High-level secretion of biologically active aprotinin from the yeastPichia pastoris

Summary A synthetic gene encoding aprotinin (bovine pancreatic trypsin, inhibitor) was fused to theSaccharomyces cerevisiae prepro alpha mating factor leader sequence at the dibasic amino acid processing site.Pichia pastoris strains were developed to'express one or multiple copies of a methanol-inducible expression cassette containing the gene fusion.P. pastoris containing a single copy of the vector secreed approximately 150 mg/l of immunoreactive protein. A construct bearing five copies of the expression cassette secreted 930 mg/l of aprotinin. The purified aprotinin molecule was equipoten with the native molecule in a trypsin inhibition assay. Protein sequence analysis showed that the alpha factor-aprotinin fusion was not processed at the basic amino acid residues Lys-Arg. Instead, recombinant aprotinin had additional N-terminal amino acids derived from prepro alpha factor. The N-terminal extension was variably 11 or 4 amino acids. Inclusion of the spacer DNA sequence encoding Glu and Ala between aprotinin and the Lys-Arg processing site led to the secretion of a biologically active aprotinin containing only a Glu-Ala N-terminal extension.     

Correlation of glycosylation forms with position in amino acid sequence

We surveyed published reports on about 50 glycoproteins whose amino acid sequence, glycosylation sites, and type of glycosylation at a particular site have been established. We note that high-mannose substances were rarely found at the N-terminal side of a previously glycosylated complex site. There was a very definite distribution of complex sites about the N-terminal region. Furthermore, secreted glycoproteins usually contained only complex oligosaccharides whereas membrane proteins contained both types. We suggest that the position of the glycosylation site with respect to the N-terminus affects the extent of oligosaccharide processing and subsequent presentation of complex or high-mannose structures in the mature glycoprotein. This review relates glycosylation type to its position in the known sequence of given proteins and discusses these observations in light of known glycosylation processing reactions.     

Effects of pre- and pro-sequence of thaumatin on the secretion by Pichia pastoris

Thaumatin is a 22-kDa sweet-tasting protein containing eight disulfide bonds. When thaumatin is expressed in Pichia pastoris using the thaumatin cDNA fused with both the alpha-factor signal sequence and the Kex2 protease cleavage site from Saccharomyces cerevisiae, the N-terminal sequence of the secreted thaumatin molecule is not processed correctly. To examine the role of the thaumatin cDNA-encoded N-terminal pre-sequence and C-terminal pro-sequence on the processing of thaumatin and efficiency of thaumatin production in P. pastoris, four expression plasmids with different pre-sequence and pro-sequence were constructed and transformed into P. pastoris. The transformants containing pre-thaumatin gene that has the native plant signal, secreted thaumatin molecules in the medium. The N-terminal amino acid sequence of the secreted thaumatin molecule was processed correctly. The production yield of thaumatin was not affected by the C-terminal pro-sequence, and the pro-sequence was not processed in P. pastoris, indicating that pro-sequence is not necessary for thaumatin synthesis.     

Expression of the human salivary alpha-amylase gene in yeast and characterization of the secreted protein

Recombinant plasmids were constructed in which the human salivary alpha-amylase gene, with or without the N-terminal signal sequence for secretion, was placed under control of the APase (PHO5) promoter of Saccharomyces cerevisiae. In yeast cells transformed with the alpha-amylase gene having the human signal sequence for secretion, the gene was expressed and the enzyme was secreted into the medium in three different glycosylated forms. The amylase gene without the signal sequence was also expressed in yeast, but the products were neither secreted nor glycosylated. Determination of the N-terminal amino acid (aa) sequence revealed that the 15-aa signal sequence had been cleaved from the secreted enzyme, and that the N-terminal residue, glutamine, had been modified into pyroglutamate, as is commonly observed with the mammalian salivary alpha-amylase. Thus, the human salivary alpha-amylase signal sequence for secretion was correctly recognized and processed by the yeast secretory pathway. The C-terminal residue was identified as leucine, which is predicted from the nucleotide sequence data to be located at position 511 in front of the termination codon. Therefore, there is no post-translational processing in formation of the C terminus.     

Analysis of signals for secretion in the staphylococcal protein A gene.

