Fusions of secreted proteins to alkaline phosphatase: an approach for studying protein secretion

Aeromonas spp. can secrete Escherichia coli alkaline phosphatase into the culture supernatant, and its release requires a functional general secretion pathway

Aerolysin is a channel-forming protein secreted by Aeromonas hydrophila. To determine if regions of aerolysin could direct the secretion of another protein, portions of aerA were fused to phoA, the Escherichia coli alkaline phosphatase gene and cloned into E. coli, Aeromonas salmonicida, and A. hydrophila. We were surprised to find that secretion of the enzyme by both Aeromonas spp. was independent of the aerolysin segments fused to it. The smallest fusion product contained only the signal sequence and two amino acids of aerolysin. The largest had more than 90% of the aerolysin molecule. The fusion proteins were found in the periplasms of E. coli and A. salmonicida grown in LB medium containing glucose, as well as in the shocked cells. Aerolysin itself was secreted by A. salmonicida under these conditions. In contrast, when A. salmonicida containing any of the fused genes was grown in LB medium without glucose, most of the alkaline phosphatase activity was extracellular, whereas beta-lactamase remained in its normal periplasmic location. Similar results were obtained with A. hydrophila. The change in location of the enzyme in A. salmonicida appeared to be related to the pH of the growth medium. A. salmonicida and A. hydrophila also secreted native E. coli alkaline phosphatase, but A. hydrophila strains with mutations in the general secretion pathway were unable to release the enzyme. We conclude that the Aeromonas secretion system can recognize the E. coli enzyme as an extracellular protein and direct it outside the cell.

Formation of several bacterial c-type cytochromes requires a novel membrane-anchored protein that faces the periplasm

We report here the discovery of a novel bacterial gene (cycH) whose product is involved in the biogenesis of most of the cellular cytochromes c. The cycH gene was detected in the course of characterizing a cytochrome oxidase-deficient Bradyrhizobium japonicum Tn5 mutant (strain COX3) in which the transposon insertion disrupted cycH. All of the c-type cytochromes detectable in aerobically grown B. japonicum wild-type cells were absent in the COX3 mutant, with the exception of cytochrome c1. A secondary phenotypic effect was the spectroscopic absence of the aa3-type cytochrome c oxidase. The nucleotide sequence of the cloned wild-type cycH gene predicted a membrane-bound 369-amino-acid protein with an M(r) of 39727. Results from studies on its membrane topology suggested that approximately 110 N-terminal amino acids are involved in anchoring the protein in the membrane, whereas the remaining two-thirds of the protein are exposed to the periplasm. We postulate that the CycH protein plays an essential role in an as yet unidentified periplasmic step in the biogenesis of holocytochromes c, except that of cytochrome c1.

Analysis of the structure and subcellular location of filamentous phage pIV

The gene IV protein of filamentous bacteriophages is an integral membrane protein required for phage assembly and export. A series of gene IV::phoA fusion, gene IV deletion, and gene IV missense mutations have been isolated and characterized. The alkaline phosphatase activity of the fusion proteins suggests that pIV lacks a cytoplasmic domain. Cell fractionation studies indicate that the carboxy-terminal half of pIV mediates its assembly into the membrane, although there is no single, discrete membrane localization domain. The properties of gene IV missense and deletion mutants, combined with an analysis of the similarities between pIVs from various filamentous phage and related bacterial export-mediating proteins, suggest that the amino-terminal half of pIV consists of a periplasmic substrate-binding domain that confers specificity to the assembly-export system.

Use of the "blue halo" assay in the identification of genes encoding exported proteins with cleavable signal peptides: cloning of a Borrelia burgdorferi plasmid gene with a signal peptide

We have recently reported a phoA expression vector, termed pMG, which, like TnphoA, is useful in identifying genes encoding membrane-spanning sequences or signal peptides. This cloning system has been modified to facilitate the distinction of outer membrane and periplasmic alkaline phosphatase (AP) fusion proteins from inner membrane AP fusion proteins by transforming pMG recombinants into Escherichia coli KS330, the strain utilized in the "blue halo" assay first described by Strauch and Beckwith (Proc. Natl. Acad. Sci. USA 85:1576-1580, 1988). The lipoprotein mutation lpp-5508 of KS330 results in an outer membrane that is leaky to macromolecules, and its degP4 mutation greatly reduces periplasmic proteolytic degradation of AP fusion proteins. pMG AP fusions containing cleavable signal peptides, including the E. coli periplasmic protein beta-lactamase, the E. coli and Chlamydia trachomatis outer membrane proteins OmpA and MOMP, respectively, and Tp 9, a Treponema pallidum AP recombinant, diffused through the leaky outer membrane of KS330 and resulted in blue colonies with blue halos. In contrast, inner membrane AP fusions derived from E. coli proteins, including leader peptidase, SecY, and the tetracycline resistance gene product, as well as Tp 70, a T. pallidum AP recombinant which does not contain a signal peptide, resulted in blue colonies without blue halos. Lipoprotein-AP fusions, including the Borrelia burgdorferi OspA and T. pallidum Tp 75 and TmpA showed halo formation, although there was significantly less halo formation than that produced by either periplasmic or outer membrane AP fusions. In addition, we applied this approach to screen recombinants constructed from a 9.0-kb plasmid isolated from the B31 virulent strain of B. burgdorferi. One of the blue halo colonies identified produced an AP fusion protein which contained a signal peptide with a leader peptidase I cleavage recognition site. The pMG/KS330r- cloning and screening approach can identify genes encoding proteins with cleavable signal peptides and therefore can serve as a first step in the identification of genes encoding potential virulence factors.

