Positional and additive effects of basic amino acids on processing of precursor proteins within the constitutive secretory pathway

Thrombin activation of protein C requires prior processing by a liver proprotein convertase

Protein C, a secretory vitamin K-dependent anticoagulant serine protease, inactivates factors Va/VIIIa. It is exclusively synthesized in liver hepatocytes as an inactive zymogen (proprotein C). In humans, thrombin cleavage of the propeptide at PR²²¹↓ results in activated protein C (APC; residues 222-461). However, the propeptide is also cleaved by a furin-like proprotein convertase(s) (PCs) at KKRSHLKR¹⁹⁹↓ (underlined basic residues critical for the recognition by PCs), but the order of cleavage is unknown. Herein, we present evidence that at the surface of COS-1 cells, mouse proprotein C is first cleaved by the convertases furin, PC5/6A, and PACE4. In mice, this cleavage occurs at the equivalent site, KKRKILKR¹⁹⁸↓, and requires the presence of Arg¹⁹⁸ at P1 and a combination of two other basic residues at either P2 (Lys¹⁹⁷), P6 (Arg¹⁹³), or P8 (Lys¹⁹¹) positions. Notably, the thrombin-resistant R221A mutant is still cleaved by these PCs, revealing that convertase cleavage can precede thrombin activation. This conclusion was supported by the fact that the APC-specific activity in the medium of COS-1 cells is exclusively dependent on prior cleavage by the convertases, because both R198A and R221A lack protein C activity. Primary cultures of hepatocytes derived from wild-type or hepatocyte-specific furin, PC5/6, or complete PACE4 knock-out mice suggested that the cleavage of overexpressed proprotein C is predominantly performed by furin intracellularly and by all three proprotein convertases at the cell surface. Indeed, plasma analyses of single-proprotein convertase-knock-out mice showed that loss of the convertase furin or PC5/6 in hepatocytes results in a ∼30% decrease in APC levels, with no significant contribution from PACE4. We conclude that prior convertase cleavage of protein C in hepatocytes is critical for its thrombin activation.

Effect of C-terminal sequence on competitive semaphorin binding to neuropilin-1

Neuropilins (Nrp) are type I transmembrane proteins that function as receptors for vascular endothelial growth factor (VEGF) and class III Semaphorin (Sema3) ligand families. Sema3s function as potent endogenous angiogenesis inhibitors but require proteolytically processing by furin to compete with VEGF for Nrp binding. This processing liberates a C-terminal arginine (CR) that is necessary for binding to the b1 domain of Nrp, a common feature shared by Nrp ligands. The CR is necessary but not sufficient for potent Nrp inhibition, and the role of upstream residues is unknown. We demonstrate that the second-to-last residue (C-1), immediately upstream of the CR, plays a significant role in controlling competitive ligand binding by orienting the C-terminus for productive Nrp binding. With the use of a peptide library derived from Sema3F, C-1 residues that preferentially adopt an extended bound-like conformation, including proline and β-branched amino acids, were found to produce the most avid competitors. Consistent with this, analysis of the binding thermodynamics revealed that more favorable entropy is responsible for the observed binding enhancement of C-1 proline. We further tested the effect of the C-1 residue on Sema3F processing by furin and found an inverse relationship between processing and inhibitory potency. Analysis of all Sema3 family members reveals two non-equivalent furin processing sites differentiated by the presence of either a C-1 proline or a C-1 arginine and resulting in up to a 40-fold difference in potency. These data reveal a novel regulatory mechanism of Sema3 activity and define a fundamental mechanism for preferential Nrp binding.

Proprotein convertases process Pmel17 during secretion

Pmel17 is a melanocyte/melanoma-specific protein that traffics to melanosomes where it forms a fibrillar matrix on which melanin gets deposited. Before being cleaved into smaller fibrillogenic fragments the protein undergoes processing by proprotein convertases, a class of serine proteases that typically recognize the canonical motif RX(R/K)R↓. The current model of Pmel17 maturation states that this processing step occurs in melanosomes, but in light of recent reports this issue has become controversial. We therefore addressed this question by thoroughly assessing the processing kinetics of either wild-type Pmel17 or a secreted soluble Pmel17 derivative. Our results demonstrate clearly that processing of Pmel17 occurs during secretion and that it does not require entry of the protein into the endocytic system. Strikingly, processing proceeds even in the presence of the secretion inhibitor monensin, suggesting that Pmel17 is an exceptionally good substrate. In line with this, we find that newly synthesized surface Pmel17 is already quantitatively cleaved. Moreover, we demonstrate that Pmel17 function is independent of the sequence identity of its unconventional proprotein convertase-cleavage motif that lacks arginine in P4 position. The data alter the current view of Pmel17 maturation and suggest that the multistep processing of Pmel17 begins with an early cleavage during secretion that primes the protein for later functional processing.

Evaluation of germline BMP4 mutation as a cause of colorectal cancer

Transforming growth factor-â (TGF-â) signalling plays a key role in colorectal cancer (CRC). Bone morphogenetic protein-4 (BMP4) is a member of the TGF-â family of signal transduction molecules. To examine if germline mutation in BMP4 causes CRC we analysed 504 genetically enriched CRC cases (by virtue of early-onset disease, family history of CRC) for mutations in the coding sequence of BMP4. We identified three pathogenic mutations, p.R286X (g.8330C>T), p.W325C (g.8449G>T) and p.C373S (g.8592G>C), amongst the CRC cases which were not observed in 524 healthy controls. p.R286X localizes to the N-terminal of the TGF-â1 prodomain truncating the protein prior to the active domain. p.W325C and p.C373S mutations are predicted from protein homology modelling with BMP2 to impact deleteriously on BMP4 function. Segregation of p.C373S with adenoma and hyperplastic polyp in first-degree relatives of the case suggests germline mutations may confer a juvenile polyposis-type phenotype. These findings suggest mutation of BMP4is a cause of CRC and the value of protein-based modelling in the elucidation of rare disease-causing variants. © 2010 Wiley-Liss, Inc.

Expression, purification and characterization of human IFN-lambda1 in Pichia pastoris

Interferon-lambda (IFN-lambda) is a newly identified IFN family which belongs to the class II cytokines. The three members of this family represent antiviral activities like other IFNs. In the present study, recombinant human IFN-lambda1 (rhIFN-lambda1) was produced by using the methylotrophic yeast Pichia pastoris (P. pastoris) expression system. cDNAs encoding amino acids 23-200 or 20-200 of human IFN-lambda1 were cloned and joined to sequence encoding the leader region (prepro segment) of the precursor of Saccharomyces cerevisiae alpha-factor. The two hybrid genes were subcloned into yeast integrative vector pAO815 separately to construct expression plasmids bearing four tandem copies of IFN-lambda1 expression cassettes. The expression plasmids were then used to transform into P. pastoris strain GS115, resulting in recombinant strains GS115/IFNlambda1P and GS115/IFNlambda1G with Mut(+) or Mut(s) phenotype. rhIFN-lambda1 was secreted into the medium upon methanol induction. In GS115/IFNlambda1P, however, KEX2 cleavage for mature rhIFN-lambda1 generation was inhibited by a proline at [Formula: see text] and the products were different from anticipation. GS115/IFNlambda1G strain secreted two forms of mature rhIFN-lambda1 with the same N-terminal sequence and different molecular weight. Periodic acid-Schiff (PAS) staining indicated that these proteins were glycosylated. The yield of low-glycosylated rhIFN-lambda1 in GS115/IFNlambda1G strain was approximately 65 mg l(-1) in shaking flasks, representing around 57% of the total secreted proteins. rhIFN-lambda1 was purified by cation exchange chromatography and gel filtration. The purified rhIFN-lambda1 showed specific efficiency to activate signal transducer and activator of transcription 1 (STAT1) and STAT2 that was comparable to that of commercial IFNalpha2a.

Mesothelin variant 1 is released from tumor cells as a diagnostic marker

The mesothelin family comprises (at least) three variants and includes the precursor for megakaryocyte potentiating factor (MPF). Assaying soluble mesothelin-related protein (SMRP) molecules in serum and other body fluids from patients with certain cancers can provide diagnostically useful information. We have constructed fusion proteins of mesothelin variants 1, 2, and 3, made monoclonal antibodies, and investigated the binding specificity of these and three previously generated monoclonal antibodies to each of the three mesothelin variants. According to flow cytometry, the molecule that is most frequently expressed at the surface of cells from ovarian carcinomas and certain other tumors is mesothelin variant 1. Similarly, SMRP released into ascites from a patient with ovarian carcinoma was shown to have a molecular weight of approximately 40 kDa and, according to sequencing, to be variant 1. A published sandwich ELISA was shown to detect variants 1 and 3 and to be much more sensitive than a newly constructed ELISA, which detects only variant 3, the former being positive in 28 of 41 (68%) sera from patients with ovarian cancer as compared with 6 of 41 sera (15%). A standard curve was constructed to measure SMRP with a limit of detection of 200 pg/mL to facilitate future quantitative studies.