Different constructs of the gene encoding staphylococcal protein A were introduced in Staphylococcus aureus and S. xylosus as well as Escherichia coli. The product of the gene without the cell wall anchoring domain was efficiently secreted in all three hosts. N-terminal sequencing of the affinity-purified mature protein revealed a common processing site after the alanine residue at position 36. In contrast, when an internal IgG-binding fragment of protein A (region B) was inserted after the protein A signal sequence, the product was poorly secreted and N-terminal sequencing revealed no processing at the normal site. This demonstrates that the structure of the polypeptide chain beyond the signal peptide cleavage site can affect cleavage. Another construct, containing the N-terminal IgG-binding part of the mature protein A (region E) followed by region B, gave correct processing and efficient secretion. Unexpectedly, the gene product, EB, was not only secreted and correctly processed, but was also excreted to the culture medium of E. coli. Secretion vectors containing the protein A signal sequence were constructed to facilitate secretion of foreign gene products. Insertion of the E. coli gene phoA, lacking its own promoter and signal sequence, led to efficient secretion of alkaline phosphatase both in E. coli and S. aureus.     

Processing of proapolipoprotein AI requires specific conformation

Apolipoprotein AI of human high-density lipoproteins is secreted by hepatocytes as a proapolipoprotein with a N-terminal hexapeptide sequence (Arg-His-Phe-Trp-Gln-Gln-) which differs from the prosequence of rat apolipoprotein AI (Trp-Asp-Phe-Trp-Gln-Gln). The two proteins have in common the unusual cleavage site -Gln-Gln-Asp-Glu-. It is hydrolysed by a specific serum proteinase with the release of mature apo AI. We synthesized a model substrate for the study of the final processing of pro-apo AI by the serum proteinase. It is an undecapeptide embracing the human pro-hexapeptide sequence and the first five N-terminal residues of apo AI, covalently linked to a hydrophilic resin. The N-terminal arginine residue was 3H-labelled. [formula; see text] This sequence was not cleaved by human serum under the conditions under which rat serum processes the pro-form of apo AI secreted by rat hepatocytes. Pepsin and chymotrypsin fragmented the undecapeptide at sites characteristic for these proteinases. We conclude that the proteolytic cleavage at the specific site (-Gln-Gln-Asp-Glu-) requires the correct conformation in addition to the specific amino-acid sequence.     

Synthesis, transport, and processing of apolipoproteins of high density lipoproteins

Cell biology methods have greatly influenced the elucidation of the biosynthetic pathways of apolipoproteins. In vitro and tissue culture systems allow the study, to a large extent, of the process of synthesis, intracellular processing, secretion, and extracellular processing of the major high density lipoprotein apoproteins apoA-I and A-II and also of a minor component, apoA-IV. Whereas the latter apoprotein is equipped only with a signal sequence, the primary translation products of apoA-I and apoA-II carry N-terminal extensions of preprosequence of 24 amino acids for apoA-I and 23 amino acid residues for apoA-II. The pro-form of apoA-I characterized by a hexapeptide extension is completely stable intracellularly and is secreted as such. The pro-form is further processed by a serum protease specific for an unusual -Gln-Gln-Asp-Glu-sequence site. Pro-apoA-II, a pentapeptide sequence, is partially processed intracellularly to its mature form and secreted together with the residual pro-form. The cleavage site of pro-apoA-II is characterized by two basic amino acid residues Arg-Arg, present also in other known pro-proteins. The biological function of the N-terminal pro-sequences and details of their final processing by the serum protease(s) have yet to be established.     

Modifications of a signal sequence for antibody secretion from insect cells

Monoclonal antibodies and antibody fragments have recently been developed for use in diverse diagnostic and therapeutic applications. Insect cells can efficiently secrete recombinant proteins such as antibody molecules through post-translational processing and modifications that are similar to those performed in mammalian cells. In eukaryotic cells, the signal sequence in a nascent polypeptide is recognized by the signal recognition particle, and the polypeptide is then folded and modified in the endoplasmic reticulum. The signal sequence consists of three regions, a positively charged N-terminus, a hydrophobic core, and a polar C-terminus. In the present study, we examined the substitutions of the characteristic amino acids of a Drosophila immunoglobulin heavy chain binding protein signal sequence, and investigated the effect on the secretory production of an antibody Fab fragment from lepidopteran insect cells in transient expression. A modification of the signal sequence for the heavy chain resulted in a twofold increase in the secreted Fab fragment, while the modification for the light chain led to a more than 3.6-fold increase.     

