Http://www.abbs.info e-mail:[email protected]

ISSN 0582-9879                               ACTA BIOCHIMICA et BIOPHYSICA SINICA 2001, 33(6): 591-599                                     CN 31-1300/Q

 

The Inhibitory Activities of Recombinant Eglin C Mutants on Kexin and Furin, Using Site-directed Mutagenesis and Molecular Modeling

FEI Hao, LUO Ming-Juan, YE Yu-Zheng1, DING Da-Fu1, CHI Cheng-Wu*

( State Key Laboratory of Molecular Biology, Shanghai Institute of Biochemistry and Cell Biology, the Chinese Academy of Sciences,Shanghai 200031, China; 1Shanghai Institute of Biochemistry and Cell Biology, the Chinese Academy of Sciences, Shanghai 200031, China )

 

Abstract    Mammalian furin and yeast kexin are members of the proprotein convertase family involved in the proteolytic processing of many important precursor proteins. Here the gene coding for the subtilisin inhibitor eglin C was totally synthesized and expressed in E.coli. Substitution of residues at each position P1, P2 and P4 of eglin C with a basic residue using protein engineering could make eglin C a very strong inhibitor for furin (Ki around 10–9 mol/L), and even more strong for kexin (Ki around 10–11 mol/ L). Results indicated that: (1) A basic residue Lys or Arg at P1 site is prerequisite for the inhibitor. (2) The second mutation with basic residue at P4 site drastically increase the inhibitory activity by two orders of magnitude. (3) A basic residue at P2 site is favorable for the binding to the enzyme, but unfavorable for the stability of the inhibitor, resulting in a temporary inhibition. (4) A hydrophobic residue is preferential at P3 site. Based on the known crystal structures of subtilisin and eglin C, the interaction between the enzyme and inhibitor was modeled, and their involved residues were predicted which gave a good explanation to the experimental results.

Key words    proprotein convertase; proteinase inhibitor; site-directed mutagenesis; molecular modeling

 

The processing of precursor proteins via limited proteolysis at paired basic amino acids is an important and widely used cellular mechanism for the generation of biological active peptides and proteins. The enzymes responsible for this intracellular cleavage have been molecularly and functionally characterized, and formed a family of precursor convertases[1]. Among them, the Saccharomyces cerevisiae member of this family, kexin, was first cloned and elucidated[2,3], while the mammalian member, furin, was the most well studied[4]. Because of the homology of their catalytic domains to that of the bacterial serine protease, subtilisin, these enzymes are also called subtilisin-like proprotein convertases (SPCs)[5].

The important substrates for proprotein convertase include not only most peptide hormones and neuropeptides, but also many growth factors, cell-surface receptors, adhesion molecules, serum proteins, plasma proteases, matrix metalloproteinas-es[6-10]. In addition to endogenous proproteins, many pathogens of bacterial exotoxins and viral-envelope glycoproteins also require these enzymes for the activation[11-15].

Many attempts have been made to develop protein-based furin inhibitors, since tissue- or cell-type-specific expression of these inhibitors controlled by a characterized promoter could be therapeutically valuable. A variant of a1-antitrypsin (a1-PDX) has been bioengineered to be an inhibitor highly selective for furin and has potential application to be an antipathogenic agent[16,17]. A widely expressed ovalbumin-type serpin, human proteinase inhibitor 8(PI8), was also demonstrated to be an inhibitor for furin in vitro with a Ki of 53.8 pmol/L[18]. Besides, histidine-rich human salivary peptides were found to be moderately potent inhibitors for the furin-mediated cleavage of a fluorogenic peptide substrate. The inhibition was reversible and competitive, with an estimated Ki in a range of micromole[19]. In an earlier paper, a variant of the ovomucoid third domain with a consensus furin site Arg-Cys-Lys-Arg could also inhibit furin with a Ki in a range of submicromole[20].

eglin C, a proteinase inhibitor isolated from the leech Hirudo medicinalis belongs to the patato I inhibitor family and strongly inhibits human leukocyte elastase, cathepsin G, a-chymotrypsin and substilisin. It was indicated that its inhibitory specificity could be changed from inhibiting elastase to trypsin by a point mutation at its reactive site L45R[21]. Due to the flexibility of its reactive site loop and absence of disulfide bond easy to be folded, this inhibitor was regarded as a good model for the structure and function study using protein engineering technology.

