Export of Escherichia coli alkaline phosphatase attached to an integral membrane protein, SecY

The SecY protein is a membrane-bound factor required for bacterial protein export and embedded in the cytoplasmic membrane by its 10 transmembrane segments. We previously proposed a topology model for this protein by adapting the Manoil-Beckwith TnphoA approach, a genetic method to assign local disposition of a membrane protein from the enzymatic activity of the alkaline phosphatase (PhoA) mature sequence attached to the various regions. SecY-PhoA hybrid proteins with the PhoA domain exported to the periplasmic side of the membrane have been obtained at the five putative periplasmic domains of the SecY sequence. We now extended this method to apply it to follow export of the newly synthesized PhoA domain. Trypsin treatment of detergent-solubilized cell extracts digested the internalized (unfolded) PhoA domain but not those exported and correctly folded. One of the hybrid proteins was cleaved in vivo after export to the periplasm, providing a convenient indication for the export. Results of these analyses indicate that export of the PhoA domain attached to different periplasmic regions of SecY occurs rapidly and requires the normal functioning of the secY gene supplied in trans. Thus, this membrane protein with multiple transmembrane segments contains multiple export signals which can promote rapid and secY-dependent export of the PhoA mature sequence attached to the carboxyl-terminal sides.     

β-lactamase as a probe of membrane protein assembly and protein export

The enzyme TEM beta-lactamase constitutes a versatile gene-fusion marker for studies on membrane proteins and protein export in bacteria. The mature form of this normally periplasmic enzyme displays readily detectable and distinctly different phenotypes when localized to the bacterial cytoplasm versus the periplasm, and thus provides a useful alternative to alkaline phosphatase for probing the topology of cytoplasmic membrane proteins. Cells producing translocated forms of beta-lactamase can be directly selected as ampicillin-resistant colonies, and consequently a beta-lactamase fusion approach can be used for positive selection for export signals, and for rapid assessment of whether any protein expressed in Escherichia coli inserts into the bacterial cytoplasmic membrane. The level of ampicillin resistance conferred on a cell by an extracytoplasmic beta-lactamase derivative depends on its level of expression, and therefore a beta-lactamase fusion approach can be used to directly select for increased yields of any periplasmic or membrane-bound gene products expressed in E. coli.     

Autodisplay: functional display of active beta-lactamase on the surface of Escherichia coli by the AIDA-I autotransporter

Members of the protein family of immunoglobulin A1 protease-like autotransporters comprise multidomain precursors consisting of a C-terminal autotransporter domain that promotes the translocation of N-terminally attached passenger domains across the cell envelopes of gram-negative bacteria. Several autotransporter domains have recently been shown to efficiently promote the export of heterologous passenger domains, opening up an effective tool for surface display of heterologous proteins. Here we report on the autotransporter domain of the Escherichia coli adhesin involved in diffuse adherence (AIDA-I), which was genetically fused to the C terminus of the periplasmic enzyme beta-lactamase, leading to efficient expression of the fusion protein in E. coli. The beta-lactamase moiety of the fusion protein was presented on the bacterial surface in a stable manner, and the surface-located beta-lactamase was shown to be enzymatically active. Enzymatic activity was completely removed by protease treatment, indicating that surface display of beta-lactamase was almost quantitative. The periplasmic domain of the outer membrane protein OmpA was not affected by externally added proteases, demonstrating that the outer membranes of E. coli cells expressing the beta-lactamase AIDA-I fusion protein remained physiologically intact.     

Requirements for translocation of periplasmic domains in polytopic membrane proteins.

Periplasmic domains of cytoplasmic membrane proteins require export signals for proper translocation. These signals were studied by using a MalF-alkaline phosphatase fusion in a genetic selection that allowed the isolation of mislocalization mutants. In the original construct, alkaline phosphatase is fused to the second periplasmic domain of the membrane protein, and its activity is thus confined exclusively to the periplasm. Mutants that no longer translocated alkaline phosphatase were selected by complementation of a serB mutation. A total of 11 deletions in the amino terminus were isolated, all of which spanned at least the third transmembrane segment. This domain immediately precedes the periplasmic domain to which alkaline phosphatase was fused. Our results obtained in vivo support the model that amino-terminal membrane-spanning segments are required for translocation of large periplasmic domains. In addition, we found that the inability to export the alkaline phosphatase domain could be suppressed by a mutation, prlA4, in the secretion apparatus.     

