INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The Sec apparatus has long been viewed as the sole system for translocation of proteins across the inner bacterial membrane. However, recently it became clear that in addition to the Sec system many bacteria have the Tat system for the export of proteins which seem to be transported in a folded form (reviewed in references 4, 38, and 51). Evidence for a novel pathway of bacterial protein translocation came from studying protein import into chloroplast thylakoids. A pH-driven pathway was shown to require a signal peptide with an arginine pair (11), and the maize protein Hcf106, involved in this pathway, was found to have homologs in bacteria (45). This fostered the discovery of the new secretory pathway in Escherichia coli, termed mtt, for membrane targeting and transport (48), or tat, for twin-arginine translocation (43).

Supportive evidence for the Tat pathway also came independently from observations of a new type of bacterial signal peptide. On determining several primary structures of nitrous oxide reductase (N₂OR), a sequence motif with an arginine pair in an unusually long signal peptide of about 50 amino acids (aa) was recognized to be conserved not only in this enzyme but also in hydrogenase (29, 56). We reported the existence of the novel motif in a variety of exported and cofactor-carrying proteins (2, 17, 18). As initially found with hydrogenase (34), substitution of the first arginine residue in the conserved sequence prevented translocation of N₂OR to the periplasm in an R20D mutant (17, 18). The significance of the twin-arginine motif has been studied now for other enzymes (16, 26, 27, 40), and allowed variations have been specified by site-directed mutagenesis (46). A data bank survey for the twin-arginine type of signal peptide revealed a large number of redox proteins with cofactors and led to the suggestion that these proteins may be folded in the cytoplasm and acquire their cofactors prior to transport. A heptameric consensus motif in Tat-specific signal peptides was defined as S(T)RRXFLK (3).

Current sequence information from genome projects has revealed a wide, though not ubiquitous, distribution of tat genes among prokaryotes, chloroplasts, and plant mitochondria (51). However, with only a few exceptions (5, 18, 26, 27), studies of the Tat system in bacteria have been confined to enzymes of E. coli. The tatC gene encodes an integral membrane protein composed of six transmembrane helices, which is thought to form the functional export complex in association with other tat-encoded components (4). A mutation in tatC blocks protein translocation (6). Recently, a 600-kDa complex has been described as consisting of TatA, TatB, and presumably the TatC component (7). TatC is rapidly degraded in the absence of TatB (44), suggesting that TatB stabilizes TatC, which provides another, albeit indirect, argument for the presence of TatC in the translocation complex. A further tat locus of E. coli, tatD, encodes a cytoplasmic protein with DNase activity but apparently is not an obligatory factor of the translocation pathway (50).

Utilization of N₂O is a respiratory mode, found usually as part of bacterial denitrification. N₂OR provides a terminal electron sink by reducing N₂O to N₂ and is essential for global N cycling by preventing the accumulation of the greenhouse gas N₂O. In the denitrifying cell, N₂O is generated by the consecutive actions of the periplasmic respiratory nitrite reductase (cytochrome cd₁ in Pseudomonas stutzeri) (53) and the membrane-bound NO reductase, a structural and functional homolog of subunit II of cytochrome c oxidase (24, 31). N₂OR carries six copper atoms per subunit, which are arranged in two types of centers: the binuclear electron transfer site, Cu_A (13, 33), and the tetranuclear catalytic site, Cu_Z (10). The latter was recently identified as the first example of a biologically active copper-sulfide cluster (37). Cytochrome cd₁ and N₂OR both have to be translocated across the cytoplasmic membrane to reach their functional sites in the periplasm (reviewed in reference 53).

In this study we have investigated N₂OR translocation by the Tat system and the possibility of Cu cofactor insertion in the cytoplasm when the translocation pathway is interrupted. Towards these objectives, we have isolated tat loci from P. stutzeri and generated mutations in tatC and tatE. Since E. coli does not denitrify, it cannot provide the physiological environment to study the requirements for the translocation of denitrification enzymes. Our data support conformational constraints imposed by the Tat system on N₂OR transport. However, metal cofactor insertion into the enzyme, contrary to a widely held view, did not take place in the cytoplasm of P. stutzeri. We have also investigated a broader effect of tatC on denitrification, which led to the finding that heme D₁ synthesis or processing to establish a functional cytochrome cd₁ nitrite reductase is Tat dependent.