Bacillus subtilis PrsA is required in vivo as an extracytoplasmic chaperone for secretion of active enzymes synthesized either with or without pro-sequences

In prsA (protein secretion) mutants of Bacillus subtilis, decreased levels of exoproteins, including alpha-amylase and subtilisins, are found extracellularly. The effect of prsA on subtilisin secretion is elaborated here. Extracytoplasmic folding and secretion of active subtilisin is assisted by the N-terminal pro-sequence of its precursor. In this paper we present evidence that the product of the prsA gene is additionally required for these processes in vivo. We examined inducible expression of different subtilisin-alkaline phosphatase fusion genes in the prsA3 mutant. We found massive degradation of the fusion proteins, and a lack of enzymatic activity in the protein secreted. We suggest that PrsA is a novel chaperone with a predicted extracytoplasmic location, and is important in vivo for the proper conformation of various exoproteins, including those with pro-sequence (like subtilisin) and those without (like alpha-amylase).

Highly selective binding of nascent polypeptides by an Escherichia coli chaperone protein in vivo

Chaperone proteins bind to newly synthesized polypeptides and assist in various assembly reactions. The Escherichia coli chaperone protein SecB binds precursors of exported proteins and assists in export. In vitro, SecB can bind to many unfolded proteins. In this report, we demonstrate that SecB binding in vivo is highly selective; the major polypeptides that are bound by SecB are nascent precursors of the exported proteins maltose-binding protein (MBP), LamB, OmpF, and OmpA. These results support the hypothesis that the primary physiological function of SecB is to stimulate protein export. By interacting with nascent polypeptides, SecB probably stimulates their cotranslational association with the membrane-bound protein translocation apparatus.

The topological analysis of integral cytoplasmic membrane proteins

We review three general approaches to determining the topology of integral cytoplasmic membrane proteins. (i) Inspection of the amino acid sequence and use of algorithms to predict membrane spanning segments allows the construction of topological models. For many proteins, the mere identification of such segments and an analysis of the distribution of basic amino acids in hydrophilic domains leads to correct structure predictions. For others, additional factors must come into play in determining topology. (ii) Gene fusion analysis of membrane proteins, in many cases, leads to complete topological models. Such analyses have been carried out in both bacteria and in the yeast Saccharomyces cerevisiae. Conflicts between results from gene fusion analysis and other approaches can be used to explore details of the process of membrane protein assembly. For instance, anomalies in gene fusion studies contributed evidence for the important role of basic amino acids in determining topology. (iii) Biochemical probes and the site of natural biochemical modifications of membrane proteins give information on their topology. Chemical modifiers, proteases and antibodies made to different domains of a membrane protein can identify which segments of the protein are in the cytoplasm and which are on the extracytoplasmic side of the membrane. Sites of such modifications as glycosylation and phosphorylation help to specify the location of particular hydrophilic domains. The advantages and limitations of these methods are discussed.

Membrane topology model of Escherichia coli alpha-ketoglutarate permease by phoA fusion analysis

Escherichia coli alpha-ketoglutarate permease (KgtP) is a 432-amino-acid protein that symports alpha-ketoglutarate and protons. KgtP was predicted to contain 12 membrane-spanning domains on the basis of a calculated hydropathy profile. The membrane topology model of KgtP was analyzed by using kgtP-phoA gene fusions and measuring alkaline phosphatase activities in cells expressing the chimeric proteins. Comparisons of the phosphatase activity levels and the locations of the KgtP-PhoA junctions are consistent with the predicted membrane topology model of KgtP.

TnphoA and TnphoA' elements for making and switching fusions for study of transcription, translation, and cell surface localization

We describe a set of elements based on the transposon TnphoA for making transcriptional fusions to the lacZ gene and for making translational fusions to the phoA or lacZ structural gene. Each element can be switched, one for another, by homologous recombination, thereby allowing testing for transcription, translation, or cell surface localization determinants at the same site within a gene. We describe three kinds of elements for making each fusion type. Two kinds are transposition proficient (Tnp+): one encodes kanamycin resistance, and the other encodes tetracycline resistance. The third kind is transposition defective (Tnp-) and encodes kanamycin resistance. In addition, we describe one Tnp- element that has no reporter gene and encodes chloramphenicol resistance; this element is used primarily as a tool to aid in switching fusions. Switching is efficient because each element has in common 254 bp of DNA at the phoA end and 187 bp (or more) of DNA at the IS50R end of TnphoA, and switching is straightforward because individual elements encode different drug resistances. Thus, switched recombinants can be selected as drug-resistant transductants, and they can be recognized as ones that have lost the parental drug resistance and fusion phenotype. Further, switching Tnp+ elements to Tnp- elements reduces problems due to transposition that can arise in P1 crosses or cloning experiments. Some TnphoA and TnphoA' elements cause polar mutations, while others provide an outward promoter for downstream transcription. This feature is especially useful in the determination of operon structures. Strategies for the use of TnphoA and TnphoA' elements in gene analysis are also described.

PhoA gene fusions in Legionella pneumophila generated in vivo using a new transposon, MudphoA

To enable effective use of phoA gene fusions in Legionella pneumophila, we constructed MudphoA, a derivative of the mini-Mu phage Mu dII4041, which is capable of generating gene fusions to the Escherichia coli alkaline phosphatase gene (EC 3.1.3.1). Although an existing fusion-generating transposon, TnphoA, has been a useful tool for studying secreted proteins in other bacteria, this transposon and other Tn5 derivatives transpose inefficiently in Legionella pneumophila, necessitating the construction of a more effective vector for use in this pathogen. Using MudphoA we generated fusions to an E. coli gene encoding a periplasmic protein and to an L. pneumophila gene encoding an outer membrane protein; both sets of fusions resulted in alkaline phosphatase activity. We have begun to use MudphoA to mutate secreted proteins of L. pneumophila specifically, since this subset of bacterial proteins is most likely to be involved in host-bacterial interactions. This modified transposon may be useful for studies of other bacteria that support transposition of Mu, but not Tn5, derivatives.

