Evidence for interactions between domains of TatA and TatB from mutagenesis of the TatABC subunits of the twin-arginine translocase Claire M. L. Barrett and Colin Robinson Department of Biological Sciences, University of Warwick, Coventry, UK The twin-arginine translocation (Tat) system operates in the plasma membranes of a wide range of bacteria as well as the thylakoid membrane in plant chloro- plasts (reviewed in [1,2]). Working in parallel with the Sec system, it is responsible for the export of a subset of proteins into the periplasm, outer membrane or extracellular medium, and the primary defining attrib- ute of the system is its ability to transport proteins in a fully folded state [3,4]. Particular attention has centred on a series of periplasmic proteins that are exported only after binding redox cofactors such as FeS or molybdopterin centres [5–8] although it should also be emphasized that the system also transports proteins that do not bind cofactors [1,2]. Substrates for the Tat pathway are exported post- translationally [8] after synthesis with cleavable, N-ter- minal signal peptides that almost invariably contain an essential twin-arginine motif in the N-terminal domain [9,10]. They then interact with a translocon in the inner membrane that consists, minimally, of three sub- units (TatABC) in Escherichia coli and several other Gram-negative bacteria studied to date. Genetic stud- ies indicate that the tatABC genes are all important for Tat activity although a fourth gene, tatE, encodes Keywords green fluorescent protein (GFP); Tat system; twin-arginine; protein transport; signal peptide Correspondence C. Robinson, Department of Biological Sciences, University of Warwick, Coventry, CV4 7AL, UK Fax: +44 2476523701 Tel: +44 2476523557 E-mail: Crobinson@bio.warwick.ac.uk (Received 13 December 2004, revised 25 February 2005, accepted 8 March 2005) doi:10.1111/j.1742-4658.2005.04654.x The twin-arginine translocation (Tat) system transports folded proteins across the bacterial plasma membrane. Three subunits, TatA, B and C, are known to be involved but their modes of action are poorly understood, as are the inter-subunit interactions occurring within Tat complexes. We have generated mutations in the single transmembrane (TM) spans of TatA and TatB, with the aim of generating structural distortions. We show that sub- stitution in TatB of three residues by glycine, or a single residue by proline, has no detectable effect on translocation, whereas the presence of three gly- cines in the TatA TM span completely blocks Tat translocation activity. The results show that the integrity of the TatA TM span is vital for Tat activity, whereas that of TatB can accommodate large-scale distortions. Near-complete restoration of activity in TatA mutants is achieved by the simultaneous presence of a V12P mutation in the TatB TM span, strongly implying a direct functional interaction between the TatA ⁄ B TM spans. We also analyzed the predicted amphipathic regions in TatA and TatB and again find evidence of direct interaction; benign mutations in either subunit completely blocked translocation of two Tat substrates when present in combination. Finally, we have re-examined the effects of previously ana- lyzed TatABC mutations under conditions of high translocation activity. Among numerous TatA or TatB mutations tested, TatA F39A alone blocked translocation, and only substitutions of P48 and F94 in TatC blocked translocation activity. Abbreviations GFP, green fluorescent protein; Tat, twin-arginine translocation; TM, transmembrane; TMAO, trimethylamine N-oxide; TorA, TMAO reductase. FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS 2261 a TatA paralog of minor importance in some species [7,8,11–13]. Only two tat genes (designated tatC and tatA) are thought to be important in some Gram- positive species, with a single gene product apparently fulfilling both TatA and TatB functions [14–16]. Studies on the Tat mechanism are at an early stage. The Tat subunits are not related to any proteins in the database and most studies point to a mechanism that is unique among known protein transport systems. However, recent studies have begun to unravel some salient features of this system. Protein expression ⁄ puri- fication approaches have resulted in the characteriza- tion of two distinct complexes in E. coli: a TatABC complex and homo-oligomeric TatA complex. The TatABC complex has a mass of  600 kDa in deter- gent and contains multiple copies of TatABC; within this complex, TatB and TatC are in stoichiometric amounts and the two subunits appear to function as a unit [17]. Approximately equal numbers of TatA sub- units are present [17,18] but the vast majority of TatA is found as separate, apparently homo-oligomeric com- plexes [17,19,20]. In vitro cross-linking studies on the plant thylakoid [21] or E. coli Tat system [22] have shown that substrates initially bind to the TatB and TatC subunits, and it thus appears that these subunits cooperate to form the substrate binding site. In plants, the TatA homolog was only found to cross-link to the Hcf106 ⁄ cpTatC