Tat-dependent export, indicating that TatB and TatC act as a unit in both structural and functional terms. The purified Tat complex contains varying levels of TatA, suggesting a gradual loss during isolation and a looser association. The molecular mass of the complex is ∼600 kDa, demonstrating the presence of multiple copies of TatA, B, and C. Co-immunoprecipitation experiments show that TatC is required for the interaction of TatA with TatB, suggesting that TatA may interact with the complex via binding to TatC.

In bacteria, the vast majority of extracytoplasmic proteins are transported across the cytoplasmic membrane by the Sec apparatus (for review, see Ref. 1). In addition to several specialized protein export systems, another major protein export system has been discovered that, in contrast to the Sec apparatus, is able to transport folded proteins. This translocation pathway, denoted the twin-arginine translocation (Tat) 1 pathway, is closely related to the ⌬pH-dependent pathway of thylakoid membranes (reviewed in Ref. 2). Substrates for this pathway are characterized by an essential twin-arginine motif in their signal peptides (3,4).
In Escherichia coli, four genes have been shown to encode components of the Tat pathway (5)(6)(7)(8)(9). Three of these, tatA, tatB, and tatC, are located in one operon, whereas the fourth gene, tatE, is monocistronic. TatA, TatB, and TatE are homologous proteins that are predicted to contain a single transmembrane helix at their amino termini followed by a cytoplasmic domain. It has been shown that TatA and TatE fulfill similar functions, whereas TatB has a distinct role in the translocation process (8). The fourth protein, TatC, is predicted to contain six transmembrane helices, with the amino and carboxyl termini located at the cytoplasmic face of the membrane. TatB and TatC are essential components of the Tat pathway (6 -8), and translocation requires in addition either TatA or TatE. However, TatA seems to be far more important than TatE, most likely because of a greater abundance (5,10).
Only a limited amount of information is available on the composition of the Tat translocase. Recently, it was shown that TatA and TatB interact with each other and are present in a large complex of ϳ600 kDa (11). Notably, it was found that not all TatA was bound to TatB, suggesting that TatA might be present in a separate complex. Thus far, it has not been established whether TatC forms part of the 600-kDa TatA/B complex, and the possible involvement of other components could not be excluded.
To analyze the composition of the Tat system in more detail we have purified the complex from E. coli membranes. Here, we show that the Tat system purifies as a TatABC complex, and we find no evidence for the presence of additional, hitherto unidentified subunits, although the existence of further subunits cannot be excluded because the activity of the complex has not been tested. Within the Tat complex, TatB and TatC are the most stable components and are present in a 1:1 ratio, suggesting a structural association. We also show that cells containing a fusion of the TatB and TatC proteins are able to support Tat-dependent transport, indicating that TatB and TatC also form a functional unit and act in concert. Finally, although some TatA is tightly associated with TatB/C, we find that the vast majority of TatA does not co-purify with the complex, suggesting a looser association or a separate function in the overall translocation process.
The tatABC operon was amplified with the primers AB.tatA1 (atac-cATGGGTGGTATCAGTATTTG; nucleotides identical to genomic DNA are capitalized, and restriction sites are underlined) and AB.tatC-s (atattctagattatttttcaaactgtgggtgcgaccaattcgaTTCTTCAGTTTTTT-CGCTTTCTGC; nucleotides in bold specify the Strep-tag II peptide SNWSHPQFEK; Ref. 12). The resulting product was digested with NcoI and XbaI and cloned into plasmid pBAD24 (13), generating pABC-s.
A vector encoding a fusion of the TatB and TatC proteins was generated as follows. First, the tatB gene was amplified with the primers JT.B1 (gccatgccATGGCGTTTGATATCGGTTTTAGC) and JT.B2 (gctc-tagagttgttgttattgttattgttgttgttgttCGGTTTATCACTCGACGA; nucleotides in bold specify a linker consisting of 10 Asn residues). The tatC gene was amplified with the primers JT.C3 (gctctagaatcgaaggtcgtT-CTGTAGAAGATACTCAA; nucleotides in bold specify a factor Xa clea-vage site; IEGR) and JT.C4 (ccaatgcattggttctgcagttaTTATTCTTCAGT-TTTTTCG). Next, the resulting products were digested with NcoI and XbaI and with XbaI and PstI, respectively, and cloned in one step into pBAD24, generating pJDT7.