The Rex system of bacteriophage lambda: tolerance and altruistic cell death

The rexA and rexB genes of bacteriophage lambda encode a two-component system that aborts lytic growth of bacterial viruses. Rex exclusion is characterized by termination of macromolecular synthesis, loss of active transport, the hydrolysis of ATP, and cell death. By analogy to colicins E1 and K, these results can be explained by depolarization of the cytoplasmic membrane. We have fractionated cells to determine the intracellular location of the RexB protein and made RexB-alkaline phosphatase fusions to analyze its membrane topology. The RexB protein appears to be a polytopic transmembrane protein. We suggest that RexB proteins form ion channels that, in response to lytic growth of bacteriophages, depolarize the cytoplasmic membrane. The Rex system requires a mechanism to prevent lambda itself from being excluded during lytic growth. We have determined that overexpression of RexB in lambda lysogens prevents the exclusion of both T4 rII mutants and lambda ren mutants. We suspect that overexpression of RexB is the basis for preventing self-exclusion following the induction of a lambda lysogen and that RexB overexpression is accomplished through transcriptional regulation.

Molecular cloning, nucleotide sequence, and characterization of a 40,000-molecular-weight lipoprotein of Haemophilus somnus

A gene of Haemophilus somnus encoding the major 40,000-molecular-weight antigen (LppA) was cloned on a 2-kb Sau3AI fragment. The nucleotide sequence of the entire DNA insert was determined. One open reading frame, encoding a 247-residue polypeptide with a calculated molecular weight of 27,072, was identified. This reading frame was confirmed by sequencing the fusion joint of two independent IppA::TnphoA gene fusions. The 21 amino-terminal amino acids of the deduced polypeptide showed strong sequence homology to the signal peptide of secreted proteins, and the sequence Leu-Leu-Ala-Ala-Cys at the putative cleavage site is identical to the consensus cleavage sequence of lipoproteins from gram-negative bacteria. The presence of the lipid moiety on the protein was shown by incorporation of radioactive palmitic acid into the natural H. somnus protein. Palmitic acid could also be incorporated into the recombinant protein in Escherichia coli. Synthesis of the mature LppA lipoprotein was inhibited by globomycin, showing that cleavage of the signal peptide is mediated by signal peptidase II in both organisms. By using site-directed mutagenesis, the cysteine residue at the cleavage site was changed to glycine. Radiolabelled palmitate was not incorporated into the mutated protein, showing that lipid modification occurs at the Cys-22 residue.

Bacterial cytochromes c biogenesis

We report the primary sequence analyses of two loci, hel and ccl, whose gene products are required specifically for the biogenesis of c-type cytochromes in the Gram-negative photosynthetic bacterium Rhodobacter capsulatus. Genetic and molecular analyses show that the hel locus contains at least four genes, helA, helB, helC, and orf52, and the ccl locus contains two genes, ccl1 and ccl2, that are essential for cytochromes c biogenesis. HelA is homologous to a class of proteins called ABC transporters and helA, helB, and helC are proposed to encode an export complex. Cytochrome c2-alkaline phosphatase gene fusions were used to show that apocytochrome c2 synthesis and secretion are not affected by the hel and ccl defects. Ccl1 and Ccl2 possess typical signal sequences to direct them to the periplasm. The periplasmic orientation of Ccl1 was confirmed using a Ccl1-alkaline phosphatase gene fusion. The Ccl1-alkaline phosphatase gene fusion analysis also demonstrated that Ccl1 does not require hel genes for its synthesis and secretion. Ccl1 is homologous to proteins encoded by chloroplast and mitochondrial genes, suggesting analogous functions in these organelles. Taken together, these results support the hypothesis that the hel-encoded proteins are required for the export of heme to the periplasm where it is subsequently ligated to the c-type apocytochromes.

Role of the two-component leader sequence and mature amino acid sequences in extracellular export of endoglucanase EGL from Pseudomonas solanacearum

The egl gene of Pseudomonas solanacearum encodes a 43-kDa extracellular endoglucanase (mEGL) involved in wilt disease caused by this phytopathogen. Egl is initially translated with a 45-residue, two-part leader sequence. The first 19 residues are apparently removed by signal peptidase II during export of Egl across the inner membrane (IM); the remaining residues of the leader sequence (modified with palmitate) are removed during export across the outer membrane (OM). Localization of Egl-PhoA fusion proteins showed that the first 26 residues of the Egl leader sequence are required and sufficient to direct lipid modification, processing, and export of Egl or PhoA across the IM but not the OM. Fusions of the complete 45-residue leader sequence or of the leader and increasing portions of mEgl sequences to PhoA did not cause its export across the OM. In-frame deletion of portions of mEGL-coding sequences blocked export of the truncated polypeptides across the OM without affecting export across the IM. These results indicate that the first part of the leader sequence functions independently to direct export of Egl across the IM while the second part and sequences and structures in mEGL are involved in export across the OM. Computer analysis of the mEgl amino acid sequence obtained from its nucleotide sequence identified a region of mEGL similar in amino acid sequence to regions in other prokaryotic endoglucanases.

