Escherichia coli Twin Arginine (Tat) Mutant Translocases Possessing Relaxed Signal Peptide Recognition Specificities*

The twin arginine (Tat) secretion pathway allows the translocation of folded proteins across the cytoplasmic membrane of bacteria. Tat-specific signal peptides contain a characteristic amino acid motif ((S/T)RRXFLK) including two highly conserved consecutive arginine residues that are thought to be involved in the recognition of the signal peptides by the Tat translocase. Here, we have analyzed the specificity of Tat signal peptide recognition by using a genetic approach. Replacement of the two arginine residues in a Tat-specific precursor protein by lysine-glutamine resulted in an export-defective mutant precursor that was no longer accepted by the wild-type translocase. Selection for restored export allowed for the isolation of Tat translocases possessing single mutations in either the amino-terminal domain of TatB or the first cytosolic domain of TatC. The mutant Tat translocases still efficiently accepted the unaltered precursor protein, indicating that the substrate specificity of the translocases was not strictly changed; rather, the translocases showed an increased tolerance toward variations of the amino acids occupying the positions of the twin arginine residues in the consensus motif of a Tat signal peptide.

Transport of proteins across biological membranes is a fundamental process in all living cells. In almost all cases, proteins that have to exit the cytosol are initially synthesized as larger precursor proteins possessing an amino-terminal signal peptide (1). In eubacteria, the export of the vast majority of extracytosolic proteins is mediated by the general (Sec) secretion system. Powered by the translocation motor SecA, Sec-dependent proteins are translocated across the plasma membrane in a more or less unfolded state through a protein-conducting channel (SecYEG) (for a recent review, see Ref. 2). Besides the Sec system, many bacteria possess another protein export pathway for the translocation of a subset of proteins, which in many cases, contains a tightly bound cofactor (for reviews, see Refs. [3][4][5][6]. This Sec-independent mechanism, which translocates its substrates in a fully folded or even oligomeric form across the plasma membrane, has been designated Tat (for twin-arginine translocation) due to the fact that a characteristic amino acid motif ((S/T)RRXFLK) including two consecutive arginine residues can be identified in the signal peptide of the respective precursor proteins (7,8).
Various site-directed mutagenesis studies have manifested the importance of the two arginines for Tat-dependent export. In most cases the substitution of either arginine by other amino acids resulted in a complete block of transport or a severe reduction in the export efficiency (Ref. 9 and references therein). These results, together with their high conservation, strongly suggests that the two arginine residues are a crucial part of the signal by which Tat-dependent precursors are specifically recognized by one or more components of the Tat export machinery.
In Escherichia coli four genes (tatA, tatB, tatC, and tatE) have been found that encode the membrane-integral components of the Tat translocation apparatus (8,10). Based mostly on biochemical evidence from Escherichia coli (11) and the thylakoidal Tat (⌬pH) pathway (12), the current model of Tat-dependent protein translocation predicts that the precursor proteins are recognized by a complex consisting of TatB and TatC. After this initial recognition step, multiple copies of the pore component TatA are recruited to the TatBC-precursor complex, a step that is dependent on the presence of a transmembrane H ϩ (⌬pH) gradient. Subsequently, the substrate is translocated across the membrane, and after cleavage of the signal peptide, is released on the trans side of the membrane. Redissociation of TatA from TatBC resets the Tat system for further rounds of substrate recognition and translocation. The tatE gene encodes a paralogue of TatA that is expressed at a very low level and, due to this fact, is currently regarded as a cryptic gene duplication of tatA (13).
A detailed cross-linking study from E. coli in which a reactive amino acid residue was introduced at different positions into the signal peptide of the Tat substrate preSufI has shown that a hierarchy exists in the steps of substrate-translocase interactions (11). First, the precursor contacts the primary signal peptide receptor TatC by virtue of the region encompassing the highly conserved RR motif. Subsequently, the precursor is passed to TatB, which in addition to the RR consensus motif region also contacts the hydrophobic region of the signal peptide. No signal peptide-TatB/C cross-links were obtained when the two arginines of the RR motif were conservatively replaced by lysine residues, stressing again the importance of these amino acids in substrate recognition.