Bridging neuropeptidomics and genomics with bioinformatics: Prediction of mammalian neuropeptide prohormone processing

Neuropeptides are an important class of cell to cell signaling molecules that are difficult to predict from genetic information because of their large number of post-translational modifications. The transition from prohormone genetic sequence information to the determination of the biologically active neuropeptides requires the identification of the cleaved basic sites, among the many possible cleavage sites, that exist in the prohormone. We report a binary logistic regression model trained on mammalian prohormones that is more sensitive than existing methods in predicting these processing sites, and demonstrate the application of this method to mammalian neuropeptidomic studies. By comparing the predictive abilities of a binary logistic model trained on molluscan prohormone cleavages with the reported model, we establish the need for phyla-specific models.

Processing of alpha4 integrin by the proprotein convertases: histidine at position P6 regulates cleavage

The proprotein convertases (PCs) participate in the limited proteolysis of integrin alpha4 subunit at the H(592)VISKR(597) downward arrow ST site (where underlined residues indicate positively charged amino acids important for PC-mediated cleavage and downward arrow indicates the cleavage site), since this cleavage is inhibited by the serpin alpha1-PDX (alpha1-antitrypsin Portland). Co-expression of alpha4 with each convertase in LoVo (furin-deficient human colon carcinoma) cells revealed that furin and proprotein convertase 5A (PC5A) are the best pro-alpha4 convertases. In agreement, processing of endogenous pro-alpha4 in human lymphoblastoid CEM-T4 cells was enhanced greatly in stable transfectants overexpressing either enzyme. In many leucocyte cell lines, the expression of furin closely correlated with the endogenous processing efficacy, suggesting that furin is a candidate pro-alpha4 convertase. Mutational analysis showed that replacement of P1 Arg(597) with alanine (R597A) abrogated cleavage, whereas the P6 mutant H592R is even better processed by the endogenous convertases of Chinese-hamster ovary CHO-K1 cells. In vitro kinetic studies using synthetic peptides confirmed the importance of a positively charged residue at P6 and showed that wild-type alpha4 processing is performed best by furin and PC5A at acidic and neutral pHs, respectively. Biosynthetic analysis of pro-alpha4 and its H592R and H592K mutants in the presence or absence of the weak base, NH(4)Cl, revealed that the P6 histidine residue renders its processing by furin sensitive to cellular pH. This suggests that pro-alpha4 cleavage occurs preferentially in acidic compartments. In conclusion, although the accepted furin processing motif is Arg-Xaa-(Lys/Arg)-Arg downward arrow, our data further extend it to include a regulatory histidine residue at P6 in precursors that lack a basic residue at P4.

Human mannose 6-phosphate-uncovering enzyme is synthesized as a proenzyme that is activated by the endoprotease furin

N-Acetylglucosamine-1-phosphodiester alpha-N-acetylglucosaminidase, also known as "uncovering" enzyme (UCE), is localized in the trans-Golgi network, where it removes a covering N-acetylglucosamine from the mannose 6-phosphate recognition marker on lysosomal acid hydrolases. Here we show that UCE is synthesized as an inactive proenzyme that is activated by the endoprotease furin, which cleaves an RARLPR/D sequence to release a 24-amino acid propiece. As furin is localized in the trans-Golgi network, newly synthesized UCE is inactive until it reaches this terminal Golgi compartment. LoVo cells (derived from a human colon adenocarcinoma) lack furin activity and have extremely low UCE activity. Addition of furin to LoVo cell extracts restores UCE activity to normal levels, demonstrating that the UCE proenzyme is stable in this cell type. LoVo cells secrete acid hydrolases with phosphomannose diesters as a consequence of the deficient UCE activity. This demonstrates for the first time that UCE is the only enzyme in these cells capable of efficiently uncovering phosphomannose diesters. UCE also hydrolyzes UDP-GlcNAc, a sugar donor for Golgi N-acetylglucosaminyltransferases. The fact that UCE is not activated until it reaches the trans-Golgi network may ensure that the pool of UDP-GlcNAc in the Golgi stack is not depleted, thereby maintaining proper oligosaccharide assembly.

Localization of functional domains of the mitogenic toxin of Pasteurella multocida

The locations of the catalytic and receptor-binding domains of the Pasteurella multocida toxin (PMT) were investigated. N- and C-terminal fragments of PMT were cloned and expressed as fusion proteins with affinity tags. Purified fusion proteins were assessed in suitable assays for catalytic activity and cell-binding ability. A C-terminal fragment (amino acids 681 to 1285) was catalytically active. When microinjected into quiescent Swiss 3T3 cells, it induced changes in cell morphology typical of toxin-treated cells and stimulated DNA synthesis. An N-terminal fragment with a His tag at the C terminus (amino acids 1 to 506) competed with full-length toxin for binding to surface receptors and therefore contains the cell-binding domain. The inactive mutant containing a mutation near the C terminus (C1165S) also bound to cells in this assay. Polyclonal antibodies raised to the N-terminal PMT region bound efficiently to full-length native toxin, suggesting that the N terminus is surface located. Antibodies to the C terminus of PMT were microinjected into cells and inhibited the activity of toxin added subsequently to the medium, confirming that the C terminus contains the active site. Analysis of the PMT sequence predicted a putative transmembrane domain with predicted hydrophobic and amphipathic helices near the N terminus over the region of homology to the cytotoxic necrotizing factors. The C-terminal end of PMT was predicted to be a mixed alpha/beta domain, a structure commonly found in catalytic domains. Homology to proteins of known structure and threading calculations supported these assignments.

Familial British dementia: expression and metabolism of BRI

Vidal et al. (1999. Nature 399: 776-778) discovered that the underlying genetic lesion in familial British dementia (FBD) is a T-A transversion at the termination codon of a membrane protein, termed BRI. The mutation creates an arginine codon; translational read-through generates a novel protein, termed BRI-L, that is extended by 11 amino acids at the carboxyl-terminus. BRI-L is the precursor of the ABri peptide, a component of amyloid deposits in FBD brain. We demonstrate that both BRI and its mutant counterpart are constitutively processed by furin, resulting in the secretion of carboxyl-terminal peptide derivatives that correspond to all, or part of, ABri. Notably, elevated levels of peptides are generated from the mutant BRI precursor, suggesting that subtle conformational alterations at the carboxyl-terminus may influence furin-mediated processing. We have examined BRI/BRI-L processing by other members of the prohormone convertase (PC) family (PACE4, LPC, PC 5/6) and found that these enzymes also process BRI, albeit inefficiently. Moreover, BRI-L processing by the other PC members is severely compromised. Finally, our electron microscopic studies reveal that synthetic ABri peptides assemble into insoluble beta-pleated fibrils. Collectively, our results support the view that enhanced furin-mediated processing of mutant BRI generates amyloidogenic peptides that initiate the pathogenesis of FBD.

Monoclonal antibodies to the carboxy-terminal Ea sequence of pro-insulin-like growth factor-IA (proIGF-IA) recognize proIGF-IA secreted by IM9 B-lymphocytes

Insulin-like growth factor-I (IGF-I) circulates in human serum as a 7 kDa peptide but analysis of IGF-I cDNAs predicts two pro-hormone precursors (proIGF-IA and proIGF-IB) with distinct C-terminal E domains. The function of these precursors, and the E peptides generated on cleavage to mature IGF-I, is unknown, largely because of a lack of tools for distinguishing precursors from constituent peptides. We used a synthetic Ea peptide to develop monoclonal antibodies (MAbs) which can recognize the carboxy-terminal sequence of proIGF-IA. These were characterized using proIGF-IA generated by transfected HEK293 cells. The anti-proIGF-IA MAbs immunoprecipitated two peptides (19--21 and 14 kDa) which were also recognized by MAbs to mature IGF-I. The proIGF-IA MAbs could also detect peptides of 9 and 4 kDa predicted to be Ea peptides. Treatment with N -glycosidase proved the 19--21 kDa and 9 kDa bands to be glycosylated proIGF-IA and Ea peptide respectively. Using these antibodies, we have identified proIGF-IA secreted from the IM9 B-lymphocyte cell line. This work paves the way for studies on proIGF-IA and Ea peptide regulation and function.

