Signal peptide–chaperone interactions on the twin-arginine protein transport pathway
- Kostas Hatzixanthis*,
- Thomas A. Clarke*,
- Arthur Oubrie*,
- David J. Richardson*,
- Raymond J. Turner†, and
- Frank Sargent*,‡
- *School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom; and †Department of Biological Sciences, University of Calgary, Calgary, AB, Canada T2N 1N4
-
Edited by Linda L. Randall, University of Missouri, Columbia, MO (received for review January 28, 2005)
Abstract
The twin-arginine transport (Tat) system is a protein-targeting pathway of prokaryotes and chloroplasts. Most Escherichia coli Tat substrates are complex metalloenzymes that must be correctly folded and assembled before transport, and a preexport chaperone-mediated “proofreading” process is therefore in operation. The paradigm proofreading chaperone is TorD, which coordinates maturation and export of the key respiratory enzyme trimethylamine N-oxide reductase (TorA). It is demonstrated here that purified TorD binds tightly and with exquisite specificity to the TorA twin-arginine signal peptide in vitro. It is also reported that the TorD family constitutes a hitherto unexpected class of nucleotide-binding proteins. The affinity of TorD for GTP is enhanced by initial signal peptide binding, and it is proposed that GTP governs signal peptide binding-and-release cycles during Tat proofreading.
The twin-arginine transport (Tat) system is a protein-targeting pathway found in the cytoplasmic membranes of many prokaryotes and the thylakoid membranes of chloroplasts (1, 2). Tat-targeted substrate proteins are synthesized as precursors with N-terminal twin-arginine signal peptides that exhibit distinctive SRRxFLK amino acid motifs (3). In Escherichia coli, the TatA, TatB, and TatC membrane proteins form the core components of the Tat translocase (1). The TatBC unit is the twin-arginine signal peptide recognition module (4), whereas multiple copies of TatA generate the protein-conducting channel (5–7). Protein translocation is powered solely by the transmembrane proton gradient, and there is some evidence to suggest the Tat translocase acts as a genuine protein–proton antiporter (8).
Most remarkably, the Tat system only transports fully folded proteins (9), and most E. coli Tat substrates must bind complex cofactors, fold, and, in some cases, oligomerize before transport can proceed (1). The Tat translocase itself is thought to accept or reject proteins for transport through a quality control activity (9). Although this Tat quality control mechanism would be applied to all proteins presented for export, some complex Tat substrates are subjected to an earlier chaperone-mediated screen that operates to prevent premature targeting of immature proteins before cofactor-loading, correct folding, or docking of partner proteins has occurred. To distinguish from Tat quality control, this important complementary system is termed “Tat proofreading” (10).
The E. coli trimethylamine N-oxide reductase (TorA) is a Tat-dependent periplasmic redox enzyme that binds a molybdenum cofactor. TorA is the archetypal model Tat substrate, its twin-arginine signal peptide being the most heavily studied and exploited in the field. Loading of the cofactor is an essential prerequisite to TorA transport (11) and the TorD protein was originally identified as a cytoplasmic accessory protein required for efficient cofactor loading into the TorA apoenzyme (12–14). The 3D structure of a TorD homolog from Shewanella massilia showed that this family of proteins comprise two separate α-helical domains connected by a short hinge region (15). Overproduced S. massilia TorD was isolated as a mixture of monomers, domain-swapped homodimers, and homotrimers (16), and the crystal structure is of the dimer form (15). TorD was initially not expected to bind the twin-arginine signal peptide directly, however genetic experiments suggested E. coli TorD could indeed recognize the TorA signal peptide during the preexport proofreading process (10). Thus, TorD performs dual roles by assisting cofactor insertion into the TorA enzyme (10, 12–14) and facilitating Tat proofreading by recognition of the TorA twin-arginine signal peptide (10).
Twin-arginine signal peptides exhibit common core tripartite structures, including polar n-regions, relatively hydrophobic h-regions, and polar c-regions that usually contain signal peptidase cleavage sites (3). The twin-arginine motif is always located at the boundary between the n- and h-regions, with the arginine pair being critical for successful translocation (1). Definitive physiological functions have not yet been assigned to the n-regions of any Tat signal peptides, although c-regions sometimes contain a positively charged Sec-avoidance motif thought to prevent mistargeting of some Tat substrates (17).