Escherichia coli alkaline phosphatase fails to acquire disulfide bonds when retained in the cytoplasm

The cysteines of the Escherichia coli periplasmic enzyme alkaline phosphatase, which are involved in disulfide bonds in the native enzyme, were found to be fully reduced when the protein was retained in the cytoplasm. Under these circumstances the cysteines remained reduced for at least several minutes after the synthesis of the protein was completed. This contrasted with the normally exported protein, wherein disulfide bonds formed rapidly. Disulfide bond formation accompanied export and processing. The implications of these findings for the inactivity of the enzyme in the cytoplasm are discussed.

Membrane topology analysis of Escherichia coli mannitol permease by using a nested-deletion method to create mtlA-phoA fusions

The Escherichia coli mannitol permease catalyzes the concomitant transport and phosphorylation of D-mannitol. This 68-kDa protein consists of a membrane-bound, N-terminal domain involved in mannitol binding and translocation and a C-terminal, cytoplasmic domain responsible for mannitol phosphorylation. Secondary-structure prediction methods suggest that the N-terminal half of the permease spans the membrane approximately seven times in alpha-helical segments, but these data cannot conclusively predict the structure. We have used gene fusions between mtlA (encoding the permease) and 'phoA (encoding alkaline phosphatase lacking its signal sequence) to further investigate the topology of the mannitol permease. Initially, fusions were constructed by using a lambda TnphoA vector and in vitro cloning of 'phoA into naturally occurring restriction sites in mtlA. However, the former method gave severe problems with insertion "hot-spots" in our vector systems, and the latter method was limited by the number of useful restriction sites. Therefore, we developed a nested-deletion method for creating mtlA-phoA fusions. 'phoA was first cloned downstream from the part of mtlA encoding the membrane-bound half of the permease. This construct was then treated with the appropriate restriction enzymes and with exonuclease III to create random fusions. An analysis of greater than 40 different fusion clones constructed by these methods provides strong evidence for six membrane-spanning regions in the mannitol permease with three relatively short periplasmic loops and two large cytoplasmic loops in the membrane-bound half of the protein.

Identification of Treponema pallidum subspecies pallidum genes encoding signal peptides and membrane-spanning sequences using a novel alkaline phosphatase expression vector

Treponema pallidum subspecies pallidum is a pathogenic spirochaete for which there are no systems of genetic exchange. In order to provide a system for the identification of T. pallidum surface proteins and potential virulence factors, we have developed a novel expression vector which confers the utility of TnphoA transposition. The relevant features of this plasmid vector, termed pMG, include an inducible tac promoter, a polylinker with multiple cloning sites in three reading frames, and an alkaline phosphatase (AP) gene lacking the signal sequence-encoding region. Library construction with Sau3A-digested T. pallidum genomic DNA resulted in the creation of functional T. pallidum-AP fusion proteins. Analysis of fusion proteins and their corresponding DNA and deduced amino acid sequences demonstrated that they could be grouped into three categories: (i) those with signal peptides containing leader peptidase I cleavage sites, (ii) those with signal peptides containing leader peptidase II cleavage sites, and (iii) those with non-cleavable hydrophobic membrane-spanning sequences. Triton X-114 detergent phase partitioning of individual T. pallidum-AP fusions revealed several clones whose AP activity partitioned preferentially into the hydrophobic detergent phase. Several of these fusion proteins were subsequently shown to be acylated by Escherichia coli following [3H]-palmitate labelling, indicating their lipoproteinaceous nature. DNA and amino acid sequence analysis of one acylated fusion protein, Tp75, confirmed the presence of a hydrophobic N-terminal signal sequence containing a consensus leader peptidase II recognition site. The DNA sequence of Tp75 also indicates that this is a previously unreported T. pallidum lipoprotein. T. pallidum-AP fusion proteins which partitioned into the hydrophobic detergent phase but did not incorporate palmitate were also identified. DNA and amino acid analysis of one such clone, Tp70, showed no cleavable signal but had a significant hydrophobic region of approximately 20 residues, consistent with a membrane-spanning domain. Immunoblot analysis of T. pallidum-AP fusions detected with a monoclonal antibody specific for AP identified several fusion proteins which migrated as doublets separated in apparent electrophoretic mobility by no more than 3 kDa. [35S]-methionine pulse-chase incorporation showed that the doublet AP fusions represented precursor and processed forms of the same protein. DNA and amino acid sequence analysis of clones expressing processed fusion proteins demonstrated hydrophobic N-terminal signal sequences containing consensus leader peptidase I recognition sites.

Topographic analysis of the toxic Gef protein from Escherichia coli

The chromosomal gef gene of Escherichia coli is a member of the gef gene family which encodes strongly toxic proteins of about 50 amino acids. We demonstrate here that the Gef protein is detectable by anti-peptide antibodies. Furthermore, we show that Gef is anchored in the cytoplasmic membrane by the N-terminal part of the protein, and that the C-terminal part is localized in the periplasm in a dimeric form with at least one disulphide bond. By mutagenesis of gef it is shown that the periplasmic portion of Gef encodes the toxic domain and that the dimerization of Gef is not essential for the toxic effect.