(A preliminary report of this work appeared previously [28].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. The bacterial strains and plasmids used in this study are listed in Table 1. E. coli DH10B and JM110 were grown in Luria-Bertani (LB) medium at 37°C and 240 rpm on a gyratory shaker. For recombinant DNA work, Pseudomonas strains were grown in LB medium at 30°C in a gyratory shaker at 240 (aerobic growth) or 120 (O₂-limited growth) rpm. For strain maintenance, when necessary, kanamycin (Km), ampicillin (Ap), or streptomycin (Str) was added at a final concentration of 50, 100, or 200 µg ml¹, respectively. Growth studies were done in an asparagine- and citrate-containing synthetic medium (AC) (13); when necessary, AC medium was supplemented with the appropriate N oxide. Cells for subcellular fractionation were grown as described previously (18).

DNA techniques. Genomic DNA was extracted from cells by the one-step chloroform method (14); plasmid DNA was prepared by alkaline cell lysis (22). For cloning purposes, plasmid DNA was purified in a preparative agarose gel with crystal violet (36) combined with the QIAquick gel extraction method (Qiagen). Standard procedures were used for agarose gel electrophoresis, dephosphorylation, ligation, PCR amplification of DNA fragments, and transformation of E. coli by electroporation (41). Restriction enzymes were used as recommended by the manufacturers.

Gene probes. Probes for tatC were deduced from sequence information of Pseudomonas aeruginosa. Data were obtained from the web site of the Pseudomonas Genome Project (http://www.pseudomonas.com). The probe TC1 was amplified from genomic DNA of P. aeruginosa PAO with the primers 5'-GGTCTGGGGCTTCATCGC-3' and 5'-CCATGCCGACCACGAAAC-3', which have the start positions 5707528 and 5707874, respectively, in the P. aeruginosa chromosome. The probe TC2 was generated with the primers 5'-GGCGGCGATCTTCCTGATCT-3' and 5'-AGGAACAGCGCCACCACCAT-3' at positions 5707330 and 5707503, respectively. The annealing temperature was 57°C. Labeling was done with digoxigenin (DIG) by PCR (Boehringer Mannheim).

Subcloning and sequencing of the tatC fragment. A P. stutzeri cosmid library (8) was screened for tatC by Southern hybridization with the probe TC1. DNA was blotted to nitrocellulose membranes by downward alkaline capillary transfer (15); hybridization was done at 55°C (20). A 3-kb PstI fragment was subcloned into pBluescript II SK(+) with E. coli DH10B as the host to give plasmid pBTAT3. A 1,831-bp region of this plasmid was sequenced on both strands by primer walking using a dye terminator kit or Cy5-labeled primers with an ALFexpress sequencer according to the instructions of the manufacturer (Amersham Pharmacia Biotech).

Mutagenesis procedures. A 150-bp BglII fragment, covering the sequence positions 1078 to 1227, was excised from the tatC gene in plasmid pBTAT3 and replaced by a Km^r cassette derived from the BamHI-digested plasmid pBSL15. The orientation of the cassette in the construct pBTAT3K was verified to be opposite to that of tatC by sequencing. pBTAT3K was transferred to P. stutzeri MK21 by electroporation (21). The transformed cells were plated on LB plates containing 200 µg of kanamycin ml¹ and 200 µg of streptomycin ml¹. After 2 days of incubation at 30°C, colonies were picked and tested for ampicillin sensitivity (200 µg ml¹). Genomic DNA from Ap^s strains was prepared, digested with PstI, and analyzed by Southern hybridization with the TC1 probe. The tatC strain MK4T4 was obtained by this procedure.