Here, recent advances in understanding the relationship between TorD and the twin-arginine signal peptide in the Tat proofreading process are reported. By using in vitro molecular techniques, tight binding of a synthetic TorA twin-arginine signal peptide to purified TorD is demonstrated. Signal peptide binding by TorD is not critically dependent on an intact twin-arginine motif, thus establishing that the preexport Tat proofreading process stands alone from Tat translocation. Finally, it is revealed that TorD is a guanosine nucleotide binding protein and that initial binding of the TorA signal peptide by TorD enhances the chaperone's affinity for GTP.
Materials and Methods
Plasmid Construction and Protein Methods. E. coli torD was amplified by PCR and cloned as an NcoI/BglII fragment into pQE60 (ampicillin-resistant, Qiagen, Valencia, CA), which incorporates a C-terminal hexahistidine tag onto recombinant proteins to give pAAF3. Modified torD genes were generated by QuickChange (Stratagene) with pUNI-torD1 (ampicillin-resistant) (10) or pAAF3 as template. Cloning fidelity was verified by DNA sequencing.
Protein production and purification was as described in ref. 18, and protein concentrations were determined by the Lowry method (19). SDS/PAGE and Western immunoblotting were as described in ref. 10.
The in vivo Tat proofreading assay was performed with the reporter strain RJ607 (ϕtorA::hybO, ΔhybA) as described in ref. 10. Briefly, RJ607 harboring TorD-producing plasmids was cultured anaerobically in 0.5% (vol/vol) glycerol/0.4% fumarate (wt/vol), harvested, resuspended in 20 mM Tris·HCl, pH 7.5, to 0.1 gram of cells per milliliter and assayed for hydrogen::benzyl viologen oxidoreductase activity (10). Custom peptides were synthesized by Alta Bioscience (Birmingham, United Kingdom), Severn Biotech (Kidderminster, U.K.), Sigma–Genosys (The Woodlands, TX), or Mimotopes (Raleigh, NC).
Analytical Ultracentrifugation. Sedimentation-equilibrium experiments were performed in a Beckman Optima XL-I analytical ultracentrifuge equipped with scanning absorbance optics and an An50Ti rotor. TorD proteins were dialysed against 20 mM Tris·HCl, pH 7.6, 200 mM NaCl, and 110 μl of aliquots placed in charcoal-filled Epon double-sector cells fitted with quartz windows. Reference sectors were filled with dialysis buffer (120 μl), and centrifugation was performed at 15,000 rpm and 20°C. Data were collected at 280 nm, and equilibrium was confirmed by the absence in any difference between values collated 4 h or more apart. Partial specific volumes for TorD proteins were calculated from the amino acid sequences by using the program sednterp and data were analyzed with the software ultrascan 6.2 and chelated by using origin 5.0 software.
Isothermal Titration Calorimetry (ITC). ITC was carried out at 28°C by using a VP-ITC titration calorimeter (Microcal, Amherst, MA) and by following standard procedures (20). TorD proteins were dialysed against 20 mM phosphate buffer, pH 7.5, or the same buffer containing 500 mM NaCl and degassed before being loaded into the sample cell (1.4-ml working volume). Peptide ligands were prepared as 1 mM stocks in dialysis buffer and titrated from a 290-μl syringe into the sample cell stirred at 300 rpm. The heat change for the dilution of peptide in the absence of TorD was recorded for each experiment and subtracted from the measured heat change of ligand binding to protein. Data analysis was performed with the origin software supplied by Microcal.
Fluorimetry. Measurements of intrinsic tryptophan fluorescence emission were made by using a PerkinElmer LS-55 spectrofluorimeter thermostated at 25°C. Concentrated (50 mg/ml) TorD was dialysed into 20 mM phosphate buffer, pH 7.5, and degassed dialysis buffer was used for all fluorescence experiments. A working TorD concentration of 0.3 μM and 3-ml quartz cuvettes were used. The excitation wavelength was 295 nm (slit width, 5 nm) and emission spectra were recorded from 310 to 390 nm (slit width, 10 nm). Dissociation constants were calculated by plotting relative fluorescence against ligand concentration and curve-fitting with the Solver add-in available for Microsoft excel software.
