Role of membrane potential in protein folding and domain formation during secretion in Escherichia coli

The synthesis and processing of the periplasmic components of the leucine transport system of E coli have been studied to determine the role played by transmembrane potential in protein secretion. Both the leucine-isoleucine-valine binding protein and the leucine-specific binding protein are synthesized as precursors with 23 amino acid N-terminal leader sequences. The processing of these precursors is sensitive to the transmembrane potential. Since the amino acid sequence and the crystal structure have been determined for the leucine-isoleucine-valine binding protein, it and the closely related leucine-specific binding protein represent convenient models in which to examine the mechanism of protein secretion in E coli. A model for secretion has been proposed, suggesting a role for transmembrane potential. In this model, the N-terminal amino acid sequence of the precursor is assumed to form a hairpin of two helices. The membrane potential may orient this structure to make it accessible to processing. In addition, the model suggests that a negatively charged, folded domain of the secretory protein may electrophorese toward the trans-positive side of the membrane, thus providing an additional role for the transmembrane potential.     

The leucine binding proteins of Escherichia coli as models for studying the relationships between protein structure and function

The genes encoding the leucine binding proteins in E coli have been cloned and their DNA sequences have been determined. One of the binding proteins (LIV-BP) binds leucine, isoleucine, valine, threonine, and alanine, whereas the other (LS-BP) binds only the D- and L-isomers of leucine. These proteins bind their solutes as they enter the periplasm, then interact with three membrane components, livH, livG, and livM, to achieve the translocation of the solute across the bacterial cell membrane. Another feature of the binding proteins is that they must be secreted into the periplasmic space where they carry out their function. The amino acid sequence of the two binding proteins is 80% homologous, indicating that they are the products of an ancestral gene duplication. Because of these characteristics of the leucine binding proteins, we are using them as models for studying the relationships between protein structure and function.     

Nucleotide sequence and genetic characterization reveal six essential genes for the LIV-I and LS transport systems of Escherichia coli

The nucleotide sequence of the genes encoding the high affinity, branched-chain amino acid transport systems LIV-I and LS has been determined. Seven genes are present on a 7568-base pair DNA fragment, six of which participate directly in branched-chain amino acid transport. Two periplasmic amino acid-binding proteins are encoded by the livJ (LIV-BP) and livK (LS-BP) genes. These two proteins confer specificity on the LIV-I and LS transport systems. livK is the first gene in a polycistronic message that includes four genes encoding membrane components, livHMGF. The protein products of the livHMGF genes are shared by the two systems. An analysis of the livH and livM DNA sequences suggests that they encode hydrophobic proteins capable of spanning the membrane several times. The LivG and LivF proteins are less hydrophobic, but are also tightly associated with the membrane. Both LivG and LivF contain the consensus sequence for adenine nucleotide binding observed in many other transport proteins. A deletion strain that does not express any of the liv genes was constructed. This strain was used to show that each of the membrane component genes is required for high affinity leucine transport, including two genes, livM and livF, for which no previous genetic evidence had been obtained.     

Secretory production of human granulocyte colony-stimulating factor in Escherichia coli.

Human granulocyte colony-stimulating factor (hG-CSF) is a glycoprotein, consisting of 174 amino acids, which plays an important role in hematopoietic cell proliferation, differentiation of hemopoietic precursor cells, and activation of mature neutrophilic granulocytes. In this study, secretory production of hG-CSF in the periplasmic space of Escherichia coli using the Bacillus sp. endoxylanase signal peptide was examined. For the efficient expression of hG-CSF gene, the first five codons at the N-terminal were altered based on the E. coli high-frequency codon database. The hG-CSF gene fused to the endoxylanase signal sequence was expressed using an inducible trc promoter. However, recombinant E. coli cells were completely lysed after induction with 1 mM isopropyl-beta-D-thiogalactopyranoside. Insertion of a small oligopeptide (13 amino acids) containing the histidine hexamer and factor Xa cleavage site between the signal peptide and the mature hG-CSF protein allowed successful secretion of hG-CSF into the periplasm without cell lysis. Among the several E. coli strains examined, E. coli BL21(DE3) and E. coli MC4100 allowed production of hG-CSF to the highest levels (20-22% of total proteins) with the secretion efficiencies greater than 98%. The circular dichroism spectra showed that the conformation of purified hG-CSF is almost identical to native hG-CSF.     