To construct a vector encoding both TatA and the Tat(BC) fusion protein, pJDT7 and pABC-s were both digested with SstII and PstI. Next, the DNA fragment encoding the Tat(BC) fusion protein (derived from pJDT7) was cloned into the vector fragment derived from pABC-s, thereby replacing the tatB and tatC genes with the tat(BC) gene. The resulting plasmid was denoted pJDT12.
To construct a vector encoding the Tat(BC) fusion protein with a Strep-tag II peptide at its carboxyl terminus, the tat(BC) gene was amplified using the primers AB.tatB1 (ataccATGGTGTTTGATATCG-GTTTTAG and AB.tatC-s (see above) and plasmid pJDT7 as template DNA. Next, the resulting DNA fragment was cloned into pBAD24, generating p(BC-s).
SDS-PAGE and Western Blot Analysis-Proteins were separated by SDS-polyacrylamide gel electrophoresis and immunoblotted and visualized with specific antibodies (11) and horseradish peroxidase (HRP) anti-rabbit IgG conjugates, using the ECL detection system (Amersham Pharmacia Biotech). TatC with a Strep-tag II was visualized directly using a Streptactin-HRP conjugate (Institut fü r Bioanalytik, Göttingen, Germany).
Purification of the TatABC Complex-Cells were grown anaerobically in TY-GT medium to the end of the exponential growth phase and then spheroplasted and sonicated, and the membranes were isolated as described before (11). Membranes were solubilized in buffer I (20 mM Tris-HCl, pH 8.0, 20% glycerol) plus 50 mM KCl and 1% digitonin (Calbiochem). To prevent degradation of proteins, a protease-inhibitor mixture (Complete, Roche Molecular Biochemicals) was added to buffers for spheroplasting, membrane isolation, and membrane solubilization. Solubilized membranes were loaded on a Q-Sepharose column. The column was washed with 1 column volume buffer I containing 100 mM KCl and 0.1% digitonin, and proteins were eluted with 2 column volumes of buffer I containing 300 mM KCl and 0.1% digitonin. Avidin (2 g/ml) was added to the sample to block any biotin-containing proteins, and the sample was loaded on a 2-ml Streptactin-Sepharose (Institut fü r Bioanalytik) column. The column was washed with 5 column volumes of buffer I containing 200 mM KCl and 0.1% digitonin. Proteins were eluted with 5 ϫ 1.5 ml of the same buffer containing 2.5 mM desthiobiotin (Institut fü r Bioanalytik). The first three fractions were pooled and concentrated to 0.5 ml using a Centricon centrifugal filter (YM-100; 100,000 molecular weight cut-off; Amicon). This sample was loaded on a Superose 6HR gel filtration column (Amersham Pharmacia Biotech). Elution was with buffer I containing 200 mM KCl and 0.1% digitonin. Fractions of 0.6 ml were collected, and those containing TatABC were pooled, diluted with buffer I to give a KCl concentration of 50 mM, and loaded on a MonoQ column (Amersham Pharmacia Biotech). Proteins were eluted with a linear KCl gradient (50 -400 mM) in 20 column volumes.
For purification of the radiolabeled TatABC-s complex, cells were grown in minimal medium supplemented with all amino acids except methionine and cysteine. At the mid-exponential growth phase, the cells were labeled with [ 35 S]methionine for 15 min. Purification of the radiolabeled TatABC-s complex was performed as described above.
Samples were analyzed by SDS-PAGE and fluorography, and the levels of TatA, TatB, and TatC-s were quantified using a PhosphorImager and ImageQuant software (Molecular Dynamics).
Isolation of Membranes and Immunoprecipitation-Membranes isolated from cells grown aerobically in TY medium were solubilized with 1% digitonin, and proteins were immunoprecipitated with anti-TatA or anti-TatB serum as described before (11). Peak fractions from the MonoQ column of radiolabeled TatABC-s complex were immunoprecipitated using anti-TatA, anti-TatB, or an irrelevant serum (raised against spinach photosystem II subunit W) and visualized by SDS-PAGE and fluorography.