Molecular Modeling. A theoretical TorD-GTP binary model was generated by using the available coordinates for S. massilia TorD (15). Guanine was manually imposed onto one oxidized DTT molecule by using the program O (21), and an initial model was chosen on the basis of the potential formation of a hydrogen bond between the guanine N 7 atom and the main chain carbonyl oxygen of S. massilia TorD E32. Replacing guanine with guanosine triphosphate, the initial model was refined with three rounds of manual model building with the program O alternated with conjugate gradient positional refinement using the program cns (22).
Results
TorD Binds the TorA Twin-Arginine Signal Peptide in Vitro. Genetic evidence suggests that the Tat proofreading activity of TorD depends on direct recognition of the TorA signal peptide (10). To characterize in vitro the TorA signal peptide binding activity, E. coli TorD was purified. A single species was isolated with a molecular mass determined as 23.6 kDa by analytical ultracentrifugation (Fig. 1A), which compares well with the predicted 23,567 Da for the recombinant protein.
TorD monomer contains a single signal peptide binding site. Determination of the native molecular mass of E. coli TorD in the absence and presence of a twin-arginine signal peptide by sedimentation equilibrium. Each sample contained 0.4 mg/ml (≈15 μM) TorD, and sedimentation was monitored at 280 nm in the absence of any ligands, resulting in a 23.6-kDa species (A) and, in the presence of 133 μM TorA-SP10–36 synthetic peptide, a 27.0-kDa species (B). (Lower) The sedimentation data for each species together with simulated single-component sedimentation curves for the molecular masses given. (Upper) Residuals between the experimental data and the fitted curves.
Binding of the TorA signal peptide by TorD was investigated by using ITC. Because of its excessive length, the TorA signal peptide was synthesized in two portions (Table 1). Peptides covering residue N2 through T22, covering the extended n-region and the twin-arginine motif (TorA-SP2–22), and S10 through R36, covering the twin-arginine motif, the h-region, and the Sec-avoidance motif (TorA-SP10–36), were synthesized. Titration of TorD with TorA-SP10–36 generated negative enthalpy values and a sigmoidal binding curve that reached a clear saturation point (Fig. 2B), indicating binding of this peptide by TorD. Analysis of the binding isotherm gave a best fit to a noncooperative model with a binding stoichiometry (n) of 0.78 ± 0.02 and a K D of ≈1 μM for this datum (Fig. 2B). Titration of TorD with the alternative TorA-SP2–22 peptide suggested much weaker binding (Fig. 2 A). TorD, therefore, binds tightly in vitro to the central core structure of the TorA twin-arginine signal peptide over and above the variable n-region.
TorD binds the core structure of the TorA signal peptide. Calorimetric titration of TorD with synthetic peptides. (Upper) Raw data for the heat effect during titration. (Lower) The binding isotherms. All titrations were performed at 28°C with 46 μM TorD in 20 mM phosphate buffer, pH 7.5. Injections (25 × 10 μl) of peptide ligands (1 mM stocks) were made to the stoichiometric excesses shown. (A) Weak binding of the TorA-SP2–22 peptide to TorD. (B) Strong binding of the TorA-SP10–36 peptide to TorD. The best fit to the TorA-SP10–36 datum gave n = 0.78 ± 0.02 peptide binding sites, ΔH = -14.6 ± 1.0 kcal/mol (1 cal = 4.18 J), and TΔS = -6.6 ± 0.9 kcal/mol.
To gain corroborative evidence of signal peptide binding, TorD was subjected to a sedimentation equilibrium procedure in the presence of excess TorA-SP10–36 peptide (Fig. 1B). The mass recorded by following TorD sedimentation was 27.0 kDa (Fig. 1B), an observed mass increase of 3.4 kDa when compared with TorD alone (Fig. 1 A) representative of 1.17 times the theoretical mass of the TorA-SP10–36 peptide (2.9 kDa). This finding, taken together with the ITC experiment (Fig. 2B), points strongly to a single signal peptide binding site on each TorD monomer.