complex (corresponding to bacterial TatBC) in the presence of substrate and a proton motive force [23]. On the basis of these studies, it has been proposed that binding of substrate to the TatBC subunits triggers the recruitment of the separate TatA complex to form an active translocation system. In an effort to pinpoint important regions of the Tat subunits, the three proteins have been subjected to site-specific mutagenesis and a number of key regions or residues have been identified [24–27]. TatA and TatB are single-span proteins with C-terminal, cyto- plasmic domains and each has also been truncated from the C-terminus in order to delineate the regions important for activity [28]. Site-specific mutagenesis has also been used to assess the importance of residues in the predicted amphipathic domains and cytoplasmic regions of TatA and TatB, and the highly conserved residues of TatC have also been probed [24–27]. In this report we have analyzed the transmembrane regions of TatA and TatB, in an effort to analyze their import- ance for Tat function. We show that mutations designed to destabilize the TatB TM span through sub- stitution by proline or multiple glycine residues have no detectable effect, whereas some of the TatA mutants are severely affected or blocked in transloca- tion activity. We also present evidence for interactions between the TM spans of TatA and TatB, and between the amphipathic regions. Finally, we have re-examined the numerous mutations made previously in TatA, B and C and we present new information on potentially important TatA and TatC mutants. Results Analysis of TatA and TatB mutants The overall structures of the TatA and TatB subunits are similar: both contain a single TM span, with very short periplasmic N-terminal regions and cytoplasmic domains that are relatively small in the case of TatA ( 40 residues) and larger in TatB ( 90 residues). The TM spans and cytoplasmic domains are separated by regions that are strongly predicted to form amphi- pathic a-helices [19,20]. In the present study we have generated mutations in the TM and amphipathic regions of TatA and TatB (see below) in order to probe the importance of this region, especially with respect to a possible role for TatA as the translocation channel. In order to present a comprehensive analysis we have analyzed in parallel the translocation activity of TatABC mutants described in several previous stud- ies [24–27]. This was considered important because one of the mutants exhibited unexpected properties when compared with previous findings. The effects of the mutations were analyzed using two types of export assay. The first involves expression of the mutated tatABC operon in the arabinose-indu- cible pBAD24 vector in a tat null background (Dtat- ABCDE strain). The cells were fractionated and the distribution of a known Tat substrate, trimethylamine N-oxide (TMAO) reductase (TorA) was analysed using a native gel activity assay. This assay is not quantita- tive but defects in translocation are usually apparent through an increased accumulation of TorA in the cytoplasm. It should be noted that this vector expres- ses the tatABC operon by a factor of  10–20 fold higher compared with wild-type TatABC levels. This means that minor, and even moderate defects in trans- location activity may not be revealed because the higher levels of Tat apparatus might be able to com- pensate for defects. Moreover, this assay is qualitative rather than quantitative because the appearance of the TorA signal in the native gels is not linear with time. In summary, this assay is best suited for identification of major defects in translocation activity. The second assay involves synthesis of a construct comprising the presequence of TorA linked to green fluorescent protein (GFP), which is efficiently exported by the Tat pathway under these conditions [24,26,28]. Mutagenesis of TatABC C. M. L. Barrett and C. Robinson 2262 FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS In these experiments, synthesis of TorA-GFP was induced for 2 h using the pBAD vector, after which the arabinose was removed and IPTG was added to induce expression of tatABC from the compatible pEXT22 plasmid (although this plasmid is relatively leaky, thus TatABC are synthesized at appreciable rates throughout the growth of the cells). Membranes were isolated after a 3 h induction with IPTG. This assay is effectively semiquantitative (as the cells are again analyzed over a relatively long period) but is more reproducible than the TorA export assay and the pEXT22 plasmid produces TatABC at lower levels (we estimate approximately three- to fivefold more than wild-type [28]). Most of the data presented below involve the use of this assay but it should be empha- sized that all mutants were tested several times using the both types of export assay. TorA export data are shown only where there were minor discrepancies with the TorA-GFP data. The experimental system varied from those of previ- ous studies [24,26] in important respects. We recently found [28] that the Tat system is inhibited by the pres- ence of