TMAO Reductase Activity Assay-Cells were grown anaerobically in TY-GT medium until the mid-exponential growth phase, and periplasm and spheroplasts were prepared by the EDTA/lysozyme/cold osmoshock procedure (14). Spheroplasts were lysed by sonication, and intact cells and cellular debris were removed by centrifugation (5 min at 10,000 ϫ g). Membranes were separated from the cytoplasmic fraction by centrifugation (30 min at 250,000 ϫ g. Protein fractions were separated on a 10% nondenaturing polyacrylamide gel, and TMAO reductase activity was visualized in the gel using a methyl-viologen-linked TMAO reduction as described before (15).

TatA Requires TatC for Interaction with TatB-Previously,
we have shown TatA and TatB co-immunoprecipitate with each other and participate in a large complex of ϳ600 kDa (11). However, the complex was not purified during this work, and we did not establish whether TatC was associated with this complex. To address these points, we first tested whether TatC is required for interaction of TatA and TatB. Membranes were isolated from E. coli MC4100 and the ⌬tatAE, ⌬tatB, and ⌬tatC strains, and, as shown in Fig. 1, TatA could be immunoprecipitated from wild-type cells (lanes WT) using the anti-TatB serum, as found previously (11). In control assays, TatA was also immunoprecipitated from wild-type cells using an anti-TatA serum (lane ␣-TatA). The immunoprecipitation of TatA using the TatB serum is not due to recognition of TatA by the anti-TatB antibodies, because TatA was not co-immunoprecipitated in cells lacking TatB (lane ⌬B). As expected, no immunoprecipitation is observed in cells lacking TatA and TatE (lane ⌬AE). Significantly, TatA was also not co-immunoprecipitated from cells lacking TatC (lane ⌬C), even though these cells contain normal levels of TatA and TatB (11), showing for the first time that TatC is involved in the complex and, moreover, indicating that TatC is required for interaction of TatA and TatB.
Controlled Expression of the tatABC Operon-To analyze the composition of the Tat complex more directly, an expression vector was constructed encoding TatC with a Strep-tag II fusion on the carboxyl terminus (denoted TatC-s). The Strep-tag II peptide enables detection using Streptactin (an engineered streptavidin; Ref. 16)-HRP conjugate and purification on Streptactin-Sepharose affinity columns. A major advantage of the latter system is that proteins can be purified using very mild conditions, because elution can be achieved by simply adding low concentrations of a biotin derivative, desthiobiotin, to a physiological buffer. Because the stability of TatB and TatC depends on the presence of TatA and TatB, respectively (8,11), the expression vector (pABC-s) encoded TatA and TatB as well as TatC-s. All three tat genes were under the control of the arabinose-inducible P BAD promoter.
To test whether the plasmid-borne expression of TatA, TatB, and TatC-s was fully functional, pABC-s was transformed into an E. coli strain lacking the tatABCD operon and tatE. The latter strain is unable to grow anaerobically in minimal glycerol/TMAO medium (5,7,8). Plasmid-borne tatABC-s was able to restore this growth defect of E. coli ⌬tatABCDE, showing that tatA, tatB, and tatC-s are expressed and, moreover, that the Strep-tagged derivative of TatC is functional (data not shown).
To verify the expression level of the plasmid-borne tatABC genes, cells of E. coli strain ⌬tatABCDE containing plasmid pABC-s were grown in the presence of 5 or 100 M arabinose, and TatB levels were compared with those in wild-type cells (E. coli MC4100) by Western blotting. In the presence of 5 M arabinose, the expression level of TatB is similar to that of wild-type cells (Fig. 2). In the presence of 100 M arabinose, TatB was overproduced ϳ50-fold. Similar values were obtained for the cellular levels of TatA (data not shown).
