Strategies for optimal synthesis and secretion of heterologous proteins in the methylotrophic yeast Pichia pastoris.

Numerous heterologous proteins have been produced at greater than gram per liter levels in the methylotrophic yeast, Pichia pastoris, using the methanol oxidase promoter. The factors that drastically influence protein production in this system include: copy number of the expression cassette, site and mode of chromosomal integration of the expression cassette, mRNA 5'- and 3'-untranslated regions (UTR), translational start codon (AUG) context, A+T composition of cDNA, transcriptional and translational blocks, nature of secretion signal, endogenous protease activity, host strain physiology, media and growth conditions, and fermentation parameters. All these factors should be considered in designing an optimal production system. The inherent ability of P. pastoris to convert the zymogen (pro-enzyme) form of matrix metalloproteinases (MMP) into active mature forms (which tend to self-degrade, and in some instances also cause damage to cells), largely limits the use of this system for the production of MMP. However, this problem can be partly alleviated by co-expression of tissue inhibitor of MMP (TIMP-1).

Expression of a synthetic gene encoding the anticoagulant-antimetastatic protein ghilanten by the methylotropic yeast Pichia pastoris.

Ghilantens are a family of cysteine-rich inhibitors of the clotting enzyme, factor Xa, that are produced in the salivary glands of the South American leech Haementeria ghilianii. In this study, a gene, designed from the amino acid sequence of a specific ghilanten isoform, was assembled from eight double-stranded oligonucleotide fragments. A yeast expression plasmid, pPIC9HG-2, was constructed by making an inframe fusion of the ghilanten-coding sequences with the region encoding the pre-pro alpha-mating factor signal sequence for secretion. The expression of ghilanten in pPIC9HG-2 was under the control of the methanol-inducible, alcohol oxidase (AOX1) promoter. Pichia pastoris yeast strains KM 71 and SMD 1168 were transformed with linearized pPIC9HG-2 to target integration of the plasmid to the chromosomal 5'-AOX1 locus via homologous recombination. Both strains yielded His+ transformants that secreted a potent anticoagulant activity into the medium. Product yield was improved by using buffered media (pH 6.0) supplemented with either casamino acids or a mixture of yeast extract and peptone. The protease-deficient strain, SMD 1168, secreted about a twofold higher level of r-ghilanten than KM 71. Significant clonal variation in the expression of r-ghilanten was found among the His+ transformants. A high producing clone was selected for production at the 2-liter shake flask and 10-liter bioreactor scales. r-Ghilanten was recovered from the fermentation broths in a single step by heparin Sepharose affinity chromatography. Protein sequence analysis of the amino terminus showed that the correct processing to yield mature ghilanten varied with the fermentation conditions.

Crystallographic structure of human gamma-thrombin.

In an effort to prepare crystals and determine the structure of alpha-thrombin complexed to a synthetic peptide inhibitor (MDL-28050) of the hirudin 54-65 COOH-terminal region, it was discovered that the crystals were not those of the complex but of gamma-thrombin. Gel electrophoresis studies revealed that autolytic degradation had occurred prior to crystallization. NH2-terminal sequence analysis of these autolytic fragments confirmed the gamma-thrombin product (cleavages at Arg75-Tyr76 and/or Arg77A-Asn78, and Lys149E-Gly150; chymotrypsinogen numbering) with a minor amount of another autolysis product, beta-thrombin (first two cleavages only). The final structure has an R-factor of 0.156 for 7.0-2.5-A data, and includes 186 water molecules. A comparison of gamma-thrombin with the thrombin structure in the alpha-thrombin-hirugen complex revealed that the two structures agreed well (r.m.s. delta = 0.39 A for main chain atoms). These structures possess uninhibited active sites where the disposition of the catalytic triad residues is nearly identical. The electron density in the vicinity of the gamma-thrombin cleavage regions is poor, and only becomes well-defined several residues prior to and after the actual cleavage sites. The extensive disorder evoked by beta-cleavage(s) in the Lys70-Glu80 loop region indicates that this part of the molecule is severely disrupted by autolysis and is the reason exosite functions are dramatically impaired in beta-and gamma-thrombin. Since autolysis did not lead to a major reorganization of the folded structure of alpha-thrombin, the likely structural features of the interaction of thrombin substrate with thrombin enzyme during beta-cleavage have been modeled by docking the exosite region of one molecule at the active site of another.