In the present paper, a gene coding for eglin C was totally synthesized and several mutants were constructed and expressed in the GST gene fusion system, their inhibitory activity against the recombinant human furin and kexin examined. The Ki values of the strong mutants for furin and kexin were around 10–9 mol/L and 10–11 mol/L, respectively. Meanwhile, using the known crystal structures of subtilisin and eglin C, the interaction between the reactive site loop of the inhibitor and the catalytic domain of the enzymes was modeled. The involved binding residues of furin corresponding to the P1, P2, P3 and P4 residues of eglin C, respectively, were predicted. The prediction gave a good explanation to the experimental results.

1    Materials and Methods

1.1  Materials

Fluorogenic substrate Boc-Arg-Val-Arg-Arg-MCA and pGlu-Arg-Thr-Lys-Arg-MCA were both purchased from Bachem. Glutathione S-transferase (GST) gene fusion system was the product of Pharmacia. PCR kit, restriction enzymes and T4 DNA ligase were from Gibco or Sino-American Biological Co. DNA sequencing kit and [a-32P] dATP were obtained from Amersham Life Science. Cell cultures and transfection reagents were all Gibco products. Host cell, S.cerevisiae strain CB018, and the expression vector pG5-prokexin were kindly given by Dr. R.S. Fuller. The full-length gene of human furin was a kind gift from Dr. G. Thomas. The ultrasphere C18 column was from Beckman. 510 HPLC pumps and 996 photodiode array detector were from Waters-Millipore. Fluorescent spectrometer was manufactured by Hitachi.

1.2  Fragment synthesis and assembly of the eglin C gene

The full-length gene of eglin C was subdivided into four fragments and synthesized (Figure 1). Duplex I of fragment 1,2 and duplex II of fragment 3,4 were constructed by incubating 100 pmole of each oligomer pair in 20 ml of solution (25 mmol/L Tris-HCl, pH 8.0, 5 mmol/L MgCl2, 25 mmol/L NaCl, 0.25 mmol/L dNTPs) at 80 ℃ for 3 min, cooling to 15 ℃ within 10 min. Three units of DNA polymerase I (Klenow) were added and the samples were incubated for 30 min at room temperature. Two ml aliquots of each duplexes I and II with a partially overlapped segment were transferred to a PCR system, which was then incubated at 94 ℃ for 5 min, and annealed to 50 ℃ before 1 unit of Taq DNA polymerase was added. Reaction was carried out at 50 ℃ for 5 min, then 72 ℃ for 10 min for a full-length template synthesis. One hundred pmole of fragment 1 and 4 was finally added to the reaction mixture and a typical PCR reaction was performed at an annealing temperature of 60 ℃ for 30 cycles. Reaction product was analyzed by horizontal agarose electrophoresis.

 

Fig.1       Nucleotide and amino acid sequence of the synthetic eglin C gene

The eglin gene was synthesized according to the published DNA sequence of eglin C[22]. The preferential codons of E. coli were used. The whole gene was subdivided into 4 fragments for the convenience of gene synthesis. The flanking BamHI and EcoRI restriction sites were designed for cloning into the pGEX2T vector.

 

1.3  Gene cloning of the wild type and mutants of eglin C

The full-length gene of the wild type eglin C was cloned through the flanking BamHI and EcoRI restriction sites into the expression vector pGEX-2T provided with the GST kit. The nucleotide sequence of eglin C was verified by DNA sequencing. Site-directed mutagenesis was performed using PCR method with one pair of primers for each mutant. All mutants were verified by DNA sequence determination.

1.4  Gene expression of the wild type eglin C and its mutants

pGEX-eglin and its mutants were transformed into E.coli strain DH5a. Single colony containing the corresponding plasmid was picked and cultured at 37 ℃ in LB/Amp medium (5 g/L trypton, 10 g/L yeast extract, 5 g/L NaCl, 100 mg/L ampicillin) with vigorous vibration over night. The cultures were then diluted 1:10 into fresh prewarmed GLB/Amp medium (20 g/L glucose in LB/Amp) and grew at 37 ℃ with shaking until A600 reached 1-2. After addition of IPTG to a final concentration of 0.2 mmol/L, incubation was continued for another 5 hours. Cells were harvested by centrifugation at 7 000 g, 5 min, and washed once with 1×PBS (140 mmol/L NaCl, 2.7 mmol/L KCl, 10 mmol/L Na2HPO4, 1.8 mmol/L KH2PO4, pH 7.3) before preparation of bacterial sonicates.