Defective export in Escherichia coli caused by DsbA'-PhoA hybrid proteins whose DsbA' domain cannot fold into a conformation resistant to periplasmic proteases.

The disulfide bond-forming factor DsbA and the alkaline phosphatase are stable in the Escherichia coli periplasmic space and can be overproduced without significant perturbation of the cell's physiology. By contrast, DsbA'-PhoA hybrid proteins resulting from TnphoA insertions into different regions of a plasmid-borne dsbA gene could become toxic (lethal) to bacteria. Toxicity was concomitant with an impairment of some step of the export mechanism and depended on at least three parameters, i.e., (i) the rate of expression of the hybrid protein, (ii) the ability of the amino-terminal DsbA' domain of the hybrid protein to fold into a protease-resistant conformation in the periplasmic space, and (iii) the activity of the DegP periplasmic protease. Even under viable conditions of low expression, DsbA' folding-deficient hybrid proteins accumulated more than the folding-proficient ones in the insoluble material and this was aggravated in a strain lacking the DegP protease. When production was more elevated, the folding-deficient hybrid proteins became lethal, but only in strains lacking the DegP activity, while the folding-proficient ones were not. Under conditions of very high production by degP+ or degP strains, both types of hybrid proteins accumulated as insoluble preproteins. Meanwhile, the export machinery was dramatically handicapped and the cells lost viability. However, the folding-deficient hybrid proteins had a higher killing efficiency than the folding-proficient ones. Free DsbA'-truncated polypeptides, although not toxic, were processed more slowly when they could not fold into a protease-resistant form in the periplasmic space. This provides indications in E. coli for a direct or indirect influence of the folding of a protein in the periplasmic environment on export efficiency.     

Multiple roles of the pilus biogenesis protein pilD: involvement of pilD in excretion of enzymes from Pseudomonas aeruginosa.

In Pseudomonas aeruginosa, the genes pilB, pilC, and pilD encode proteins necessary for posttranslational modification and assembly of pilin monomers into pilus organelles (D. Nunn, S. Bergman, and S. Lory, J. Bacteriol. 172:2911-2919, 1990). We show that PilD, encoding a putative pilin-specific leader peptidase, also controls export of alkaline phosphatase, phospholipase C, elastase, and exotoxin A. pilD mutants accumulate these proteins in the periplasmic space, while secretion of periplasmic and outer membrane proteins appears to be normal. The periplasmic form of exotoxin A was fully mature in size, contained all cysteines in disulfide bonds, and was toxic in a tissue culture cytotoxicity assay, suggesting that in pilD mutants, exotoxin A was folded into its native conformation. The function of the other two accessory proteins, PilB and PilC, appears to be restricted to pilus biogenesis, and strains carrying mutations in their respective genes do not show an export defect. These studies show that in addition to cleaving the leader sequence from prepilin, PilD has an additional role in secretion of proteins that are released from P. aeruginosa into the surrounding media. PilD most likely functions as a protease that is involved in processing and assembly of one or more components of the membrane machinery necessary for the later stages of protein extracellular localization.     

A family of modular type mannuronan C-5-epimerase genes controls alginate structure in Azotobacter vinelandii

The ToxR protein is a transmembrane protein that regulates the expression of several virulence factors of Vibrio cholerae. Previous analysis of fusion proteins between ToxR and alkaline phosphatase (ToxR-PhoA) suggested that ToxR was active as a dimer. In order to determine whether dimerization of the ToxR periplasmic domain was essential for activity, this domain was replaced by monomeric and dimeric protein domains. Surprisingly, PhoA (dimeric), β-lactamase (monomeric, ToxR–Bla), or the leucine zipper of GCN4 (dimeric, ToxR-GCN4-M) could substitute functionally for the ToxR periplasmic domain. ToxR-GCN4 fusion proteins, in which the ToxR trans-membrane domain was eliminated (ToxR-GCN4-C), were inactive, but an additional fusion protein that contained a heterologous membrane-spanning domain retained activity. Strains containing each of these ToxR fusion proteins were analysed for in vivo colonization properties and response to in vitro growth conditions that are known to affect expression of the ToxR regulon. Strains containing ToxR-GCN4-M and ToxR-Bla responded like wild-type strains to in vitro growth conditions. In the infant-mouse colonization model, strains containing ToxR fusion proteins were all deficient in colonization relative to strains containing wild-type ToxR, and strains containing monomeric ToxR-Bla were most severely outcompeted. These results suggest that, under in vitro conditions, ToxR does not require a dimerized periplasmic domain, but that, under in vivo conditions, the correct conformation of the ToxR periplasmic domain may be more important for function.     