Structure of human des(1-45) factor Xa at 2.2 A resolution.

The structure of a large molecular fragment of factor Xa that lacks only a Gla (gamma-carboxyglutamic acid) domain (N-terminal 45 residues) has been solved by X-ray crystallography and refined at 2.2 A resolution to a crystallographic R-value of 0.168. The fragment identity was clearly established by automated Edman degradation. X-ray structure analysis confirmed the biochemical characterization and also revealed that the N-terminal epidermal growth factor (EGF)-like domain is flexibly disordered in crystals. The second EGF module, however, is positionally ordered making contacts with the catalytic domain. The overall folding of the catalytic domain is similar to that of alpha-thrombin, excluding the insertion loops of the latter with respect to simpler serine proteinases. The C-terminal arginine of the A-chain interacts in a substrate-like manner with the S1 specificity site of the active site of a crystallographically neighboring molecule. Based on this interaction and the structure of D-PheProArg methylene-thrombin, a model of the commonly used dansylGluGlyArg methylene inhibitor-factor Xa interaction is proposed. The region of factor Xa corresponding to the fibrinogen recognition site of thrombin has a reversed electrical polarity to the anion binding fibrinogen recognition site of thrombin but possesses a site similar to the Ca2+ binding site of trypsin and other serine proteinases. The structure of the C-terminal EGF domain of factor Xa is the first to be determined crystallographically. Its folding has been comprehensively compared with similar domains determined by NMR. Although the A-chain makes 44 contacts at less than 3.5 A with the catalytic domain, only 16 involve the EGF module. In addition, the A-chain makes 30 intermolecular contacts with a neighboring catalytic domain.

Mechanism of enhanced kinin release from high molecular weight kininogen by plasma kallikrein after its exposure to plasmin.

When purified high molecular weight kininogen was incubated with streptokinase-activated plasmin and kallikrein, a larger amount of kinin was released than would have been predicted from the effect of either enzyme alone. To determine the mechanism of this enhancement, high molecular weight kininogen was digested sequentially with these enzymes, and the rates of kinin release and sites of cleavage were determined. Conversion of 133 kd native high molecular weight kininogen to two-chain 112 kd or 102 kd derivatives by plasmin more than doubled the rate of kinin release by kallikrein. Conversely, digestion of high molecular weight kininogen by kallikrein and then plasmin did not enhance the rate of kinin release. The kallikrein cleavage points that provided 112 kd and 102 kd two-chain high molecular weight kininogen were after residues 437 (Arg-Lys) and 389 (Arg-Ser), whereas those for plasmin were after 438 (Lys-His) and 389 (Arg-Ser). epsilon-Aminocaproic acid, which competes for lysine residues that are critical to the binding of plasminogen or plasmin to substrates, inhibited the digestion of high molecular weight kininogen by plasmin, which is consistent with the evidence that the 438-439 Lys-His was a primary site of plasmin attack on high molecular weight kininogen. Furthermore, this cleavage was observed when plasminogen activation was induced in normal and in prekallikrein or Hageman factor-deficient plasmas. We suggest that the generation of fibrinolytic activity in blood could result in enhanced kinin release by kallikrein in regions of inflammation as a result of the collaborative actions of plasmin and kallikrein on high molecular weight kininogen.

Stilbene disulfonic acids. CD4 antagonists that block human immunodeficiency virus type-1 growth at multiple stages of the virus life cycle.