Construction of nos gene expression vectors. The complete nos gene cluster was assembled in vitro from an Eco47III-SmaI fragment of cosmid cDEN1 (8), carrying nosRZD', and a fragment amplified by PCR. The latter was designed to carry the missing 3' part of nosD together with either nosYL or the nosD'LY part and, additionally, the tatE gene. The genes were assembled in plasmid pUCP22, whose replicon is functional in Pseudomonas species. The first step consisted of cloning the cDEN1 fragment into Ecl136II- and SmaI-digested pUCP22 to yield plasmid pUCP22RZ. Next, PCR amplification was done with cDEN1 as the template. The forward primer, 5'-ACGTGCGCAGATCAGCAATAACC-3', was located 34 bp upstream of the SmaI site in nosD used for the construction of pUCP22RZ. The reverse primers, 5'-gtactatctagaGCCGAACAGCATGACGAC-3' and 5'-gtactatctagaCGCGCAGTCTTGTAGAGG-3', were designed to add XbaI sites (lowercase) 174 and 45 bp downstream of nosL and tatE, respectively. The PCR products were digested with SmaI and XbaI and ligated into the likewise digested pUCP22RZ, yielding pUCP22RL (nosRZDFYL) and pUCP22RE (nosRZDFYL tatE).

Purification of cytochrome cd₁. After high-speed centrifugation, cell extract was loaded onto a DE-52 cellulose column (2.5 by 12 cm; Whatman) and eluted with a linear gradient (300 ml; 0 to 0.35 M NaCl). The buffer was 25 mM Tris-HCl, pH 7.5, throughout the procedure. Cytochrome cd₁-containing fractions were detected immunochemically, concentrated, and applied to a fast-protein liquid chromatography Superdex 200 column (1.6 by 60 cm; Amersham Pharmacia Biotech), which was developed at 0.5 ml/min with buffer (0.2 M in NaCl). The concentrated cytochrome cd₁ was applied to a 1-ml High-Trap anion-exchange column (Amersham Pharmacia Biotech), eluted with a linear gradient (30 ml; 0 to 0.25 M NaCl), and concentrated by ultrafiltration (Centricon 10,000).

Cell fractionation, gel electrophoresis, and enzyme assays. The periplasm, cytoplasm, and membrane fraction were prepared by an osmotic-shock procedure (32). The protein concentration was determined by the Lowry method. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis with a 12.5% acrylamide gel was used for protein separation. Immunochemical detection of N₂OR and nitrite reductase was done with polyclonal antisera and protein A-horseradish peroxidase conjugate (18). The activities of nitrate, nitrite, NO, and N₂O reductases of whole cells and cell extract were measured by gas chromatography or by colorimetric determination of nitrite (23, 35).

Nucleotide sequence accession number. The nucleotide sequence data reported here have been deposited in the EMBL nucleotide sequence data bank under the accession number AJ299712.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of tat genes from P. stutzeri. As a starting point, we used the sequence information provided by the Pseudomonas Genome Project to search the P. aeruginosa genome for tat homologs with E. coli sequences. A putative tatABC gene cluster was found in the P. aeruginosa PAO1 chromosome at the sequence position 5706551 to 5708038 (Fig. 1). Tat-mediated protein translocation in this bacterium has not been addressed; however, the high degree of positional identity of the predicted P. aeruginosa Tat proteins with the corresponding E. coli components allowed a clear assignment to tat genes (Fig. 2). The annotated sequence of the P. aeruginosa chromosome was published recently; a TatE component is not apparent in this sequence (47).

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Physical maps of the DNA regions carrying tat genes in P. stutzeri and P. aeruginosa. Transcriptional directions are indicated by the open arrows. The effected deletion in tatC, strain MK4T4, and the insertion site of the Kmr cassette in tatE, strain MK4E1, are illustrated. Downstream of tatE, the nnrS gene, formerly orf396, shows similarity to nnrS of Rhodobacter sphaeroides (accession number AF016258). The bars indicate the positions of the gene probes used in this work.

View larger version (61K): [in this window] [in a new window]   FIG. 2.   Multiple sequence alignments showing the similarity of Tat proteins from P. stutzeri (Pst) and P. aeruginosa (Pae) to their homologs from E. coli (Eco). The alignments were done by the BESTFIT algorithm of the Genetics Computer Group programs. Identical residues are shaded.