Purification of a TatABC-s Complex-To identify the proteins that co-purify with TatC-s, digitonin-solubilized membranes (Fig. 3A, lane 1) isolated from cells grown in the presence of 100 M arabinose were first subjected to Q-Sepharose chromatography. Proteins eluting in a buffer with 300 mM KCl (Fig. 3A, lane 2) were further purified on a Streptactin column. The eluate from this step contains two major bands of ϳ27 and 31 kDa (Fig. 3A, lane 3). Western blotting showed that these bands are TatB and TatC-s, respectively (Fig. 3, B and C). Furthermore, a band running at ϳ17 kDa was visible, which, as confirmed by Western blotting, is TatA. Finally, a number of other bands were visible that are multimers of TatC (indicated by C*) and a degradation product of TatB (indicated by B*), because these bands were also detected with Streptactin-HRP and anti-TatB, respectively (data not shown).
The TatABC proteins were further purified on a Superose 6HR gel filtration column. TatABC eluted in a single peak corresponding to a molecular mass of ϳ600 kDa (Fig. 4), which is in agreement with the molecular mass found for the solubilized TatA/B-containing complex in our earlier studies (11). A final purification step was performed using anion-exchange chromatography. Silver staining of peak fractions revealed only the presence of TatB and TatC-s (Fig. 3A, lane 5), but Western blotting showed that these fractions also contained TatA protein (Fig. 3D). Concentration of peak fractions using a Centricon filter (100,000 molecular weight cut-off) and analysis by SDS-PAGE and Coomassie staining showed a similar result: only TatB and TatC-s were detectable; the levels of TatA are once again too low for visualization (not shown).
In the experiments described above, the TatABC-s proteins were overproduced to prepare sufficient quantities of material. However, it was important to purify the complex from cells expressing wild-type levels of the Tat system to test whether the complex contains any other proteins (because these would not be overexpressed from the tatABC plasmid). This was achieved by purifying the complex from cells grown on low levels of arabinose (5 M), which corresponds to wild-type levels of TatA and TatB, using the first three steps of the purification strategy outlined above. After these purification steps, only bands corresponding to TatB and TatC were visible (data not shown). The presence of TatA was again confirmed by Western blotting. No other bands were detectable, strongly suggesting that the core components of the twin-arginine translocase complex are TatA, TatB, and TatC.
The Tat Complex Contains TatB/C in a Fixed Ratio Together with Varying Amounts of TatA-The data shown above indicate FIG. 2. Expression of TatB from pABC-s. Cells of E. coli MC4100 (WT) containing plasmid pBAD24 (vector without insert), or ⌬tatAB-CDE containing plasmid pABC-s were grown in TY medium until the end of exponential growth. The latter strain was grown in the presence of 5 or 100 M arabinose (lanes indicated by 5 and 100). The cells were collected by centrifugation, and TatB was visualized by SDS-PAGE and Western blotting. To prevent overexposure from the sample containing overexpressed TatABC, only 1 or 10% of the sample was loaded on the gel. that TatB and TatC are easily detected in the purified complex, but TatA is difficult to detect and quantify using these procedures. We have found that this protein stains aberrantly with Coomassie and very poorly with silver (not shown). We therefore used an alternative strategy to obtain a clearer picture of the subunit ratios in the purified complex. The Tat complex was isolated from cells grown in the presence of [ 35 S]methionine, and Fig. 5A shows the elution of the radiolabeled complex from the penultimate (gel filtration) and final ion-exchange chromatography stages. Only three bands are visible, corresponding to TatA, TatB, and TatC-s. Using this procedure, TatA is now clearly detected, and this is clear evidence that this subunit is indeed an integral component of the purified complex. To analyze the subunit ratios in the various fractions, the band intensities were quantified using a PhosphorImager. Furthermore, to be able to compare the ratios accurately, amino-terminal sequencing was performed to determine whether the amino-terminal methionines were present in the TatA, TatB, and TatC proteins. The sequences obtained were MGGISI, MFXIG, and XVEDT, respectively. The TatA and TatB sequences are exactly as predicted from the gene sequences, demonstrating that TatA and TatB do contain their