Signal and propeptide processing of human tissue plasminogen activator: activity of a pro-tPA derivative

We have explored the heterogeneity in the proteolytic processing of the N-terminus of human tissue plasminogen activator. We demonstrate that normal propeptide processing occurs following Arg-4, preceding the sequence Gly-Ala-Arg-Ser+1. Generation of the previously designated Ser+1 occurs via secondary proteolysis following secretion. By site-directed mutagenesis, we have eliminated this cleavage site resulting in a derivative containing the propeptide sequence. N-terminal sequence analysis of this form indicated that signal peptide cleavage occurs following Ser-13. The pro-tPA derivative had near normal serine protease and plasminogen activating activities, and could be stimulated by fibrin. An additional derivative, containing the tribasic sequence from the human protein C propeptide preceding Ser+1, was secreted with full processing of the propeptide. Our data have defined the cleavages for the signal peptide and propeptide and demonstrate that a tribasic sequence can be used to eliminate N-terminal heterogeneity in this molecule. In addition, we demonstrate that, unlike several other serine proteases, a propeptide sequence does not alter the activity of this enzyme.     

Purification of a novel extracellular laccase from solid-state culture of the edible mushroom <Emphasis Type="Italic">Lentinula edodes</Emphasis>

The laccases (EC 1.10.3.2) secreted into solid-state culture by Lentinula edodes were analyzed. The fungus secreted at least two laccases in the solid-state culture. One laccase was purified to a homogeneous preparation using anion-exchange, hydrophobic, and size-exclusion chromatography. SDS-PAGE analysis showed that the purified laccase, Lcc6, was a monomeric protein of 58.5 kDa. The optimum pH for enzyme activity was about 3.5, and the laccase was most active at 40°C. The N-terminal amino acid sequence of Lcc6 did not correspond to the sequence of Lcc1, which was previously purified from L. edodes. Lcc6 had decolorization activity to some chemical dyes.     

Comparison of barley malt α-amylase isozymes 1 and 2: construction of cDNA hybrids by in vivo recombination and their expression in yeast

Germinating barley produces two α-amylase isozymes, AMY1 and AMY2, having 80% amino acid (aa) sequence identity and differing with respect to a number of functional properties. Recombinant AMY1 (re-AMYI) and AMY2 (re-AMY2) are produced in yeast, but whereas all re-AMYI is secreted, re-AMY2 accumulates within the cell and only traces are secreted. Expression of AMY1::AMY2 hybrid cDNAs may provide a means of understanding the difference in secretion efficiency between the two isozymes. Here, the efficient homologous recombination system of the yeast, Saccharomyces cerevisiae, was used to generate hybrids of barley AMY with the N-terminal portion derived from AMY1, including the signal peptide (SP), and the C-terminal portion from AMY2. Hybrid cDNAs were thus generated that encode either the SP alone, or the SP followed by the N-terminal 21, 26, 53, 67 or 90 aa from AMY1 and the complementary C-terminal sequences from AMY2. Larger amounts of re-AMY are secreted by hybrids containing, in addition to the SP, 53 or more aa of AMY1. In contrast, only traces of re-AMY are secreted for hybrids having 26 or fewer aa of AMY1. In this case, re-AMY hybrid accumulates intracellularly. Transformants secreting hybrid enzymes also accumulated some re-AMY within the cell. The AMY1 SP, therefore, does not ensure re-AMY2 secretion and a certain portion of the N-terminal sequence of AMY1 is required for secretion of a re-AMYI::AMY2 hybrid.     

High-level expression of nonglycosylated human pancreatic lipase-related protein 2 in Pichia pastoris

The human pancreatic lipase-related protein 2 (HPLRP2) was produced in the methylotrophic yeast Pichia pastoris. The HPLRP2 cDNA corresponding to the protein coding sequence including the native signal sequence, was cloned into the pPIC9K vector and integrated into the genome of P. pastoris. P. pastoris transformants secreting high-level rHPLRP2 were obtained and the expression level into the liquid culture medium reached about 40mg/L after 4 days of culture. rHPLRP2 was purified by a single anion-exchange step after an overnight dialysis. N-terminal sequence analysis showed that the purified rHPLRP2 mature protein possessed a correct N-terminal amino acid sequence indicating that its signal peptide was properly processed. Mass spectrometry analysis showed that the recombinant HPLRP2 molecular weight was 52,532Da which was 2451Da greater than the mass calculated from the sequence of the protein (50,081Da) and 1536Da greater than the mass of the native human protein (50,996Da). In vitro deglycosylation experiments by peptide:N-glycosidase F (PNGase F) indicated that rHPLRP2 secreted from P. pastoris was N-glycosylated. Specific conditions were setup in order to obtain a recombinant protein free of glycan chain. We observed that blocking glycosylation in vivo by addition of tunicamycin in the culture medium during the production resulted in a correct processing of the rHPLRP2 mature protein. The lipase activity of glycosylated or nonglycosylated rHPLRP2, which was about 800U/mg on tributyrin, was inhibited by the presence of bile salts and not restored by adding colipase. In conclusion, the experimental procedure which we have developed will allow us to get a high-level production in P. pastoris of glycosylated and nonglycosylated rHPLRP2, suitable for subsequent biophysical and structural studies.     