TatC is required for N₂O respiration. If N₂OR were translocated by the Tat system, the mutational inactivation of tatC should prevent export of the reductase, since inactivation of tatC blocks translocation of proteins with the twin-arginine leader peptide in E. coli (6). We have generated a tatC deletion strain and studied its growth behavior under conditions where different respiratory substrates are utilized. The activities of the denitrification enzymes of whole cells and, when appropriate, of cell extract were also followed (Fig. 3). The following traits of the tatC deletion strain MK4T4 of P. stutzeri were observed. Since the aerobic growth rate of the mutant was only slightly decreased, we conclude that no vital component of aerobic respiration had been affected (Fig. 3A). When cells were grown under denitrifying conditions, i.e., with nitrate under O₂ limitation, growth was clearly retarded, indicating that one or more functions of denitrification had been targeted by mutation of tatC (Fig. 3B). Nevertheless, nitrate reductase, the first enzyme of the anaerobic denitrification pathway, was active, and nitrite accumulated in a nitrate-containing medium. Hence, the cause of growth inhibition had to be elsewhere in the further reduction of nitrite. The most striking effect was found when cells were grown anaerobically with N₂O as the electron acceptor. While the wild type, represented by MK21 (Str^r), grew readily on N₂O, MK4T4 had lost this capability (Fig. 3C). We also assayed cells for the reduction of the denitrification substrates, nitrite, NO, and N₂O. The tatC mutation caused the loss of N₂O reductase activity (Fig. 3D). Cell extract was not assayed in this case because N₂OR is usually inactive in crude extract due to an unspecified interference (55). There was also a striking loss of nitrite reductase activity both in vivo and in vitro (Fig. 3E). The underlying causes of the observed Nos and Nir phenotypes of MK4T4 were studied in more detail.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Growth characteristics of the tatC mutant and denitrification activities in vivo and in vitro of tatC and tatE mutants. (A to C) Growth curves. , MK4T4; , wild-type behavior represented by MK21. (A) Aerobic conditions; 150 ml of AC medium in 500-ml flasks shaken on a gyratory incubator at 240 rpm and 30°C. (B) O2-limited, denitrifying conditions; AC medium containing nitrate (1 g/liter); shaker speed, 120 rpm. (C) Incubation with N2O; 100 ml of AC medium in 150-ml flasks inoculated with overnight cultures to give a beginning optical density at 660 nm (OD660) of 0.05. The flasks were sparged with filter-sterilized N2O and incubated at 30°C. (D to F) Activity measurements of MK4T4 (; ), MK4E1 (), and MK21 (). (D) N2OR activity of whole cells. (E) Nitrite reductase activity of whole cells and additionally MK4T4 in vitro (). (F) NO reductase activity of whole cells and additionally MK4T4 in vitro (). For the assay conditions, see Materials and Methods. The reaction was started by the addition of the substrate. Gases were analyzed by gas chromatography.

tatC inactivation results in a mislocated, cytoplasmic N₂OR lacking the Cu chromophores. The location of N₂OR in MK4T4 was investigated by fractionating cells into cytoplasm, periplasm, and membranes and analyzing them for the enzyme immunochemically. N₂OR of MK4T4 was found to be mislocated in the cytoplasmic cell compartment, mainly in its unprocessed precursor form (Fig. 4A). No N₂OR was detected in the periplasm or bound to membranes of MK4T4. Under all growth conditions, only one band, corresponding to processed enzyme, was detected in the cell extract of MK21 representing wild-type character. Thus, translocation and processing of N₂OR was not rate limiting in relation to synthesis.