In this work we have analyzed the specificity of signal peptide recognition by the Tat translocase using a genetic approach. Replacement of the two arginine residues in the signal peptide of a Tat-specific, selectable reporter protein (TorA-MalE) by a lysine-glutamine pair resulted in an export-defective precursor protein (TorA(KQ)-MalE) that is no longer accepted by the wild-type Tat translocase. Selection for restored export of the TorA(KQ)-MalE reporter allowed for the isolation of mutant Tat translocases, containing amino acid alterations in either the amino-terminal domain of TatB or the first cytosolic domain of TatC. Besides the finding that the mutant Tat translocases have gained the ability to handle a Tat precursor protein possessing a drastic alteration in the highly conserved twin arginine motif in the signal peptide, they fully retained their ability to recognize and translocate the unaltered TorA-MalE precursor. Our results strongly indicate that, in the mutant Tat translocases, the specificity of the twin arginine binding site in the TatBC substrate receptor complex has been significantly relaxed.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Culture Conditions-The bacterial strains and plasmids used in this study are listed in Table 1. Bacterial strains were grown at 37°C in Luria Bertani medium (14), minimal medium (15) supplemented with 0.4% maltose, or MacConkey agar base medium (Difco) supplemented with 1% maltose. If required, isopropyl-␤-D-thiogalactopyranoside was used at a 0.1 mM concentration. Antibiotic supplements were at the following concentrations: kanamycin, 50 mg/liter; chloramphenicol, 25 mg/liter; tetracycline, 15 mg/liter.
In vitro mutagenesis of pHSG-TatABCE using hydroxylamine (causing GC 3 AT transitions) as a mutagenic agent was performed as described by Humphreys et al.  concentration) and washed several times with 70% ethanol. The mutagenized plasmid pools were independently transformed into GSJ101 (pTorA(KQ)-MalE) by electroporation. The transformed cells were plated on minimal medium agar plates containing 0.4% maltose and incubated at 37°C for up to 5 days. Some of the single mutant colonies that appeared on the selection plates were randomly picked and restreaked on the same medium, and those isolates that showed reproducible growth were chosen for further characterization. From these isolates, plasmid pHSG-TatABCE was isolated, and the tatABCE gene fragment was transferred into fresh, non-mutagenized pHSG575 vector to eliminate possible mutations that might have occurred in the vector part of pHSG-TatABCE. The resulting pHSG-TatABCE plasmids were subsequently used for DNA sequence analysis and further functional characterizations. The E. coli tatABCE genes were also mutagenized via errorprone PCR (ep-PCR) as described by Jaeger et al. (22). A standard amplification reaction that resulted in a frequency of 1 to 7 point mutations per kilobase contained 20 ng of plasmid pHSG-TatABCE as a template, 5 pmol each of primers ep-for (5Ј-TTA CGA ATT CCC AAT TCG AGC TCG G-3Ј) and ep-rev (5Ј-GCA GGT CGA CGG ATC CCC-3Ј), 6 mM MgCl 2 , 0.1-0.3 mM MnCl 2 , 0.2 mM dNTPs, and 3 units of Taq polymerase (MBI Fermentas) in a total volume of 50 l. The primers used contained EcoRI (ep-for) and SalI (ep-rev) restriction sites. After completion of the PCR, the amplified tat genes were cut with EcoRI and SalI and ligated into EcoRI/SalI-digested pHSG575. After transformation of the ligation products into E. coli DH5␣, about 20000 clones were obtained from which a library of mutagenized, plasmid-borne E. coli tatABCE genes was isolated. Small aliquots of this pool were transformed into GSJ101 (pTorA(KQ)-MalE) by electroporation. As described above, the transformed cells were plated on minimal medium agar plates containing 0.4% maltose and incubated at 37°C for up to 5 days. Again, some of the single mutant colonies that appeared on the selection plates were randomly picked and restreaked on the same medium, and those isolates that showed reproducible growth were chosen for further characterization. From these isolates, plasmid pHSG-TatABCE was isolated and subsequently used for DNA sequence analysis and further functional characterizations.