SocA is a novel periplasmic binding protein for fructosyl amino acid

Bacterial periplasmic proteins (bPBPs) undergo drastic conformational changes upon binding substrate, making them appealing as novel molecular recognition tools for biosensing. A putative bPBP-encoding gene, socA, belongs to the soc operon responsible for santhopine (fructosyl glutamine, FQ) catabolism of Agrobacterium tumefaciens. The socA gene was isolated and expressed in Escherichia coli as a soluble 28.8kDa periplasmic protein to investigate its properties as a potential bPBP for fructosyl amino acid (FA). The autofluorescence of SocA was used to monitor the protein's conformational change resulting from substrate binding. The fluorescence intensity changed upon binding FQ in a concentration dependent manner with a calculated K(d) of 2.1muM, but was unaffected by the presence of sugars or amino acid. Our results demonstrate that SocA is a novel FA bPBP that can be utilized as a novel molecular recognition element for the monitoring of FA.     

Structural homology between rbs repressor and ribose binding protein implies functional similarity.

The deduced amino acid sequence of the rbs repressor, RbsR, of Escherichia coli is homologous over its C-terminal 272 residues to the entire sequence of the periplasmic ribose binding protein. RbsR is also homologous to a family of bacterial repressor proteins including LacI. This implies that the structure of the repressor consists of a two-domain binding protein portion attached to a DNA-binding domain having the four-helix structure of the LacI headpiece. The implications of these relationships to the mechanism of this class of repressors are discussed.     

Synthesis of Escherichia coli heat-stable enterotoxin STp as a pre-pro form and role of the pro sequence in secretion.

Escherichia coli heat-stable enterotoxin STp is presumed from its DNA sequence to be synthesized in vivo as a 72-amino-acid residue precursor that is cleaved to generate mature STp consisting of the 18 carboxy-terminal amino acid residues. There are two methionine residues in the inferred STp sequence in addition to the methionine residue at position 1. In order to confirm production of the STp 72-amino-acid residue precursor, we substituted the additional methionine residues by oligonucleotide-directed site-specific mutagenesis. Since these substitutions did not cause a significant change in STp production, it can be concluded that STp is normally synthesized as the 72-amino-acid residue precursor. The length of the STp precursor indicated the existence of a pro sequence between the signal peptide and the mature protein. In order to identify the pro sequence and determine its role in protein secretion, deletion and fusion proteins were made. A deletion mutant in which the gene fragment encoding amino acid residues 22 to 53 of STp was removed was made. STp activity was found in the culture supernatant of cells. Amino acid sequence analysis of the purified STp deletion mutant revealed that the pro sequence encompasses amino acid residues 20 to 54. A hybrid protein consisting of STp amino acids 1 to 53 fused in frame from residue 53 to nuclease A was not secreted into the culture supernatant. These results indicate that the pro sequence does not function to guide periplasmic protein into the extracellular milieu.     

Extracellular recombinant protein production from an Escherichia coli lpp deletion mutant

E. coli is one of the most commonly used host strains for recombinant protein production. However, recombinant proteins are usually found intracellularly, in either cytoplasm or periplasmic space. Inadequate secretion to the extracellular environment is one of its limitations. This study addresses the outer membrane barrier for the translocation of recombinant protein directed to the periplasmic space. Specifically, using recombinant maltose binding protein (MalE), xylanase, and cellulase as model proteins, we investigated whether the lpp deletion could render the outer membrane permeable enough to allow extracellular protein production. In each case, significantly higher excretion of recombinant protein was observed with the lpp deletion mutant. Up to 90% of the recombinant xylanase activity and 70% of recombinant cellulase activity were found in the culture medium with the deletion mutant, whereas only 40-50% of the xylanase and cellulase activities were extracellular for the control strain. Despite the weakened outer membrane in the mutant strain, cell lysis did not occur, and increased excretion of periplasmic protein was not due to cell lysis. The lpp deletion is a simple method to generate an E. coli strain to effect significant extracellular protein production. The phenotype of extracellular protein production without cell lysis is useful in many biotechnological applications, such as bioremediation and plant biomass conversion.     