It is immediately obvious from these data that TorD does not oligomerize into dimer or trimer forms in the presence of signal peptide. Indeed, the physiological significance of the oligomerization exhibited by S. massilia TorD (16) remains somewhat ambiguous and is clearly not central to the signal peptide binding process.
Having established the presence of a single peptide binding site on TorD, a one-site, noncooperative, binding model was used to determine the dissociation constants of various synthetic peptides by ITC. This model underscored the tight binding of the TorA-SP10–36 peptide by TorD (Table 1), and the specificity of this interaction was tested next. Peptides covering N14 through K34 of the E. coli hydrogenase-2 Tat signal peptide (HybO-SP14–24) and N2 through N28 of the TorA signal peptide from Shewanella oneidensis (Shew-SP2–28) were synthesized (Table 1). Despite the similar overall physicochemical properties displayed by the alternative peptides, the E. coli TorD protein bound only to the peptide modeled on E. coli TorA (Table 1). These experiments therefore highlight the exquisite specificity of this tandem chaperone/signal peptide system.
Characteristics of Signal Peptide Binding. The signal peptide arginine pair is essential for Tat transport of the TorA precursor (23); however, it is not known whether the twin-arginine motif has a role in the preexport proofreading process. A synthetic TorA signal peptide essentially identical to TorA-SP10–36 but with twin-arginine residues replaced by twin-lysine residues was synthesized (Table 1). The twin-lysine peptide (TorA-SPKK) displayed identical TorD binding characteristics to its twin-arginine counterpart (Table 1). Thus, the twin-arginine motif itself is clearly not the overarching signal recognition factor.
To further dissect the signal peptide recognition activity, the TorA-SP10–36 peptide was synthesized in two truncated forms. First, the positively charged twin-arginine motif was completely omitted to give peptide TorA-SP15–36, and, second, a peptide (TorA-SP23–36) was prepared that covered the C-terminal stretch of the TorA signal missing from the TorA-SP2–22 peptide (Table 1). Binding of the shorter TorA-SP23–36 peptide was poor and under increased ionic strength titration of TorD with the TorA-SP23–36 peptide elicited only a weak response that did not reach saturation and could not be fitted convincingly to any model (Table 1). Complete omission of the SRRRF stretch from the N terminus of the synthetic signal peptide (peptide TorA-SP15–36) did not prevent peptide binding by TorD but did reduce the binding affinity 5- to 9-fold (Table 1). Thus, a polar N terminus is probably required to maintain a high degree of signal peptide occupancy by TorD. Alternatively, although it is not likely that the structures of the synthetic peptides in aqueous solution will change as the peptide sequences are changed (24–26), it is certainly possible that the ability of the peptides to attain a secondary structure once bound by TorD could be altered, and this could affect the binding efficacies.
In summary, recognition and stable binding of the TorA signal peptide by TorD requires all of the central core features of the TorA twin-arginine signal peptide. The ideal peptide recognized by E. coli TorD probably has a sequence close to that of the TorA-SP10–36 peptide, with positively charged n- and c-regions and relatively hydrophobic h-region (of a definite length and sequence) between the two.
The TorD hinge region contains a DH dipeptide that is conserved in all TorD family proteins (13), and substitution of E. coli D124 or H125 with alanine abolished Tat proofreading activity (10). The TorDD124A and TorDH125A variant proteins were purified, and ITC experiments established that protein–peptide affinities only decreased between three and six times for each (Table 1). It is possible that this decrease is enough to impair signal recognition in vivo, but it probably suggests that the function of the variant proteins is blocked at a downstream biochemical step.
TorD Binds Guanosine Nucleotides. Having established that TorD binds the TorA signal peptide with reasonable affinity, we then investigated what next drives release of the signal peptide, an action that is obviously required to allow TorA export and complete the TorD cycle. It is notable that nucleotides are common players in other signal peptide-dependent processes. For example, the E. coli SecA-dependent protein targeting system (27) and the E. coli signal recognition particle-dependent protein targeting system (28) utilize ATP and GTP, respectively, to facilitate signal peptide release. Indeed, general chaperones (e.g., DnaK and GroEL–GroES) also exhibit nucleotide-binding activity (29). Therefore, despite the lack of canonical motifs (e.g., ref. 30), the nucleotide-binding activity of TorD was tested.