Miscellaneous Procedures-Fractionation of cells into a fraction containing the cytosol and membranes (C/M) 6 and a periplasmic fraction (P) was done by using an EDTA-lysozyme spheroplasting method as described earlier (18) with slight modifications. 50 ml of Luria Bertani medium were inoculated with 700 l from 5 ml of overnight cultures of the respective E. coli strains that were subsequently grown to an A 600 of about 2. The cells were harvested by centrifugation and washed once with 30 mM ice-cold Tris-HCl, pH 8.0. Subsequently, the cells were resuspended in 40 l of ice-cold sucrose solution (20% sucrose, 30 mM Tris-HCl, pH 8.0), after which 20 l of ice-cold EDTA-lysozyme solution (1 mg of lysozyme/ml, 0.1 M EDTA) was added. After incubation on ice for 5 min, the resulting spheroplasts were spun down at 17,600 ϫ g for 15 min at 4°C. After sampling of the supernatant (which represents the P fraction), the pellet was washed with sucrose solution, resuspended in 1 ml of ice-cold 30 mM Tris-HCl, pH 8.0, and disrupted in a bead mill by stirring with glass beads (0.1-0.25 mm in diameter). The glass beads were removed by centrifugation, and the supernatant (which corresponds to the C/M fraction) was sampled. C/M and P fractions corresponding to an identical number of cells were subjected to SDS-PAGE and Western blotting using anti-MalE or anti-transaldolase B antibodies as described earlier (23).
For the preparation of membranes, the cells were grown overnight in 10 ml of Luria Bertani medium. The cells were harvested by centrifugation, washed once with 10 mM Tris-HCl, pH 7.5, and subsequently broken by ultrasonification (UP 200 S; Dr. Hielscher GmbH, Teltow, Germany; amplitude, 60%; cycle, 0.6). After removal of residual cell debris by a low speed centrifugation step (15 min at 4°C at 17,600 ϫ g), the membranes were isolated by ultracentrifugation (1 h at 4°C at 200,000 ϫ g). The membrane pellet was washed once with 1 ml of 1 M potassium acetate and subsequently resuspended in 200 l of 1% Triton X-100, 10 mM Tris-HCl, pH 7.5. Protein concentrations in the samples were determined by the method of Bradford (24).
SDS-PAGE and Western blotting using anti-TatA, anti-TatB, or anti-TatC antibodies were done as described earlier (23). Western blotting using anti-MalE was performed by using the ECL Western blotting detection kit (GE Healthcare) according to the manufacturer's instructions. The chemiluminescent protein bands were recorded and quantified using the CCD camera and image analyzing system Fujifilm LAS-1000 (Fuji Photo Film) together with the software AIDA 2.41 (Raytest).

Substitution of the Twin Arginine Residues in the Signal Peptide of a TorA-MalE Reporter Protein by Lysine-Glutamine Prevents Its Tat-dependent Membrane Translocation-Previously,
we have established a sensitive Tat-specific reporter system (TorA-MalE) that allows an easy in situ detection of Tat-dependent protein export on indicative media (17,18). The plasmid-encoded TorA-MalE reporter protein consists of the mature part of the periplasmic maltose-binding protein (MalE) fused to the signal peptide of the Tat-dependent periplasmic trimethylamine N-oxide reductase (TorA). The presence of MalE in the periplasm is absolutely required for growth of E. coli on minimal agar plates containing maltose as the sole carbon source (25) and, in addition, for the formation of red colonies on MacConkey agar plates containing maltose (26). When plasmid pTorA-MalE is transformed into GSJ101 (a malE-negative derivative of the tat deletion strain DADE (17,27), growth on maltose minimal medium and red colonies on MacConkey maltose agar plates are only observed when plasmid pHSG-TatABCE (containing the known tat genes cloned in an operon-like fashion) but not when the empty vector pHSG575 is cotransformed into the same strain, showing that TorA-MalE export strictly requires the presence of a functional Tat system (Table 2) (17,18).