The ClpE protein involved in biogenesis of the CS31A capsule-like antigen is a member of a periplasmic chaperone family in Gram-negative bacteria

Abstract The putative chaperone-like protein ClpE, required for biogenesis of the Escherichia coli capsule-like antigen CS31A, was compared with ten known periplasmic chaperones from E. coli, Klebsiella pneumoniae, Bordetella pertussis, Haemophilus influenzae and Yersinia pestis . The amino acid sequence alignment was superimposed onto the three-dimensional structure of the PapD chaperone of uropathogenic E. coli , and amino acid residues involved in maintaining the structure integrity of the suggested binding site were found identical in most of the 11 chaperones. Construction of a phylogenetic tree to investigate the relationship within the chaperone family has revealed interesting degrees of relatedness between the different proteins.     

A periplasmic intermediate in the extracellular secretion pathway of Pseudomonas aeruginosa exotoxin A.

Pseudomonas aeruginosa exotoxin A is synthesized with a secretion signal peptide typical of proteins whose final destination is the periplasm. However, exotoxin A is released from the cell without a detectable periplasmic pool, suggesting that additional determinants in this protein are important for recognition by a specialized machinery of extracellular secretion. The role of the N terminus of the mature exotoxin A in this recognition was investigated. A series of exotoxin A proteins with amino acid substitutions for the glutamic acid pair at the +2 and +3 positions were constructed by mutagenesis of the exotoxin A gene. These N-terminal acidic residues of the mature exotoxin A protein were found to be important not only for efficient processing of the precursor protein but also for extracellular localization of the toxin. The mutated exotoxin A proteins, in which a glutamic acid at the +2 position was replaced by a lysine or a double substitution of lysine and glutamine for the pair of adjacent glutamic acids, accumulated in precursor forms in the mixed cytoplasmic and membrane fractions, which was not seen with the wild-type exotoxin A. The processing of the precursor form of one exotoxin A mutant, in which the glutamic acid at the +2 position was replaced with a glutamine, was not affected. Moreover, a substantial fraction of the mature forms of all three mutants of exotoxin A accumulated in the periplasm, while wild-type exotoxin A could be detected only extracellularly. The periplasmic pools of these variants of exotoxin A could therefore represent the intermediate state during extracellular secretion. The signal for extracellular localization may be located in a small region near the amino terminus of the mature protein or could consist of several regions that are brought together after the polypeptide has folded. Alternatively, the acidic residues may be important for ensuring a conformation essential for exotoxin A to traverse the outer membrane.     

Stabilization of SecA ATPase by the primary cytoplasmic salt of Escherichia coli

Much is known about the structure, function, and stability of the SecA motor ATPase that powers the secretion of periplasmic proteins across the inner membrane of Escherichia coli. Most studies of SecA are carried out in buffered sodium or potassium chloride salt solutions. However, the principal intracellular salt of E. coli is potassium glutamate (KGlu), which is known to stabilize folded proteins and protein‐nucleic acid complexes. Here we report that KGlu stabilizes SecA, including its dimeric state, and increases its ATPase activity, suggesting that SecA is likely fully folded, stable, and active in vivo at 37°C. Furthermore, KGlu also stabilizes a precursor form of the secreted maltose‐binding protein.     