Characterization of trbC, a new F plasmid tra operon gene that is essential to conjugative transfer

We have characterized a previously unidentified gene, trbC, which is contained in the transfer region of the Escherichia coli K-12 fertility factor, F. Our data show that the trbC gene is located between the F plasmid genes traU and traN. The product of trbC was identified as a polypeptide with an apparent molecular weight (Ma) of 23,500 that is processed to an Ma-21,500 mature protein. When ethanol was present, the Ma-23,500 polypeptide accumulated; the removal of ethanol resulted in the appearance of the processed mature protein. Subcellular fractionation experiments demonstrated that the processed, Ma-21,500 mature protein was located in the periplasm. DNA sequence analysis showed that trbC encodes a 212-amino-acid Mr-23,432 polypeptide that could be processed to a 191-amino-acid Mr-21,225 mature protein through the removal of a typical amino-terminal signal sequence. We also constructed two different Kmr gene insertion mutations in trbC and crossed these onto the transmissible F plasmid derivative pOX38. We found that cells carrying pOX38 trbC mutant plasmids were transfer deficient and resistant to infection by F-pilus-specific phages. Transfer proficiency and bacteriophage sensitivity were restored by complementation when a trbC+ plasmid clone was introduced into these cells. These results showed that trbC function is essential to the F plasmid conjugative transfer system and suggested that the TrbC protein participates in F-pilus assembly.

Use of alkaline phosphatase fusions to study protein secretion in Bacillus subtilis

We have constructed a vector designed to facilitate the study of protein secretion in Bacillus subtilis. This vector is based on a translational fusion between the expression elements and signal sequence of Bacillus amyloliquefaciens alkaline protease and the mature coding sequence for Escherichia coli alkaline phosphatase (phoA). We show that export of alkaline phosphatase from B. subtilis depends on a functional signal sequence and that alkaline phosphatase activity depends upon secretion. The vector design facilitates the insertion of heterologous coding sequences between the signal and phoA to generate three-part translational fusions. Such phoA fusions are easily analyzed by monitoring alkaline phosphatase activity on agar plates or in culture supernatants or by immunological detection. Exploitation of this methodology, which has proven to be extremely useful in the study of protein secretion in E. coli, has a variety of applications for studying protein secretion in B. subtilis.

Sequence and TnphoA analysis of a Mycoplasma hyorhinis protein with membrane export function

Proteins translocated across the single plasma membrane of mycoplasmas (class Mollicutes) represent important components likely to affect several interactions of these wall-less microbes with their respective hosts. However, identification and functional analysis of such proteins is hampered by the lack of mutational systems in mycoplasmas and by a perceived limitation in translating recombinant mycoplasma genes containing UGA (Trp) codons in other eubacteria. Here we directly analyze a gene encoding a Mycoplasma hyorhinis protein capable of promoting its membrane translocation. It was initially detected by screening a recombinant phage genomic library with antibody from a host with M. hyorhinis-induced arthritis and was localized by Tn5 and deletion mutations affecting expression of antigenic translational products. Sequence analysis of the isolated gene predicted a hydrophilic protein, P101, containing three UGA codons and a putative signal peptide with an uncharacteristic cluster of positively charged amino acids near its C terminus. Nevertheless, lambda::TnphoA transposon mutagenesis of an Escherichia coli plasmid bearing the p101 gene resulted in p101::TnphoA fusions expressing products that could translocate as much as 48 kDa of the P101 sequence (up to the first UGA codon) across the E. coli plasma membrane. Fusion proteins containing mature P101 sequences expressed mycoplasma epitopes and were found by cell fractionation and detergent phase partitioning to be integral membrane proteins in E. coli, suggesting a lack of signal peptide cleavage in this system. Importantly, identification of P101 by direct analysis of its export function relied neither on prior identification of the mycoplasmal product nor on complete expression of the product from the cloned mycoplasma gene.

Secretion of hybrid proteins by the Yersinia Yop export system

After incubation at 37 degrees C in the absence of Ca2+ ions, pathogenic strains of Yersinia spp. release large amounts of a set of plasmid-encoded proteins called Yops. The secretion of these proteins, involved in pathogenicity, occurs via a mechanism that involves neither the removal of a signal sequence nor the recognition of a C-terminal domain. Analysis of deletion mutants allowed the secretion recognition domain to be localized within the 48 N-terminal amino acids of protein YopH, within the 98 N-terminal residues of protein YopE, and within the 76 N-terminal residues of YopQ. Comparison of these regions failed to reveal any sequence similarity, suggesting that the secretion signal of Yop proteins is conformational rather than sequential. Hybrid proteins containing the amino-terminal part of YopH fused to either the alpha-peptide of beta-galactosidase or to alkaline phosphatase deprived of its signal sequence were efficiently secreted to the Yersinia culture medium. This observation opens new prospects in using Yersinia spp. as chimeric-protein producers and as potential live carriers for foreign antigens.

An O antigen can interfere with the function of the Yersinia pseudotuberculosis invasin protein

Escherichia coli strains harbouring the Yersinia pseudotuberculosis inv gene are able to enter cultured mammalial cells. We show here that this property is not shared by all enteric bacteria, since Shigella flexneri 2a cured of its virulence-associated plasmid and harbouring the inv gene is unable to enter mammalian cells efficiently. Mapping studies showed that the region of the chromosome responsible for this phenotype includes rfaB, a locus involved in the production of O antigen. S. flexneri 2a strains that express O antigen were unable to enter mammalian cells, even though invasin was efficiently expressed and localized, showing that this structure interferes with invasin activity. The O antigen either masks invasin or sterically hinders the ability of the mammalian cell receptor to bind this protein.

Activity of the nicotinamide mononucleotide transport system is regulated in Salmonella typhimurium

Transport of nicotinamide mononucleotide (NMN) requires two functions, NadI(T) and PnuC. The PnuC protein is membrane associated, as judged by isolation of active TnphoA gene fusions and demonstration that the fusion protein is membrane associated. The PnuC function appears to be the major component of the transport system, since mutant alleles of the pnuC gene permit NMN transport in the absence of NadI(T) function. We present evidence that the activity of the NMN transport system varies in response to internal pyridine levels (presumably NAD). This control mechanism requires NadI(T) function, which is provided by a bifunctional protein encoded by the nadI gene (called nadR by Foster and co-workers [J. W. Foster, Y. K. Park, T. Fenger, and M. P. Spector, J. Bacteriol. 172:4187-4196]). The nadI protein regulates transcription of the nadA and nadB biosynthetic genes and modulates activity of the NMN permease; both regulatory activities respond to the internal pyridine nucleotide level.