The stilbene disulfonic acids 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS), 4,4'-diisothiocyanatodihydrostilbene-2,2'-disulfonic acid and, 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulfonic acid bound the variable-1 immunoglobulin-like domain of CD4 on JM cells. The interaction blocked the binding of the anti-CD4 monoclonal antibody OKT4A and the envelope glycoprotein gp120 of the human immunodeficiency virus type-1 (HIV-1). DIDS inhibited the acute infection of CD4+ cells by HIV-1 with a potency (IC50 approximately 30 microM) similar to that which blocked gp120 binding (IC50 approximately 20 microM) to the cellular antigen. Pretreating uninfected CD4+ C8166 cells with DIDS blocked their fusion with chronically infected gp120+ cells. DIDS covalently and selectively modified lysine 90 of soluble CD4 and abolished the gp120-binding and antiviral properties of the recombinant protein. When added to cells productively infected with HIV-1, DIDS blocked virus growth and cleared cultures of syncytia without inhibiting cellular proliferation. The stilbene disulfonic acids are a novel class of site-specific CD4 antagonists that block multiple CD4-dependent events associated with acute and established HIV-1 infections.

Studies on the anticoagulant, antimetastatic and heparin-binding properties of ghilanten-related inhibitors.

The purpose of this study was to investigate the structure-activity relationships of ghilanten, an anticoagulant-antimetastatic protein of the South American leech Haementeria ghilianii. Five sequence-related variants of ghilanten, termed P1-P5, were purified and were shown to potently block the active-site hydrolysis of methoxycarbonyl-D-cyclohexylglycyl-glycyl-arginine-p-nitroanilide acetate by the human blood coagulation enzyme factor Xa; inhibition was rapid and stoichiometric. The amino acid sequence of P5 revealed a consensus sequence for heparin-binding at the carboxy-terminus. A synthetic peptide homologous to this region (93P-N-G-L-K-R-D-K-L-G-C-E-Y-C-E-C-R-P-K-R-K-L-I-P-R-L-S119) bound 125I-labelled heparin maximally at physiological pH and salt concentration. When administered intravenously to mice, the peptide suppressed lung metastases although less potentially than whole ghilanten. These findings suggest that the carboxy-terminal heparin-binding region may play a role in the antimetastatic action of the inhibitor.

Amino acid sequence of ghilanten: anticoagulant-antimetastatic principle of the South American leech, Haementeria ghilianii.

This study reports the amino acid sequence of ghilanten, an anticoagulant-antimetastatic principle of the hematophagous leech, Haementeria ghilianii. Ghilanten consists of 119 amino acids with twenty cysteines and a consensus sequence for heparin-binding at its carboxyl-terminus. Arginine-34 represents the reactive residue involved in the active-site inhibition of trypsin and Factor Xa. Immunoreactivity data suggest that heterogeneity among ghilantens is due in part to amino acid substitutions at their carboxyltermini.

Curtatoxins. Neurotoxic insecticidal polypeptides isolated from the funnel-web spider Hololena curta.

Three polypeptide neurotoxins (curtatoxins) were isolated from the venom of the spider Hololena curta by reverse-phase high performance liquid chromatography, gel permeation, and ion-exchange chromatography. The purified toxins induced an immediate paralysis in the cricket Acheta domestica that resulted in desiccation and death of the insect within 24-48 h (LD50 congruent to 4-20 micrograms/g); this toxic effect is consistent with irreversible presynaptic neuromuscular blockade. Curtatoxins are a class of sequence-related, cysteine-rich, carboxyl-terminal amidated polypeptides of 36 to 38 amino acid residues. The cysteine residues are conserved at identical sequence positions among these polypeptides and form 4 intramolecular disulfide bonds. Hydropathy calculations show that the toxins have an internal hydrophobic region flanked by hydrophilic and oppositely charged amino- and carboxyl-terminal ends. By analogy to other cysteine-rich arthropod venom proteins, the folded structure of the curtatoxins is likely important for their target specificity and mode of action at the neuromuscular junction.

Ghilantens: anticoagulant-antimetastatic proteins from the South American leech, Haementeria ghilianii.