Compartmentalization of mammalian proteins produced in Escherichia coli

We have examined the patterns of compartmentalization of several mammalian proteins in Escherichia coli which do not have signal peptides or functional signal peptide equivalents. These proteins include (i) human proapolipoprotein A-I (proapoA-I), a 249-residue protein which contains a hexapeptide NH2-terminal prosegment plus a mature domain of 243 residues comprised of tandemly arrayed, docosapeptide repeats with predicted amphipathic alpha-helical structure; (ii) the mature apoA-I molecule without its prosegment; (iii) mouse interleukin-1 beta (IL-1 beta), a 17-kDa protein which is composed of 12 beta strands that form a tetrahedral structure; and (iv) the 31-kDa precursor of IL-1 beta, proIL-1 beta. Efficient expression of these proteins in E. coli was achieved using a plasmid that contains the nalidixic acid-inducible recA promoter and ribosome binding site from the gene 10 leader of bacteriophage T7. In induced cultures the mammalian proteins represented up to 20% of the total bacterial protein mass. Surprisingly, cell fractionation using cold (osmotic) shock indicated that proapoA-I, apoA-I, and IL-1 beta, but not its 31-kDa precursor, were segregated into the periplasmic space with high efficiency: the ratio of periplasmic space/spheroplast distribution ranged from 0.6 to 1.1 in cells harvested 60-180 min after nalidixic acid induction. Not only was this compartmentalization efficient but it was also selective: analysis of the osmotic shock fractions revealed that the periplasmic space preparations were not contaminated with cytoplasmic proteins (e.g. phosphoglycerate dehydrogenase). Sequential Edman degradation showed that these proteins had not undergone any NH2-terminal proteolytic processing. The mammalian proteins did not affect the export of a prototypic bacterial preprotein, beta-lactamase. Together the data suggest that osmotic shock fractionation of E. coli may facilitate the purification of functional foreign proteins produced in this prokaryote. They also raise the possibility that structural elements in these proteins other than conventional signal peptides may effect periplasmic targeting in E. coli.     

Functional Properties of Pro Region of Escherichia coli Heat-Stable Enterotoxin

Escherichia coli heat-stable enterotoxin Ip (STp) is synthesized as the 72-amino-acid residue precursor consisting of three regions: pre region (amino acid residues 1 to 19), pro region (amino acid residues 20 to 54), and mature ST (mST) region (amino acid residues 55 to 72). We examined the role of the pro sequence of STp in enterotoxigenicity of a strain by deleting the gene fragment encoding amino acids 22 to 57. This deletion caused a remarkable reduction of its enterotoxic activity of culture supernatant. In order to analyze the sequence responsible for the function of the pro region, two additional deletion mutants were made. The deletion of the sequence covering amino acids 29 to 38, which is conserved in all sequences of ST reported, brought about a significant reduction of enterotoxic activity but the deletion of the non-conserved sequence (amino acids 40 to 53) did not. This result shows that conserved sequence is mainly responsible for the function. Subsequently, to examine the mechanism of action of the pro region, plasmids carrying DNA sequences of hybrid proteins consisting of pre-pro-nuclease, pre-mST-nuclease, pre-pro-mST-nuclease and pre-pro-nuclease-mST were constructed. Amino acid sequence determination and SDS-polyacrylamide gel analysis revealed that these fusion proteins were cleaved between pre sequence and pro sequence during secretion and the cleaved fusion proteins were accumulated in periplasmic space. But the amount of hybrid protein accumulated in the periplasmic space varied among the strains. That is, the amount of the pre-pro-nuclease gene product that accumulated in the periplasmic space was the highest of all fusion gene products. These results indicate that the existence of the mST region strongly interferes with the translocation of the gene product into the periplasmic space and that the pro region functions to guide the mST region into the periplasmic space.     

