Potential receptor function of three homologous components, TatA, TatB and TatE, of the twin‐arginine signal sequence‐dependent metalloenzyme translocation pathway in Escherichia coli

This is the authors' abstract. We don't add key points for this paper.

Sir, It is well established that plant thylakoids possess a Sec-independent, ΔpH-driven import pathway for which the substrates are synthesized with cleavable targeting signals with an essential twin-arginine motif (reviewed by Robinson and Mant, 1997, Trends Plant Sci2: 431–437). Growing evidence strengthens the idea that a similar mechanism is used in the targeting and translocation of a subset of bacterial metalloenzymes across the cytoplasmic membrane (Bogsch et al., 1998, J Biol Chem273: 18003–18006; Santini et al., 1998, EMBO J17: 101–112; Sargent et al., 1998, EMBO J17: 3640–3650; Weiner et al., 1998, Cell93: 93–101). A component of this pathway, HCF106, was first identified in maize (Settles et al., 1997, Science278: 1467–1470). Database searches with the HCF106 sequence revealed homology to open reading frames from all completely sequenced bacterial genomes, and related sequences are found in organisms ranging from archaebacteria to higher eukaryotes, suggesting that the twin-arginine pathway is a ubiquitous system for membrane targeting and translocation of a subset of proteins (Settles et al., 1997, ibid; Weiner, 1998, ibid). HCF106 contains 243 amino acids with a predicted membrane-spanning domain (MSD) from amino acid 68 to 89. The conserved region is restricted to the MSD and approximately 40 residues of the adjacent C-terminal domain, which is predicted to form an amphipathic alpha helix. The two Escherichia coli homologues of HCF106 are located in yigTUW (86 min) and ybeC (14 min) loci (Settles et al., 1997, ibid). The first sequence revision of the yig region assigned three genes mttABC (for membrane targeting and translocation, corresponding to yigTUW with several sequence corrections in yigT ) to this locus (Weiner, 1998, ibid). The MttA protein was predicted to contain 277 amino acids with two MSDs and a long C-terminal periplasmic domain. The reported HCF106 homologous region comprises 43 amino acids from residues 17 to 59, including the first MSD. Substitution of proline 128 of MttA by leucine (P128L) results in cytoplasmic accumulation of the catalytic subunit DmsA of the membrane-bound DMSO reductase (Weiner et al., 1998, ibid). Likewise, active TMAO reductase TorA accumulated in the cytoplasm in the mutant strain, but the normally periplasmic nitrate reductase NapA become associated with the membrane fraction. Weiner et al. proposed that the periplasmic domain of MttA might play a role in gating the pore, with MttB protein as the pore for protein transport and MttC as a receptor function for twin-arginine proteins. A recent revision of the DNA sequence and gene product identification revealed the presence of a stop codon in the mttA gene (Sargent et al., 1998, ibid). The mttA gene is thus divided into tatA and tatB (for twin-arginine translocation). TatA is composed of 89 amino acids corresponding to the region from the methionine 14 to the valine 103 of MttA. The C-terminal 171 residues of MttA was renamed TatB. The new nomenclatures for mttB (yigU ), mttC (yigW ) and ybeC are tatC, tatD and tatE respectively (Table 1). TatA and TatE share about 50% sequence identity at the protein level. Disruption of tatA, tatE and tatC results in different effects depending on the enzymes analysed. The effects of tat mutations on seven enzymes known to use this pathway are summarized in Table 1. We cloned tatABCD and constructed a tatB mutant by the insertion of a KanR cassette at the SacII site, which interrupts the TatB at amino acid position 88. The downstream region of the KanR cassette does not have a transcription terminator but contains an optimal ribosome binding site and, thus, does not have a polar effect on the expression of tatCD. Like the mttA P128L mutation, tatB ::KanR results in the accumulation of active TMAO reductase in the cytoplasm. In addition, no precursor of this enzyme is translocated into the periplasm, which is consistent with our previous observation that only the cofactor-containing active forms are