Five proteins with anticoagulant and antimetastatic activities were isolated from the salivary glands of the Amazon leech, Haementeria ghilianil. These proteins, designated ghilantens, were co-purified on DEAE-cellulose and heparin-agarose, and were purified by microbore C-18 reverse-phase HPLC. Each variant had a similar molecular weight (18,000), amino acid composition, and a blocked amino terminus. Ghilantens caused a dose-dependent prolongation of the prothrombin time of normal human plasma and blocked the factor Xa-mediated hydrolysis of methoxycarbonyl-D-cyclohexylglycyl-glycl-arginine-p-nitro anillide acetate. Ghilantens were quantitatively absorbed to bovine factor Xa-AffiGel-15 and were eluted with 0.1 mol/L benzamidine, an active-site reversible inhibitor of factor Xa. These findings show that ghilantens can form a reversible association with the enzyme. When administered intravenously to mice by tall vein injection, ghilantens potently suppressed lung metastases of B16-F10 melanoma cells. These findings suggest that ghilantens may have therapeutic value in the treatment of metastatic disease.

High-sensitivity amino acid analysis by derivatization with O-phthalaldehyde and 9-fluorenylmethyl chloroformate using fluorescence detection: applications in protein structure determination.

An amino acid analysis method using a commercially available analyzer that accurately quantitates protein-derived amino acids in the 10-100 pmol range is described. The method utilizes the robotic capability of the analyzer's autosampler to perform precolumn derivatization of both primary and secondary amino acids with o-phthalaldehyde and 9-fluorenylmethyl chloroformate, respectively. The derivatized amino acids are then separated on a C-18 reverse-phase amino acid column and quantitated in a single run by fluorescence detection. The characterization of beta-lactoglobulin and two tryptic peptides from the bacterial enzyme diaminopimelic acid epimerase is used to demonstrate the sensitivity and utility of this method.

Isolation and cDNA sequence of human postheparin plasma hepatic triglyceride lipase.

Hepatic triglyceride lipase (H-TGL) was isolated from human postheparin plasma by column chromatography on heparin-Sepharose and phenyl-Sepharose and immunoaffinity chromatography with monoclonal antibodies. The purified enzyme had an apparent molecular weight of 65,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis and an amino-terminal sequence of Leu-Gly-Gln-Ser-Leu-Lys-Pro-Glu. Partial amino acid sequences of seven cyanogen bromide peptides were obtained. A human hepatoma cDNA library was screened with synthetic oligonucleotides derived from the partial protein sequence. The cloned H-TGL cDNA of 1569 nucleotides predicts a mature protein of 477 amino acids plus a leader sequence of 22 amino acids. Blot hybridization analysis of poly(A)+ mRNA with a putative H-TGL cDNA clone gave a single hybridizing band of 1.7 kilobases. The protein contains four consensus N-glycosylation sequences based on the cDNA sequence. Comparison of the enzyme sequence with that of other lipases reveals highly conserved sequences in regions of putative lipid and heparin binding. The carboxyl terminus of H-TGL contains a highly basic sequence which is not reported to be present in rat H-TGL or other members of the lipase gene family.

Isolation and characterization of four heparin-binding cyanogen bromide peptides of human plasma apolipoprotein B.