Ferric hydroxamate binding protein FhuD from Escherichia coli: mutants in conserved and non-conserved regions

Uptake of iron complexes into the Gram-negative bacterial cell requires highly specific outer membrane receptors and specific ATP-dependent (ATP-Binding-Cassette (ABC)) transport systems located in the inner membrane. The latter type of import system is characterized by a periplasmic binding protein (BP), integral membrane proteins, and membrane-associated ATP-hydrolyzing proteins. In Gram-positive bacteria lacking the periplasmic space, the binding proteins are lipoproteins tethered to the cytoplasmic membrane. To date, there is little structural information about the components of ABC transport systems involved in iron complex transport. The recently determined structure of the Escherichia coli periplasmic ferric siderophore binding protein FhuD is unique for an ABC transport system (Clarke et al. 2000). Unlike other BP's, FhuD has two domains connected by a long -helix. The ligand binds in a shallow pocket between the two domains. In vivo and in vitro analysis of single amino acid mutants of FhuD identified several residues that are important for proper functioning of the protein. In this study, the mutated residues were mapped to the protein structure to define special areas and specific amino acid residues in E. coli FhuD that are vital for correct protein function. A number of these important residues were localized in conserved regions according to a multiple sequence alignment of E. coli FhuD with other BP's that transport siderophores, heme, and vitamin B₁₂. The alignment and structure prediction of these polypeptides indicate that they form a distinct family of periplasmic binding proteins.     

Extracellular secretion of pullulanase is unaffected by minor sequence changes but is usually prevented by adding reporter proteins to its N- or C-terminal end.

Linker insertions in the pullulanase structural gene (pulA) were examined for their effects on pullulanase activity and cell surface localization in Escherichia coli carrying the cognate secretion genes from Klebsiella oxytoca. Of the 23 insertions, 11 abolished pullulanase activity but none were found to prevent secretion. To see whether more drastic changes affected secretion, we fused up to five reporter proteins (E. coli periplasmic alkaline phosphatase, E. coli periplasmic maltose-binding protein, periplasmic TEM beta-lactamase, Erwinia chrysanthemi extracellular endoglucanase Z, and Bacillus subtilis extracellular levansucrase) to three different positions in the pullulanase polypeptide: close to the N terminus of the mature protein, at the C terminus of the protein, or at the C terminus of a truncated pullulanase variant lacking the last 256 amino acids. Only 3 of the 13 different hybrids were efficiently secreted: 2 in which beta-lactamase was fused to the C terminus of full-length or truncated pullulanase and 1 in which maltose-binding protein was fused close to the N terminus of pullulanase. Affinity-purified endoglucanase-pullulanase and pullulanase-endoglucanase hybrids exhibited apparently normal levels of pullulanase activity, indicating that the conformation of the pullulanase segment of the hybrid had not been dramatically altered by the presence of the reporter. However, pullulanase-endoglucanase hybrids were secreted efficiently if the endoglucanase component comprised only the 60-amino-acid, C-terminal cellulose-binding domain, suggesting that at least one factor limiting hybrid protein secretion might be the size of the reporter.     

Role of the mature protein sequence of maltose-binding protein in its secretion across the E. coli cytoplasmic membrane

Secretion of amber fragments of an E. coli periplasmic protein, the maltose-binding protein, was studied to determine if the mature portion of the protein is required for its export across the cytoplasmic membrane. A fragment lacking 25–35 amino acid residues at the C terminus is secreted at normal levels, suggesting that this sequence is not required for secretion. This is in contrast to the results obtained with the periplasmic protein β-lactamase. In studying another fragment of one-third the molecular weight of the intact protein, we found that the majority of the fragment is not recovered from the periplasmic fraction. However, a small amount of secretion of this polypeptide was observed. This fragment is synthesized as a larger molecular weight form when cells are induced for the synthesis of a maltose-binding protein-β-galactosidase hybrid protein, which was previously shown to block the proper localization and processing of envelope proteins. This result is consistent with the idea that the larger form is a precursor with an unprocessed signal sequence, whereas in the absence of the hybrid protein the fragment is a processed mature form. Thus secretion of the smaller fragment may be occurring up to the point where the signal sequence is removed. That this fragment has passed through the cytoplasmic membrane is further supported by its accessibility to externally added trypsin. We suggest that the fragment may be secreted to the periplasm, but cannot assume a water-soluble conformation; the majority of the polypeptide may be associated with the external surface of the cytoplasmic membrane. Thus the mature sequence of maltose-binding protein, at least its C-terminal two thirds, may not be required for its export across the cytoplasmic membrane.     