The importance of the conserved twin arginine residues (RR) in the TorA signal peptide for Tat-dependent export of TorA-MalE was analyzed by replacing them with either two lysine (KK) residues or a lysine-glutamine (KQ) pair. Tat ϩ (but not Tat Ϫ ) cells expressing the TorA(KK)-MalE reporter were still able to form red colonies on MacConkey maltose plates and to grow on maltose minimal medium although at a slower rate when compared with Tat ϩ cells expressing the unaltered reporter (Table 2). These results show that the conservative replacement of RR by KK obviously reduced but not completely abolished the Tat-dependent export of TorA-MalE into the periplasm. In contrast, a less conservative change of RR to KQ resulted in an export-defective TorA(KQ)-MalE mutant precursor protein, since Tat ϩ cells expressing this reporter variant showed pale colonies on MacConkey maltose plates and did not grow on maltose minimal medium ( Table 2).
Taken together, these results clearly demonstrate that the TorA signal peptide mediates Tat-dependent export of MalE into the periplasm of E. coli and, furthermore, underscores the importance of the conserved twin arginine residues for the productive recognition and membrane translocation of Tat precursor proteins by the Tat export machinery.

Isolation of Tat Mutants Restoring Membrane Translocation of the Otherwise Export-defective TorA(KQ)-MalE Precursor
Protein-As shown in the previous chapter, the replacement of the conserved RR residues by KQ in the TorA signal peptide completely prevented recognition and/or membrane translocation of the corresponding TorA(KQ)-MalE precursor protein.
Next, we asked whether mutant Tat translocases can be identified that possess a relaxed specificity with respect to the requirement for the RR amino acid residues in the consensus motif of Tat signal peptides and, therefore, might allow export of TorA(KQ)-MalE. Plasmid pHSG-TatABCE was mutagenized either in vivo using 2-aminopurine or 5-azacytidine or in vitro using hydroxylamine as mutagens. The differently mutagenized pools of pHSG-TatABCE were transformed into GSJ101 containing pTorA(KQ)-MalE, and the resulting transformants were plated onto maltose minimal agar plates. After 2 days of incubation, the formation of single colonies could be observed. 14 randomly chosen colonies that after restreaking showed reproducible growth were selected for further characterization. Plasmid pHSG-TatABCE was isolated from the mutant bacteria, and the entire tatABCE region was recloned into fresh, non-mutagenized pHSG575 to eliminate possible mutations in the vector part. The reconstructed pHSG-TatABCE plasmids were retransformed into GSJ101(pTorA(KQ)-MalE), and the resulting transformants were retested for growth on maltose minimal agar. In all cases the corresponding cells were able to grow, indicating that in fact mutations in the tatABCE genes must be responsible for the observed growth and, furthermore, suggesting that these mutations must have created mutant Tat translocases that now allow export of at least some MalE into the periplasm.
DNA sequencing of the entire tatABCE region of the 14 mutant clones revealed that two different types of mutant plasmids encoding altered Tat translocases (subsequently named KQS for KQ suppressor; Table 3 and Fig. 1) could be identified. In KQS100, isolated 13 times in independent experiments, the

Phenotype of bacterial strains on MacConkey maltose and maltose minimal medium
Bacterial strains were streaked on minimal medium agar plates containing 0.4% maltose as the sole carbon source or on MacConkey agar plates containing 1% maltose and incubated at 37°C. ϩϩϩ, fast growth; ϩϩ medium growth; ϩ, slow growth; Ϫ, no growth.