Transposon tagging of genes for cell-cell interactions in Myxococcus xanthus

The prokaryote Myxococcus xanthus is a model for cell interactions important in multicellular behavior. We used the transposon TnphoA to specifically identify genes for cell-surface factors involved in cell interactions. From a library of 10,700 insertions of TnphoA, we isolated 36 that produced alkaline phosphatase activity. Three TnphoA insertions tagged cell motility genes, called cgl, which control the adventurous movement of cells. The products of the tagged cgl genes could function in trans upon other cells and were localized primarily in the cell envelope and extracellular space, consistent with TnphoA tagging genes for extracellular factors controlling motility.

Construction of a bifunctional Escherichia coli heat-stable enterotoxin (STb)-alkaline phosphatase fusion protein

A fusion between the genes encoding the Escherichia coli STb heat-stable enterotoxin (estB) and alkaline phosphatase (phoA) was constructed, and the expressed protein product was characterized. The STb-alkaline phosphatase protein (STb-PhoA) had an apparent molecular mass of 50,000 daltons and was detected with both monoclonal anti-alkaline phosphatase and polyclonal anti-STb antibodies. Expression of the gene fusion resulted in high-level production of alkaline phosphatase activity, indicating that STb-PhoA was processed and exported into the periplasm of the E. coli host strain. Amino acid sequence analysis of the hybrid protein yielded the sequence Ser-Thr-Gln-Ser-Asn-Lys-Lys, indicating that STb-PhoA was processed during export in a fashion identical to that of native STb (Y. M. Kupersztoch, K. Tachias, C. R. Moomaw, L. A. Dreyfus, R. G. Urban, C. Slaughter, and S. Whipp, J. Bacteriol. 172: 2427-2432, 1990). STb-PhoA was purified from an expressed bacterial lysate by preparative isoelectric focusing. In a rat ligated intestinal loop model, purified STb-PhoA induced highly significant (P less than 0.002) fluid secretion. In addition, the specific activity of STb-PhoA was nearly identical to that of purified STb. Thus, the STb-PhoA hybrid protein represents a readily obtainable source of biologically active (STb) enterotoxin that may prove useful in studies to determine the mode of toxin action.

Expression of recombinant proteins in Escherichia coli using an oxygen-responsive promoter

The oxygen-dependent promoter of the Vitreoscilla hemoglobin (VHb) gene has been shown to be functional in E. coli. Earlier studies established that the promoter is maximally induced under microaerobic conditions and that its activity is also influenced by the cAMP-CAP complex. We demonstrate here that the promoter can be used for regulated, high-level expression of recombinant proteins in two-stage fed-batch fermentations. The promoter is maximally induced at dissolved oxygen levels lower than 5% air saturation. Despite the influence of catabolite repression, glucose and glycerol-containing media give comparable product levels under carbon-limited conditions such as those encountered in typical fed-batch fermentations. The possibility of a third level of control of promoter activity is also indicated. This mode of induction can be repressed by addition of a complex nitrogen source such as yeast extract to the medium. The observed promoter activity can be modulated at least 30-fold over the course of high-cell density fermentations producing either cloned beta-galactosidase or cloned chloramphenicol acetyltransferase (CAT). Densitometer scanning of SDS-polyacrylamide gels revealed that beta-galactosidase was expressed to a level of approximately 10% of total cellular protein.

Genetic analysis of membrane protein topology by a sandwich gene fusion approach

We describe a cloning vector that allows the construction of phoA sandwich fusions in which mature alkaline phosphatase is inserted into target proteins. In contrast to previous fusions obtained using the TnphoA transposon, the entire amino acid sequence of the target protein is present in the fusion product. We have constructed a series of sandwich fusions of alkaline phosphatase to the multispanning cytoplasmic membrane protein MalF. Despite the fact that the alkaline phosphatase was tethered to MalF at both its N and its C terminus, the enzyme exhibited high activity when it was fused to a periplasmic domain of the membrane protein. Cells harboring an alkaline phosphatase sandwich fusion to the end of the first membrane-spanning segment of MalF exhibited both MalF and alkaline phosphatase activity. When alkaline phosphatase was inserted into a cytoplasmic domain of MalF, its specific activity was very low. Our results suggest that the alkaline phosphatase activity of phoA sandwich fusions provides a more sensitive monitor than previous methods of the cellular localization of the domain of the target protein to which the enzyme is fused. Thus, the sandwich fusion approach can give a more accurate picture of membrane protein topology.

DNA sequence analysis of pglA and mechanism of export of its polygalacturonase product from Pseudomonas solanacearum

The pglA gene encodes a 52-kilodalton extracellular polygalacturonase (PGA) which is associated with the phytopathogenic virulence of Pseudomonas solanacearum. The nucleotide sequence of pglA and the putative amino acid sequence of the PGA protein were determined. A computer search identified a 150-residue region of PGA which was similar (41%) to the amino acid sequence of a region of the PG-2A polygalacturonase from tomato. Comparison of the amino terminus of the pglA open reading frame with the actual amino-terminal sequence of purified extracellular PGA suggested that pglA is initially translated as a higher-molecular-mass precursor with a 21-residue amino-terminal signal sequence. Localization of various pglA-phoA fusion proteins in Escherichia coli and P. solanacearum indicated that the 21-residue leader sequence directs the export of PhoA only as far as the periplasm of both bacteria. Deletion of the last 13 residues of PGA eliminated its catalytic activity, as well as its ability to be exported outside of the P. solanacearum cell. Our results suggest that PGA excretion occurs in two steps. The first step involves a signal sequence cleavage mechanism similar to that used for periplasmic proteins and results in export of PGA across the inner membrane; the second step (transit of the outer membrane) occurs by an unknown mechanism requiring sequences from the mature PGA protein and biochemical factors absent from E. coli.