Processing of the prepropeptide portions of the Bacillus amyloliquefaciens neutral protease fused to Bacillus subtilis alpha-amylase and human growth hormone during secretion in Bacillus subtilis.

A set of nested 3'-terminal deletions of the prepropeptide of the Bacillus amyloliquefaciens neutral protease gene was constructed. Alpha-amylase and human growth hormone were secreted using these truncated genes in Bacillus subtilis. The level of the secreted alpha-amylase varied with the region for the truncated prepropeptide contained in the fusion gene but was independent of its length. Even though length of the prepropeptide varied, the mobilities of secreted alpha-amylases were the same as that of the control alpha-amylase derived from the alpha-amylase clone, pTUB4 (Yamazaki et al., 1983). Analyses of the secreted N-terminal amino acid sequences confirmed that they were all identical to that of the authentic one. Precursor proteins of the alpha-amylase were found in the cell-associated fraction, suggesting that the prepropeptide portion was processed during secretion. On the other hand, the N-terminus of hGH secreted using one of these prepropeptide portions varied by 1 to 4 additional N-terminal amino acid residues derived from the junction sequence between the sequence for propeptide portion and mature hGH or from C-terminal region of the propeptide portion. These results suggest that the prepropeptide portion can be generally processed even in the heterogeneous fusion. A probable mechanism of processing and maturation of the fusion gene products is also discussed.     

The DNA replication inhibitor microcin B17 is a forty-three-amino-acid protein containing sixty percent glycine

Microcin B17 is a low-molecular-weight protein that inhibits DNA replication in a number of enteric bacteria. It is produced by bacterial strains which harbor a 70-kilobase plasmid called pMccB17. Four plasmid genes (named mcbABCD) are required for its production. The product of the mcbA gene was identified by labelling minicells. The mcbA gene product was slightly larger when a mutation in any of the other three production genes was present. This indicates that these genes are involved in processing the primary mcbA product to yield the active molecule. The mcbA gene product predicted from the nucleotide sequence has 69 amino acids including 28 glycine residues. Microcin B17 was extracted from the cells by boiling in 100 mM acetic acid, 1 mM EDTA, and purified to homogeneity in a single step by high-performance liquid chromatography through a C18 column. The N-terminal amino acid sequence and amino acid composition demonstrated that mcbA is the structural gene for microcin B17. The active molecule is a processed product lacking the first 26 N-terminal residues. The 43 remaining residues include 26 glycines. While microcin B17 is an exported protein, the cleaved N-terminal peptide does not have the characteristic properties of a "signal sequence", which suggests that it is secreted by a mechanism different from that used by most secreted proteins of E. coli.     

A novel relaxase homologue is involved in chromosomal DNA processing for type IV secretion in Neisseria gonorrhoeae

The Neisseria gonorrhoeae type IV secretion system secretes chromosomal DNA that acts in natural transformation. To examine the mechanism of DNA processing for secretion, we made mutations in the putative relaxase gene traI and used nucleases to characterize the secreted DNA. The nuclease experiments demonstrated that the secreted DNA is single-stranded and blocked at the 5' end. Mutation of traI identified Tyr93 as required for DNA secretion, while substitution of Tyr201 resulted in intermediate levels of DNA secretion. TraI exhibits features of relaxases, but also has features that are absent in previously characterized relaxases, including an HD phosphohydrolase domain and an N-terminal hydrophobic region. The HD domain residue Asp120 was required for wild-type levels of DNA secretion. Subcellular localization studies demonstrated that the TraI N-terminal region promotes membrane interaction. We propose that Tyr93 initiates DNA processing and Tyr201 is required for termination or acts in DNA binding. Disruption of an inverted-repeat sequence eliminated DNA secretion, suggesting that this sequence may serve as the origin of transfer for chromosomal DNA secretion. The TraI domain architecture, although not previously described, is present in 53 uncharacterized proteins, suggesting that the mechanism of TraI function is a widespread process for DNA donation.     