arabinose (for unknown reasons) and TorA export assays were conducted in a slightly different manner compared to previous studies: only 50 lm ara- binose was used for induction (instead of 200 lm). With the TorA-GFP export assays, TorA-GFP was induced with arabinose for 2 h, after which the arabi- nose was removed and the cells incubated with IPTG for 3 h to induce expression of the mutated tatABC operons. We have found that these conditions give more reproducible results and the Tat pathway of wild-type cells is shown below to be highly active at the time of analysis. First we analyzed TatABC levels in cells expressing the various mutated subunits, and the data for the TatA and TatB mutants (expressed using the pBAD vector) are shown in Fig. 1. The expression of the wild-type tatABC from the pBAD-ABC is illustrated in the indicated lane, with wild-type cells in the adja- cent MC4100 lane; it is evident that the TatA and TatB proteins are produced from pBAD-ABC at eleva- ted levels as described above. No TatC signal is evi- dent in wild-type cells because this protein is detected using antibodies to the Strep-tag II on the TatC sub- unit in the pBAD-ABC vector. In the other control lane, no Tat components are detected using mem- branes from DtatABCDE cells (denoted DABCDE in Fig. 1 and other figures) as expected. The remaining lanes contain membrane samples from DtatABCDE cells expressing pBAD-ABC in which mutations are present in the TatA or TatB subunit as indicated. Fig. 1. TatABC expression levels in cells expressing wild-type or mutated TatA ⁄ B subunits. Membranes were isolated from wild-type MC4100 cells, DtatABCDE cells (DABCDE)andDtatABCDE cells expressing pBAD-ABC containing mutations in the TatA or TatB subunit as indicated. Samples were immunoblotted using antibodies to TatA, TatB or the Strep-tag II on TatC. Asterisks denote strains in which a Strep-tag II is not detected by immunoblotting. C. M. L. Barrett and C. Robinson Mutagenesis of TatABC FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS 2263 Although the expression levels vary to some extent, all of the mutant proteins are present at similar levels, with the possible exception of TatA ⁄ G2A which exhib- its a relatively low TatB signal for unknown reasons. One anomaly is, however, evident with two newly gen- erated TatB mutants, E8Q and L63A: TatA and TatB are formed at typical levels but the TatC signal is com- pletely absent (lanes denoted by asterisks). This is reproducible and, because these mutants display high levels of Tat activity (see below), we assume that the C-terminal Strep-tag II has been removed from the proteins after expression. Some of the other mutants also have this property (see below). Although the primary aim was to analyze new mutants affected in the amphipathic or TM regions, previously analysed TatA mutants containing single amino acid changes were also analyzed in terms of their ability to export TorA and TorA-GFP, and the data for the TatA mutants are shown in Fig. 2. With the TorA export assays in Fig. 2A, it is observed that the bulk of the activity is found in the periplasm in wild-type cells and cells expressing pBAD-ABC, as expected (lanes P) with very little cytoplasmic signal evident. TorA is found exclusively in the cytoplasm in DtatABCDE cells, where it migrates more slowly in the gel system (denoted by an asterisk). The TatA mutants all export TorA with high efficiency with the exception of F39A, where all of the TorA is present as the cyto- plasmic form. Some cytoplasmic TorA is also evident with L25A. The TorA-GFP export assays are in good agreement with the TorA data (Fig. 2B). In pEXT-ABC-expres- sing cells, the bulk of GFP is found as mature-size protein in the periplasm (P), whereas GFP is found only in the cytoplasm and membrane fractions (C, F) in cells expressing the pEXT22 vector. Some of this protein is present as precursor form (TorA-GFP) and some mature-size GFP is also present, presumably due to proteolytic clipping. No signal is observed in DtatABCDE cells that do not synthesize TorA-GFP (a control for the specificity of the GFP antibodies). All of the TatA mutants export TorA-GFP with high effi- ciency except F39A, which is again completely defect- ive in translocation. Whereas some cytoplasmic TorA A B C Fig. 2. The TatA F39A mutant is inactive whereas other TatA ⁄ B mutants show no detectable loss of translocation activity. (A) Dtat- ABCDE cells expressing pBAD-ABC or the same vector containing mutations in tatA were induced using 50 l M arabinose for 3.5 h, and cytoplasmic (C) and periplasmic fractions (P) were prepared as detailed in Experimental procedures. These fractions were electro- phoresed on native polyacrylamide gels that were subsequently stained for TMAO reductase (TorA) activity. Asterisk denotes slower-migrating cytoplasmic form of TorA. (B) A TorA-GFP con- struct was expressed in DtatABCDE cells using an arabinose- inducible vector for 2 h. The arabinose was then washed out and wild-type or tatABC operons containing the same tatA mutations as in (A) were expressed for 3 h using the isopropyl