Mapping of export signals of Pseudomonas aeruginosa pilin with alkaline phosphatase fusions.

Pili of Pseudomonas aeruginosa are assembled from monomers of the structural subunit, pilin, after secretion of this protein across the bacterial membrane. These subunits are initally synthesized as precursors (prepilin) with a six-amino-acid leader peptide that is cleaved off during or after membrane traversal, followed by methylation of the amino-terminal phenylalanine residue. This report demonstrates that additional sequences from the N terminus of the mature protein are necessary for membrane translocation. Gene fusions were made between amino-terminal coding sequences of the cloned pilin gene (pilA) and the structural gene for Escherichia coli alkaline phosphatase (phoA) devoid of a signal sequence. Fusions between at least 45 amino acid residues of the mature pilin and alkaline phosphatase resulted in translocation of the fusion proteins across the cytoplasmic membranes of both P. aeruginosa and E. coli strains carrying recombinant plasmids, as measured by alkaline phosphatase activity and Western blotting. Fusion proteins constructed with the first 10 amino acids of prepilin (including the 6-amino-acid leader peptide) were not secreted, although they were detected in the cytoplasm. Therefore, unlike that of the majority of secreted proteins that are synthesized with transient signal sequences, the membrane traversal of pilin across the bacterial membrane requires the transient six-amino-acid leader peptide as well as sequences contained in the N-terminal region of the mature pilin protein.     

Purification and processing of cellulose-binding domain-alkaline phosphatase fusion proteins

Fusion of the leader peptide and the cellulose-binding domain (CBD) of endoglucanase A (CenA) from Cellulomonas fimi, with of without linker sequences, to the N-terminus of alkaline phosphatase (PhoA) from Escherichia coli leads to the accumulation of significant amounts of the CBD-PhoA fusion proteins in the supernatants of E. coli cultures. The fusion proteins can be purified from the supernatants by affinity chromatography on cellulose. The fusion protein can be desorbed from the cellulose with water or guanidine-HCl. If the sequence IEGR in present between the CBD and PhoA, the CBD can be cleaved from the PhoA with factor Xa. The efficiency of hydrolysis by factor Xa is strongly in fluenced by the amino acids on either side of the IEGR sequence. The CBD released by factor Xa is removed by adsorption to cellulose. A nonspecific proteases from C. fimi, which hydrolyzes native CenA between the CBD and the catalytic domain, may be useful for removing the CBD from some fusion proteins. (c) 1994 John Wiley & Sons, Inc.     

Linkers for improved cleavage of fusion proteins with an engineered alpha-lytic protease.

Addition of an N-terminal fusion partner can greatly aid the expression and purification of a recombinant protein in Escherichia coli. We investigated two genetically engineered proteases designed to remove the fusion partner after the protein of interest has been expressed. Recombinant human insulin-like growth factor-II (hIGF-II) has been produced from E. coli-derived fusion proteins using a novel enzymatic cleavage system that uses a mutant of alpha-lytic protease. Initially, two potential fusion protein linkers were designed, Pro-Ala-Pro-His (PAPH) and Pro-Ala-Pro-Met (PAPM), and were tested as substrates in the form of synthetic dodecapeptides. Using mass spectrometry and reverse-phase HPLC, the position of cleavage was confirmed and the kinetics of synthetic peptide cleavage were examined. Use of the linkers in hIGF-II fusion proteins produced in E. coli was then evaluated. The fusion proteins constructed consist of the first 11 amino acids of porcine growth hormone linked N-terminally to hIGF-II by six amino acids that include the dipeptide Val-Asn followed by a variable tetrapeptide protease cleavage motif. Mass spectrometry and N-terminal sequencing confirmed that proteolytic cleavage of the fusion proteins had occurred at the predicted sites. Using the fusion proteins as substrates, the cleavage of the rationally designed motifs by the alpha-lytic protease mutant was compared. The fusion protein containing the motif PAPM had a k(cat)/K(M) ratio indicating a 1.6-fold preference over the PAPH fusion protein for cleavage by this enzyme. Furthermore, when hIGF-II fusion proteins containing the designed cleavable linkers were processed with the engineered alpha-lytic protease, they gave greatly improved yields of native hIGF-II compared to an analogous fusion protein cleaved by H64A subtilisin. Comparison of the peptide and protein cleavage studies shows that the efficient proteolysis of the cleavage motifs is an inherent property of the designed sequences and is not determined by secondary or tertiary structure in the fusion proteins.     