The Tsr chemosensory transducer of Escherichia coli assembles into the cytoplasmic membrane via a SecA-dependent process

The Tsr protein of Escherichia coli is a chemosensory transducer that mediates taxis toward serine and away from certain repellents. Like other bacterial transducers, Tsr spans the cytoplasmic membrane twice, forming a periplasmic domain of about 150 amino acids and a cytoplasmic domain of about 300 amino acids. The 32 N-terminal amino acids of Tsr resemble the consensus signal sequence of secreted proteins, but they are not removed from the mature protein. To investigate the function of this N-terminal sequence in the assembly process, we isolated translational fusions between tsr and the phoA and lacZ genes, which code for the periplasmic enzyme alkaline phosphatase and the cytoplasmic enzyme beta-galactosidase, respectively. All tsr-phoA fusions isolated code for proteins whose fusion joints are within the periplasmic loop of Tsr, and all of these hybrid proteins have high alkaline phosphatase activity. The most N-terminal fusion joint is at amino acid 19 of Tsr. Tsr-lacZ fusions were found throughout the tsr gene. The beta-galactosidase activity of the LacZ-fusion proteins varies greatly, depending on the location of the fusion joint. Fusions with low activity have fusion joints within the periplasmic loop of Tsr. The expression of these fusions is most likely reduced at the level of translation. In addition, one of these fusions markedly reduces the export and processing of the periplasmic maltose-binding protein and the outer membrane protein OmpA, but not of intact PhoA or of the outer membrane protein LamB. A temperature-sensitive secA mutation, causing defective protein secretion, stops expression of new alkaline phosphatase activity coded by a tsr-phoA fusion upon shifting to the nonpermissive temperature. The same secA mutation, even at the permissive temperature, increases the activity and the level of expression of LacZ fused to the periplasmic loop of Tsr relative to a secA+ strain. We conclude that the assembly of Tsr into the cytoplasmic membrane is mediated by the machinery responsible for the secretion of a subset of periplasmic and outer membrane proteins. Moreover, assembly of the Tsr protein seems to be closely coupled to its synthesis.     

The normally periplasmic enzyme β-lactamase is specifically and efficiently translocated through the Escherichia coli outer membrane when it is fused to the cell-surface enzyme pullulanase

Hybrid proteins were constructed in which C-terminal regions of the bacterial cell surface and extracellular protein pullulanase were replaced by the mature forms of the normally periplasmic Escherichia coli proteins beta-lactamase or alkaline phosphatase. In E. coli strains expressing all pullulanase secretion genes, pullulanase-beta-lactamase hybrid protein molecules containing an N-terminal 834-amino-acid pullulanase segment were efficiently and completely transported to the cell surface. This hybrid protein remained temporarily anchored to the cell surface, presumably via fatty acids attached to the N-terminal cysteine of the pullulanase segment, and was subsequently specifically released into the medium in a manner indistinguishable from that of pullulanase itself. These results suggest that the C-terminal extremity of pullulanase lacks signal(s) required for export to the cell surface. When beta-lactamase was replaced by alkaline phosphatase, the resulting hybrid also became exposed at the cell surface, but exposition was less efficient and specific release into the medium was not observed. We conclude that proteins that do not normally cross the outer membrane can be induced to do so when fused to a permissive site near the C-terminus of pullulanase.     

A bacterial two-hybrid system based on the twin-arginine transporter pathway of E. coli

We have developed a bacterial two-hybrid system for the detection of interacting proteins that capitalizes on the folding quality control mechanism of the Twin Arginine Transporter (Tat) pathway. The Tat export pathway is responsible for the membrane translocation of folded proteins, including proteins consisting of more than one polypeptide, only one of which contains a signal peptide ("hitchhiker export"). Here, one protein (bait) is expressed as a fusion to a Tat signal peptide, whereas the second protein (prey) is fused to a protein reporter that can confer a phenotype only after export into the bacterial periplasmic space. Since the prey-reporter fusion lacks a signal peptide, it can only be exported as a complex with the bait-signal peptide fusion that is capable of targeting the Tat translocon. Using maltose-binding protein as a reporter, clones expressing interacting proteins can be grown on maltose minimal media or on MacConkey plates. In addition, we introduce the use of the cysteine disulfide oxidase DsbA as a reporter. Export of a signal peptide-prey:bait-DsbA complex into the periplasm allows complementation of dsbA(-) mutants and restores the formation of active alkaline phosphatase, which in turn can be detected by a chromogenic assay.     