thio-b- D-galacto- side-inducible pEXT22 plasmid as detailed in Experimental proce- dures (pEXT-ABC plasmid contains wild-type tatABC operon). Cytoplasmic, membrane and periplasmic fractions (C, M, P), were isolated and immunoblotted using antibodies to GFP. The mobility of mature-size GFP is indicated. (C) TatB mutants described in Fig. 1 were analyzed for export of TorA-GFP (B) exactly as des- cribed in (B) for TatA mutants. Mutagenesis of TatABC C. M. L. Barrett and C. Robinson 2264 FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS was evident with the L25A mutation, as described above, no defect is apparent using TorA-GFP as a substrate. Because signal strengths are not linearly related to protein activity in the native gel TorA assay, we are more inclined to regard the TorA-GFP data as evidence of high translocation activity, although other possibilities can not be excluded. Similar tests on the TatB mutants are shown in Fig. 2C. The results show that all of the strains effi- ciently export TorA-GFP, including the E8Q and L63A mutants that exhibited no signal with the Strep- tag II immunoblots for TatC in Fig. 1. Inhibitory effects of mutations in the predicted amphipathic regions of TatA and TatB Considerable attention has centred on possible roles of conserved predicted amphipathic regions that are highly conserved in both TatA and TatB. These regions effectively bridge the transmembrane helices and soluble cytoplasmic domains, and truncation ana- lysis [29] has shown that they are essential for translo- cation activity. Their sequences are shown in Fig. 3A. In a previous report [24], we analyzed the effects of changing three lysine residues in TatA (residues 37, 40 and 41) to glutamine, and in a second mutant we addi- tionally changed K24 to alanine. These mutants, denoted TatA ⁄ )3K and TatA ⁄ )4K previously [24] were shown to be active albeit with reduced efficien- cies. These TatA mutants have been re-made (and re- named TatA ⁄ 3K > Q and TatA ⁄ K24A,3K > Q) after finding several revertants in recent studies of previ- ously analyzed mutants. In the case of TatB, it was previously found that changing two arginines (residues 37 and 40) to asparagine (TatB ⁄ 2R > N), or three lysines (residues 65, 67 and 68) to glutamine (TatB ⁄ 3K > Q) had little effect on the efficiency of Tat-dependent export [24]. In the present report we have re-assessed these mutants and another mutant combining the two sets of mutations in TatB (TatB ⁄ 2R > N,3K > Q). This new mutant is indica- ted by ‘Ù’ in Fig. 3. The expression profiles are shown in Fig. 3B, which shows TatA and TatB to be synthes- ized in all cases, although again at slightly varying lev- els. Strep-tagged TatC is formed in every case except TatA ⁄ K24A,3K > Q, but because this mutant is act- ive (see below) we again believe that the TatC protein is present but lacking the Strep-tag II. Export assays using these mutants are shown in Fig. 4A. The data show that all three TatB mutants TatA TatA TatB TatC TatA TatB TatC pBAD-ABCs pBAD-ABCs 2R>N K24A K24A K24A K24A K24A 3K>Q 3K>Q 3K>Q 3K>Q 3K>Q 3K>Q/ 2R>N 2R>N 3K>Q 3K>Q 3K>Q 3K>Q 2R>N 3K>Q 2R>N 3K>Q ∆ABCDE ∆ABCDE TatB 16 25 30 * A BC Fig. 3. Mutations in the predicted amphi- pathic regions of TatA and TatB. (A) Primary sequences of the amphipathic regions with the targeted residues indicated by arrows and numbered. The changes introduced in the various mutants are indicated and under- lined. (B) pBAD-ABC cells, DtatABCDE cells and DtatABCDE cells expressing pBAD-ABC containing the ‘amphipathic’ mutations shown in (A) were analyzed by immunoblot- ting with antibodies to TatA, TatB and the Strep-tag II on TatC. Asterisks denote strains that exhibit no signal with Strep-tag II antibodies. (C) as in (B), except with strains expressing combinations of muta- tions in TatA + TatB. ‘^’ denotes new muta- tions analyzed in this study. Mutations in TatA are shown underlined and in grey font. Mobilities of molecular mass markers are given on the left. C. M. L. Barrett and C. Robinson Mutagenesis of TatABC FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS 2265 exhibit efficient export of TorA-GFP, in that the vast majority of protein is found in the periplasmic fraction. This includes the new TatB mutant (TatB ⁄ 2R > N, 3K > Q) in which five basic residues are substituted; surprisingly, there is little evident effect. Among the TatA mutants, TatA ⁄ 3K > Q was previously described as active [24] but Fig. 4A shows this not to be the case with the newly generated mutant, which fails to export TorA-GFP to any detectable extent. No export of TorA was observed either (data not shown). Surprisingly, the presence of an additional mutation (K24A) enables export to take place, with the TatA ⁄ K24A,3K > Q quadruple mutant exhibiting moderate export activity. Some precursor form of TorA-GFP is present in the cytoplasm and ⁄ or membrane fractions but considerable amounts of periplasmic TorA and GFP are present. It is likely that the original TatA ⁄ 3K > Q mutant [24] simi- larly acquired an additional mutation prior to analysis that enabled translocation to