Bacterial strain
Growth on maltose minimal medium

TABLE 3 Amino acid alterations present in mutant translocases
Mutations present in the multiple mutants KQS101,KQS102, and KQS103 that allow suppression of the TorA(KQ)-MalE export defect also in a single context are indicated in bold. leucine at position 9 of TatC (located in the cytoplasmically localized extreme amino-terminal part) was replaced by a phenylalanine. Mutation KQS200 resulted in the replacement of the glutamate at position 8 of TatB (located in the short periplasmic domain) by a lysine residue. Besides the chemical mutageneses described above, we additionally used the method of ep-PCR for the introduction of mutations into the tat genes. The ep-PCR conditions were adjusted such that 1-7 mutations were introduced per kilobase of template DNA, which is a significantly higher mutation frequency than the one we had obtained by chemical mutagenesis (1-3 nucleotide alterations per kilobase). The tatABCE ep-PCR fragment was cloned into pHGS575, and the resulting pool of mutagenized pHSG-TatABCE plasmids was used in the selection for suppressors of the TorA(KQ)-MalE export defect as described above. Three different types of mutant plasmids, all containing multiple amino acid changes affecting TatB, TatC, and/or TatE, could be identified (Table 3 and Fig. 1). In KQS101, three mutations (TatC(K18M,P194S) and TatB(A135T)) were found. Likewise, three mutations were present in mutant plasmid KQS102 (TatC(N22I,V35A) and TatE(K41T)). Five mutations (TatC(K18E,Y195H,V198A,G238D) and TatB(Q23R)) were identified in mutant KQS103. Interestingly, all three multiple mutants had one amino acid alteration that, like the TatC(L9F) mutation present in KQS100, is located in the cytoplasmically located, extreme amino-terminal end of TatC. To test whether the cor-responding mutations at these positions are, in fact, the ones that exert the suppressing activity in the multiple mutants, we separately introduced these mutations into the tatC gene by site-directed mutagenesis, resulting in mutant translocases KQS104 (TatC(K18M)), KQS105 (TatC(K18E)), and KQS106 (TatC(N22I)).

Mutant translocase Amino acid alterations
Next, the amounts of the proteins TatA, TatB, and TatC in the membrane fraction of the corresponding strains were analyzed by Western blotting using specific antibodies (Fig. 2). In all cases, the amounts of the Tat proteins present in the membrane fractions of the cells expressing the mutant Tat translocases were equal or slightly lower when compared with the amounts of Tat proteins in the membrane fraction of the cells expressing a Tat wild-type translocase. This finding excludes that the observed effects are due to increased amounts of Tat proteins in the strains expressing the Tat mutant translocases and are indeed caused by the corresponding amino acid alterations.

The Mutant Tat Translocases Mediate Export of TorA(KQ)-MalE into the Periplasm to Various Degrees-Next, TorA(KQ)-MalE export in the strains expressing the various mutant Tat translocases was analyzed (i) indirectly by plate assays and (ii)
directly by determining the amount of MalE in the periplasm. As shown in Table 2, GSJ101(pTorA(KQ)-MalE, pHSG-TatABCE) did not grow on maltose minimal medium and formed pale colonies on MacConkey agar plates containing maltose. In contrast, GSJ101 (pTorA(KQ)-MalE, pKQS100) showed efficient growth on maltose minimal medium and the formation of red colonies on MacConkey maltose agar plates. Significant, but somewhat slower growth and the formation of pink colonies was observed with the strains expressing the other KQS mutant Tat translocases. Interestingly, the strains expressing TorA(KQ)-MalE in combination with the single mutant Tat translocases KQS104, KQS105, and KQS106, generated by site-directed mutagenesis, all showed identical behavior in the plate assays as the strains expressing the multiple mutant translocases KQS101, KQS102, and KQS103, from which they were derived, indicating that the amino acid alterations, present in the cytosolic extreme amino-terminal part of TatC, are in fact the ones that are primarily responsible for the suppressing activity of the multiple mutated Tat translocases.
Because the plate assays only indirectly reflect the presence of MalE in the periplasm in a semiquantitative manner, we directly and quantitatively analyzed the export of TorA(KQ)-MalE in the corresponding cells by determining the subcellular localization of MalE-derived polypeptides after EDTA-lysozyme spheroplasting (18). The fractions encompassing C/M and P, respectively, were separated by SDS-PAGE followed by Western blotting using MalE-specific antibodies. As a control for the quality of the fractionation experiments, the subcellular distribution of the cytoplasmic enzyme transaldolase B was ana-  lyzed in parallel. As expected, transaldolase B was found exclusively in the C/M fraction of all cells examined (data not shown).