The normally periplasmic enzyme beta-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 phoA structural gene mutation that conditionally affects formation of the enzyme bacterial alkaline phosphatase

The phoA503 mutant was identified as a mutant that shows a novel phoA regulatory phenotype. The phoA503 allele dramatically reduces the synthesis of bacterial alkaline phosphatase activity during Pi starvation in an otherwise wild-type host and during the logarithmic growth phase in a phoR or phoU background. Near-normal amounts of enzyme activity are found in phoR phoA503 or phoU phoA503 mutants when starved for carbon, nitrogen, or sulfur or during the stationary phase, however. Marker rescue and DNA sequence analysis located the phoA503 mutation to the phoA coding region. It is a C-to-T transition that would cause a substitution of Val for Ala-22 in the mature protein. Transcriptional and translational lacZ fusions to both wild-type and mutant alleles demonstrated that phoA gene expression is unaltered. Also, the mutant protein was secreted and processed as efficiently as the wild type. Furthermore, the subunits appeared to dimerize and to be stable in the periplasm. But, greater than 98% of the dimers were inactive and found exclusively as isozyme 1. An activation of preformed phoA503 dimers occurred during the stationary phase with the concomitant conversion into isozymes 2 and 3. We propose that the phoA503 mutation affects a late stage in the formation of active enzyme. An unknown change when Pi is present during stationary-phase growth leads to formation of active dimers, which is responsible for this new conditional phenotype.

Alkaline, acid, and neutral phosphatase activities are induced during development in Myxococcus xanthus

One of the signals that has been reported to be important in stimulating fruiting body formation of Myxococcus xanthus is starvation for phosphate. We therefore chose to study phosphatase activity during M. xanthus development. Many phosphatases can cleave the substrate p-nitrophenol phosphate. Using this substrate in buffers at various pHs, we obtained a profile of phosphatase activities during development and germination of M. xanthus. These experiments indicated that there are five patterns of phosphatase activity in M. xanthus: two vegetative and three developmental. The two uniquely vegetative activities have pH optima at 7.2 and 8.5. Both require magnesium and both are inhibited by the reducing agent dithiothreitol. The developmental (spores) patterns of activity have pH optima of 5.2, 7.2, and 8.5. All three activities are Mg independent. Only the alkaline phosphatase activity is inhibited by dithiothreitol. The acid phosphatase activity is induced very early in development, within the first 2 to 4 h. Both the neutral and alkaline phosphatase Mg-independent activities are induced much later, about the time that myxospores become evident (24 to 30 h). The three activities are greatly diminished upon germination; however, the kinetics of loss differ for all three. The acid phosphatase activity declines very rapidly, the neutral activity begins to decline only after spores begin to convert to rods, and the alkaline phosphatase activity remains high until the time the cells begin to divide. All three developmental activities were measured in the developmental signalling mutants carrying asg, csg, and dsg. The pattern of expression obtained in the mutants was consistent with that of other developmentally regulated genes which exhibit similar patterns of expression during development. The ease with which phosphatases can be assayed should make the activities described in this report useful biochemical markers of stages of both fruiting body formation and germination.

A lipopeptide-encoding sequence upstream from the lysA gene of Pseudomonas aeruginosa

An open reading frame (ORF) of 141 bp was observed upstream from the Pseudomonas aeruginosa lysA gene. The translation product of this ORF contains a signal peptide with a lipoprotein box, Ile-Ala-Ala-Cys, at the predicted signal peptidase cleavage site. The Escherichia coli phoA gene without its signal sequence was fused in frame to this ORF in a broad host-range plasmid. The resulting construct expressed a hybrid protein exhibiting alkaline phosphatase activity in phoA mutants of both E. coli and P. aeruginosa. This indicates that the ORF encodes a peptide, part of which acts as an export signal. The hybrid peptide was identified by immunoblotting with alkaline phosphatase antiserum. The accumulation of a precursor form was observed when P. aeruginosa cells carrying this gene fusion on a plasmid were treated with globomycin. Moreover, the mature form could be labelled with 2-[3H]-glycerol, indicating that lipidic residues may be linked to the hybrid protein. Taken together, these results strongly suggest that the ORF encodes a lipopeptide. We propose that the gene is called IppL.