occur, although this has yet to be confirmed. Other TatABC mutations described previously [24,26] have been remade and shown to have unchanged properties. We also expressed tatABC operons in which these multiple mutations in the amphipathic regions were combined; in the relevant Figures the TatA mutations are shown underlined for simplicity. We previously reported data [24] on two such combinations: (TatA ⁄ K24A, 3K > Q + TatB ⁄ 3K > Q) and (TatA ⁄ 3K > Q + TatB ⁄ 2R > N), and both mutants were described as active. In this report we have reassessed the effects of these mutations (as the former TatA mutant TatA ⁄ 3K > Q is now known to be inactive on its own, as shown above) and have constructed several new permutations as detailed in Fig. 3. These new mutations are again denoted by ‘Ù’. Figure 3C shows that strains synthesizing all of the multiple TatAB mutations contain TatABC at similar levels and activ- ity assays are shown in Fig. 4(B,C). These ‘mixed amphipathic’ mutations exhibit very interesting properties. Figures 2 and 4A showed that the TatA ⁄ K24A and TatB ⁄ 2R > N mutations support wild-type levels of export activity but Fig. 4B shows that the combined (TatA ⁄ K24A + TatB ⁄ 2R > N) mutations severely disrupt activity, with no periplasmic A B C Fig. 4. Combinations of mutations in the TatA ⁄ TatB amphipathic regions have partic- ularly severe effects on translocation activ- ity. (A) Mutants containing alterations in the predicted amphipathic regions of either TatA or TatB (as detailed in Fig. 3) were analyzed for export of TorA-GFP using the assay pro- tocols detailed in Fig. 2. For clarity, muta- tions in TatA are shown underlined and in grey. (B) Combinations of mutations in the amphipathic regions of TatA and TatB, whose structures and expression profiles are illustrated in Fig. 3, were tested for the export of TorA-GFP. Mutations in TatA are shown underlined and in grey font; addition- ally, labels on the right indicate whether mutations are in TatA or TatB. (C) Dtat- ABCDE cells, or DtatABCDE expressing pBAD-ABC or the same vector containing mutations in both TatA and TatB as indica- ted, were assayed for export of TorA using the protocol described in Fig. 2. Mutagenesis of TatABC C. M. L. Barrett and C. Robinson 2266 FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS mature-size GFP detected at all after synthesis of TorA-GFP. Note that mature-size GFP accumulates in the cytoplasm, and not the full precursor protein; this is due to proteolytic clipping of the signal peptide when export is blocked [28,30]. Figure 4C shows TorA export assays in which the vast majority of TorA is found in the cytoplasm. A minor fraction of TorA is found in the periplasm, indicating that the mutant is not completed blocked in Tat-dependent transloca- tion, but this mutant is clearly badly compromised. An almost identical result is obtained with (TatA ⁄ K24A + TatB ⁄ 3K > Q), and it should again be noted that the individual TatA and TatB mutants show no appar- ent defects. These data show that the mutations have synergistic effects and the particularly severe effects on TorA-GFP export raise the intriguing possibility that these mutations somehow affect the export of GFP differently when compared to TorA. The next two mutants in Fig. 4B (from left to right) contain TatA ⁄ 3K > Q in combination with either TatB ⁄ 2R > N or TatB ⁄ 3K > Q. Both combinations exhibit no detectable Tat activity and this is unsurpris- ing as we showed above that the ‘new’ TatA ⁄ 3K > Q mutant is inactive on its own. Of the remaining mutants (TatA ⁄ K24A, 3K > Q + TatB ⁄ 3K > Q) is completely inactive, although the individual TatA and TatB mutants did exhibit activity, and the final mutant containing five changes in TatA plus four in TatB is likewise totally inactive. In all, these mutants empha- size the importance of the amphipathic region but they also show for the first time that combinations of TatA ⁄ TatB mutations can have far more dramatic effects than the individual mutations. Mutations in the transmembrane spans of TatA and TatB Deletion of the TM spans of TatA or TatB leads to a loss of activity [31] but the important characteristics of these regions have not been probed. We constructed a series of new TatA ⁄ B mutants containing changes within the TM spans, for two reasons. First, it has been suggested that TatA may form the translocation channel, in which case drastic structural alterations may be expected to selectively block the translocation event. Secondly, mutations affecting the structure and orientation of the TM span may be expected to disrupt the interactions with other Tat components, and this would provide information on the inter-subunit associ- ations occurring within and between Tat complexes. Both of these areas are poorly understood at present. It has been shown with other proteins that the intro- duction of proline residues has a marked effect on the structures of TM spans [32,33], usually introducing dis- tortions of major proportions, and such substitutions were made in the TM spans of TatA and TatB in the present study. We also