Immobilization and purification of enzymes with staphylococcal protein A gene fusion vectors.

Two improved plasmid vectors, containing the gene coding for staphylococcal protein A and adapted for gene fusions, have been constructed. These vectors allow fusion of any gene to the protein A moiety, giving fusion proteins which can be purified, in a one-step procedure by IgG affinity chromatography. One vector, pRIT2, is designed for temperature-inducible expression of intracellular fusion proteins in Escherichia coli and the other pRIT5, is a shuttle vector designed for secretion. The latter gives a periplasmatic fusion protein in E. coli and an extracellular protein in Gram-positive hosts such as Staphylococcus aureus. The usefulness of these vectors is exemplified by fusion of the protein A gene and the E. coli genes encoding the enzymes beta-galactosidase and alkaline phosphatase. High amounts of intact fusion protein are produced which can be immobilized on IgG-Sepharose in high yield (95-100%) without loss of enzymatic activity. Efficient secretion in both E. coli and S. aureus, was obtained for the alkaline phosphatase hybrid, in contrast to beta-galactosidase which was only expressed efficiently using the intracellular system. More than 80% of the protein A alkaline-phosphatase hybrid protein can be eluted from IgG affinity columns without loss of enzymatic activity.     

Export of FepA::PhoA fusion proteins to the outer membrane of Escherichia coli K-12.

A library of fepA::phoA gene fusions was generated in order to study the structure and secretion of the Escherichia coli K-12 ferric enterobactin receptor, FepA. All of the fusion proteins contained various lengths of the amino-terminal portion of FepA fused in frame to the catalytic portion of bacterial alkaline phosphatase. Localization of FepA::PhoA fusion proteins in the cell envelope was dependent on the number of residues of mature FepA present at the amino terminus. Hybrids containing up to one-third of the amino-terminal portion of FepA fractionated with their periplasm, while those containing longer sequences of mature FepA were exported to the outer membrane. Outer membrane-localized fusion proteins expressed FepA sequences on the external face of the outer membrane and alkaline phosphatase moieties in the periplasmic space. From sequence determinations of the fepA::phoA fusion joints, residues within FepA which may be exposed on the periplasmic side of the outer membrane were identified.     

Proteome-based identification of fusion partner for high-level extracellular production of recombinant proteins in Escherichia coli

Extracellular production of recombinant proteins in Escherichia coli has several advantages over cytoplasmic or periplasmic production. However, nonpathogenic laboratory strains of E. coli generally excrete only trace amounts of proteins into the culture medium under normal growth conditions. Here we report a systematic proteome-based approach for developing a system for high-level extracellular production of recombinant proteins in E. coli. First, we analyzed the extracellular proteome of an E. coli B strain, BL21(DE3), to identify naturally excreted proteins, assuming that these proteins may serve as potential fusion partners for the production of recombinant proteins in the medium. Next, overexpression and excretion studies were performed for the 20 selected fusion partners with molecular weights below 40 kDa. Twelve of them were found to allow fused proteins to excrete into the medium at considerable levels. The most efficient excreting fusion partner, OsmY, was used as a carrier protein to excrete heterologous proteins into the medium. E. coli alkaline phosphatase, Bacillus subtilis alpha-amylase, and human leptin used as model proteins could all be excreted into the medium at concentrations ranging from 5 to 64 mg/L during the flask cultivation. When only the signal peptide or the mature part of OsmY was used as a fusion partner, no such excretion was observed; this confirmed that these proteins were truly excreted rather than released by outer membrane leakage. The recombinant protein of interest could be recovered by cleaving off the fusion partner by enterokinase as demonstrated for alkaline phosphatase as an example. High cell density cultivation allowed production of these proteins to the levels of 250-700 mg/L in the culture medium, suggesting the good potential of this approach for the excretory production of recombinant proteins.     