amino-terminal methionines. In contrast, the tatC gene sequence predicts MSVEDT, indicating that the aminoterminal methionine is not present in the purified TatC. Fig. 5B shows the A:B:C ratios of the lanes shown in Fig. 5A, calculated according to the number of methionine residues in each subunit (8 in TatC, 3 in TatB, and 2 in TatA). The figures show that the ratio of B:C is close to 1 in each eluate, indicating that these subunits are present in equimolar amounts. Interestingly, the ratio of A:B/C varies considerably in the gel filtration eluate. All fractions contain at least 1 TatA per TatB/C, but in some fractions the ratio is much greater. The peak fractions from the final ion-exchange column contain A, B, and C in a ratio that is very close to 1:1:1. To test how tightly TatA was bound to TatB/C, the peak fractions from the anion-exchange column were immunoprecipitated with anti-TatA (lane ␣A), anti-TatB (lane ␣B), or an irrelevant antibody (lane ␣W), and after SDS-PAGE, band intensities were quantified using a PhosphorImager (Fig. 5C). The ratio of B:C after immunoprecipitation with either anti-TatA or anti-TatB was again very close to 1 (1.0 and 1.1, respectively). This again suggests a close association of TatB and TatC. Strik-ingly, the ratio of A:B/C in these two lanes is very different, being 8.8 after immunoprecipitation with anti-TatA and 0.53 after immunoprecipitation with anti-TatB. In conclusion, these data show that TatB and TatC remain in a fixed 1:1 ratio throughout four different chromatography stages and an extensive immunoprecipitation procedure, strongly indicating that they are tightly associated and might form the core of the Tat complex. In contrast, TatA is also an integral element of the complex, but most of this subunit is gradually lost during the purification stages, and much of it may not even be bound to Tat(BC) at the point of membrane solubilization. Thus, TatA seems to be more loosely associated with the TatB/C complex.
A TatB-TatC Chimera Supports Tat-dependent Protein Translocation-The data described above indicate an intimate structural link between the TatB and TatC proteins and raise the possibility that the two subunits may act in concert. This has been confirmed by further experiments in which we tested whether a translational fusion between these two proteins would be functional in E. coli cells. For this purpose, a plasmid (pJDT12) was constructed containing the tatA gene followed by a translational fusion of the tatB and tatC genes, all under the control of the arabinose-inducible P BAD promoter. The tatBC gene encodes a TatBC fusion protein, denoted Tat(BC), in which the TatB and TatC domains are separated by a spacer region of 10 asparagine residues and a factor Xa protease cleavage site (IEGR; Fig. 6). This Tat(BC) protein comprises 443 residues and is predicted to contain seven transmembrane helices: one transmembrane helix in the amino-terminal domain (corresponding to the amino terminus in TatB) and six transmembrane helices in the carboxyl-terminal domain (corresponding to the TatC domain). Furthermore, it contains a large cytoplasmic domain corresponding to the carboxyl-terminal domain of TatB (including the predicted amphipathic helix), the spacer region, and the first amino-terminal residues of the TatC protein.
To examine expression levels of the tat(BC) gene, cells of E. coli strain ⌬tatABCDE containing plasmid pJDT12 were grown in the presence of varying amounts of arabinose, and the protein was visualized by Western blotting using an anti-TatB serum. The Tat(BC) protein has a mobility corresponding to its predicted molecular mass of 48 kDa and was readily detected in samples prepared from whole cells cultured with increasing concentra- FIG. 4. The TatABC-s complex has a molecular mass of ϳ600 kDa. A chromatogram of the gel filtration of the TatABC-s complex is shown. A sample concentrated to 0.5 ml after purification of the Tat complex on a Streptactin column was loaded on a Superose 6HR gel filtration column, and the proteins were eluted as described under "Experimental Procedures." Protein elution was monitored by absorbance (A) at 280 nm. The column was calibrated with the following proteins: thyroglobulin (669 kDa), ferritin (440 kDa), catalase (232 kDa), and aldolase (158 kDa). MW, molecular weight; V o , void volume; V e , elution volume. tions of arabinose (Fig. 7A). However, in the absence of arabinose it was difficult to detect the protein above background bands in whole cells. Because we noted some toxic effects of arabinose on some of the E. coli strains