Engineered eglin c variants inhibit yeast and human proprotein processing proteases, Kex2 and furin

We engineered eglin c, a potent subtilisin inhibitor, to create inhibitors for enzymes of the Kex2/furin family of proprotein processing proteases. A structural gene was synthesized that encoded "R(1)-eglin", having Arg at P(1) in the reactive site loop in place of Leu(45). Ten additional variants were created by cassette mutagenesis of R(1)-eglin. These polypeptides were expressed in Escherichia coli, purified to homogeneity, and their interactions with secreted, soluble Kex2 and furin were examined. R(1)-eglin itself was a modest inhibitor of Kex2, with a K(a) of approximately 10(7) M(-)(1). Substituting Arg (in R(4)R(1)-eglin) or Met (in M(4)R(1)-eglin) for Pro(42) at P(4) created potent Kex2 inhibitors exhibiting K(a) values of approximately 10(9) M(-)(1). R(4)R(1)-eglin inhibited furin with a K(a) of 4.0 x 10(8) M(-)(1). Introduction of Lys at P(1), in place of Arg in R(4)R(1)-eglin reduced affinity only approximately 3-fold for Kex2 but 15-fold for furin. The stabilities of enzyme-inhibitor complexes were characterized by association and dissociation rate constants and visualized by polyacrylamide gel electrophoresis. R(4)R(1)-eglin formed stable 1:1 complexes with both Kex2 and furin. However, substitution of Lys at P(2) in place of Thr(44) resulted in eglin variants that inhibited both Kex2 and furin but which were eventually cleaved (temporary inhibition). Surprisingly, R(6)R(4)R(1)-eglin, in which Arg was substituted for Gly(40) in R(4)R(1)-eglin, exhibited stable, high-affinity complex formation with Kex2 (K(a) of 3.5 x 10(9) M(-)(1)) but temporary inhibition of furin. This suggests that enzyme-specific interactions can alter the conformation of the reactive site loop, converting a permanent inhibitor into a substrate. Eglin variants offer possible avenues for affinity purification, crystallization, and regulation of proprotein processing proteases.

Egg envelope glycoprotein gp37 as a Xenopus homolog of mammalian ZP1, based on cDNA cloning

The egg envelope is a kind of extracellular matrix, which surrounds growing oocytes, ovulated eggs and early embryos. Among the glycoprotein components of the Xenopus laevis egg envelope, gp43/gp41 and gp69/64 have already been shown to be frog homologs of the mammalian zona pellucida components ZP3 and ZP2, respectively. To determine the structure of another major component of egg envelope, gp37, the peptides isolated from the lysyl endopeptidase digests of gp37 were sequenced for amino acids to design degenerate primers for polymerase chain reaction. By reverse transcription-polymerase chain reaction with a poly(A)+ RNA from the ovary of a postovulated female Xenopus, a specifically amplified band was obtained and sequenced. The upstream and downstream sequences of the sequenced region were completed by 5'- and 3'-rapid amplification of cDNA ends, respectively. The gp37 cDNA comprises 1674 bp and contains one open reading frame encoding a polypeptide with 543 amino acids. The predicted amino acid sequence of the gp37 cDNA has a close similarity to that of mammalian ZP1. Northern blot and in situ hybridization studies indicated that the transcript (1.8 kb) is exclusively expressed in the oocytes, particularly in the previtellogenic young oocytes, just like the expression pattern of gp43 mRNA, suggesting a coordinate transcription of the gp43 and gp37 genes in Xenopus.

Mutational analysis of bli-4/kpc-4 reveals critical residues required for proprotein convertase function in C. elegans

Kex2/subtilisin-like proteinase activity is required for the production of the adult cuticle in the nematode Caenorhabditis elegans. Deletion of the carboxy termini of four of the bli-4/kpc-4 convertase isoforms results in blistering of the adult cuticle. The blisters vary in severity (expressivity) and are not evident in all individuals (reduced penetrance). We have isolated 13 bli-4/kpc-4 mutants that arrest development in late embryogenesis. Using a PCR-based heteroduplex technique, we have identified nucleotide changes responsible for eight of these lethal mutations. The lesions reside within the first 12 exons that are shared by all of the bli-4/kpc-4 gene products, with the majority of mutations clustered within the protease domain. This finding suggests that the protease domain represents a large mutable target. Among these mutations, allele h384 represents a molecular null mutant in which the catalytically essential serine residue (Ser415) is replaced by phenylalanine. Novel missense mutations that change the identity of amino acids evolutionary conserved in all kex2/subtilisin-convertases highlight critical residues essential for activity. We examined the functional activity of BLI-4/KPC-4 products expressed from several lethal mutants by testing their effect on the variable penetrance of blistering exhibited by the e937 allele. We found that the combination of a bli-4/kpc-4 lethal mutation in trans to the bli-4(e937) mutation was sufficient to cause severe blistering in heteroallelic progeny, even in the presence of a known dominant suppressor.

Cleavage of the respiratory syncytial virus fusion protein is required for its surface expression: role of furin

The fusion (F) glycoprotein of respiratory syncytial virus (RSV) is synthesized as a nonfusogenic precursor protein (F(0)), which during its migration to the cell surface is activated by cleavage into the disulfide-linked F(1) and F(2) subunits. In the present study, soluble secreted human furin produced by a recombinant baculovirus cleaved RSV F(0) into proteins the size of F(1) and F(2). Furthermore, cleavage of F(0) was partially inhibited in the furin defective LoVo cell line, in calcium depleted HEp-2 cells, and in HEp-2 cells treated with the furin inhibitor decanoyl-R-V-K-R-chloromethylketon. These findings strongly suggest an important role for furin in activation of the RSV F protein. The F(0) protein could not be detected on the surface of cells, in which F protein activation was inhibited, and RSV particles did not appear to be released from these cells. It thus seems that in contrast to the F proteins of most other paramyxoviruses, the RSV F(0) protein is very inefficient in reaching the cell surface or is unable to reach the cell surface and therefore cannot be incorporated into virus particles.

Furin mediates enhanced production of fibrillogenic ABri peptides in familial British dementia

The genetic lesion underlying familial British dementia (FBD), an autosomal dominant neurodegenerative disorder, is a T-A transversion at the termination codon of the BRI gene. The mutant gene encodes BRI-L, the precursor of ABri peptides that accumulate in amyloid deposits in FBD brain. We now report that both BRI-L and its wild-type counterpart, BRI, were constitutively processed by the proprotein convertase, furin, resulting in the secretion of carboxyl-terminal peptides that encompass all or part of ABri. Elevated levels of peptides were generated from the mutant BRI precursor. Electron microscopic studies revealed that synthetic ABri peptides assembled into irregular, short fibrils. Collectively, our results support the view that enhanced furin-mediated processing of mutant BRI generates fibrillogenic peptides that initiate the pathogenesis of FBD.

Molecular basis for oviductin-mediated processing from gp43 to gp41, the predominant glycoproteins of Xenopus egg envelopes

Acquisition of fertilizability in Xenopus coelomic eggs is correlated with the conversion from coelomic to vitelline envelope during passage of the eggs through the pars recta portion of oviduct. The conversion includes processing of a major envelope constituent gp43 of coelomic envelopes to gp41 of vitelline envelopes by a trypsin-type protease, oviductin, which is secreted from the pars recta. Our recent sequencing analyses [Kubo et al., (1997): Dev Growth Diff 39:405-411] strongly suggested that the N-terminal portion of gp41 is exposed as a result of oviductin digestion. In this study, a monoclonal antibody specific to the predicted N-terminus of gp41 was raised by immunizing mice with a synthetic N-terminal hexapeptide (QLPVSP) coupled to keyhole limpet hemocyanin. The antibody specifically reacted to gp41, but not to gp43, indicating that Gln62 is exposed as the N-terminal amino acid of gp41 by oviductin-mediated cleavage of gp43 at Arg61 in GSR61. The C-terminal sequencing of gp43 and gp41 indicated that Arg373 in GSR373 as the C-terminus of gp41 is generated by cleavage of three amino acid (WNQ) residues from the C-terminus of gp43. The resulting polypeptide moiety of gp41 has a molecular mass of 33900 Da with 312 amino acid residues. We propose that oviductin possessing the substrate specificity of GSR simultaneously digests gp43 at Arg residues in GSR61 and GSR373 to generate the N- and C-terminus of gp41, respectively.