A topological analysis of subunit alpha from Escherichia coli F1F0-ATP synthase predicts eight transmembrane segments

The membrane topology of subunit alpha from the Escherichia coli F1F0-ATP synthase was studied using a gene fusion technique. Fusion proteins linking different amino-terminal fragments of the alpha subunit with an enzymatically active fragment of alkaline phosphatase were constructed by both random transposition of TnphoA and site-directed mutagenesis. Those proteins with high levels of alkaline phosphatase activity are predicted to define periplasmic domains of alpha, and this was confirmed by testing for cell growth in minimal medium supplemented with polyphosphate (P greater than 75) as the sole source of phosphate. The enzymatic activity of some fusion proteins was shown to be sensitive to glucose present in the growth medium. Results from subcellular fractionation experiments suggest that these fusion proteins may be inactive even though they have a periplasmic alkaline phosphatase. The enzymatic activity appears dependent upon proteolytic release of the alkaline phosphatase moiety from its alpha subunit membrane anchor and suggests the target of glucose repression may be a protease present in the periplasm. For the topological analysis of the alpha subunit, a total of 28 unique fusion proteins were studied and the results were consistent with a model of alpha containing eight transmembrane segments, including periplasmic amino and carboxyl termini. Surprisingly, separate periplasmic domains were identified near amino acids 200, 233, and 270. These results suggest the flanking membrane spans are only 10-15 amino acids in length and not able to span a standard 30 A bilayer in an alpha-helical conformation. These short spans may have interesting mechanistic implications for the function of F0, because they contain several amino acids which appear critical for proton translocation. Finally, a fusion of alkaline phosphatase at amino acid 271, the carboxyl-terminal residue, but not at amino acid 260, was able to complement the strain RH305 (uncB-) for growth on succinate and suggests the last 11 amino acids of the alpha subunit are critical to the function of F1F0-ATP synthase.     

Plasmid expression vectors with elements of the E. coli alkaline phosphatase gene

A family of plasmid expression vectors, containing fragments of the E. coli alkaline phosphatase gene (phoA), has been constructed for the lambda PR promoter-directed thermoinducible superproduction of fusions of heterologous polypeptides to the N- or C-terminus of E. coli alkaline phosphatase and its leader peptide. Effective expression and export to periplasm of resulting fusion proteins, which may retain enzymatic activity of the phosphatase, has been shown.     

In vivo biotinylation of fusion proteins expressed in Escherichia coli with a sequence of Propionibacterium freudenreichii transcarboxylase 1.3S biotin subunit.

Biotinylation of fusion proteins in E. coli was studied using a sequence of Propionibacterium freudenreichii transcarboxylase 1.3S biotin subunit. As the biotinylation sequence, we examined two sequences: one was of amino acid residues [84-123] of 1.3S, a partial sequence containing a region from a conserved tetrapeptide (Ala-Met-Bct-Met) around the biotinyl lysine (Bct) to the carboxyl terminal; the other was of an almost entire sequence [18-123]. We constructed recombinant plasmids for fusion proteins of beta-galactosidase, of chloramphenicol acetyltransferase, and of alkaline phosphatase. We found the biotinylation in the [18-123] sequence fused to alkaline phosphatase.     

Membrane topology analysis of Escherichia coli K-12 Mtr permease by alkaline phosphatase and beta-galactosidase fusions.

The mtr gene of Escherichia coli K-12 encodes an inner membrane protein which is responsible for the active transport of trypotophan into the cell. It has been proposed that the Mtr permease has a novel structure consisting of 11 hydrophobic transmembrane spans, with a cytoplasmically disposed amino terminus and a carboxyl terminus located in the periplasmic space (J.P. Sarsero, P. J. Wookey, P. Gollnick, C. Yanofsky, and A.J. Pittard, J. Bacteriol. 173:3231-3234, 1991). The validity of this model was examined by the construction of fusion proteins between the Mtr permease and alkaline phosphatase or beta-galactosidase. In addition to the conventional methods, in which the reporter enzyme replaces a carboxyl-terminal portion of the membrane protein, the recently developed alkaline phosphatase sandwich fusion technique was utilized, in which alkaline phosphatase is inserted into an otherwise intact membrane protein. A cluster of alkaline phosphatase fusions to the carboxyl-terminal end of the Mtr permease exhibited high levels of alkaline phosphatase activity, giving support to the proposition of a periplasmically located carboxyl terminus. The majority of fusion proteins produced enzymatic activities which were in agreement with the positions of the fusion sites on the proposed topological model of the permease. The synthesis of a small cluster of hybrid proteins, whose enzymatic activity did not agree with the location of their fusion sites within putative transmembrane span VIII or the preceding periplasmic loop, was not detected by immunological techniques and did not necessitate modification of the proposed model in this region. Slight alterations may need to be made in the positioning of the carboxyl-terminal end of transmembrane span X.     