Production of an active recombinant thrombomodulin derivative in transgenic tobacco plants and suspension cells

Thrombomodulin is a membrane-bound protein that plays an active role in the blood coagulation system by binding thrombin and initiating the protein C anticoagulant pathway. Solulin™ is a recombinant soluble derivative of human thrombomodulin. It is used for the treatment of thrombotic disorders. To evaluate the production of this pharmaceutical protein in plants, expression vectors were generated using four different N-terminal signal peptides. Immunoblot analysis of transiently transformed tobacco leaves showed that intact Solulin™ could be detected using three of these signal peptides. Furthermore transgenic tobacco plants and BY2 cells producing Solulin™ were generated. Immunoblot experiments showed that Solulin™ accumulated to maximum levels of 115 and 27 μg g⁻¹ plant material in tobacco plants and BY2 cells, respectively. Activity tests performed on the culture supernatant of transformed BY2 cells showed that the secreted Solulin™ was functional. In contrast, thrombomodulin activity was not detected in total soluble protein extracts from BY2 cells, probably due to inhibitory effects of substances in the cell extract. N-terminal sequencing was carried out on partially purified Solulin™ from the BY2 culture supernatant. The sequence was identical to that of Solulin™ produced in Chinese hamster ovary cells, confirming correct processing of the N-terminal signal peptide. We have demonstrated that plants and plant cell cultures can be used as alternative systems for the production of an active recombinant thrombomodulin derivative.     

A Novel Approach for Secretion of Heterologous Proteins with Correct N-Terminal Processing by Using α-Factor Pre Sequence in Pichia pastoris

Numerous proteins have been secreted in P. pastoris by fusing the target gene with α-factor pre-pro sequence at Kex2 endopeptidase cleavage site. However, in some instances the product cannot be correctly processed due to aberrant cleavage by Kex2 endopeptidase such as aprotinin. In this study, an aprotinin gene was cloned into pPIC9K at the signal peptidase cleavage site through a single NheI restriction site designed at the 3'end of the α-factor signal sequence preregion, and transformed into GS115 host cell. By G418 resistance and ELISA assay, a high-yield recombinant was selected. After fed-batch cultivation in a 7-L bioreactor, the product was efficiently secreted into culture medium and accumulated up to ～ 4.7 mg L-1. MALDI-TOF/MS and N-terminal analyses confirmed its authenticity. Thus, a novel cloning strategy for secretion of aprotinin with correct N-terminal processing in P. pastoris has been developed which can be potentially applied to other proteins.     

Characteristics of a monoclonal beta-endorphin antibody recognizing the N-terminus of opioid peptides

The properties of a mouse monoclonal antibody to beta-endorphin secreted by a clone of hybrid myelomas (3-E7) are described. The antibody displays virtually complete cross-reactivity to met-enkephalin and leu-enkephalin, but no cross-reactivity to beta-lipotropin, alpha-N-acetyl-beta-endorphin and des-Tyr1-beta-endorphin. Substantial cross-reactivity is seen with some other naturally occurring opioid peptides bearing the enkephalin sequence, such as dynorphin, alpha-neo-endorphin and BAM 22, but cross-reactivity is lacking in the case of certain synthetic enkephalin derivatives possessing a D-amino acid in position 2. The data indicate that for the binding of an antigen to the antibody the N-terminal tyrosine moiety is essential. The antibody recognizes, thus, a site which is of functional significance for the interaction of many naturally occurring opioid peptides with the opiate receptor.     

Expression inEscherichia coli of active human alcohol dehydrogenase lacking N-terminal acetylation

Human alcohol dehydrogenase (ADH, tiff isozyme of class I) was expressed in Escherichia coli, purified to homogeneity, and characterized regarding N-terminal processing. The expression system was obtained by ligation of a cDNA fragment corresponding to the fl-subunit of human liver alcohol dehydrogenase into the vector pKK 223-3 containing the tac promoter. The enzyme, detected by Western-blot analysis and ethanol oxidizing activity, constituted up to 3 ~o of the total amount of protein. Recombinant ADH was separated from E. coli ADH by ion-exchange chromatography and the isolated enzyme was essentially pure as judged by SDS-polyacrylamide gel electrophoresis and sequence analysis. The N-terminal sequence was identical to that of the authentic fl-subunit except that the N-terminus was non-acetylated, indicating a correct removal of the initiator methionine, but lack of further processing.