Use of phoA fusions to study the topology of the Escherichia coli inner membrane protein leader peptidase.

A topology of the Escherichia coli leader peptidase has been previously proposed on the basis of proteolytic studies. Here, a collection of alkaline phosphatase fusions to leader peptidase is described. Fusions to the periplasmic domain of this protein exhibit high alkaline phosphatase activity, while fusions to the cytoplasmic domain exhibit low activity. Elements within the cytoplasmic domain are necessary to stably anchor alkaline phosphatase in the cytoplasm. The amino-terminal hydrophobic segment of leader peptidase acts as a weak export signal for alkaline phosphatase. However, when this segment is preceded by four lysines, it acts as a highly efficient export signal. The coherence of in vitro studies with alkaline phosphatase fusion analysis of the topology of leader peptidase further indicates the utility of this genetic approach to membrane protein structure and insertion.     

Mutagenesis affecting the carboxyl terminus of the biotinyl subunit of transcarboxylase. Effects on biotination

Biotin is added to biotin-containing enzymes as a post-translational modification catalyzed by holoenzyme synthetase. This reaction is fairly general in that synthetase from one organism will modify enzymes from heterologous sources. This suggests that the polypeptides share some structural characteristic(s) that define(s) them as biotin enzymes. We have reported previously that when the gene coding for the 1.3 S biotinyl subunit of transcarboxylase is expressed in Escherichia coli, the polypeptide produced is biotinated by the cellular synthetase. Using in vitro mutagenesis of this gene, we have begun to define the primary structure involved in the enzymatic addition of biotin to a lysine residue. We show here that the carboxyl terminus of the 1.3 S subunit is critical in biotination. Mutations affecting the COOH-terminal residue do not influence the modification, but elimination of the hydrophobic side chain of the penultimate residue abolishes biotin addition.     

Lack of effect of leader peptidase overproduction on the processing in vivo of exported proteins in Escherichia coli

The kinetics of maturation of certain exported proteins were analysed in Escherichia coli strains that also concomitantly overproduce either a periplasmic protein or the leader peptidase. The results led to three conclusions. Overproduction of leader peptidase has no effect on the rate of maturation of at least two exported proteins, one periplasmic (TEM beta-lactamase), one outer membrane (PhoE); therefore, the quantity of leader peptidase is not rate-limiting for normal export. Overproduction of PhoS reduces the rate of maturation of two other periplasmic proteins (beta-lactamase and PhoA) and itself, presumably by competing for the rate-limiting component of the export apparatus. Overproduction of leader peptidase in a strain overproducing PhoS has no effect on the retarded maturation of PhoS. Therefore even in these conditions, leader peptidase is not rate limiting.     

A genetic screen for suppressors of Escherichia coli Tat signal peptide mutations establishes a critical role for the second arginine within the twin-arginine motif

The Escherichia coli Tat protein export pathway transports folded proteins synthesized with N-terminal twin-arginine signal peptides. Twin-arginine signal sequences contain a conserved SRRxFLK "twin-arginine" amino acid sequence motif which is required for protein export by the Tat pathway. The E. coli trimethylamine N-oxide reductase (TorA) is a Tat-dependent periplasmic molybdoenzyme that facilitates anaerobic respiration with trimethylamine N-oxide as terminal electron acceptor. Here, we describe mutant strains constructed with modified TorA twin-arginine signal peptides. Substitution of the second arginine residue of the TorA signal peptide twin-arginine motif with either lysine or aspartate, or the simultaneous substitution of both arginines with lysine residues, completely abolished export. In each case, the now cytoplasmically localised TorA retained full enzymatic activity with the artificial electron donor benzyl viologen. However, the mutant strains were incapable of anaerobic growth with trimethylamine N-oxide and the non-fermentable carbon-source glycerol. The growth phenotype of the mutant strains was exploited in a genetic screen with the aim of identifying second-site suppressor mutations that allowed export of the modified TorA precursors.     