Titration of TorD with various nucleotides in the microcalorimeter did not yield convincing data (results not shown) and our attention turned, therefore, to alternative biophysical methods. The E. coli TorD protein contains four tryptophan residues and, when irradiated at 295 nm, has a fluorescence emission maximum at 350 nm (Fig. 3A). Very interestingly, titration of TorD with GTP caused a clear quench of the tryptophan fluorescence, whereas the emission maximum remained unchanged (Fig. 3A). The GTP-induced fluorescence quench was saturable, and the calculated dissociation constant suggested weak binding of GTP (apparent K D ≈ 370 μM; Table 2) and this “loose” binding probably explains the failure of ITC to detect the interaction. Fluorimetry indicated no binding of ATP by TorD (Table 2), and titration with GDP, GMP, and cGMP resulted in similar dissociation constants to GTP (Table 2). These data suggest that the guanosine moiety, rather than the phosphate groups, is recognized by TorD. Indeed, titration of TorD with guanosine alone returned a K D of 336 ± 1 μM. Given that cytoplasmic levels of GTP are in the region of 0.5–1 mM, whereas concentrations of GDP and cGMP, for example, are 10 and 1,000 times less, respectively, these data indicate that TorD will only bind GTP in vivo.
TorD is a GTP-binding protein. Intrinsic tryptophan fluorescence-quenching experiments (λexcitation = 295 nm) carried out at 25°C in degassed 20 mM phosphate buffer, pH 7.5. (A) Fluorescence emission spectra of 0.3 μM TorD without GTP (line 1) or with 100 μM GTP (line 2) or 200 μM GTP (line 3) (final concentrations). (B) The unmodified (native) and R22A variant TorD proteins were separately preincubated in increasing amounts of TorA-SP10–36 peptide (0–25 μM) before dissociation constants for GTP were calculated by using intrinsic tryptophan fluorescence quenching as an indicator of GTP binding. Error bars indicate SEM (n = 3).
To test the independence of the two ligand-binding events, calorimetric titration of TorD with the TorA10–36 peptide was carried out in the presence of GTP or the nonhydrolysable GTP analog GMPPNP. Note that fluorimetry indicated that TorD bound GMPPNP with a K D of 375 ± 22 μM and that analytical ultracentrifugation experiments in the presence of 1 mM GTP did not lead to oligomerization of the protein (data not shown). The K D for the TorA10–36 peptide interaction with TorD was 3.8 ± 0.1 μM in the presence of 1 mM GTP and 1.9 ± 0.1 μM in the presence of 1 mM GMPPNP. Taken together with previous data (Table 1), binding of the TorA-SP10–36 peptide is independent of GTP binding.
Next, in the presence of the TorD protein, the TorA-SP10–36 peptide was added to the fluorescence cell. Under these conditions, no significant quenching of the intrinsic tryptophan fluorescence of TorD was observed. However, increasing levels of TorA-SP10–36 peptide served to decrease significantly the TorD dissociation constant for GTP (Fig. 3B). A minimal apparent K D for GTP of ≈200 μM was reached at ≈10 times the K D for the TorA-SP10–36 peptide (Fig. 3B). It seems likely, therefore, that initial binding of the TorA twin-arginine signal peptide modifies the conformation of TorD in such a way that enhances and assists the subsequent binding of GTP.