Apolipoprotein B-100 (apoB-100) is the major protein constituent of human plasma low-density lipoproteins (LDL). On the basis of its amino acid sequence [Chen, S.-H., Yang, C.-Y., Chen, P.-F., Setzer, D., Tanimura, M., Li, W.-H., Gotto, A. M., Jr., & Chan, L. (1986) J. Biol. Chem. 261, 12918-12921], apo B-100 is one of the largest monomeric proteins known with a calculated molecular weight of 512937. Heparin binds to the LDL surface by interacting with positively charged amino acid residues of apoB-100, forming soluble complexes in the absence of divalent metals and insoluble complexes in their presence. The purpose of this study was to isolate and characterize the heparin-binding domain(s) of apoB-100. Human plasma LDL were fragmented with cyanogen bromide (CNBr). After delipidation and reduction-carboxymethylation, the CNBr peptides were fractionated by sequential chromatography on DEAE-Sephacel, Mono S, and high reactive heparin (HRH) AffiGel-10; HRH was purified by chromatography of crude bovine lung heparin on LDL AffiGel-10. Heparin-binding peptides were further purified by reverse-phase high-performance liquid chromatography. Heparin-binding activity was monitored by a dot-blot assay with 125I-HRH. The amino-terminal sequences of four CNBr heparin-binding peptides (CNBr-I-IV) were determined. CNBr-I-IV correspond to residues 2016-2151, 3109-3240, 3308-3394, and 3570-3719, respectively, of the amino acid sequence of apoB-100. Each CNBr peptide contains a domain(s) of basic amino acid residues which we suggest accounts for their heparin-binding activity.(ABSTRACT TRUNCATED AT 250 WORDS)

In vitro processing of rANF7-28-NH2 by rat kidney homogenates.

The major rat kidney in vitro degradation products of the rat atrial natriuretic factor analog, H-Cys7-Phe-Gly-Gly-Arg-Ile-Asp-Arg-Ile-Gly-Ala-Gln-Ser-Gly-Leu-Gly -Cys-Asn-Ser-Phe-Arg-Tyr28-NH2 (with a Cys27-Cys23 disulfide bridge), have been identified. The degradation products are the C-terminal modified compounds rANF7-27, rANF7-26, and rANF7-25.

Rapid purification and revised amino-terminal sequence of hirudin: a specific thrombin inhibitor of the bloodsucking leech.

Hirudin is a specific polypeptide thrombin inhibitor consisting of 65 amino acids that is produced by the leech, Hirudo medicinalis. We describe a rapid method for the purification of hirudin from a leech extract. Crude hirudin, purchased from a commercial source, was first fractionated on a DEAE-HPLC column using a salt gradient. Hirudin activity was monitored by inhibition of the thrombin-mediated hydrolysis of a synthetic substrate H-D-Phenylalanyl-Pipecolyl-Arginine-p-Nitroanilide. The fractions containing antithrombin activity were pooled and further purified by reverse-phase HPLC. The homogeneity of purified hirudin was confirmed by a single amino-terminal sequence for 43 residues with Val-Val as the first two amino acids. Residue 33 was Asn rather than Asp as reported previously.

Structural organization of apolipoprotein B-100 of human plasma low density lipoproteins. Comparison to B-48 of chylomicrons and very low density lipoproteins.

The present study characterizes the substructural organization of apolipoprotein B-100 (B-100) of human plasma low density lipoproteins (LDL) and its relation to B-74 and B-26 of LDL and B-48 of chylomicrons and very low density lipoproteins. LDL were digested with human kallikrein and thrombin to yield two major fragments: K1 (Mr 410,000) and K2 (Mr 145,000) and T1 (Mr 385,000) and T2 (Mr 170,000), respectively. The antigenic sequences, Mr, and amino acid compositions of K1 and K2 were identical to those of plasma B-74 and B-26; B-26 and K2 had identical NH2-terminal sequences and correspond to the NH2-terminal region of B-100. K1 was further degraded by kallikrein to give K3 (Mr 235,000) and K4 (Mr 170,000); these peptides correspond immunochemically to two newly discovered plasma LDL peptides B-44 and B-30 and are assigned as complementary fragments of B-74. The thrombin cleavage fragments, T1 and T2, did not correspond to B-74 and B-26. Neither kallikrein nor thrombin generated a fragment from B-100 corresponding to B-48 in chylomicrons. However, B-48 showed antigenic homology with B-26 and to the of B-74 adjoining B-26, indicating that its structure is represented in the NH2-terminal half of B-100. The structural studies further clarify the relatedness among the B-100, B-74, B-26, and B-48 polypeptides and should now make possible the delineation of the functional domains mediating the interactions of apolipoprotein B in the circulation and arterial wall.