Translational control of exported proteins in Escherichia coli

We recently described the suppression of export of a class of periplasmic proteins of Escherichia coli caused by overproduction of a C-terminal truncated periplasmic enzyme (GlpQ'). This truncated protein was not released into the periplasm but remained attached to the inner membrane and was accessible from the periplasm. The presence of GlpQ' in the membrane strongly reduced the appearance in the periplasm of some periplasmic proteins, including the maltose-binding protein (MBP), but did not affect outer membrane proteins, including the lambda receptor (LamB) (R. Hengge and W. Boos, J. Bacteriol., 162:972-978, 1985). To investigate this phenomenon further we examined the fate of MBP in comparison with the outer membrane protein LamB. We found that not only localization but also synthesis of MBP was impaired, indicating a coupling of translation and export. Synthesis and secretion of LamB were not affected. The possibility that this influence was exerted via the level of cyclic AMP could be excluded. Synthesis of MBP with altered signal sequences was also reduced, demonstrating that export-defective MBP which ultimately remains in the cytoplasm abortively enters the export pathway. When GlpQ' was expressed in a secA51(Ts) strain, the inhibition of MBP synthesis caused by GlpQ' was dominant over the precursor accumulation usually caused by secA51(Ts) at 41 degrees C. Therefore, GlpQ' acts before or at the level of recognition by SecA. For LamB the usual secA51(Ts) phenotype was observed. We propose a mechanism by which GlpQ' blocks an yet unknown membrane protein, the function of which is to couple translation and export of a subclass of periplasmic proteins.     

Analysis of factors affecting the periplasmic production of recombinant proteins in Escherichia coli

Five fusion proteins between Z domains derived from Staphylococcal Protein A and Green Fluorescent Protein or Human Proinsulin were produced on the periplasm of Escherichia coli. The effects of the molecular weight and amino acid composition of the translocated peptide, culture medium composition, and growth phase of the bacterial culture were analyzed regarding the expression and periplasmic secretion of the recombinant proteins. It was found that secretion was not affected by the size of the translocated peptide (17-42 kDa) and that the highest periplasmic production values were obtained on the exponential phase of growth. Moreover, the highest periplasmic values were obtained in minimal medium, showing the relevance of the culture medium composition on secretion. In silico prediction analysis suggested that with respect to the five proteins used in this study, those that are prone to form alpha-helix structures are more translocated to the periplasm.     

Hyperproduction of phosphate-binding protein, phoS, and pre-phoS proteins in Escherichia coli carrying a cloned phoS gene

A large amount of phosphate-binding protein, the phoS gene product, accumulated in the periplasmic space of the cells when an Escherichia coli strain carrying a multicopy plasmid containing a chromosomal fragment of the phoS-phoT region (pSN507) was grown in a low-phosphate medium. When the same strain carrying a plasmid containing only the phoS gene (pSN518 or pSN5182) was grown in low-phosphate medium, phosphate-binding protein accumulated in the periplasm, and in addition a larger protein accumulated in the non-periplasmic fraction. The apparent Mr of this protein and the phosphate-binding protein were 39000 and 35000 respectively, as judged by sodium dodecyl sulfate/polyacrylamide gel electrophoresis. This larger protein showed immunological cross-reaction with the phosphate-binding protein. The 39000-Mr protein was also detected in cells carrying pSN507 when the proteins were pulse-labeled with radioactive amino acids. From these findings, together with the fact that this protein is recovered from the membrane fraction, we conclude that this protein is an unsecreted precursor protein of the phosphate-binding protein. Kinetics and regulation of accumulation of these proteins were studied. This system will be useful for preparation and purification of the precursor protein for biochemical studies in relation to the mechanism of protein secretion.     

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.