As shown in Fig. 3B, various MalE-derived polypeptides can be detected in the C/M fraction of cells coexpressing the unaltered TorA-MalE together with the wild-type Tat translocase (lane 1). As described previously (18), these bands correspond to the unprocessed precursor protein and various cytosolic degradation products of it. In the P fraction (Fig. 3A, lane 1) of the same cells, a mature-sized MalE-derived polypeptide is present that corresponds to the MalE protein that has been translocated across the cytoplasmic membrane in a Tat-dependent manner (18). In the C/M fractions of the strains expressing the TorA(KQ)-MalE reporter in combination with either the wildtype or the KQS mutant Tat translocases, a similar pattern of bands and a pronounced accumulation of unprocessed precursor is observed (Fig. 3B, lanes 2-10). As expected from the plate assays, no mature MalE is visible in the P fraction of GSJ101(pTorA(KQ)-MalE; pHSG-TatABCE), directly demonstrating that the TorA(KQ)-MalE mutant reporter is practically not exported by the wild-type Tat translocase (Fig. 3A, lane 2). In contrast, various amounts of mature MalE can be detected in the P fraction of the strains expressing TorA(KQ)-MalE together with the KQS mutant translocases (Fig. 3A, lanes  3-10). Quantification revealed that the amount of MalE translocated into the periplasm of GSJ101(pTorA(KQ)-MalE; pHSG-TatABCE) is extremely low (ϳ0.8%) compared with the positive control GSJ101(pTorA-MalE; pHSG-TatABCE) that was set to 100%. In the strains expressing the mutant KQS Tat translocases, values between 2.9% (KQS103) and 44% (KQS100) were observed (Fig. 3C). Together with the results obtained from the plate assays, these results also directly show that, in contrast to the Tat wild-type translocase, the KQS mutant translocases have gained the ability to recognize and translocate the normally export-defective TorA(KQ)-MalE precursor to various degrees.

The KQS Mutant Tat Translocases Still Allow Efficient Membrane Translocation of the Unaltered TorA-MalE Reporter-
Subsequently, we asked whether the KQS mutant translocases have completely changed their substrate specificity or whether they are still able to accept the unaltered TorA-MalE precursor. Plating of GSJ101 coexpressing TorA-MalE together with the KQS mutant translocases on maltose minimal medium showed that, in all cases, growth of the corresponding strains was indistinguishable from growth of GSJ101 (pTorA-MalE, pHSG-TatABCE). Furthermore, all strains formed red colonies on MacConkey maltose agar plates ( Table 2). The results from the plate assays, already indicative for efficient export of MalE into the periplasm, were directly confirmed by cell fractionation experiments. As described above, the cells were fractionated by EDTA-lysozyme spheroplasting (18), and the fractions encompassing C/M and P, respectively, were separated by SDS-PAGE followed by Western blotting using MalE-specific antibodies. As a control for the quality of the fractionation experiments, the subcellular distribution of the cytoplasmic enzyme transaldolase B was analyzed in parallel also in these experiments. Transaldolase B was found exclusively in the C/M fraction of all cells examined (not shown). As shown in Fig. 4B, a similar pattern of MalE-derived bands, corresponding to the unprocessed precursor and cytosolic degradation products of it, were detected in the C/M fraction of all strains. Whereas no mature MalE was detectable in the P fraction of the negative control GSJ101 expressing the export-defective TorA(KQ)-MalE together with the wild-type Tat translocase (Fig. 4A, lane 2), similar amounts of mature MalE protein are present in the periplasmic fractions of the strains coexpressing the unaltered TorA-MalE precursor together with either the wild-type or the KQS mutant Tat translocases (Fig. 4, A, lanes 1 and 3-10, and  C). Taken together, these results clearly show that the KQS mutant Tat translocases have acquired a broadened substrate specificity with respect to the nature of the amino acid residues

DISCUSSION
In the present work we have shown that the replacement of the conserved twin arginine residues in the TorA signal peptide by a lysine-glutamine pair completely abolished Tat-dependent membrane translocation of a TorA(KQ)-MalE reporter pro-tein. Mutant Tat translocases were identified that showed a restoration of TorA(KQ)-MalE export to various degrees. Because the mutant Tat translocases still efficiently accepted the TorA-MalE reporter protein possessing the unaltered TorA(RR) signal peptide, the substrate specificity of the translocases was not entirely changed from RR to KQ; rather, the translocases showed an increased tolerance toward variations of the amino acids occupying the positions of the twin arginine residues in the consensus motif of a Tat signal peptide.