A short N-proximal region of prochymosin inhibits the secretion of hybrid proteins from Escherichia coli

A gene encoding bovine prochymosin (PC) was fused to the coding sequence (phoA) for the Escherichia coli alkaline phosphatase (AP) signal peptide and expressed in E. coli under the control of the phoA promoter. Upon induction, an AP-PC fusion protein was produced which was neither processed nor exported into the periplasm. We investigated this lack of secretion by constructing a series of gene fusions in which different regions of the PC gene were inserted between the coding regions of the AP leader and mature protein. Analysis of the cellular location of the proteins encoded by these fusions revealed that a region of PC (between amino acids 6 and 29) prevented processing and secretion of an AP-PC fusion when inserted near to the AP signal peptide. In contrast, when this 'blocking sequence' was inserted elsewhere in AP the hybrid proteins were efficiently processed and translocation was initiated.     

Evidence for two structural genes for alkaline phosphatase in Bacillus subtilis.

Two secreted alkaline phosphatase proteins were purified from cultures of Bacillus subtilis JH646MS. The two proteins showed slight differences in subunit molecular weight, substrate specificity, and charge characteristics. A total of 62% of the first 22 amino-terminal amino acids were identical. Both sequences showed conservation of structural features identified in Escherichia coli and human alkaline phosphatases. One alkaline phosphatase was a monomer and the other was a dimer. Southern analysis of genomic DNA with degenerative oligomers based on the amino acid sequences suggest that there are two structural genes for alkaline phosphatase in the genome of B. subtilis.     

Comparison of SUMO fusion technology with traditional gene fusion systems: enhanced expression and solubility with SUMO

Despite the availability of numerous gene fusion systems, recombinant protein expression in Escherichia coli remains difficult. Establishing the best fusion partner for difficult-to-express proteins remains empirical. To determine which fusion tags are best suited for difficult-to-express proteins, a comparative analysis of the newly described SUMO fusion system with a variety of commonly used fusion systems was completed. For this study, three model proteins, enhanced green fluorescent protein (eGFP), matrix metalloprotease-13 (MMP13), and myostatin (growth differentiating factor-8, GDF8), were fused to the C termini of maltose-binding protein (MBP), glutathione S-transferase (GST), thioredoxin (TRX), NUS A, ubiquitin (Ub), and SUMO tags. These constructs were expressed in E. coli and evaluated for expression and solubility. As expected, the fusion tags varied in their ability to produce tractable quantities of soluble eGFP, MMP13, and GDF8. SUMO and NUS A fusions enhanced expression and solubility of recombinant proteins most dramatically. The ease at which SUMO and NUS A fusion tags were removed from their partner proteins was then determined. SUMO fusions are cleaved by the natural SUMO protease, while an AcTEV protease site had to be engineered between NUS A and its partner protein. A kinetic analysis showed that the SUMO and AcTEV proteases had similar KM values, but SUMO protease had a 25-fold higher kcat than AcTEV protease, indicating a more catalytically efficient enzyme. Taken together, these results demonstrate that SUMO is superior to commonly used fusion tags in enhancing expression and solubility with the distinction of generating recombinant protein with native sequences.     