View larger version (57K): [in this window] [in a new window]   FIG. 4.   Localization and processing of N2OR in tat and CuA mutants. Immunoblot analysis with polyclonal antisera was done for cell extract (CE) and cells fractionated into periplasm (P), cytoplasm (C), and membranes (M). Enzyme (E) purified from MK21 was used as a size marker for the mature forms of N2OR (NosZ) and cytochrome cd1 (NirS). Cell extract from the signal peptide mutant R20D provided unprocessed N2OR (pre-NosZ); cell extract from MK21 shows the wild-type situation. (A) Cytoplasmic location of a predominantly unprocessed N2OR in MK4T4; cell extract from MK21 shows only processed N2OR, whereas extract from MK4E1 also shows unprocessed enzyme. (B) Processing and periplasmic location of cytochrome cd1 nitrite reductase in MK4T4. The cell fractions of MK4T4 were the same as those used in panel A. The cytoplasmic fraction was free of contamination by periplasm, as is evident from the absence of a cytochrome cd1 signal. In panels A and B, no N2OR or cytochrome cd1 was found bound to membranes. (C) Localization and processing of N2OR in CuA mutants. N2OR in the cell extracts of C618D and C622D was completely processed. C618V synthesized a predominantly unprocessed, cytoplasmic N2OR; the small contribution of processed enzyme was periplasmic. No unprocessed material was found in the cell extract of C622V; H583 harbored both species, with the mature form being predominant.

View larger version (20K): [in this window] [in a new window]   FIG. 5.   Lack of incorporation of Cu into N2OR of the tatC mutant. (A) Cell extract of MK4T4 was separated on a Sephacryl S-300 HR gel column (2.5 by 90 cm) equilibrated with 25 mM Tris-HCl (pH 7.5)-0.1 M KCl (flow rate, 25 ml/h; 2.5-ml fractions). (B) The same experiment using extract from MK21. The eluate was analyzed immunochemically for N2OR () and for Cu () by atomic absorption spectroscopy (18). The data were normalized with respect to the highest concentrations of N2OR and Cu in each data set.

Role for TatE in N₂OR translocation. tatE (formerly orf57) of P. stutzeri is part of the nos region encoding assembly factors. We have generated a tatE mutant (see Materials and Methods) and studied in the knockout strain MK4E1 whether TatE has a role specific for the translocation or maturation of its N₂OR substrate. Mutant MK4E1 grew aerobically with the growth rate of strain MK21. It also grew on N₂O like the wild type and, consequently, intact cells had N₂OR activity (Fig. 3D). We also found nitrite reductase and NO reductase activities, which were only slightly reduced, if at all, compared to those of MK21 (Fig. 3E and F). However, cell extract of MK4E1 consistently showed a small fraction of the unprocessed form of N₂OR (Fig. 4A), whereas in MK21, no unprocessed enzyme was detectable.

View larger version (46K): [in this window] [in a new window]   FIG. 6.   TatE supports transport and processing of N2OR. (A) Expression vector, pUCP22RE, for the entire nos region, including tatE; vector pUCP22RL was identical but without tatE. The rep function is required for stable maintenance of the vector in Pseudomonas spp. (49). (B) Immunoblot analysis for N2OR of cell extracts of MK21 (lane 1), nosD mutant MK401 (lane 2), MK401 complemented with pUCP22RL (nosRZDFL) (lane 3), and MK401 complemented with pUCP22RE (nosRZDFYtatE) (lane 4). The protein standard was Seeblue (Novex). (C) Southern blot evidence for a tatE homolog in P. stutzeri. Genomic DNA was digested by PstI and hybridized at 45°C stringency by the downward capillary method; detection was done with DIG Easy Hyb granules (Roche). The observed fragments correspond to the tatABC cluster (3 kb), tatE (1.6 kb), and a putative homolog of tatE (0.35 kb).