competent for translocation (Rodrigue et al., 1996, FEBS Lett392: 81–86; Santini et al., 1998, ibid). Surprisingly, the interruption of tatB has no effect on the membrane targeting and translocation of hydrogenase 2 (HYD2), which does not support the idea that mttA is involved in the translocation of all twin-arginine signal-bearing redox enzymes. As the reported region of homology between YigT and HCF106 is located in TatA, it was assumed that TatB would be unrelated to HCF106 (Sargent et al., 1998, ibid). The phenotypes of the mttA and tatB ::KanR mutants suggest, however, that TatB is another counterpart of TatA and TatE and is involved in the translocation of a subset of twin-arginine redox enzymes. To test this idea, we first analysed the local similarity between HCF106 and TatE, TatA and TatB by using the SIM program (Huang and Miller, 1991, Adv Appl Methods12: 337–357). To facilitate direct comparison, the three Tat proteins were artificially fused together and compared as a single polypeptide with HCF106. Three regions, corresponding to the N-termini of TatE, TatA and TatB, show the same similarity score with the conserved region of HCF106 (Fig. 1). In addition, a segment covering the amino acids from 170 to 220 shares some similarity with the C-terminal region of TatB (data not shown). We then analysed the statistical significance of the similarities between HCF106 and the three Tat proteins using the RDF2 program (Pearson and Lipman, 1988, Proc Natl Acad Sci USA85: 2444–2448). The comparison score between HCF106 and TatB was as good as that between HCF106 and TatE (Fig 1, 8.1 vs. 8.4). The probability of getting such a score by chance is less than 10−16. These results confirm that HCF106 is homologous to all three Tat proteins and strongly support the idea that they have a common evolutionary origin. . Local similarity was analysed using BLOSUM62 as comparison matrix, and 20 alignments were computed. The gap open penalty and gap extension penalty were 12 and 4 respectively. RDF2 comparison scores were obtained by 500 local random shuffles (window size of 10) of the three entire Tat proteins with a ktup of 2 against Hcf106. MaxH values were calculated with a window size of 17. MSD, membrane-spanning domain. The cellular location of TatA, TatB and TatE was predicted according to Boyd et al. (1998, Prot Sci7: 201–205) and von Heijne (1992, J Mol Biol225: 487–494). Their MaxH values are significantly higher than the discriminator value of 1.53 (Fig. 1) and allow us to assign them unambiguously as membrane proteins. Furthermore, their C-terminal regions are most likely to be at the cytoplasmic side of the membrane. Therefore, TatA, TatB and TatE seem to function, independently or together as a trimer, as membrane receptors for different subsets of redox proteins, each one showing different affinities for the substrates, which accounts for the effects of the same mutation on various enzymes. The heterotrimer model would further explain the observation that interruption of any one of the three genes severely affects the translocation of TorA. These three branches of the TAT pathway would join together at the TatC channel level, as the tatC mutation affects the translocation of all enzymes analysed (Table 1). The amphipathic alpha-helix domain in the conserved region might be the best candidate for the site at which these receptors interact with each other and/or with TatC. The tatD (yigW, mttC ) gene encodes a soluble protein that shows sequence homology with the products of E. coli ORFs yjjV and ycfH and whose function has not been assigned experimentally (Sargent et al., 1998, ibid; Weiner et al., 1998, ibid). The MaxH value of 1.36 and the absence of a signal sequence from TatD suggests that this potential Tat component is cytoplasmic and could play a role in the early stage of protein translocation. Taken together, these results suggest that the homologous components TatA, TatB and TatE might constitute three branches at the membrane targeting/reception stage of the TAT pathway and that the three different branches join together at the channel level. We are grateful to B. Ize and G. Giordano for valuable discussion, and Y. Quentin for computer assistance in calculation of the MaxH value.