Engineering subtilisin BPN' for site-specific proteolysis

A combination of protein engineering and substrate optimization was used to create variants of the serine protease, subtilisin BPN', which efficiently and specifically cleave a designed target sequence in a fusion protein. The broad substrate specificity of wild-type subtilisin BPN' is greatly restricted by substitution of the catalytic histidine-containing of the catalytic histidine 64 with alanine (H64A) so that certain histidine-containing substrates are preferentially hydrolysed (Carter, P., Wells, J.A. Science 237:394-399, 1987). The catalytic efficiency, (kcat/Km), of this H64A variant was increased almost 20-fold by judicious choice of substrate and by installing three additional mutations which increase the activity of wild-type subtilisin. The most favorable substrate sequence identified was introduced as a linker in a fusion protein between a synthetic IgG binding domain of Staphylococcus aureus protein A and Escherichia coli alkaline phosphatase. The fusion protein (affinity purified on an IgG column) was cleaved by the prototype H64A enzyme and its improved variant, efficiently and exclusively at the target site, to liberate an alkaline phosphatase product of the expected size and N-terminal sequence. Several features of H64A variants of subtilisin make them attractive for site-specific proteolysis of fusion proteins: they have exquisite substrate specificity on the N-terminal side of the cleavage site and yet are broadly specific on the C-terminal side; they can be produced in large quantities and remain highly active even in the presence of detergents, reductants (modest concentrations), protease inhibitors, at high temperatures, or when specifically immobilized on a solid support.     

Use of gene fusions to determine a partial signal sequence of alkaline phosphatase.

We have isolated strains of Escherichia coli in which an amino-terminal portion of the cytoplasmic enzyme beta-galactosidase is replaced by an amino-terminal portion of the periplasmic enzyme alkaline phosphatase. The synthesis of these hybrid proteins is regulated by inorganic phosphate and they are located in the cytoplasm. One of these proteins was purified, and 14 amino acids of the amino-terminal sequence were determined. The first five amino acids, Met-Lys-Gln-Ser-Thr, appear to represent a portion of the signal sequence of the precursor of alkaline phosphatase, and the remaining sequence corresponds to that of beta-galactosidase, beginning at amino acid residue 20. The approach described here could be used for the analysis of signal sequences of exported proteins and for partial amino acid sequence determination of certain of certain other proteins.     

Bacillus subtilis alkaline phosphatases III and IV. Cloning, sequencing, and comparisons of deduced amino acid sequence with Escherichia coli alkaline phosphatase three-dimensional structure

Bacillus subtilis has an alkaline phosphatase multigene family. Two members of this gene family, phoAIII and phoAIV, were cloned, taking advantage of in vitro constructed strains containing a plasmid insertion within one or the other of the structural genes. The DNA sequences of the two genes showed approximately 64% identity at the DNA level and 63% identity in the deduced primary amino acid sequences. The phoAIII and phoAIV genes code for predicted proteins of 47,149 and 45,935 Da, respectively. Comparison of the deduced primary amino acid sequence of the mature proteins with other sequenced alkaline phosphatases from Escherichia coli, yeast, and humans shows 25-30% identity. Based on the refined crystal structure of E. coli alkaline phosphatase, it appears that the active site and the core of the structure are retained in both Bacillus alkaline phosphatases. However, both proteins are truncated at the amino terminus compared with other mature alkaline phosphatases, three sizable surface loops of E. coli are deleted, and a minidomain is replaced with a larger domain in the model. Neither Bacillus alkaline phosphatase sequenced contains any cysteine residues, an amino acid implicated in intrachain disulfide bond formation in other alkaline phosphatases.     

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.     

Enzymatic labeling of a single chain variable fragment of an antibody with alkaline phosphatase by microbial transglutaminase

Functional cross-linking of a single chain Fv fragment of anti-hen egg-white lysozyme antibody (scFv) and alkaline phosphatase (AP) was explored using microbial transglutaminase (MTG) from Streptomyces mobaraensis. A specific peptidyl linker for MTG was genetically fused to the N-terminus of each protein and the resultant proteins were obtained separately by bacterial expression. The recombinant peptide-tagged scFv and AP were site-specifically cross-linked by MTG through the extra peptidyl linkers in vitro, which mainly yielded the heterodimer (i.e., scFv-AP conjugate). The enzymatic cross-linking reaction had little influence on either the antigen-binding ability of the scFv moiety or the enzymatic activity of the AP moiety of the conjugate, allowing use within an enzyme-linked immunosorbent assay. The results obtained suggest that the enzymatic approach with MTG facilitates the posttranslational construction of functional fusion proteins.