Modeling a GTP Binding Site on TorD. Previous in vivo studies of E. coli TorD function uncovered four functionally important side chains (10); however, none of these are essential for GTP binding by TorD (Table 2). In an attempt to identify a GTP binding site on TorD, a theoretical model of a TorD–GTP binary complex was generated by using the structure of S. massilia TorD (15). A shallow cavity on the protein surface was focused on that, under crystallization conditions, contained one molecule of oxidized and, therefore, cyclic DTT (15). The final model suggests that GTP could be bound in a well conserved hydrophobic pocket at the N terminus of TorD and is lined with S. massilia TorD residues (with E. coli equivalents in parentheses) L12 (C11), L16 (W15), L33 (I32), F41 (W40), and W42 (F41) (Fig. 4). The binary model suggests that hydrogen bonds could be formed between the guanine N 7 and the E32 (Q31) backbone carbonyl oxygen and the guanine C 2 amino group and the backbone carbonyl oxygen of L12 (C11). Note that ATP does not display an NH2 group at position C 2 and is therefore incapable of replacing GTP in this model. Interestingly, the model also predicts the presence of putative phosphate ligands. The R29 (Q28) NH 1 group is within hydrogen-bonding distance to the GTP α-phosphate O 1A oxygen, and, most intriguingly, an ion pair interaction is possible between the γ-phosphate and the R23 (R22) guanidinium group.
A putative GTP binding site on TorD. Theoretical model of an S. massilia TorD-GTP complex shown as stereo diagrams. (A) Surface representation of S. massilia TorD colored by electrostatic potential, with GTP shown as sticks. (B) Ribbon representations of the two protomers of the TorD dimer are shown in pink and purple, respectively. GTP and specific amino acids thought to be involved in GTP binding are shown as sticks. Carbon atoms of GTP are white, those of amino acids are pink, and nitrogen, oxygen, and phosphate atoms are in blue, red, and purple, respectively. Amino acid numbering is from the S. massilia sequence (equivalent E. coli TorD numbering is in parentheses): L12 (C11), Q15 (A14), L16 (W15), S19 (Q18), R23 (R22), R29 (Q28), E32 (Q31), L33 (I32), F41 (W40), and W42 (F41). The images were generated with pymol (31).
Mutagenesis of the Putative GTP Binding Pocket. By using the theoretical model as a guide for further experimentation, the aromatic amino acids that could be responsible for the fluorescence effects observed upon GTP binding and the polar side chains in the vicinity, including those postulated to contribute hydrogen or other bonds, were targeted for mutagenesis within the equivalent region of E. coli TorD. Amino acids W15, Q18, R22, Q28, Q31, W40, and F41 were separately substituted with alanine, and Western immunoblotting established that each variant protein, with the exception of F41A, was produced at levels equivalent to native TorD in the same system (Fig. 5A).
Mutagenesis of the putative GTP-binding pocket. (A) The RJ607 (ϕtorA::hybO, ΔhybA) reporter strain (Basal) was transformed with pUNI-torD1 (TorD) and seven pUNI-torD1 derivatives as indicated. Cell pellets were resuspended at 0.1 g/ml, and whole-cell proteins were separated by SDS/PAGE (14% wt/vol acrylamide), blotted, and challenged with anti-TorD serum. Identical proportions of cellular protein were loaded. (B) Tat proofreading assay. Intact cells were assayed for hydrogen::benzyl viologen oxidoreductase activity in units of micromole of benzyl viologen reduced per minute per gram of cells. Error bars indicate SEM (n = 3).
An E. coli reporter strain has recently been described that expresses a hydrogenase enzyme bearing the TorA twin-arginine signal peptide (10). Because of a loss of Tat proofreading activity, this strain normally has low hydrogenase activity; however, increasing cellular levels of biologically active TorD results in a concomitant increase in hydrogenase activity as TorD recognizes the TorA signal peptide and restores Tat proofreading to this system (10). The Tat proofreading assay showed that the W40A and F41A variants were inactive in vivo (Fig. 5B). Of the putative GTP phosphate ligands, the Q28A variant protein retained Tat proofreading activity in vivo demonstrating that this side chain does not play a critical role in ligand-binding; however, substitution of R22 with alanine knocked-out the Tat proofreading function of TorD (Fig. 5B). Although the instability of the F41A protein could possibly account for its inactivity in this assay, it is abundantly clear that the R22 and W40 side chains, which are located at either end of the putative GTP binding pocket (Fig. 4), do play critical roles in the TorD-dependent Tat proofreading process in vivo.