Export of beta-lactamase is independent of the signal recognition particle

In Escherichia coli, three different types of proteins engage the SecY translocon of the inner bacterial membrane for translocation or insertion: 1) polytopic membrane proteins that prior to their insertion into the membrane are targeted to the translocon using the bacterial signal recognition particle (SRP) and its receptor; 2) secretory proteins that are targeted to and translocated across the SecY translocon in a SecA- and SecB-dependent reaction; and 3) membrane proteins with large periplasmic domains, requiring SRP for targeting and SecA for the translocation of the periplasmic moiety. In addition to its role as a targeting device for membrane proteins, a function of the bacterial SRP in the export of SecB-independent secretory proteins has also been postulated. In particular, beta-lactamase, a hydrolytic enzyme responsible for cleavage of the beta-lactam ring containing antibiotics, is considered to be recognized and targeted by SRP. To examine the role of the SRP pathway in beta-lactamase targeting and export, we performed a detailed in vitro analysis. Chemical cross-linking and membrane binding assays did not reveal any significant interaction between SRP and beta-lactamase nascent chains. More importantly, membrane vesicles prepared from mutants lacking a functional SRP pathway did block the integration of SRP-dependent membrane proteins but supported the export of beta-lactamase in the same way as that of the SRP-independent protein OmpA. These data demonstrate that in contrast to previous results, the bacterial SRP is not involved in the export of beta-lactamase and further suggest that secretory proteins of Gram-negative bacteria in general are not substrates of SRP.     

Protein targeting towards the thylakoid lumen of chloroplasts: proper localization of fusion proteins is only observed in vivo.

Routeing of fusion proteins to the thylakoid lumen of the chloroplast was compared in vitro and in vivo. The Escherichia coli protein beta-lactamase was used as a passenger to study this intraorganellar sorting process. The first step, translocation of beta-lactamase into the chloroplast stroma, occurs properly both in vitro and in vivo and is dependent on the presence of a transit peptide in the protein construct. The second step, targeting towards the thylakoid lumen, is more complicated as was also observed previously when other passenger proteins were used. In vitro, the presence of a thylakoid transfer domain is not enough for routeing and proper processing. Only when the complete thylakoid lumen precursor plastocyanin was fused to beta-lactamase was the fusion protein processed adequately, but routeing was still incomplete. However, in vivo, the information present in the thylakoid transfer domain was the only requirement for proper transport towards the thylakoid lumen. These data show that in vivo, the only requirement for targeting of passenger proteins towards the thylakoid lumen is the presence of a transit peptide and a thylakoid transfer domain. Furthermore, we demonstrate that the in vitro import system does not necessarily reflect the in vivo situation with respect to intraorganellar sorting.     

Disulphide bridge formation in the periplasm of Escherichia coli: β-lactamase::human lgG3 hinge fusions as a model system

We report the construction and the expression in Escherichia coli of three different fusion genes encoding the extended human IgG3 hinge region (Hi) fused in-phase to the C-terminal end of bacterial TEM1 beta-lactamase (Bla). In the first fusion gene blahi, TEM1 beta-lactamase (Bla). In the first fusion gene blahi, the hinge sequence was directly coupled to the 3' end of the beta-lactamase gene, whereas in the two other constructs, blal1hi and blal2hi, a linker encoding 14 and 10 amino acids, respectively, was inserted between the two subunits. After expression (24 h, 20 degrees C) under control of the constitutive kanamycin phosphoribosyl transferase promoter, the fusion proteins, BlaHi, BlaL1Hi and BlaL2Hi, respectively, were almost exclusively detected in the periplasmic fraction, and they conferred carbenicillin-resistance to the cells. These results indicate that beta-lactamase can efficiently direct the export of proteins fused to its C-terminus, and moreover, at least some of the exported fusion proteins must carry the beta-lactamase moiety in a properly folded form. Analysis of their assembly, however, revealed that only a minor fraction was recovered as the expected F(ab')2-like dimer. The presence in the periplasm of 'oxidized' monomers (with intrachain disulphide bonds) as well as of several high-molecular-mass proteins, probably resulting from the association between monomers and other cysteine-rich proteins, strongly suggests that the conditions in the bacterial periplasm are insufficient to allow proper assembly of multimeric proteins with several interchain disulphide bonds.     