substituted three residues in the TatA and TatB TM spans by glycine. The presence of glycine can also lead to increased flexibility in TM helices [34]. However, glycine residues can also play important roles in modulating inter-helix interactions [35] and the effects of inserting or removing these resi- dues may therefore be less predictable than with pro- line mutations. Nevertheless, the presence of three consecutive glycines should lead to a significant struc- tural effect in either case. The proline and glycine sub- stitutions were made near the centre of the TM span in order to maximize possible structural effects. Figure 5A shows the sequences of the TatA and TatB TM spans, together those of the mutated forms. With TatA, residues 11–13 were changed to glycine in one case (TatA ⁄ 3Gly) and a single residue was chan- ged to proline in the centre of the TM span in another (TatA ⁄ I12P). We also changed three additional residues to proline in the TatA ⁄ 3Gly mutant (TatA ⁄ 3Pro3Gly) and then a further three residues to glycine in the same subunit (TatA ⁄ 3Pro6Gly). With TatB, resi- dues 11–13 were changed to glycine (TatB ⁄ 3Gly) or a single residue to proline (TatB ⁄ V12P). The expression characteristics of these mutants are shown in Fig. 5B. With the simplest TatA mutants, TatA ⁄ 3Gly and TatA ⁄ I12P, TatABC are all present at expected levels. With TatA ⁄ 3Pro3Gly, TatA and TatB are present at the usual levels but TatC is not detected at all. However, as this mutant is highly act- ive (see below) it appears that this is another example of the C-terminal Strep-tag II being clipped or modi- fied. With the most drastic of the TatA mutants, TatA ⁄ 3Pro6Gly, TatB and TatC are synthesized but no TatA is detected. Fractionation of cells synthes- izing TatA ⁄ 3Pro3Gly shows the presence of full- length protein in the cytosol, consistent with a slight defect in membrane-insertion; in contrast, smaller fragments of the TatA ⁄ 3Gly6Pro protein are found in the cytosol (data not shown). This suggests that the TatA ⁄ 3Gly6Pro protein is degraded either within the membrane or, perhaps more likely, after failure to insert into the membrane. Cells expressing the simpler of the tatB mutants, TatB ⁄ V12P, appear to contain TatABC at expected levels but the TatB ⁄ 3Gly mutations result in a much- reduced TatB signal, presumably reflecting problems in insertion and ⁄ or stability. No TatC signal is evident, but this again appears to reflect problems in detection of the Strep-tag II as this mutant is active in Tat- dependent transport (see below). C. M. L. Barrett and C. Robinson Mutagenesis of TatABC FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS 2267 Figure 6 shows TorA and TorA-GFP export assays for these TM span mutants. Of the TatA mutants, TatA ⁄ I12P shows no detectable defect because GFP and TorA are found predominantly in the periplasm. The TatA ⁄ 3Gly mutant, on the other hand is blocked in export and no translocation activity can be detected. The same applies to the TatA ⁄ 3Pro6Gly mutant, although this is expected because the immunoblots in Fig. 5 show no signal for TatA. The real surprise is the TatA ⁄ 3Pro3Gly mutant, which is shown to export both TorA and TorA-GFP with high efficiency. Given that the parent TatA ⁄ 3Gly mutant is completely inac- tive, this result indicates that the three proline residues somehow compensate for the inhibitory effects of the glycine residues and enable translocation to occur. Finally, Fig. 6 shows that the two TatB mutants, TatB ⁄ 3Gly and TatB ⁄ V12P, are highly active in export; perhaps surprisingly given the drastic effects of the 3Gly mutations in TatA. We also tested the effects of expressing tatABC operons carrying combinations of these mutations in the TM spans of both TatA and TatB, namely (TatA ⁄ 3Gly + TatB ⁄ 3Gly) (TatA ⁄ 3Gly + TatB ⁄ V12P) (TatA ⁄ I12P + TatB ⁄ 3Gly) and (TatA ⁄ I12P + TatB ⁄ V12P). Immunoblots confirmed that the TatABC were syn- thesized at approximately the same levels as the wild- type subunits generated from pBAD-ABC (data not shown), and activity assays are shown in Fig. 7. With the TorA assays, the control tests show efficient export with wild-type TatABC and a complete block in export in the tat null mutant (denoted DABCDE), as expected. The combination of (TatA ⁄ 3Gly + TatB ⁄ 3Gly) is blocked in Tat function, and this is not unexpected given that the TatA ⁄ 3Gly mutant itself shows no export activity. However, a combination of (TatA ⁄ 3Gly + TatB ⁄ V12P) displays very efficient export of TorA and this shows that the TatB ⁄ V12P mutation compensates for the drastic effects of the 3Gly muta- tions in TatA. This is confirmed by the TorA-GFP export assays, which reveal a complete block in export with the (TatA ⁄ 3Gly + TatB ⁄ 3Gly) mutant but near wild-type export efficiency with (TatA ⁄ 3Gly + TatB ⁄ V12P). The remaining (TatA ⁄ I12P + TatB ⁄ 3Gly) and (TatA ⁄ I12P + TatB ⁄ V12P) mutants export both TorA and TorA-GFP efficiently, in keeping with the finding that none of the mutations affect export to a detect- able extent when present in TatA or TatB alone (see Fig. 