Analysis of protein localization by use of gene fusions with complementary properties

This report describes a new transposon designed to facilitate the combined use of beta-galactosidase and alkaline phosphatase gene fusions in the analysis of protein localization. The transposon, called TnlacZ, is a Tn5 derivative that permits the generation of gene fusions encoding hybrid proteins carrying beta-galactosidase at their C termini. In tests with plasmids, TnlacZ insertions that led to high cellular beta-galactosidase activity were restricted to sequences encoding either cytoplasmic proteins or cytoplasmic segments of a membrane protein. The fusion characteristics of TnlacZ are thus complementary to those of TnphoA, a transposon able to generate alkaline phosphatase fusions whose high-activity insertion sites generally correspond to periplasmic sequences. The structure of TnlacZ allows the conversion of a TnlacZ fusion into the corresponding TnphoA fusion (and vice versa) through recombination or in vitro manipulation in a process called fusion switching. Fusion switching was used to generate the following two types of fusions with unusual properties: a low-specific-activity beta-galactosidase-alkaline phosphatase gene fusion and two toxic periplasmic-domain serine chemoreceptor-beta-galactosidase gene fusions. The generation of both beta-galactosidase and alkaline phosphatase fusions at exactly the same site in a protein permits a comparison of the two enzyme activities in evaluating the subcellular location of the site, such as in studies of membrane protein topology. In addition, fusion switching makes it possible to generate gene fusions whose properties should facilitate the isolation of mutants defective in the export or membrane anchoring of different cell envelope proteins.

Molecular analysis of the Haemophilus influenzae type b pilin gene

A Haemophilus influenzae DNA library was prepared in the vector lambda EMBL3, and recombinant phage were screened for the pilin gene (pil) using a synthetic oligonucleotide. Southern blot analysis of the positive clones revealed a 2.5kb PstI/PvuI fragment that hybridized with the oligonucleotide probe. This fragment was subcloned into pBR322 and sequenced. The nucleotide sequence disclosed an open reading frame of 653 bases. The deduced amino acid sequence corresponded with the known amino acid sequence of the purified pilin protein. Primer extension analysis using total RNA from piliated H. influenzae cells delineated a start site for the gene, -10 and -35 promoter regions, and a ribosome-binding site. No transcripts were seen with the RNA derived from a non-piliated strain. Southern blots of DNA from a number of H. influenzae strains revealed homology with the pil structural gene. DNA from a non-piliated strain of H. influenzae also hybridized with the pil probe. Transcriptional and translational studies were performed in Escherichia coli with plasmids containing: (i) the pil gene on the 2.5 kb PstI/PvuI fragment, (ii) the pil gene fused to the phoA gene, and (iii) the pil gene present on a 12.2 kb insert containing extensive H. influenzae DNA flanking the pil gene. The results suggest that the H. influenzae pil gene is expressed in Escherichia coli, but from a promoter other than the one used in H. influenzae.

Alkaline phosphatase fusions: sensors of subcellular location

Alkaline phosphatase fusions allow genes to be identified solely on the basis of their protein products being exported from the cytoplasm. Thus, the use of such fusions helps render biological processes which involve cell envelope and secreted proteins accessible to a sophisticated genetic analysis. Furthermore, alkaline phosphatase fusions can be used to locate export signals. Specifying such signals is an important component of studies on the structure of individual cell envelope proteins. The basis of the alkaline phosphatase fusion approach is the finding that the activity of the enzyme responds differently to different environments. Thus, the activity of the fusion protein gives evidence as to its location. This general approach of using sensor proteins which vary in their function, depending on their environment, could be extended to the study of other sorts of problems. It may be that certain enzymes will provide an assay for localization to a particular subcellular compartment, if the environment of the compartment differs from that of others. For instance, the lysosome is more acidic than other intracellular organelles. A gene fusion system employing a reporter enzyme that could show activity only at the pH of the lysosome could allow the detection of signals determining lysosomal localization. Analogous types of enzymes may be used as probes for other subcellular compartments.

Molecular characterization of pulA and its product, pullulanase, a secreted enzyme of Klebsiella pneumoniae UNF5023

The determined nucleotide sequence of the Klebsiella pneumoniae UNF5023 gene pulA comprises a single open reading frame coding for a 1090-residue precursor of the secreted protein pullulanase. The predicted sequence of this protein is highly homologous to that of pullulanase of Klebsiella aerogenes strain W70. However, the UNF5023 pullulanase lacks a collagen-like sequence present at the N-terminus of the mature W70 enzyme and differs further from the W70 pullulanase around residue 300 and at the C-terminus. Pullulanases with or without the collagen-like sequence could not be separated by gel electrophoresis under denaturing or non-denaturing conditions, and were unaffected by collagenase. A large central domain which is highly conserved in both UNF5023 and W70 polypeptides contains eight short sequences that are also found in amylases and iso-amylases. Linker mutations in the region of the UNF5023 pulA gene coding for this domain abolished catalytic activity without affecting transport of the polypeptide across the outer membrane. Hybrid proteins comprising at least the amino-terminal 656 residues of prepullulanase fused to alkaline phosphatase were partially localized to the cell surface, as judged by their accessibility to anti-pullulanase serum in immuno-fluorescence tests. On the basis of these results, we tentatively propose that secretion signals required for recognition and translocation across the outer membrane via the pullulanase-specific extension of the secretion pathway are located near the N-terminus of the pullulanase polypeptide.

Positively charged amino acid residues can act as topogenic determinants in membrane proteins

When alkaline phosphatase is fused to the periplasmic domain of a cytoplasmic membrane protein, it is efficiently exported to the periplasm. Such a hybrid protein exhibits high alkaline phosphatase enzymatic activity. When alkaline phosphatase is fused to the cytoplasmic domain of a membrane protein, it remains, for the most part, in the cytoplasm. Such fusions exhibit low enzymatic activity. However, stable retention of alkaline phosphatase in the cytoplasm requires the presence in the fusion protein of the cytoplasmic loop ordinarily present in that position in the native, unfused protein. Using oligonucleotide-directed mutagenesis, we have shown that positively charged amino acids are required for the stable cytoplasmic localization of the fused alkaline phosphatase. We propose that, in addition to hydrophobic transmembrane segments, positively charged amino acids in the hydrophilic cytoplasmic domains of a membrane protein are determinants of the protein's topology.