The close homologue of the neural adhesion molecule L1 (CHL1): patterns of expression and promotion of neurite outgrowth by heterophilic interactions

The close homologue of L1 (CHL1), a member of the L1 family of neural adhesion molecules, is first expressed at times of neurite outgrowth during brain development, and is detectable in subpopulations of neurons, astrocytes, oligodendrocyte precursors and Schwann cells of the mouse and rat. Aggregation assays with CHL1-transfected cells show that CHL1 does not promote homophilic adhesion or does it mediate heterophilic adhesion with L1. CHL1 promotes neurite outgrowth by hippocampal and small cerebellar neurons in substrate-bound and soluble form. The observation that CHL1 and L1 show overlapping, but also distinct patterns of synthesis in neurons and glia, suggests differential effects of L1-like molecules on neurite outgrowth.

Nucleic acid sequence of feline preprorelaxin and its localization within the feline placenta

The cat placenta is known to secrete large amounts of relaxin. We employed uteroplacental tissue at approximately Day 35 of gestation to determine the nucleic acid sequence of feline preprorelaxin using reverse transcription- and rapid amplification of cDNA ends-polymerase chain reaction. Feline preprorelaxin cDNA was found to consist of 540 base pairs encoding a protein of 180 amino acids (aa). We identified a signal peptide of 25 aa, a B domain of 33 aa, a C domain of 98 aa, and an A domain of 24 aa. The putative receptor binding region in the N'-terminal part of the B domain contained one substitution from the classical GRELVR motif (L-->F). Feline preprorelaxin shared highest homology with porcine and equine preprorelaxin. Northern analysis revealed a specific 1-kilobase transcript present in total RNA of feline uteroplacental tissue but not of liver tissue. Nonradioactive in situ hybridization was used to localize relaxin mRNA, and immunohistochemistry was used to localize the relaxin hormone and cytokeratin, in tissues of the feto-maternal interface recovered from two queens at Day 35 of gestation. Specific hybridization signals for relaxin mRNA were exclusively detected in cells located in the lamellar placental labyrinth but were absent from other placental and nonplacental uterine parts. The cells expressing relaxin mRNA also displayed immunoreactivity for cytokeratin and were, therefore, identified as trophoblast cells. Immunoreactive relaxin colocalized in those placental areas expressing relaxin mRNA. Trophoblast cells located at the villous chorioallantoic tips invading the endometrium and extravillous trophoblast cells in the junctional placental zone were devoid of relaxin.

Regulation of bone morphogenetic protein activity by pro domains and proprotein convertases

Bone morphogenetic proteins (BMPs) are derived from inactive precursor proteins by endoproteolytic cleavage. Here we show that processing of Nodal and Myc-tagged BMP4 is significantly enhanced by SPC1/Furin or SPC4/PACE4, providing direct evidence that regulation of BMP signaling is likely to be controlled by subtilisin-like proprotein convertase (SPC) activities. Nodal processing is dramatically enhanced if two residues adjacent to the precursor cleavage site are substituted with amino acids found at the equivalent positions of Activin, demonstrating that structural constraints at the precursor cleavage site limit the processing efficiency. However, in transfection assays, mature Nodal is undetectable either in culture supernatants or in cell lysates, despite efficient cleavage of the precursor protein, suggesting that mature Nodal is highly unstable. Domain swap experiments support this conclusion since mature BMP4 or Dorsalin are also destabilized when expressed in conjunction with the Nodal pro domain. By contrast, mature Nodal is stabilized by the Dorsalin pro domain, which mediates the formation of stable complexes. Collectively, these data show that the half-life of mature BMPs is greatly influenced by the identity of their pro regions.

Prohormones: their clinical relevance

The majority of peptide hormones and neuropeptides are synthesized as precursors, which are cleaved in a sequence-specific and tissue-specific manner to yield the biologically active peptides. There has been considerable progress in the past ten years in understanding the nature and mechanism of action of the prohormone convertases that cleave these prohormones. Evidence from knockout technology and clinical examples of gene mutations has provided functional information on disruption of prohormone cleavage and the bioactivity of prohormones. There are specific examples of the clinical relevance of circulating prohormones, such as adrenocorticotrophin (ACTH) precursors and proinsulin. The central issues that still remain are: (1) What is the relative importance of each of the different processing pathways and processing enzymes in regulating hormone action? (2) How do the serum concentrations of prohormones compare with the mature hormone levels? (3) What are the biological consequences of prohormones in the circulation?

Biosynthetic processing and quaternary interactions of proprotein convertase SPC4 (PACE4)

SPC4 (PACE4), a member of the eukaryotic family of subtilisin-like proprotein convertases, is synthesized as a proenzyme (proSPC4) which undergoes proteolytic removal of N-terminal propeptide during transit through the secretory pathway. As this propeptide processing seems to be a key event in the functional expression of SPC4, we have investigated its mechanism and the intracellular site where it occurs. In transfected fibroblast cells, the 110-kDa proSPC4 undergoes slow cleavage to generate a 103-kDa mature enzyme in the endoplasmic reticulum (ER). Site-directed mutagenesis studies demonstrate that the proteolytic activation of SPC4 occurs mainly through a unimolecular autocatalytic process and propeptide cleavage is a prerequisite for its export from the ER. Sedimentation velocity and chemical cross-linking analysis demonstrate that the precursor protein in the cells exists as both a monomer and a dimer-sized complex whereas mature SPC4 exists only as a monomer. These results suggest that the cleavage of the N-terminal propeptide of SPC4 plays a regulatory role in its activation and secretion through the change in its oligomeric state.

Biochemical analysis of the secreted and virion glycoproteins of Ebola virus

The glycoproteins expressed by a Zaire species of Ebola virus were analyzed for cleavage, oligomerization, and other structural properties to better define their functions. The 50- to 70-kDa secreted and 150-kDa virion/structural glycoproteins (SGP and GP, respectively), which share the 295 N-terminal residues, are cleaved near the N terminus by signalase. A second cleavage event, occurring in GP at a multibasic site (RRTRR downward arrow) that is likely mediated by furin, results in two glycoproteins (GP1 and GP2) linked by disulfide bonding. This furin cleavage site is present in the same position in the GPs of all Ebola viruses (R[R/K]X[R/K]R downward arrow), and one is predicted for Marburg viruses (R[R/K]KR downward arrow), although in a different location. Based on the results of cross-linking studies, we were able to determine that Ebola virion peplomers are composed of trimers of GP1-GP2 heterodimers and that aspects of their structure are similar to those of retroviruses, paramyxoviruses, and influenza viruses. We also determined that SGP is secreted from infected cells almost exclusively in the form of a homodimer that is joined by disulfide bonding.

Human ADAM 12 (meltrin alpha) is an active metalloprotease

The ADAMs (a disintegrin and metalloprotease) are a family of multidomain proteins with structural homology to snake venom metalloproteases. We recently described the cloning and sequencing of human ADAM 12 (meltrin alpha). In this report we provide evidence that the metalloprotease domain of ADAM 12 is catalytically active. We used the trapping mechanism of alpha2-macroglobulin to assay for protease activity of wild-type and mutant ADAM 12 proteins produced in a COS cell transfection system. We found that ADAM 12 is synthesized as a zymogen, with the prodomain maintaining the metalloprotease in a latent form, probably by means of a cysteine switch. The zymogen could be activated chemically by alkylation with N-ethylmaleimide. Cleavage of the prodomain at a site for a furin-like endopeptidase resulted in an ADAM 12 protein with proteolytic activity. The protease activity was sensitive to inhibition by 1,10-phenanthroline and could be eliminated by mutation of the critical glutamate residue at the active site. The demonstration that the ADAM 12 metalloprotease domain is functional may have important implications for future studies that explore the role of ADAM 12 protein in development and disease.

Ventral mesoderm induction and patterning by bone morphogenetic protein heterodimers in Xenopus embryos

Bone morphogenetic proteins (BMPs) perform diverse functions in vertebrate development. Here we demonstrate that the heterodimeric BMP-4/7 protein directly induces ventral mesoderm and blood in Xenopus animal caps, and BMP-2/7 heterodimers may function similarly. We also provide indirect evidence that BMP heterodimers function in embryos, using assays with dominant-negative BMP ligands. Homodimeric BMP-2 and BMP-4 proteins do not induce mesoderm, but they ventralize mesoderm induction by activin. In contrast, BMP-7 protein interferes with mesoderm induction by activin, but BMP-7 stimulates ventral mesoderm induction by the heterodimer, BMP-4/7. This novel property of BMP-7 distinguishes it from other BMPs. BMP-7 may therefore function in early embryogenesis to antagonize activin signals and potentiate BMP signals. We propose that BMP heterodimers convey signals for ventral mesoderm induction and patterning in Xenopus development.