6). In conjunction with the data shown in Fig. 6, these data show that the severe effects of the TatA ⁄ 3Gly mutations can be rescued by the presence of additional mutations either elsewhere in the TatA TM span or in the TatB TM span (TatB ⁄ V12P mutation). Mutagenesis of TatC TatC has also been studied in previous reports and a number of mutations have been characterized [26,27], but some apparent differences were reported in studies on the same mutants by different groups (see below). Few TatC residues are highly conserved but of these, a high proportion is clustered in the N-terminal cyto- plasmic domain and the first cytoplasmic loop. In our previous analysis [26] only two residues were found to A B Fig. 5. Expression of TatA and TatB mutants containing alterations in the transmembrane spans. (A) Sequences of the TM regions of TatA and TatB, with residues numbered and the substitutions indi- cated. (B) Expression of the various mutations constructed within the pBAD-ABC vector. Samples were analyzed by immunoblotting with antibodies to TatA, TatB and the Strep-tag II on TatC. Mobili- ties of molecular mass markers are given on the left. Mutagenesis of TatABC C. M. L. Barrett and C. Robinson 2268 FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS be absolutely essential for TatC function: R17A (N-terminal cytoplasmic region) and P48A (first peri- plasmic loop). The deletion of three residues (20–22) also disrupted function, prompting the suggestion that this cytoplasmic domain played an important role. In a separate study [27], R17 was again found to be important but two acidic residues, E103 and D211 were found to be particularly critical; E103A ⁄ Q ⁄ R mutants and D211A were completely inactive, although D211E ⁄ N mutants were partially active. These residues are located on the first cytoplasmic loop and third periplasmic loop, respectively. Substitution of F94 (at the interface between cytoplasmic loop I and the membrane bilayer) was also reported to block transport activity [27]. D211 was not analyzed in our previous study [26] but we did find that E103A showed no translocation defect at all, and so we have made new mutants in all three residues in order to analyse the effects using our expression and assay systems. As with other mutants, we carried out expression studies using all of the TatC mutants; in each case the TatABC subunits were present at similar levels and in A B Fig. 6. Effects of mutations in the TM regions of TatA and TatB. The TatA and TatB mutants containing alterations in TM spans (detailed in Fig. 5) were tested for effects on translocation activity. (A) Muta- tions within the pBAD-ABC vector were assayed for export of TorA. (B) mutations within pEXT-ABC were assayed for export of TorA-GFP, as detailed in Fig. 2. Cells expressing pBAD-ABC or pEXT-ABC without mutations were analyzed as controls, and DtatABCDE cells were analyzed for export of TorA, again as a control. Other symbols are as in Fig. 2. A B Fig. 7. A mutation in the TatB TM span can compensate for the severe effects of the TatA ⁄ 3Gly mutation. This figure illustrates the translocation activities of four DtatABCDE strains expressing pBAD- ABC or pEXT-ABC in which mutations are present in the TM spans of both TatA and TatB (mutations in TatA are shown underlined and labels on the right indicate whether the mutations are in TatA or TatB). Cells were analyzed for the export of TorA (A) or TorA-GFP (B) using protocols described in Fig. 2. C. M. L. Barrett and C. Robinson Mutagenesis of TatABC FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS 2269 these instances the TatC Strep-tag II were all found to be intact (data not shown). Mutations in the first cytoplasmic loop are charac- terized in Fig. 8(A). None of these mutants display any detectable defects in translocation efficiency, inclu- ding the E103A and E103Q mutants which do not even contain elevated levels of the precursor protein TorA-GFP. These data do not agree with reports that the two residues are critical [27], and with this appar- ent contradiction in mind we made each mutant again and again found them to be highly active (data not shown). Figure 8(B) shows the effects of mutations in TM spans two to six. Most of these mutations again have no detectable effect with the notable exception of F94A which is completely inactive. These results agree with those reported in [27]. The final study on TatC is shown in Fig. 9, where mutations in the periplasmic loops are analyzed. We have previously shown that the P48A mutation des- troys activity, and the data confirm this result with no export detected using either assay system. The K73A and Y154S mutants are active, as expected from previ- ous studies [26] and so too are the D211A and D211N mutants that were described as completely or partially inactive, respectively, in a previous study [27]. In Fig. 9, no translocation defects are apparent with either D211A or D211N and these mutants were again A B Fig. 8. Mutations in several residues of the cytoplasmic loops in TatC cause no detect- able loss of translocation activity. (A) TatC mutants carrying substitutions in the 1st cytoplasmic loops were tested for export of TorA-GFP as described in Fig. 2; mutations were made within the pEXT-ABC vector and the Figure shows assays using pEXT-ABC as a control. (B) Similar tests carried out using TatC mutants carrying substitutions in TM spans (numbered above the lanes). A B Fig. 9. Effects of mutations in the periplasmic loops of TatC. Muta- tions in the three periplasmic loops of TatC (numbers indicated above the lanes) were assessed using export assays for TorA (upper panel) or TorA-GFP (lower panel) as described earlier for other mutants (Fig. 2). Mutagenesis of TatABC C. M. L. Barrett and C. Robinson 2270 FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS [...]