when grown on minimal TMAO/ glycerol medium, resulting in extended lag phase with increasing concentrations of arabinose, it was important to show whether Tat(BC) was produced also without arabinose. Moreover, the correct localization of the Tat(BC) protein in the membrane had to be established. Therefore, periplasmic, membrane, and cytoplasmic fractions were prepared from E. coli ⌬tatABCDE (pJDT12) cells grown without arabinose and, for comparison, E. coli MC4100 cells containing vector pBAD24 (without insert). As shown in Fig. 7B, Tat(BC) could only be detected in the membrane fraction, demonstrating that it is properly located in the membrane. The level of Tat(BC) was substantially lower than that of TatB in wild-type cells, but, to avoid any toxic effects of arabinose and prevent an increased lag phase, the following experiment was performed using cells grown without arabinose. We first tested whether the pJDT12 plasmid encoding TatA and Tat(BC) could complement the growth defect of the E. coli ⌬tatABCDE strain on liquid minimal TMAO/glycerol medium. Fig. 8 shows comparisons of the growth rates of wild-type MC4100 cells with those of ⌬tatABCDE containing the pBAD24 vector alone or pJDT12 expressing the tatA(BC) operon. The results show that the growth rate of E. coli ⌬tatABCDE (pJDT12) was very similar to that of the wild-type strain (Fig. 8), whereas the control strain E. coli ⌬tatABCDE (pBAD24) did not grow at all under these conditions. These results demonstrate that, despite the low levels of Tat(BC) as compared with TatB in wildtype cells, Tat(BC) is able to restore the growth defect of a strain lacking both TatB and TatC. This shows not only that both TatB and TatC domains are functional individually but that the Tat(BC) fusion protein is also functional as a unit.
The periplasmic enzyme TMAO reductase (TorA), which is required for growth on minimal TMAO/glycerol medium, is a Tat-dependent substrate (5,17). Therefore, a final test on the functionality of the Tat(BC) fusion protein was performed by analyzing whether active TMAO reductase was present in the periplasm, using a methyl-viologen-linked TMAO reduction assay on a nondenaturing gel (Fig. 9). The results show that, in wild-type cells containing pBAD24, TMAO reductase activity is found in the periplasm as expected. Substantial activity is also found in the cytoplasm because not all of the TMAO reductase is exported under these conditions, as found in other studies (17). No periplasmic activity is detected in the ⌬tatABCDE cells containing this vector, again as expected because the export of this protein is completely dependent on the Tat pathway (5,17). Significantly, TMAO reductase is found in the periplasm of ⌬tatABCDE cells expressing the tatA(BC) operon (Fig. 9, lanes pJDT12) confirming that the Tat pathway is operational. The export efficiency is low in cells grown without arabinose, but upon induction by arabinose the export efficiency is clearly increased to the point where export efficiency is comparable with that found in wild-type cells. In conclusion, our data show that the Tat(BC) fusion protein is active in cells lacking TatB and TatC, supporting the hypothesis that TatB and TatC form a functional unit within the Tat complex.
The Tat(BC-s) Chimera Has a Molecular Mass of ϳ600 kDa-To test whether the Tat(BC) chimera forms large molecular mass complexes on its own, the Tat(BC) chimera was modified to contain a carboxyl-terminal Strep tag (as added to wildtype TatC in the purification work described above). Addition of the Strep tag does not affect activity because the chimera is active in ⌬tatB and ⌬tatC cells (not shown). The construct was expressed in the absence of TatA (in ⌬tatABCDE cells), and Tat(BC-s) was purified using the first three steps of the purification protocol as described for the TatABC-s complex. The results of the last gel filtration step are shown in Fig. 10. The Tat(BC-s) protein elutes in a peak corresponding to a molecular mass of ϳ600 kDa, demonstrating that the Tat(BC-s) protein forms a large complex even in the absence of TatA. DISCUSSION In the present study we have purified a complex from E. coli containing all three of the major known Tat components, namely TatA, TatB, and TatC. We find no evidence for the presence of novel membrane-bound proteins that would represent the products of hitherto unidentified tat genes. We emphasize, however, that the purified complex has not been shown to be active in an in vitro assay, and we cannot therefore rule out the possibility that further subunits may remain to be identified.