1.5  Purification of the wild type eglin C and its mutants

Primary purification steps followed a protocol provided with the GST kit, including the preparation of bacterial sonicates, binding of GST-eglin fusion protein to glutathione Sepharose 4B, elution of the protein of interest and hydrolysis of the fusion protein with thrombin. The hydrolysate was applied to a C18 reverse-phase HPLC column, which was then eluted with a linear gradient of 0%-70% of acetonitrile in 0.1% TFA at a flow rate of 1 ml/min for 50 min. The purified inhibitor was lyophilized for inhibitory activity assay.

1.6  Mass spectra

Wild-type eglin C and all its mutants were submitted for mass determination on a Finnigan LCQ iontrap mass spectrometer (ThermoQuest, San Jose, CA, USA) equipped with an electrospray ionization source.

1.7  Gene expression of the recombinant kexin

The transformed yeast strain CB018 with plasmid pG5-prokexin devoid of the trans-membrane domain was grown in YPD medium (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose, 50 mmol/L Tris-HCl, pH 7.2) at 30 ℃ for 2-3 days. After centrifugation, the supernatant was collected, distributed into 0.5 ml tubes and stored at -70 ℃ until used.

1.8  Gene expression of the recombinant furin

The truncated human furin, terminating at Leu713 and thus devoid of the transmembrane domain, was subcloned into the expression vector pCDNA3 through HindIII and XbaI. The monolayers of COS-7 cells with 60%-70% confluence were transfected with pCDNA-hFur713t using LipofectAmine reagent, and then cultured in 4 ml of DMEM-FBS at 37 ℃ in a CO2 incubator for 48 hours. The medium was collected and centrifuged at 10 000 g for 10 min at 4 ℃. The supernatant was distributed into 100 ml aliquots and stored in －70 ℃ until used.

1.9  Enzymatic assays

The enzyme activity of furin was monitored at 37 ℃ in a final volume of 1 ml Hepes buffer (100 mmol/L, pH 7.5) containing different amount of fluorogenic MCA substrate pGlu-Arg-Thr-Lys-Arg-MCA from 2.5 to 10 mmol/L, 1 mmol/L CaCl2, 0.5% Triton X-100 and 1 mmol/L b-mercaptoethanol. For yeast kexin, the buffer used was 25 mmol/L Bis-Tris-Cl, pH 7.0, 0.1 mmol/L CaCl2, and the substrate was Boc-Arg-Val-Arg-Arg-MCA. For each assay, an equivalent amount of enzyme was added to release 15 nmol/(L·min)-1 AMC (amino-4-methylcoumarin) in the period of 5 min of enzyme reaction. The fluorescence of the released AMC was measured on-line with a Hitachi spectrofluorimeter using an excitation and an emission wavelength of 380 nm (slit width, 10 nm) and 460 nm (slit width, 10 nm), respectively.

1.10       Measurement of kinetic parameters, Ki and IC50

The purified wild eglin C and its mutants were concentrated and dissolved in distilled water to make a stock solution for subsequent use. In order to establish the reversible and competitive nature of inhibition, their Ki values against furin or kexin were determined by Dixon's plot (1/V against I) using three different concentrations of substrate pGlu-Arg-Thr-Lys-Arg-MCA (2.5 mmol/L, 5.0 mmol/L and 10 mmol/L) for furin or two concentrations of Boc-Arg-Val-Arg-Arg-MCA (1.0 mmol/L, 3.0 mmol/L) for kexin. Their IC50 values (the concentration of inhibitor required for 50% inhibition of proteolytic activity) of several mutants were also determined by reciprocal regression results of different inhibitory curves by using Sigmaplot software. Inhibition constant Ki values were then deduced from the calculated IC50 according to the following equation:

Ki = IC50 / ( 1 + [s]/Km )

s = concentration of fluorogenic substrate, Km = Michaelis-Menten constant (0.1 mmol/L of Boc-Arg-Val-Arg-Arg-MCA for kexin, 2.7 mmol/L of pGlu-Arg-Thr-Lys-Arg-MCA for furin).

1.11       Molecular modeling of the binding interface between enzyme and inhibitor

The catalytic domains of furin and kexin were modeled with program “Pmodeling”[23]. The protein templates were subtilisin (Pdb code 1sbn) and thermitase (Pdb code 3tec). The initial models were optimized with software DISCOVER (InsightII Ver. 95) under AMBER force field. The initial binding interface between enzyme and eglin C RVKR45 variant was predicted by superimposing the modeled catalytic domain of furin or kexin onto the subtilisin in its complex structure with wild-type eglin C (Pdb code 1sbn). Structures were also optimized with DISCOVER.