In vitro cleavage of internally quenched fluorogenic human proparathyroid hormone and proparathyroid-related peptide substrates by furin. Generation of a potent inhibitor

The cleavage of parathyroid hormone (PTH) from its precursor proparathyroid hormone (pro-PTH) is accomplished efficiently by the proprotein convertase furin (Hendy, G. N., Bennett, H. P. J., Gibbs, B. F., Lazure, C., Day, R., and Seidah, N. G. (1995) J. Biol. Chem. 270, 9517-9525). We also showed that a synthetic peptide comprising the -6 to +7 sequence of human pro-PTH is appropriately cleaved by purified furin in vitro. The human pro-PTH processing site Lys-Ser-Val-Lys-Lys-Arg differs from the consensus furin site Arg-Xaa-(Lys/Arg)-Arg that is represented by Arg-Arg-Leu-Lys-Arg in the cleavage site of pro-PTH-related peptide (pro-PTHrP). An earlier study demonstrated that an internally quenched fluorogenic substrate bearing an O-aminobenzoyl fluorescent donor at the NH2 terminus and an acceptor 3-nitrotyrosine near the COOH terminus was appropriately cleaved by the convertases furin and PC1 (Jean, F., Basak, A., DiMaio, J., Seidah, N. G., and Lazure, C. (1995) Biochem. J. 307, 689-695). Here, we have synthesized a series of internally quenched fluorogenic substrates based upon the pro-PTH and pro-PTHrP sequences to determine which residues are important for furin cleavage. Purified recombinant furin and PC1 cleaved the human pro-PTH internally quenched substrate at the appropriate site in an identical manner to that observed with the nonfluorescent peptide. Several substitutions in the P6-P3 sequence were well tolerated; however, replacement of the Lys at the P6 position with Gly and replacement of the P3 Lys by an acidic residue led to markedly compromised cleavage by furin. Furin activity was very sensitive to substitution in P' positions. Replacement of Ser at P1' with Gly and Val at P2' with Ala generated substrates that were less well cleaved. Substitution at the P1' position of Val for Ser in conjunction with Ala for Val at P2', as well as a single substitution of Lys for Val at P2', generated specific inhibitors of furin cleavage. The findings of this study open the way to the rational design of inhibitors of furin with therapeutic potential.

Furin: a mammalian subtilisin/Kex2p-like endoprotease involved in processing of a wide variety of precursor proteins

Limited endoproteolysis of inactive precursor proteins at sites marked by paired or multiple basic amino acids is a widespread process by which biologically active peptides and proteins are produced within the secretory pathway in eukaryotic cells. The identification of a novel family of endoproteases homologous with bacterial subtilisins and yeast Kex2p has accelerated progress in understanding the complex mechanisms underlying the production of the bioactive materials. Seven distinct proprotein convertases of this family (furin, PC2, PC1/PC3, PC4, PACE4, PC5/PC6, LPC/PC7/PC8/SPC7) have been identified in mammalian species, some having isoforms generated via alternative splicing. The family has been shown to be responsible for conversion of precursors of peptide hormones, neuropeptides, and many other proteins into their biologically active forms. Furin, the first proprotein convertase to be identified, has been most extensively studied. It has been shown to be expressed in all tissues and cell lines examined and to be mainly localized in the trans-Golgi network, although some proportion of the furin molecules cycle between this compartment and the cell surface. This endoprotease is capable of cleaving precursors of a wide variety of proteins, including growth factors, serum proteins, including proteases of the blood-clotting and complement systems, matrix metalloproteinases, receptors, viral-envelope glycoproteins and bacterial exotoxins, typically at sites marked by the consensus Arg-Xaa-(Lys/Arg)-Arg sequence. The present review covers the structure and function of mammalian subtilisin/Kex2p-like proprotein convertases, focusing on furin (EC 3.4.21.85).

The chemorepulsive activity of secreted semaphorins is regulated by furin-dependent proteolytic processing

The semaphorins are a large group of cell surface and secreted proteins implicated in axonal pathfinding. Here we show that the secreted mouse semaphorin D (SemD) is synthesized as an inactive precursor (proSemD) and becomes repulsive for sensory and sympathetic neurites upon proteolytic cleavage. ProSemD processing can be blocked completely by an inhibitor selective for furin-like endoproteases or mutagenesis of three conserved dibasic cleavage sites. Its C-terminal pro-peptide contains a processing signal that is essential for SemD to acquire its full repulsive activity. SemD processing is regulated during the embryonic development of the mouse and determines the magnitude of its repulsive activity. Similarly to SemD, the secreted semaphorins SemA and SemE display repulsive properties that are regulated by processing. Our data suggest that differential proteolytic processing determines the repulsive potency of secreted semaphorins and implicate proteolysis as an important regulatory mechanism in axonal pathfinding.

cDNA cloning and sequence analysis of the Xenopus laevis egg envelope glycoprotein gp43

The glycoproteins of the Xenopus laevis egg envelope function in fertilization and development. As the unfertilizable coelomic egg transits the pars recta region of the oviduct, it is converted to a fertilizable egg by limited proteolysis of the envelope glycoprotein gp43 to gp41. This conversion is caused by an oviductally secreted serine active site protease, oviductin. We cloned a cDNA for gp43 from an oocyte cDNA library. The cDNA encoded a 454 amino acid protein homologous to the ZPC family of glycoproteins previously shown to be present in mammalian and fish egg envelopes. Conserved ZPC domains and motifs present in the Xenopus sequence included a signal peptide sequence, an N-linked glycosylation site, and 12 aligned Cys residues. In mammalian and Xenopus sequences, a furin-like (convertase) site and a C-terminal transmembrane domain were present reflecting the biosynthesis of ZPC in these species via the secretory glycoprotein pathway. However, fish envelope glycoproteins lack these sequences since they are synthesized via a different route (in the liver, transported to the ovary, and assembled into the egg envelope surrounding the oocyte). Consensus amino acid residues were identified by sequence comparisons of seven ZPC family members; 19% of the amino acid residues were invariant and 48% of the residues were identical in at least four of the seven sequences. The consensus sequence was used to make structure-fertilization function predictions for this phylogenetically conserved family of glycoproteins.

Tissue targeting of angiotensin peptides

Angiotensin II (Ang II) is an octapeptide generated by the sequential proteolytic action of renin and angiotensin converting enzyme on the glycoprotein angiotensinogen. While numerous mammalian tissues have been shown to express some or all of the components of the renin-angiotensin system (RAS), the function of most of these tissue RAS remains a matter of conjecture. To test for tissue-specific functions of Ang II and as an alternative to co-expressing all the components of RAS, we have engineered a fusion protein that leads to direct Ang II release within specific tissues. The angiotensin peptide is cleaved from the fusion protein within the secretory pathway by the ubiquitous endoprotease furin and is released from the cell by constitutive secretion. Direct injection of an expression vector encoding such a fusion protein into rat cardiac ventricles results in a highly localized expression of atrial natriuretic peptide mRNA (an angiotensin responsive marker of cardiac hypertrophy), demonstrating the utility of this approach for local targeting of mature peptides to tissues in animal models.