... Mutations in the TM spans of TatA and TatB Most of the previous studies on TatA ⁄ B have focused on the more highly conserved residues and we considered it important to test the effects of introducing structural rearrangements of the TM spans of these subunits It has not been established whether these TM spans simply anchor the amphipathic regions and cytosolic 2272 domains to the membrane, or whether they... translocation activity and this reinforces the importance of this region Truncation analysis has similarly pointed to essential roles for this region in both TatA and TatB [29] While the precise effects of the individual sets of mutations await further analysis, the data obtained with new combinations of TatA and TatB mutants have unexpected and interesting properties For example, the TatA K24A mutant exhibits... with each other, and that these mutations can be safely accommodated in either subunit separately, whereas the simultaneous presence of both mutations leads to an undue disturbance of the inter-subunit interaction 2271 Mutagenesis of TatABC C M L Barrett and C Robinson Table 1 Primers used for site-specific mutagenesis of TatABC Only the forward primers are shown Name of primer TatA mutants G2A 3Pro3Gly... essential The only mutant devoid of translocation activity in the present study and [25] is TatA ⁄ F39A Mutations in the predicated amphipathic region can have more drastic effects, and we previously analyzed the effects of substituting groups of basic residues in these regions of TatA and TatB [24] The present study has increased our understanding of this region in two ways First, the removal of three... levels of TorA and TorA-GFP export, as does the TatB 2R > N mutant However, a combination of the two mutations leads to a near-complete loss of TorA export and an absolute loss of TorA-GFP export The same phenomenon is observed with the combination of K24A in TatA and 3K > Q in TatB, each of which exhibits no translocation defects in isolation We propose that these regions interact with each other, and. .. in participating in the translocation channel, although further studies are certainly required to obtain a clear picture of TatB function Again, an analysis of mixed TatA ⁄ TatB mutations provides strong indications for critical interactions between these subunits The TatA 3Gly mutant is unable to export either TorA or TorA-GFP, but the simultaneous presence of the V12P mutation in TatB restores export... results described above, these data provide indirect evidence of interactions between at least two regions of TatA and TatB: the predicted amphipathic regions and the TM spans It will be important to probe these issues in detail because little is currently known about the molecular details of the proposed TatA TatB interaction under conditions where TatABC are expressed at physiological stoichiometries... h growth the cells were analyzed by fractionation as detailed below For studies on the export of TorA-GFP, cells expressing pJDT1 and pEXTABC were grown for 2 h in the presence of 50 lm arabinose to induce expression of TorA-GFP, and the cells were then pelleted and resuspended in fresh medium containing 1 mm isopropyl thio-b-d-galactoside This leads to induction of TatABC synthesis from the pEXT-ABC... studies [17,37] have shown that TatA and TatB are present within a single complex but these studies did not prove that TatA and TatB actually contact each other A more direct crosslinking approach found evidence TatA dimers and trimers, as well as TatB dimers, but no TatA- TatB cross-links were observed [31] FEBS Journal 272 (2005) 2261–2275 ª 2005 FEBS Mutagenesis of TatABC TatC mutants Most TatC mutations... methods, because in [27] the mutated subunits were generated in a background of wild-type levels of the remaining Tat subunits These residues are of real interest because acidic side-chains could participate in proton translocation or, in the case of E103, in the binding of the twin-arginine motif in the signal peptide The absence of any translocation defects leads us to conclude that these residues are . Evidence for interactions between domains of TatA and TatB from mutagenesis of the TatABC subunits of the twin-arginine translocase Claire. in the TM spans of both TatA and TatB, namely (TatA ⁄ 3Gly + TatB ⁄ 3Gly) (TatA ⁄ 3Gly + TatB ⁄ V12P) (TatA ⁄ I12P + TatB ⁄ 3Gly) and (TatA ⁄ I12P + TatB