2    Results and Discussion

2.1  Design of the gene coding for eglin C

For the gene synthesis, the two strands were subdivided into four oligomers (37 to 89 nucleotides long) overlapped with 11 to 15 pair bases. The sequence of the synthetic eglin gene is present in Figure 1. The preferential codons of E. coli were used. The synthesized gene flanked with BamHI and EcoRI restriction sites was then inserted into the GST containing expression vector pGEX2T.

2.2  Gene expression and purification of eglin C and its mutants

Either the wild type eglin C or its mutants were expressed in E.coli strain DH5a. Expression level reached around 5 mg of GST-eglin per liter of medium. The GST-eglin fusion protein was purified in one-step procedure by affinity chromatography using glutathione Sepharose 4B matrix. Cleavage of the desired eglin C from GST was achieved by using the recognition sequence of thrombin (Leu-Val-Pro-Arg↓-Gly-Ser, the arrow indicates the scissile bond) located immediately upstream from the N-terminal of eglin C. After the thrombin cleavage, the recombinant Gly-Ser-eglin C was released from GST. The cleavage mixture was then subjected to a HPLC C18 column and purified (Figure 2). eglin C and its variants were characterized by mass spectrometer. The molecular masses were in a good accordance with the calculated ones (Table 1). The wild-type eglin C showed a strong inhibition on chymotrypsin, while their variants displayed no inhibitory activities as expected (data not shown).

 

Fig.2       Chromatography of the GST-eglin C mixture on reverse-phase HPLC after thrombin cleavage

The C18 column (4.6 mm×250 mm) was eluted with the trifluoroacetic acid (TFA)-acetonitrile linear gradient system (0%-100% solution B) in 50 min. Eluting solution A was 0.1% TFA in water; solution B was 0.1% TFA, 70% acetonitrile in water.

 

 

2.3  The effect of P1, P2, and P4 basic residue on the inhibitory activity

All four eglin C variants PVTR45, RVTR45,PVKR45 and RVKR45 are demonstrated to be competitive and reversible inhibitors of kexin by dixon's plot as shown in Figure 3. Their Ki values toward kexin were determined by both curve fitting method and Dixon's plot. Results were summarized in Table 2. By introduction of basic residue at 45(P1), 44(P2), 42(P4) sites, all the four variants of eglin C displayed different degrees of inhibition activity toward kexin. With one substitution at P1 site (L45R) of the wild-type eglin C, the mutant turned out to be a weak kexin inhibitor from a chymotrypsin inhibitor. According to kexin's preferential recognition sequence of dibasic residues, the mutant with paired basic residues at P1 (L45R) and P2 (T44K) sites was engineered and displayed an obvious effect on the Ki value by 23 times. Further study showed that the double mutation at P1 (L45R) and P4 (P42R) gave even more drastic enhancement on inhibitory activity by 54 times. Namely, the P4 site played a more important role than P2 site in the inhibitor. If all three basic residues were substituted, the mutant turned out to be a very strong kexin inhibitor with Ki of 1.86×10-11 mol/L. A typical graph of inhibition curves of all the four mutants were shown as in Figure 4.

 

Fig.3       Graphical determination of inhibition constant (Ki) of eglin C mutants (PVTR45, RVTR45, PVKR45, RVKR45) against kexin using Dixon's plot

The proteolytic activity of kexin against the peptidyl fluorogenic substrate, Boc-Arg-Val-Arg-Arg-MCA was measured at pH 7.0 in the presence of various amounts of mutants using 1.0 mmol/L (S1) and 3.0 mmol/L (S2) substrate concentration, respectively.

 

 

Fig.4       Inhibition curves of the four eglin C mutants (PVTR45, RVTR45, PVKR45, RVKR45) against kexin were overlaid in one graph for a direct comparison of their inhibition activity toward kexin (in the case of the PVTR45 mutant, the concentration of the inhibitor used was far beyond the graphic scale)

 

The previous four mutants were also examined of their inhibitory activities toward furin. As expected the similar results were obtained. The mutants with one or two basic residue mutation at P1 and P2 site both displayed very weak inhibition, their Ki values were over 1×10-7 mol/L(Table 2). However, if additional basic residue was also substituted at P4 site, the mutants RVTR45 and RVKR45 both became a strong inhibitor toward furin, decreasing Ki values from over 10-7 mol/L to 10-9 mol/L (Table 2, Figure 5). It led to about two order difference in inhibitory activity. Compared with the results obtained in the case of kexin, furin seemed to exhibit more stringent requirements for basic residues at P4 than kexin. It could be explained by molecular modeling that there was a cluster of three acidic residues in the furin binding pocket to contact with P4 site of the inhibitor. Whereas in the corresponding subsite of kexin, one of these two acidic residues was replaced by other residue. As expected the mutant with paired basic residue mutation (RVKR45) displayed around 12 times more strong inhibitory effect than the mutant (RVTR45). Thus, it is not strange that the favorable substrates for furin or kexin always consist of paired basic residues.