Identification of the cleavage sites in the alpha6A integrin subunit: structural requirements for cleavage and functional analysis of the uncleaved alpha6Abeta1 integrin

The alpha6A and alpha6B integrin subunits are proteolytically cleaved during biosynthesis into a heavy chain (120 kDa) that is disulphide-linked to one of two light chains (31 or 30 kDa). Analysis of the structure of the alpha6A subunit on the carcinoma cell line T24 and human platelets demonstrated that the two light chains of alpha6 are not differentially glycosylated products of one polypeptide. Rather they possess different polypeptide backbones, which presumably result from proteolytic cleavage at distinct sites in the alpha6 precursor. Mutations were introduced in the codons for the R876KKR879, E883K884, R890K891 and R898K899 sequences, the potential proteolytic cleavage sites, and wild-type and mutant alpha6A cDNAs were transfected into K562 cells. The mutant alpha6A integrin subunits were expressed in association with endogenous beta1 at levels comparable to that of wild-type alpha6Abeta1. A single alpha6 polypeptide chain (150 kDa) was precipitated from transfectants expressing alpha6A with mutations or deletions in the RKKR sequence. Mutations in the EK sequence yielded alpha6A subunits that were cleaved once into a heavy and a light chain, whereas alpha6A subunits with mutations in one of the two RK sequences were, like wild-type alpha6A, cleaved into one heavy and two light chains. Thus a change in the RKKR sequence prevents the cleavage of alpha6. The EK site is the secondary cleavage site, which is used only when the primary site (RKKR) is intact. Microsequencing of the N-termini of the two alpha6A light chains from platelets demonstrated that cleavage occurs after Arg879 and Lys884. Because alpha6(RKKG), alpha6(GKKR) and alpha6(RGGR) subunits were not cleaved it seems that both the arginine residues and the lysine residues are essential for cleavage of RKKR. alpha6A mutants with the RKKR sequence shifted to the EK site, in such a way that the position of the arginine residue after which cleavage occurs corresponds exactly to Lys884, were partly cleaved, whereas alpha6A mutants with the RKKR sequence shifted to other positions in the alpha6A subunit, including one in which it was shifted two residues farther than the EK cleavage site, were not cleaved. In addition, alpha6A mutants with an alpha5-like cleavage site, i.e. arginine, lysine and histidine residues at positions -1, -2 and -6, were not cleaved. Thus both an intact RKKR sequence and its proper position are essential. After activation by the anti-beta1 stimulatory monoclonal antibody TS2/16, both cleaved and uncleaved alpha6Abeta1 integrins bound to laminin-1. The phorbol ester PMA, which activates cleaved wild-type and mutant alpha6Abeta1, did not activate uncleaved alpha6Abeta1. Thus uncleaved alpha6Abeta1 is capable of ligand binding, but not of inside-out signalling. Our results suggest that cleavage of alpha6 is required to generate a proper conformation that enables the affinity modulation of the alpha6Abeta1 receptor by PMA.

A novel member of the subtilisin-like protease family from Streptomyces albogriseolus

We previously isolated three extracellular endogenous enzymes from a Streptomyces albogriseolus mutant strain which were targets of Streptomyces subtilisin inhibitor (SSI) (S. Taguchi, A. Odaka, Y. Watanabe, and H. Momose, Appl. Environ. Microbiol. 61:180-186, 1995). In the present study, of the three enzymes the largest one, with a molecular mass of 45 kDa (estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis), termed SAM-P45, has been characterized in detail. The entire gene encoding SAM-P45 was cloned as an approximately 10-kb fragment from S. albogriseolus S-3253 genomic DNA into an Escherichia coli host by using a shuttle plasmid vector. The amino acid sequence corresponding to the internal region of SAM-P45, deduced from the nucleotide sequence of the gene, revealed high homology, particularly in three regions around the active-site residues (Asp, His, and Ser), with the amino acid sequences of the mature domain of subtilisin-like serine proteases. In order to investigate the enzymatic properties of this protease, recombinant SAM-P45 was overproduced in Streptomyces coelicolor by using a strong SSI gene promoter. Sequence analysis of the SAM-P45 gene and peptide mapping of the purified SAM-P45 suggested that it is synthesized as a large precursor protein containing a large C-terminal prodomain (494 residues) in addition to an N-terminal preprodomain (23 and 172 residues). A high proportion of basic amino acids in the C-terminal prodomain was considered to serve an element interactive with the phospholipid bilayer existing in the C-terminal prodomain, as found in other membrane-anchoring proteases of gram-positive bacteria. It is noteworthy that SAM-P45 was found to prefer basic amino acids to aromatic or aliphatic amino acids in contrast to subtilisin BPN', which has a broad substrate specificity. The hydrolysis by SAM-P45 of the synthetic substrate (N-succinyl-L-Gly-L-Pro-L-Lys-p-nitroanilide) most preferred by this enzyme was inhibited by SSI, chymostatin, and EDTA. The proteolytic activity of SAM-P45 was stimulated by the divalent cations Ca2+ and Mg2+. From these findings, we conclude that SAM-P45 interacts with SSI and can be categorized as a novel member of the subtilisin-like serine protease family.

Production of recombinant proteins in Chinese hamster ovary cells overexpressing the subtilisin-like proprotein converting enzyme furin

The proprotein processing enzyme furin is the mammalian prototype of a novel family of subtilisin-like serine endoproteases which possess cleavage specificity for sites involving multiple basic amino acid residues and are involved in the processing of precursor proteins of a variety of regulatory peptides and proteins. One of the limiting steps in the engineering of mammalian cells designed for the overproduction of secreted proteins is the endoproteolytic cleavage of the precursor molecule to its mature biologically active form. The extremely low level of endogenous furin is likely the reason why cells are not able to fully mature overexpressed precursor proteins to their mature form. Here, we report a CHO-derived cell line genetically engineered for the production of high levels of recombinant proteins that need such endoproteolytic maturation. First, the human furin cDNA under the control of the cytomegalovirus early promoter and enhancer was introduced and overexpressed in a DHFR-deficient CHO cell line. A permanent cell line CHO-D3-FUR was established that expressed biologically active furin. Subsequently, to demonstrate the capacity of CHO-D3-FUR cells to produce recombinant proteins in a fully matured form, two derivative cell lines were established that overexpressed the von Willebrand factor (vWF) and transforming growth factor beta 1 (TGF beta 1); CHO-D3-vWF and CHO-D3-TGF beta 1, respectively. Both derivative cell lines were able to produce relatively high levels of recombinant protein in a fully matured and biologically active form. Our results illustrate the potential of the CHO-D3-FUR cell line in the production of recombinant secretory proteins that need endoproteolytic activation at the consensus furin cleavage sequence Arg-X-Lys/Arg-Arg.

Transmembrane-deletion mutants of the membrane-type matrix metalloproteinase-1 process progelatinase A and express intrinsic matrix-degrading activity

Membrane-type matrix metalloproteinase-1 (MT-MMP-1) has been proposed to play a critical role in regulating the expression of tissue-invasive phenotypes in normal and neoplastic cells by directly or indirectly mediating the activation of progelatinase A. To begin characterizing MT-MMP-1 structure-function relationships, transmembrane-deletion mutants were constructed, and the processing of the zymogens as well as the enzymic activity of the mature proteinases was analyzed. We now demonstrate that pro-MT-MMP-1 mutants are efficiently processed to active proteinases following post-translational endoproteolysis immediately downstream of an Arg108-Arg-Lys-Arg basic motif by a proprotein convertase-dependent pathway. The secreted form of active MT-MMP-1 not only displays an N terminus identical with that described for the processed wild-type enzyme at Tyr112 (Strongin, A. Y., Collier, I., Bannikov, G., Marmer, B. L., Grants, G. A., and Goldberg, G. I. (1995) J. Biol. Chem. 270, 5331-5338), but also directly mediated progelatinase A activation via a two-step proteolytic cascade indistinguishable from that observed with intact cells. Furthermore, although the only function previously ascribed to MT-MMP-1 is its ability to act as a progelatinase A activator, purified transmembrane deletion mutants also expressed proteolytic activities against a wide range of extracellular matrix molecules. Given recent reports that MT-MMP-1 ectodomains may undergo intercellular transfer in vivo (Okada, A., Bellocq, J.-P., Rouyer, N., Chenard, M.-P., Rio, M.-C., Chambon, P., and Basset, P. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 2730-2734), our data suggest that soluble forms of the proteinase confer recipient cells with the ability to not only process progelatinase A, but also directly degrade extracellular matrix components.

Processing of mutated human proinsulin to mature insulin in the non-endocrine cell line, CHO

Heterologous genes encoding proproteins, including proinsulin, generally produce mature protein when expressed in endocrine cells while unprocessed or partially processed protein is produced in non-endocrine cells. Proproteins, which are normally processed in the regulated pathway restricted to endocrine cells, do not always contain the recognition sequence for cleavage by furin, the endoprotease specific to the constitutive pathway, the principal protein processing pathway in non-endocrine cells. Human proinsulin consists of B-Chain-C-peptide-A-Chain and cleavage at the B/C and C/A junctions is required for processing. The B/C, but not the C/A junction, is recognised and cleaved in the constitute pathway. We expressed a human proinsulin and a mutated proinsulin gene with an engineered furin recognition sequence at the C/A junction and compared the processing efficiency of the mutant and native proinsulin in Chinese Hamster Ovary cells. The processing efficiency of the mutant proinsulin was 56% relative to 0.7% for native proinsulin. However, despite similar levels of mRNA being expressed in both cell lines, the absolute levels of immunoreactive insulin, normalized against mRNA levels, were 18-fold lower in the mutant proinsulin-expressing cells. As a result, there was only a marginal increase in absolute levels of insulin produced by these cells. This unexpected finding may result from preferential degradation of insulin in non-endocrine cells which lack the protection offered by the secretory granules found in endocrine cells.