Binding of a high reactive heparin to human apolipoprotein E: identification of two heparin-binding domains.

Ligand-blotting and dot-blotting procedures were used to investigate the binding of [125I]-heparin to apolipoprotein E, its thrombin fragments E22 (residues 1-191) and E12 (residues 192-299), and to nine apolipoprotein E synthetic fragments. E22 and E12 bound [125I] heparin indicating multiple heparin-binding domains. Synthetic peptides of apoE corresponding to residues 129-169, 139-169, and 144-169, but not 148-169, bound [125I] heparin suggesting that residues 144-147 (Leu-Arg-Lys-Arg) in E22 are important for binding. Peptide 202-243 and 211-243 but not 219-243 bound [125I] heparin suggesting that residues 211-218 (Gly-Glu-Arg-Leu-Arg-Ala-Arg-Met) comprise a portion of the E12 heparin-binding domain.

Isolation, sequence determination and expression in Escherichia coli of the isopenicillin N synthetase gene from Cephalosporium acremonium.

The enzyme isopenicillin N synthetase (IPS) catalyses the oxidative condensation of delta-(L-alpha-aminoadipyl)-L-cysteinyl-D-valine (LLD-ACV) to isopenicillin N, which is a central reaction in the pathway to clinically important penicillins and cephalosporins. Here we report the cloning, characterization and expression in Escherichia coli of the gene encoding the IPS protein in Cephalosporium acremonium. The IPS gene was identified by purifying IPS protein, determining the first 23 amino-terminal amino acids, preparing a set of synthetic oligonucleotides encoding a portion of the determined amino-acid sequence, and probing a cosmid genome library with the mixed oligonucleotides. A cosmid hybridizing with the probe was isolated and the IPS gene was localized and sequenced. The IPS gene encodes a polypeptide of relative molecular mass (Mr) 38,416. When this open reading frame was cloned into an E. coli expression vector and inserted into E. coli, the recombinant E. coli produced a new protein co-migrating with authentic IPS as the major protein of the cell (approximately 20% of cell protein). Crude cell extracts condensed LLD-ACV to a penicillinase-sensitive molecule whose antibacterial activity indicated that it was isopenicillin N.

Purification and characterization of dihydrofolate reductase from methotrexate-resistant human lymphoblastoid cells.

Dihydrofolate reductase has been isolated from methotrexate-resistant WIL2 human lymphoblastoid cells. This subline produces ca. 150 times more enzyme than the parental drug-sensitive line. The reductase has been purified to homogeneity by methotrexate affinity chromatography, gel filtration, and preparative isoelectric focusing in a yield of 65%. The enzyme has a pI = 7.7 and a molecular weight of ca. 22000. The amino-terminal 27 amino acid residues have been determined, revealing the location of the single cysteine residue at position 6. Reaction of this cysteine with p-(hydroxy-mercuri)benzoate in the presence of reduced nicotinamide adenine dinucleotide phosphate (NADPH) results in a 5-fold increase in enzyme activity. Other agents including KCl, urea, and thiourea also cause enzyme activation. The Km values for NADPH and dihydrofolate are 0.25 and 0.036 microM, respectively. Mercurial activation of the enzyme results in a 27-fold increase in the Km for NADPH and a 35-fold increase in the Km for dihydrofolate.

Lysine and ornithine analogues of methotrexate as inhibitors of dihydrofolate reductase.

The ornithine (6a) and lysine (6b) analogues of methotrexate (1) have been synthesized via condensation of 4-amino-4-deoxy-N10-methylpteroic acid (2) with N gamma-carbobenzoxy-L-ornithine tert-butyl ester (3a) and N epsilon-carbobenzoxy-L-lysine tert-butyl ester (3b), respectively. Removal of the protecting groups gave 5a and 6b. Compounds 6a and 6b and their precursor Cbz acids (5a and 5b) show significant inhibition of dihydrofolate reductase.