The Tat system is sensitive to mutation of the Cu_A domain of N₂O reductase. The Tat system is thought to transport folded proteins, in contrast to the Sec system (3). We asked, therefore, whether the Tat system would be sensitive to a conformational alteration of the exported protein. We addressed this with site-directed mutants of the Cu_A center. The previously generated Cu_A mutants, H583G, C618D, C622D, H626G, and M629C, displayed a loss of catalytic activity, but all of these recombinant N₂OR derivatives were exported to the periplasm (13). Figure 4C shows cell extracts of C618D and C622D, which contained only mature N₂OR consistent with the periplasmic location. Both cysteine mutations affected the bridging residues of the Cu_A center. Replacement of the same residues by valine resulted in a differential positional effect for cysteine 618. Cell extract from the C618V mutant showed a predominance of the unprocessed N₂OR species and a small amount of processed material (Fig. 4C). Unprocessed and processed forms of enzyme C618V were located in the cytoplasm and periplasm, respectively. In contrast, the C622V recombinant form exhibited wild-type behavior. This indicated for the C618V exchange a strong qualitative and positional effect. The H583G mutation of the Cu_A center also consistently showed a small percentage of unprocessed N₂OR (Fig. 4C), and occasionally we have observed this with the M629C derivative. Thus, there are distinct structural alterations in the Cu_A domain of N₂OR to which the Tat translocation pathway is sensitive.

The synthesis of a functional cytochrome cd₁ is TatC dependent. A conspicuous trait of MK4T4 was related to cytochrome cd₁ nitrite reductase. This enzyme was detected immunochemically in MK4T4, where it was exported to the periplasm (Fig. 4B). However, the cells did not reduce nitrite, and the weak reductase activity detected in vitro could not be due to cytochrome cd₁ for the following reason: we have isolated cytochrome cd₁ from MK4T4 and found it to carry only heme C but no heme D₁, as is evident from the electronic spectrum (Fig. 7). We attribute this to the dependence of the NirD protein, involved in heme D₁ biosynthesis, on the Tat translocation pathway (see Discussion). In contrast, cytochrome cd₁ from MK4E1 exhibited wild-type properties and was synthesized as an active holoenzyme with both types of heme groups.

View larger version (24K): [in this window] [in a new window]   FIG. 7.   Electronic absorption spectrum of purified cytochrome cd1 from tatC and tatE mutants. a, dithionite-reduced semiapo-cytochrome cd1 from MK4T4 (tatC); absorption bands of only the heme C moiety are present. b, dithionite-reduced holoenzyme from MK4E1 (tatE); the arrows indicate absorption bands due to heme D1.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In gram-negative bacteria, complete nitrite respiration leading to N₂ is a periplasmic process, with nitrite reductase and N₂OR both located in the outer cell compartment together with electron carriers and ancillary proteins necessary for denitrification (53). For the extant topology of the denitrification apparatus, it is necessary to understand how the principal components are transported to their functional site. Our strategy to isolate tat genes was based on sequence information from the chromosome of P. aeruginosa and a comparison with their E. coli homologs. This enabled us to design specific probes for identification of the P. stutzeri counterparts based on the considerable sequence similarity between the two species. We anticipate that the findings with P. stutzeri have relevance for its close relative.

A brief survey among pseudomonads indicated that the Tat system occurs frequently in this bacterial group. We found tatC homologs in EcoRI-digested genomic DNA by Southern hybridization with probe TC1 at 50°C stringency in Pseudomonas aureofaciens (ATCC 13985); Pseudomonas chlororaphis (ATCC 9446); Pseudomonas fluorescens biotypes A (DSM 50091), C (DSM 50117), and F (ATCC 12983); Pseudomonas putida (DSM 50906); and Ralstonia (formerly Pseudomonas) solanacearum (DSM 50905) irrespective of whether these strains are denitrifiers or utilize N₂O (data not shown). In agreement with our evidence from Southern hybridization, we found a tatABC cluster in contig 10775 at positions 40584 to 42019 in the ongoing genome project of P. putida KT2440 (The Institute for Genomic Research website [http://www.tigr.org]).