Molecular cloning and expression of megakaryocyte potentiating factor cDNA

The human megakaryocyte potentiating factor (hMPF) has been previously purified from a culture supernatant of human pancreatic cancer cells HPC-Y5 (Yamaguchi, N., Hattori, K., Oh-eda, M., Kojima, T., Imai, N., and Ochi, N. (1994) J. Biol. Chem. 269, 805-808). We have now isolated hMPF cDNA from a HPC-Y5 cDNA library using polymerase chain reaction and plaque hybridization methods. The hMPF cDNA encodes a polypeptide consisting of 622 amino acids, including a signal peptide of 33 amino acids, and with a deduced molecular mass of 68 kDa, although HPC-Y5 cells secrete a 33-kDa form of hMPF. Human MPF does not show any significant homology with other previously described sequences. The cDNA was expressed in COS-7 and Chinese hamster ovary (CHO) cells, and megakaryocyte potentiating activity was detected in their culture supernatant. The COS-7 cells secreted only a 33-kDa recombinant hMPF, whereas an additional 30-kDa form was detected in the culture medium of CHO cells. The 33-kDa rhMPF purified from CHO cells showed megakaryocyte potentiating activity, but not the purified 30-kDa rhMPF. The difference in structure and activity between the 33- and 30-kDa forms of hMPF was ascribed to the existence in the 33-kDa form of the C-terminal 25 amino acid residues.

Peptidyl substrates containing unnatural amino acid at the P'1 position are potent inhibitors of prohormone convertases

In order to study further the importance of the P'1 residue upon the activity of human PC1 and human furin, two important members of subtilisin/kexin family of enzymes, we have prepared by solid-phase Fmoc or recently introduced FastMoc chemistry a series of 10 peptidyl substrate analogs. The structures of these analogs are based upon the core sequence of pro-mPC1(83-93) namely, D-Tyr-Lys-Glu-Arg-Ser-Lys-Arg-Xaa-Val-Gln-Lys-Asp, where D-Tyr replaces the native L-Tyr residue and Xaa, representing the P'1 position, corresponds to L-Ser or to nonproteinacous amino acids such as Tle, Sarc, MLeu, Aib, D-Tic or L-Tic. Two more analogs with L-Tic at P'1 position but with one amino acid less, namely P5 Glu or P'3 Gln, and one with a Cit residue in place of Arg at P1 site of the dodecapeptide were also obtained. These peptides were all fully characterized by a combination of MS, 1H-NMR and amino acid analysis. In contrast to the Ser analog, which is an excellent substrate for both hPC1 and hfurin, these analogs displayed moderate to strong inhibition of both hPC1 and hfurin activity in a reversible competitive manner. They all exhibited higher potency for hfurin than for hPC1, with an inhibition constant (Ki) ranging from 0.8 to 10 microM and from 1.0 to 170 microM, respectively. Incorporation of L-Tic yielded an analog with a two to four-fold increased inhibition of either enzymes when compared to its D-Tic counterpart, the effect being more pronounced for hPC1 than for hfurin. Comparison of these data with those for the corresponding N-terminal Fmoc protected peptides revealed that the highly hydrophobic N-terminal Fmoc function occupying the P8 position can contribute positively or negatively towards proteinase inhibition depending on the nature of the unnatural amino acid at P'1 and the enzyme used. Finally, none of the analogs was significantly cleaved by either enzyme. FTIR data on these analogs revealed some important structural differences between the substrate and inhibitor analogs, as there appears to be a conformational shift from a more beta-sheet-like structure for the substrates to a more alpha-helical-like structure for the inhibitors.

Direct role of furin in mammalian prosomatostatin processing

We have previously reported that rat prosomatostatin (rPSS) undergoes conversion at Arg decreases and Lys decreases monobasic sites to SS-28 and PSS-(1-10) respectively in COS-7 cells, and have proposed furin or a related enzyme of the constitutive secretory pathway as the endoproteinase responsible. Here we have tested directly the ability of furin to cleave rPSS at the two monobasic sites as well as at the RXRK dibasic site of SS-14 conversion (a furin motif, except for Lys substituting for Arg at P1). Recombinant vaccinia virus (VV) vectors were used to co-express rPSS with graded doses of furin in COS-7 cells and LoVo colon carcinoma cells deficient in furin. PSS and cleavage products in cell extracts and media were characterized by HPLC analysis and C-terminal [SS-14-like immunoreactivity (SS-14 LI)] and N-terminal [PSS-(1-10) LI] directed radioimmunoassays. There was a dose-dependent increase in SS-28 production from rPSS by furin in COS-7 cells from 29% (control) to 58% (high-dose furin) associated with a progressive decrease in unprocessed PSS from > 60% to approximately 20% of total SS-14 LI. Significant SS-14 production occurred only at high levels of furin infection. Control LoVo cells infected with VV:rPSS exhibited production of approximately 21% SS-28, approximately 15% PSS-(1-10) and 3.5% SS-14. Infection of LoVo cells with VV:hfurin (hfurin = human furin) enhanced SS-28 production to 30-34%. SS-14 synthesis also increased to 25-40%, probably by conversion from SS-28. Overexpression of furin in COS-7 or LoVo cells failed to increase PSS-(1-10) production. These results show that furin is a candidate SS-28 convertase. Arginine is the preferred residue at the P1 site of furin cleavage. Furin does not process rPSS to PSS-(1-10), suggesting the existence of another monobasic convertase with a preference for Lys rather than Arg at P1. Such an enzyme could also explain the presence of endogenous SS-28-, PSS-(1-10)- and SS-14-producing activities in LoVo cells.

Furin-dependent intracellular activation of the human stromelysin-3 zymogen

Human stromelysin-3, a new member of the matrix metalloproteinase family, is expressed in tissues undergoing the active remodelling associated with embryonic development, wound healing and tumour invasion. But like all other members of the matrix metalloproteinase gene family, stromelysin-3 is synthesized as an inactive precursor that must be processed to its mature form in order to express enzymic activity. Here we identify stromelysin-3 as the first matrix metalloproteinase to be discovered that can be processed directly to its enzymically active form by an obligate intracellular proteolytic event that occurs within the constitutive secretory pathway. Intracellular activation is regulated by an unusual 10-amino-acid insert sandwiched between the pro- and catalytic-domains of stromelysin-3, which is encrypted with an Arg-X-Arg-X-Lys-Arg recognition motif for the Golgi-associated proteinase, furin, a mammalian homologue of the yeast Kex2 pheromone convertase. A furin-stromelysin-3 processing axis not only differentiates the regulation of this enzyme from all previously characterized matrix metalloproteinases, but also identifies pro-protein convertases as potential targets for therapeutic intervention in matrix-destructive disease states.

Proparathyroid hormone is preferentially cleaved to parathyroid hormone by the prohormone convertase furin. A mass spectrometric study

Parathyroid hormone (PTH), an 84-amino acid peptide, is the major regulator of blood calcium homeostasis. Its mRNA, in addition to encoding the mature peptide, also encodes a "pre" sequence of 25 amino acids and a basic "pro" hexapeptide. To assess which of the subtilisin-like prohormone convertases can process proPTH to PTH we coinfected cells with a vaccinia virus construct expressing human preproPTH and vaccinia virus constructs expressing furin, PC1 or PC2. BSC-40 cells, having a constitutive secretory pathway, and GH4C1 cells, having a regulated secretory pathway, were used. PTH biosynthetic products in cell extracts and media were purified by high performance liquid chromatography, identified by radioimmunoassay, and unambiguously defined as either proPTH or PTH by ion-spray mass spectrometry. In both cell types, furin was the most effective in processing proPTH to PTH. In all cases only PTH was released into the medium. In addition, partially purified furin and PC1 were tested for their ability to appropriately cleave a tridecapeptide spanning the prohormone cleavage site found in proPTH. Here too furin was much more effective at cleaving at the correct site. Northern blot analysis and in situ hybridization showed that furin and preproPTH mRNA are co-expressed in the parathyroid, whereas PC1, PC2, and PC5 are not and PACE4 is expressed only at very low levels. Taken together these studies strongly suggest that furin is the enzyme responsible for the physiological processing of proPTH to PTH.