So far, the exact role of the highly conserved RR residues in the consensus motif of Tat signal peptides is not clear. Using an extensive cross-linking approach, Alami et al. (11) showed that the Tat precursor protein preSufI first binds to the primary substrate receptor TatC in an RR-dependent manner. Subsequently, the precursor is handed over to TatB that, in addition to the RR region, also closely contacts the hydrophobic region of the signal peptide. No cross-links of preSufI to the TatBC receptor complex were observed when the RR residues in preSufI were replaced by KK. Based on these findings, it has been proposed that the initial step in precursor binding to the Tat translocase occurs via the specific recognition of the RR consensus motif by the receptor TatBC.
In this study we have found that replacement of RR by KK in the TorA signal peptide reduced but not abolished membrane translocation of the very sensitive Tat reporter protein TorA-MalE. A similar finding was reported by Ize et al. (28) using a likewise very sensitive reporter system that is based on the bactericidal effect of colicin V (ColV), which is observed only when the colicin gains access to the plasma membrane from the periplasmic side. Replacement of RR by KK in the signal peptide of the TorA-ColV hybrid precursor significantly reduced but not completely abolished its Tat-specific membrane translocation. Together, these combined results suggest that an arginine residue at either position of the RR residues in Tat signal peptides is not absolutely required for precursor recognition by the Tat translocase and that a KK pair allows a weak binding of the TorA signal peptide to the TatBC substrate receptor. To completely block export of TorA-MalE, a more drastic alteration (i.e. RR to KQ) was required. In this case binding of the corresponding TorA(KQ)-MalE precursor to the TatBC substrate receptor most likely is too weak to establish a productive interaction. So far, almost nothing is known about the nature and the number of contacts that are required for the productive interaction of Tat signal peptides with the Tat translocase. A possible scenario would be that the Tat consensus motif in Tat signal peptides plays a crucial role in substrate recognition specificity and makes several contacts to a signal peptide binding pocket in the TatBC receptor complex. We have addressed this possibility by genetic means. If the export-defect of the TorA(KQ)-MalE precursor is indeed caused by an inability of the Tat translocase to recognize the altered signal peptide, then mutations in the translocase components might exist that restore the defective recognition step. In fact, several mutations in the tat genes that suppress the export defect of the TorA(KQ)-MalE precursor could be identified.
In the strain expressing the strongest suppressor translocase (KQS100) together with TorA(KQ)-MalE, the amount of containing plasmid pTorA-MalE in addition to plasmid pHSG-TatABCE (wild-type tat genes; positive control) or one of the various pHSG-TatABCE-KQS plasmids (KQS mutant translocases) was determined in three different independent experiments via quantification of the chemiluminescence signals. The negative control was GSJ101 containing plasmids pTorA(KQ)-MalE and pTat-ABCE (bar, KQ signal peptide, wild-type TatABCE). The signals were recorded by a CCD camera and subsequently analyzed by the program AIDA 2.41 (Raytest). The relative export efficiency of the positive control GSJ101 (pTorA-MalE, pHSG-TatABCE) was set to 100%.
mature MalE in the periplasm was 44% compared with the situation found in a strain expressing a wild-type Tat translocase in combination with the unaltered TorA-MalE reporter. The corresponding mutation was found to affect the TatC protein (TatC(L9F)), which is in perfect agreement with the biochemical data which showed that TatC is the primary substrate receptor that recognizes Tat signal peptides by virtue of the conserved RR residues (11). Furthermore, the position of the L9F mutation (see Fig. 1) suggests that the cytoplasmically exposed, extreme amino-terminal domain of TatC that precedes the first transmembrane segment might constitute or at least is part of the signal peptide binding pocket that interacts with the Tat-box of Tat signal peptides during the initial stages of the translocation process. This is further supported by the identification of additional mutations located in this domain of TatC that restored the defective interaction of the TorA(KQ) signal peptide with the translocase, although with less efficiency. The corresponding mutations, K18M, K18E, and N22I, were originally identified as one of several mutations in the Tat proteins. However, since the Tat translocases possessing the single mutations behaved similarly to the translocases containing the multiple amino acid alterations, these results strongly suggest that the mutations located in the amino-terminal domain of TatC are the ones that are primarily responsible for the suppressing activity.