The R22A variant protein was purified and shown to retain signal peptide- and GTP-binding activity (Tables 1 and 2). However, unlike native TorD, R22A displays no enhancement of GTP binding affinity in the presence of excess signal peptide (Fig. 3B). This result is consistent with the prediction that R22 could be a GTP ligand and underscores the intimate linkage of the ligand binding events. Taken together, these data point strongly to a ligand-binding mechanism for which the initial interaction with signal peptide results in a repositioning of the R22 side chain, which in turn confers a (physiologically essential) higher affinity for GTP on the TorD Tat proofreading protein.
Discussion
It has been demonstrated here that E. coli TorD binds directly to the E. coli TorA twin-arginine signal peptide and that the interaction is not dependent on the signal peptide extended n-region. This finding is important because many TorA homologs, including those from Shewanella, do not display n-extended twin-arginine signal peptides, despite expressing TorD homologs. The establishment that Tat proofreading chaperones could bind the central cores of Tat signal peptides potentially increases the scope of this type of proofreading mechanism to many other complex Tat substrates.
Once enzyme assembly is complete, it is important that Tat proofreading chaperones rapidly dissociate to allow transport to proceed. It is possible that the mechanism of signal peptide release by TorD follows the principles of other signal peptide or nascent chain binding proteins in that nucleotide-binding steps may govern this process. Ligand binding by TorD appears to be a reciprocal process in which the signal peptide binds first, thus increasing affinity of the chaperone for GTP. The affinity of TorD for GTP is reminiscent of the low-affinity ATP binding site on the SecA protein (K D = 340 μM) that is thought to provide a regulatory role (27). It is indeed conceivable that the GTP binding activity of TorD is regulatory rather than catalytic because we have been unable to detect any intrinsic or peptide-induced GTPase activity associated with the purified protein. However, if GTP hydrolysis is ultimately required, one possibility is that there are missing components from this highly purified system. Protein–protein interactions mediated by GTP binding proteins conjures images of eukaryotic G protein systems in which guanine nucleotide exchange factors or GTPase-activating proteins are required to activate small G proteins (32). It is tempting to speculate that a degree of continuity might be given to the cellular Tat proofreading process if the known or predicted proofreading chaperones (e.g., TorD, DmsD, HyaE, and HybE) were acting as specialist “adaptors” for a single general activating enzyme.
The signal peptide twin-arginine motif is probably not required for TorD-mediated Tat proofreading, which adds weight to the argument that TorD family proteins are not themselves Tat-targeting factors (33). The twin-arginine motif is primarily a TatC recognition motif and is too highly conserved across enzymes, species, and kingdoms to account for the very specific recognition event between E. coli TorD and the E. coli TorA signal peptide. However, interaction of the TorD/precursor complex with an additional membrane-bound recognition factor before transport cannot be ruled out. For instance, the TorD homolog DmsD is reportedly attached to the membrane under Tat transport conditions (34), and very recently two uncharacterized E. coli membrane proteins have been identified that could be important in Tat targeting (35).
Future challenges will be to generate structural information on the nature of the twin-arginine signal peptide and nucleotide binding sites on TorD and to elucidate the molecular mechanism of controlled signal peptide release by TorD that is ultimately required to complete the Tat proofreading cycle.
Acknowledgments
We thank T. Palmer for invaluable discussion, advice, and encouragement and B. C. Berks, B. Ize, R. Little, S. Bornemann, and P. Jakimowicz for technical help and discussions. This work was supported by Biotechnology and Biological Sciences Research Council Research Grant BBS/B/07780 (to F.S. and D.J.R.). F.S. is a Royal Society University Research Fellow.
Footnotes
-
↵ ‡ To whom correspondence should be addressed. E-mail: f.sargent{at}uea.ac.uk.
-
Author contributions: K.H., T.A.C., A.O., and F.S. performed research; K.H., T.A.C., A.O., D.J.R., and F.S. analyzed data; A.O., D.J.R., R.J.T., and F.S. designed research; and A.O. and F.S. wrote the paper.
-
This paper was submitted directly (Track II) to the PNAS office.
-
Abbreviations: ITC, isothermal titration calorimetry; Tat, twin-arginine transport.
-
Freely available online through the PNAS open access option.
- Copyright © 2005, The National Academy of Sciences