An important point to emerge from this study is that TatB and TatC are clearly present in a fixed 1:1 ratio, suggesting a close structural association between these subunits. This suggestion is reinforced by the finding that cells lacking tatB and tatC genes but containing instead a single polypeptide in which TatB and TatC are fused are capable of Tat-dependent protein translocation. This was demonstrated by the ability of these cells to grow by anaerobic respiration on TMAO and their ability to export active TMAO reductase into the periplasm, even when expression levels of the Tat(BC) chimera are low. These results show that TatB and TatC form a functional unit within the Tat complex, and we conclude from these data that TatB and TatC must act in concert.
These data have implications for the Tat translocation mech-anism. At present we do not know which subunits act as either the initial receptor or as the translocation channel. However, our data indicate that TatB and TatC are likely to carry out one or more particular functions together, and this in turn suggests that TatA may serve a different function. This idea is supported by the observation that the amount of the third component of the translocase, TatA, varies much more in the purification steps. This may be significant in terms of the translocation mechanism. One possibility is that the "core" TatB/C complex contains low amounts of TatA and that additional molecules are recruited to form the full, active complex. The stably bound TatA molecules might, for example, serve as a nucleation point for the binding of further TatA molecules. We have shown before that only a minor fraction of the total TatA pool is in a complex with TatB (11), and this study has confirmed the following point: the majority of the Tat(BC) is efficiently bound by the affinity column, but the vast majority of the TatA is not. However, it is important to stress that the fraction of TatA that is in a complex with Tat(BC) is fairly tightly bound, because TatA was still present after purification using a variety of columns and an immunoprecipitation step with anti-TatB serum. Co-immunoprecipitation experiments furthermore demonstrated that TatC is required for the coimmunoprecipitation of TatA using anti-TatB serum, suggesting that TatA and TatB may not interact directly, but only through TatC (although this point remains to be tested more directly, because TatC may simply affect the conformations of TatA or TatB).
The apparent molecular mass of the purified Tat complex is ϳ600 kDa, indicating that multiple TatA, B, and C subunits must be present. An exact calculation of the number of subunits is difficult to make, because the molecular mass determined is that of the TatABC complex in digitonin micelles, and the detergent could contribute to the size estimation. Thus, for an accurate measurement of the Tat complex, other methods must be used. Further studies are also required to characterize the complex in other respects. First, it has to be established whether the purified complex is active in a reconstituted sys-tem, because this would help to determine whether the TatABC proteins are indeed the only components required for Tat-dependent transport and whether the unbound TatA subunits must be recruited for activity. Finally, it is presently completely unclear how the Tat complex is able to transport large folded proteins without compromising membrane impermeability. Structural analysis of the purified complex is clearly required for a more detailed understanding of its organization and mechanism.
FIG. 9. Export of active TMAO reductase in cells expressing a tatA(BC) operon. Cells of MC4100 (pBAD24) or E. coli ⌬tatABCDE containing pJDT12 (encoding TatA and Tat(BC)) were grown anaerobically in TY-GT medium until the mid-exponential growth phase. Cells of the latter strain were grown in the presence of various concentrations of arabinose (indicated in micromolars). E. coli ⌬tatABCDE (pBAD24) was also analyzed, but because this strain is not able to grow anaerobically, these cells were grown in the same medium under aerobic conditions. Next, cells were collected by centrifugation and fractionated, and TMAO reductase (TorA) activity in the periplasm (p), membrane (m), or cytoplasm (c) was visualized by analysis on a nondenaturing polyacrylamide gel and activity staining. 5 g of protein was loaded in each lane.
FIG. 10. The Tat(BC-s) chimera has a molecular mass of ϳ600 kDa. E. coli ⌬tatABCDE cells containing p(BC-s) were grown aerobically in TY medium, and the Tat(BC-s) protein was partially purified using the same protocol as described for the TatABC-s complex. Only the results of the third step, i.e. purification on the Superose 6HR gel filtration column, are shown. The TatBC-s protein in each fraction was analyzed by SDS-PAGE and Western blotting using antibodies to TatB. The elution profiles of molecular mass markers (see legend to Fig. 4) are indicated.