In the second best suppressor translocase (KQS200), a mutation in the extreme amino-terminal domain of TatB (E8K) was identified. According to the topological model of TatB, the corresponding amino acid is located in the short periplasmic domain preceding the single transmembrane segment (Fig. 1). Because recognition of the signal peptide most likely occurs on the cis side of the plasma membrane, a direct involvement of the periplasmic amino-terminal TatB domain in this recognition event seems unlikely. If so, then the suppressing effect of the TatB(E8K) mutation, which restores significant TorA(KQ)-MalE export (i.e. 18.5% compared with the wild-type control) must be due to a long range conformational effect that is transmitted to and alters the properties of the cytosolic signal peptide binding pocket of the primary receptor subunit TatC of the TatBC receptor complex such that the TorA(KQ) signal peptide can be productively bound.
All KQS mutant translocases still allowed efficient recognition and membrane translocation of the unaltered TorA-MalE reporter protein. In addition, some of the KQS mutant translocases could also suppress the export defect of a TorA(KA)-MalE mutant precursor, in which the twin arginine residues of the TorA signal peptide have been replaced by a lysine-alanine pair. 7 Together, these findings strongly suggest that, in the mutant translocases identified in this study, the specificity of the signal peptide recognition domain has not been entirely changed from RR to KQ; rather, the specificity of signal peptide recognition has been relaxed. If binding of the signal peptide to the TatC receptor is mediated by several contacts between the conserved extended twin arginine motif ((S/T)RRXFLK) and TatC, the KQS translocase mutations either directly or indi-rectly might alter the signal peptide binding pocket in a way that a stronger binding of amino acids at a non-RR position in the extended twin arginine motif or elsewhere in the signal peptide compensates for the weakened or missing binding of the altered amino acid residues at the RR position. Thereby, a productive recognition of the TorA(KQ) signal peptide without negatively affecting recognition of the TorA(RR) wild-type signal peptide would be allowed.
Such a phenotype is reminiscent of the phenotype of the so-called prl alleles of various sec genes that have been isolated by genetic selections that were basically similar to the selection that has been performed in this study. These prl alleles were also isolated as suppressors that allow export of signal peptidedefective Sec-dependent precursor proteins (Ref. 29; for reviews, see Refs. 30 -31). prl alleles of the genes secA, secY, secE, and secG result in Sec translocases that, besides being still proficient of accepting wild-type Sec signal peptides, allow the recognition and membrane translocation of proteins containing different mutations in their signal peptides or even of proteins that entirely lack a signal peptide (32)(33)(34). Originally, it was reasoned that the Prl mutations in the Sec components would act by restoring defective signal peptide-translocase interactions. Because of a lack of allele specificity and the fact that even signal-less proteins could be exported in a prl strain background, this hypothesis was abandoned, and it was suggested that the Sec translocase possesses a proofreading activity that rejects proteins lacking a functional signal peptide and which is reduced in the Prl mutant translocases. However, recent modeling of the effects of prl mutations on the structure of the E. coli SecYEG complex suggests that the Prl variants of the channel components SecY, SecE, and SecG most likely act by destabilizing the closed state of the protein-conducting SecYEG channel or by stabilizing its open form, thereby allowing channel opening to occur without the triggering event of signal peptide binding that is required to open the channel in a wild-type SecYEG complex (35). Therefore, at present it cannot be entirely excluded that the KQS mutant translocases, isolated in this study, exert their suppressing activity by means other than restoring a defect in signal peptide recognition via the conserved twin arginine motif. Nevertheless, the mutations analyzed in this communication represent the first gain-offunction mutations described for the Tat system so far and, like the prl mutations affecting the Sec translocase components, might significantly contribute to the understanding of the molecular mechanism of Tat-mediated protein translocation across biological membranes.