Dislocation of membrane proteins in FtsH-mediated proteolysis.

Escherichia coli FtsH degrades several integral membrane proteins, including YccA, having seven transmembrane segments, a cytosolic N-terminus and a periplasmic C-terminus. Evidence indicates that FtsH initiates proteolysis at the N-terminal cytosolic domain. SecY, having 10 transmembrane segments, is also a substrate of FtsH. We studied whether and how the FtsH-catalyzed proteolysis on the cytosolic side continues into the transmembrane and periplasmic regions using chimeric proteins, YccA-(P3)-PhoA-His6-Myc and SecY-(P5)-PhoA, with the alkaline phosphatase (PhoA) mature sequence in a periplasmic domain. The PhoA domain that was present within the fusion protein was rapidly degraded by FtsH when it lacked the DsbA-dependent folding. In contrast, both PhoA itself and the TM9-PhoA region of SecY-(P5)-PhoA were stable when expressed as independent polypeptides. In the presence of DsbA, the FtsH-dependent degradation stopped at a site near to the N-terminus of the PhoA moiety, leaving the PhoA domain (and its C-terminal region) undigested. The efficiency of this degradation stop correlated well with the rapidity of the folding of the PhoA domain. Thus, both transmembrane and periplasmic domains are degraded by the processive proteolysis by FtsH, provided they are not tightly folded. We propose that FtsH dislocates the extracytoplasmic domain of a substrate, probably using its ATPase activity.     

Use of a beta-lactamase fusion vector to investigate the organization of penicillin-binding protein 1B in the cytoplasmic membrane of Escherichia coli

The coding region for the mature form of TEM beta-lactamase was fused to random positions within the coding region of the penicillin-binding protein 1B (PBP 1B) gene and the nucleotide sequences across the fusion junctions of 100 in-frame fusions were determined. All fusion proteins that contained at least the NH2-terminal 94 residues of PBP 1B provided individual cells of E. coli with substantial levels of ampicillin resistance, suggesting that the beta-lactamase moiety had been translocated to the periplasm. Fusion proteins that contained less than or equal to 63 residues of PBP 1B possessed beta-lactamase activity, but could not protect single cells of E. coli from ampicillin, indicating that the beta-lactamase moiety of these fusion proteins remained in the cytoplasm. The beta-lactamase fusion approach suggested a model for the organization of PBP 1B in which the protein is embedded in the cytoplasmic membrane by a single hydrophobic transmembrane segment (residues 64-87), with a short NH2-terminal domain (residues 1-63), and the remainder of the polypeptide (residues 88-844) exposed on the periplasmic side of the cytoplasmic membrane. The proposed model for the organization of PBP 1B was supported by experiments which showed that the protein was completely digested by proteinase K added from the periplasmic side of the cytoplasmic membrane but was only slightly reduced in size by protease attack from the cytoplasmic side of the membrane.     

Behaviour of topological marker proteins targeted to the Tat protein transport pathway

The Escherichia coli Tat system mediates Sec-independent export of protein precursors bearing twin arginine signal peptides. Formate dehydrogenase-N is a three-subunit membrane-bound enzyme, in which localization of the FdnG subunit to the membrane is Tat dependent. FdnG was found in the periplasmic fraction of a mutant lacking the membrane anchor subunit FdnI, confirming that FdnG is located at the periplasmic face of the cytoplasmic membrane. However, the phenotypes of gene fusions between fdnG and the subcellular reporter genes phoA (encoding alkaline phosphatase) or lacZ (encoding beta-galactosidase) were the opposite of those expected for analogous fusions targeted to the Sec translocase. PhoA fusion experiments have previously been used to argue that the peripheral membrane DmsAB subunits of the Tat-dependent enzyme dimethyl sulphoxide reductase are located at the cytoplasmic face of the inner membrane. Biochemical data are presented that instead show DmsAB to be at the periplasmic side of the membrane. The behaviour of reporter proteins targeted to the Tat system was analysed in more detail. These data suggest that the Tat and Sec pathways differ in their ability to transport heterologous passenger proteins. They also suggest that caution should be observed when using subcellular reporter fusions to determine the topological organization of Tat-dependent membrane protein complexes.