Proteolytic activation of bacterial toxins by eukaryotic cells is performed by furin and by additional cellular proteases

Before intoxication can occur, anthrax toxin protective antigen (PA), Pseudomonas exotoxin A (PE), and diphtheria toxin (DT) must be activated by proteolytic cleavage at specific amino acid sequences. Previously, it was shown that PA and DT can be activated by furin. In Chinese hamster ovary (CHO) cells, wild-type (RKKR) and cleavage site mutants of PA, each administered with a modified form of anthrax toxin lethal factor (the N terminus of lethal factor fused to PE domain III), had the following potencies: RKKR (wild type) (concentration causing 50% cell death [EC50] = 12 ng/ml) > or = RAAR (EC50 = 18 ng/ml) > FTKR (EC50 = 24 ng/ml) > STRR (EC50 = 49 ng/ml). In vitro cleavage of PA and cleavage site mutants of PA by furin demonstrated that native PA (RKKR) and PA with the cleavage sequence RAAR are substrates for furin. To characterize eukaryotic proteases that play a role in activating bacterial toxins, furin-deficient CHO cells were selected after chemical mutagenesis. Furin-deficient cells were resistant to PE, whose cleavage site, RQPR, constitutes a furin recognition site and to all PA cleavage site mutants, but were sensitive to DT (EC50 = 2.9 ng/ml) and PA (EC50 = 23 ng/ml), whose respective cleavage sites, RKKR and RVRR, contain additional basic residues. Furin-deficient cells that were transfected with the furin gene regained sensitivity to PE and PA cleavage site mutants. These studies provide evidence that furin can activate the three toxins and that one or more additional proteases contribute to the activation of DT and PA.

Maturation of the trans-Golgi network protease furin: compartmentalization of propeptide removal, substrate cleavage, and COOH-terminal truncation

We have cloned a bovine cDNA encoding the trans-Golgi network (TGN) protease furin and expressed it via recombinant vaccinia viruses to investigate intracellular maturation. Pulse-chase labeling reveals that the 104-kD pro-furin bearing high mannose N-glycans is rapidly processed into the 98-kD protease whose N-glycans remain sensitive to endoglycosidase H for a certain period of time. Furthermore, in the presence of brefeldin A, pro-furin cleavage occurs. From these data we conclude that the ER is the compartment of propeptide removal. Studies employing the ionophore A23187 and DTT show that autocatalysis is Ca2+ dependent and that it does not occur under reducing conditions. Pro-furin produced under these conditions never gains endo H resistance indicating that it is retained in the ER. Coexpression of furin with the fowl plague virus hemagglutinin in the presence of brefeldin A and monensin reveals that furin has to enter the Golgi region to gain substrate cleaving activity. N-glycans of furin are sialylated proving its transit through the trans-Golgi network. A truncated form of furin is found in supernatants of cells. Truncation is inhibited in the absence of Ca2+ ions and in the presence of acidotropic agents indicating that it takes place in an acidic compartment of cells. Comparative analysis with furin expressed from cDNA reveals that the truncated form prevails in preparations of biologically active, endogenous furin obtained from MDBK cells. This observation supports the concept that secretion of truncated furin is a physiological event that may have important implications for the processing of extracellular substrates.

Nucleotide and deduced amino acid sequences of the matrix (M) and fusion (F) protein genes of cetacean morbilliviruses isolated from a porpoise and a dolphin

Morbilliviruses have been isolated from stranded dolphins and porpoises. The present paper describes the cloning and sequencing of the porpoise morbillivirus (PMV) F gene and of the dolphin morbillivirus (DMV) M and F genes and their flanking regions. The gene order of the DMV genome appeared to be identical to that of other morbilliviruses. A genomic untranslated region of 837 nucleotides was found between the translated DMV M and F gene regions. The predicted DMV M protein were highly conserved with those of other morbilliviruses. Both the deduced PMV and DMV F0 proteins exhibited three major hydrophobic regions as well as a cysteine rich region, a leucine zipper motif and a cleavage motif allowing cleavage of the F0 protein into F1 and F2 subunits. Apparently the DMV F0 cleavage motif was not modified by adaptation of DMV to Vero cells. The predicted PMV and DMV F proteins were 94% identical. Comparisons with the corresponding sequences of other morbilliviruses demonstrated that the cetacean morbillivirus does not derive from any known morbillivirus but represents an independent morbillivirus lineage.

Proinsulin processing in the regulated and the constitutive secretory pathway

Proinsulin is converted to insulin in beta-cell granules. Conversion involves endoproteolytic cleavage at the two pairs of basic residues linking the insulin A- and B-chains to C-peptide. The sequence of events leading to complete conversion differs from one proinsulin species to the next. In man, the structure of the proinsulin molecule is such as to favour cleavage at the B-chain/C-peptide junction leading to the generation of des-31,32 split proinsulin as the predominant, naturally occurring conversion intermediate. Under normal circumstances, proinsulin conversion is largely completed before secretion, and neither the intact prohormone nor conversion intermediates are thus encountered in large quantities in the circulation. In some pathological situations, including non-insulin-dependent diabetes, insulinoma and familial hyperproinsulinaemia, unusually high ratios of des-31,32 split proinsulin and/or proinsulin to insulin have been reported. As we understand the biochemistry of proinsulin conversion in increasingly fine molecular detail, it should become possible to make use of such unusual ratios to provide insight into lesions underlying altered beta-cell function in disease states.

A Kex2-related endopeptidase activity present in rat liver specifically processes the insulin proreceptor

The insulin proreceptor is cleaved by limited proteolysis post-translationally at an Arg-Lys-Arg-Arg site to generate its mature alpha- and beta-subunit form. An 35S-labelled insulin proreceptor substrate preparation and a 15-mer peptide substrate that mimics the amino acid sequence around and including the insulin proreceptor processing site (IRP-peptide) has revealed an endopeptidase activity that catalyses insulin proreceptor cleavage in a rat liver subcellular fraction. Under optimal conditions, normal 35S-labelled insulin proreceptor substrate processing by this fraction was quantitative. This fraction was not able to process an 35S-labelled insulin proreceptor variant substrate (where the Arg-1 of the tetrabasic cleavage site had been replaced by Ala-1), similarly to previous in vivo observations, suggesting that this endopeptidase activity has physiological relevance. Biochemical characterization of the insulin proreceptor/IRP-peptide processing revealed this rat liver endopeptidase activity to have a broad pH range (> 70% maximal activity between pH 5.5 and 10.0) and a pH optimum of pH 8-10. It was Ca(2+)-dependent activity, maximally active between 0.5 and 5 mM Ca2+ and half-maximally activated between 50 and 90 microM Ca2+. Endoproteolytic activity was not inhibited by group-specific inhibitors of serine-, cysteinyl or aspartyl proteinases or by 1,10-phenanthroline; however, EDTA and 1,2-cyclohexanediaminetetraacetic acid did inhibit the activity, but this was accounted for by Ca2+ chelation. The IRP-peptide substrate assay enabled measurement of an apparent Km of 22 microM and a Vmax of 18.6 pmol/min for this endopeptidase activity. These biochemical characteristics suggest that insulin proreceptor processing endopeptidase activity to be a legitimate member of the Kex2-related proprotein convertase family. Immunoblotting detected furin and PACE4 proteins (both members of this family) to be present in the rat liver subcellular fraction containing insulin proreceptor processing activity. Since the biochemical characteristics of the insulin proreceptor processing endopeptidase activity mostly resembled those of furin activity, it is likely that insulin proreceptor proteolytic maturation can be catalysed by furin in the liver.

Cleavage of proalbumin peptides by furin reveals unexpected restrictions at the P2 and P'1 sites

Proalbumin is the principal substrate of the in situ hepatic convertase. Here we investigated the specificity of furin using synthetic peptides based on the N-terminal sequence of human proalbumin. The propeptide was rapidly cleaved from the normal ((-6)RGVFRR(-1)DAHKSEAVW(+9)) peptide but as expected, there was no cleavage of the proalbumin Lille analogue with a -2 His (-2H). Surprisingly, the effect of this substitution could not be corrected by introducing a -4 Arg (-4R-2H). In contrast, the peptide -4R-2A was an excellent substrate being cleaved five times faster than normal, indicaabting that His is not allowed as an P2 residue. Replacement of the -4 Val by Glu supported the expected importance of a positive charge at P4 as the cleavage rate dropped to 10% of normal after this substitution. The -6 Arg makes a small contribution to cleavage, its replacement by Ala decreased the cleavage rate to 60% of normal. The Lys-Arg propeptide was almost as good a substrate as the normal Arg-Arg peptide, but the introduction of a Lys at P1 totally abolished processing. The exclusion of P'1 positive charges would be an important requirement for preventing aberrant cleavage in the middle of tetrabasic sequences.