TatC and TatB of E. coli play crucial roles in the translocation mechanism, because deletion of either tat gene is sufficient to block protein export (6, 44). Likewise, a deletion in tatC of P. stutzeri blocked N₂OR export and caused the unprocessed enzyme to accumulate in the cytoplasm. In the denitrifier Ralstonia eutropha, a putative tatA locus has also been implicated in N₂OR translocation based on the observation of a mutational loss of N₂O-reducing activity of whole cells (5). The P. stutzeri tatC mutant exhibited the phenotype of the R20D mutant studied previously (18). Although the unprocessed N₂OR of mutant R20D lacked Cu, the electron transfer center, Cu_A, could be occupied by incubating the isolated protein with Cu. We assume this to be indicative of a correctly folded form of the enzyme in spite of its cytoplasmic location. Although a rather specific conformation seems to be required to explain the observation regarding the C618V and H583G recombinant forms of the enzyme, Cu insertion into N₂OR is not a prerequisite for the Tat pathway to function. This is different from nickel-iron hydrogenase 2, where Ni has to be incorporated into the protein prior to export (39). Our results with N₂OR are in contrast to the view that cofactor insertion prior to export is a general trait of Tat-dependent proteins. Preexport cofactor insertion may simply reflect protein folding (42) and serve metabolic efficiency by dispensing with a separate transport system for the cofactor.

On substituting the polar side chain of cysteine 618 for the hydrophobic one of valine, the Tat translocation pathway strongly rejected N₂OR. Valine fits into the site of cysteine in the N₂OR structure (10), but it is unknown precisely which structural change is imposed by the mutation on the electron transfer domain and the overall enzyme conformation. Certain changes in the Cu_A center affect the juxtaposed Cu_Z domain, as deduced from altered spectroscopic properties (13). Our observations indicate that the Tat system is constrained by distinct conformations in the folded form of N₂OR. Since cysteine 618 is positioned only 20 aa away from the C terminus of N₂OR (overall, the enzyme comprises 638 aa), it means that the Tat system recognizes the mutationally induced structural change in the completely or nearly completely synthesized N₂OR.

When a Tat system is present, the genome usually encodes one TatC protein and at least two homologs of the TatA or -E and TatB groups. It has been suggested that TatA, TatB, and TatE may function as membrane receptors for different subsets of redox proteins (12), although no difference in substrate specificity between TatA and TatE has been established (43, 44). Our data do not support a specific receptor role of TatE for N₂OR, in spite of the location of tatE in the nos gene cluster. The view that TatC provides the membrane receptor for the signal peptide and that TatE or TatA forms the translocase pore (4) is more plausible from our observations. A TatE transport channel may be the usual route for N₂OR, but a homolog can fulfill this role when TatE is missing in mutant MK4E1.

The periplasmic cytochrome cd₁ nitrite reductase of P. stutzeri has a Sec-type signal peptide (30). The translocation of cytochrome cd₁ was not affected by the tatC mutation, which strongly suggests export of this enzyme via the Sec pathway. Since nitrite reductase was not active in the tatC strain, at least one Tat-dependent factor is required to establish a functional cytochrome cd₁. We suggest that this factor is NirD. Heme D₁ is found only in the cytochrome cd₁-type nitrite reductase of denitrifying bacteria. The nirD locus is involved in heme D₁ biosynthesis or processing (35). The presequence of the NirD protein, MHIDALSRRLIDRYQHGMPLCAEPYRAMA (critical residues are shown in boldface), exhibits the characteristics of a Tat-specific signal peptide. It has an H region significantly less hydrophobic than that of a Sec signal peptide. The heptameric motif with an arginine pair overlaps the boundary of the N and H regions, and even though this motif varies from the suggested consensus for Tat-dependent transport, most of the variant positions are conservative substitutions and consistent with evidence from site-directed mutagenesis (9, 46). A critical proline residue which is commonly found in Tat signal sequences (16) occupies position 6 of the putative signal peptidase cleavage site, further supporting the notion of a Tat-type signal peptide. Cleavage may be at 27-AMAE, and a basic residue, arginine 26, is next to the consensus sequence, features which are frequent in Tat signal peptides. The likely dependence of NirD export on the Tat pathway raises interesting questions as to whether there are steps in the synthesis of heme D₁ and its delivery to cytochrome cd₁ that proceed in the periplasm or at the periplasmic side of the inner membrane and precisely which function is performed by the NirD protein. Our data on cytochrome cd₁ present an intriguing case where the Tat and Sec systems have to cooperate for the assembly of a periplasmic enzyme.