 

Fig.5       Graphical determination of inhibition constant (Ki) of eglin C mutants (RVTR45, RVKR45, RGKR45, RTKR45) against furin using Dixon's plot

The proteolytic activity of furin against the peptidyl fluorogenic substrate, pGlu-Arg-Thr-Lys-Arg-MCA was measured at pH 7.5 in the presence of various amounts of mutants using 2.5 mmol/L(S1), 5.0 mmol/L(S2) and 10 mmol/L(S3) substrate concentration, respectively.

 

The above results indicated that in both cases of furin and kexin, the inhibitory activity of the P1, P2 and P4 mutant (RVKR45) was definitely stronger than that of the P1 and P4 mutant (RVTR45). However, we noted that if the enzyme and inhibitor incubation time lasted more than one hour instead of 5 min as described in the methods, the mutant RVKR45 with a basic P2 residue became a temporary inhibitor, and could be finally degraded by the enzyme itself. With regard to this aspect, the similar results were also obtained by Fuller and his colleagues[24].

2.4  The effect of P3 residue on the inhibitory activity

Several more mutants were constructed to examine the possible effect of P3 residue on the inhibitory activity of the furin inhibitor. The P3 residue Val in the mutant (RVKR45) was replaced by Gly, Thr, and Asp, respectively. Their inhibition constants are listed in Table 3. The smallest residue Gly without side chain was regarded to be capable of enhancing the backbone flexibility of the reactive site loop of the mutant, however, such replacement did not seem to provide better fitting of the binding loop into the furin molecule. On the contrary, the binding ability decreased by 3-4 fold as compared to that of the mutant RVKR45. When the P3 site was introduced with a polar residue Thr (RTKR45) or a acidic residue Asp (RDKR45), the inhibitory activity of these mutants also obviously decreased, especially, the mutant RDKR45 became a weak furin inhibitor. It implied that a rather large hydrophobic residue at P3 site is favorable for the furin inhibitor and a acid residue is undesirable. These results were also confirmed by the molecular modeling.

 

 

2.5  The residues possibly involved in the interaction between furin and eglin C mutant

Based on the homologous structures of subtilisin and thermitase in complexes with the inhibitor eglin C (Pdb code: 1SBN and 3TEC), the catalytic domains of furin and kexin and their complexes with eglin C RVKR45 were modeled using Pmodeling[23]. Figure 6 displays the modeled complex structure of furin with eglin C RVKR45, focusing on their binding interface. It is in accordance with that of Siezen, except for some details[25].

In general, there are numerous acidic residues of furin (kexin) in the binding interface, which contribute to specificity for multiple basic residues of their corresponding substrates or inhibitors. Corresponding to P1 site, the S1 subsite was constituted by Asn188, Asn197, Glu224 and surrounded with acidic residues Glu150. They are expected to provide essential electrostatic interactions with the basic residue Arg/Lys at P1 position. Corresponding to P4 site, the S4 subsite was a large pocket with an assembly of two acidic residues Asp126, Glu129 at the bottom and one Asp151 at the rim [not shown in Figure 6(A)], which determine the preference for basic P4 residues. Similar modeling results could also be found in kexin, indicating that a basic residue at P4 site was favorable for both of two enzymes. However, in the case of kexin, the residue Thr121 was substituted for the corresponding Asp126 in furin. As a result, the effect of introducing a basic residue at P4 site in furin was more evident than that in kexin. Comparing the two mutants PVTR45 and RVTR45, the difference in their inhibitory activity was more than two orders of magnitude for furin, while only one order for kexin. Similar results were observed that kexin recognizes P4 site, with dual specificity for aliphatic and basic residues[26].

Although a basic residue at P2 is expected to provide advantageous electrostatic interactions with Asp46, Asp47 and Asn85 of S2 in furin (corresponding residues are Asp44, Asp45 and Asp80 in kexin), however, the structural context of eglin C itself disfavors such basic residue as Lys at P2. The T44K mutation at P2 destroys the hydrogen bonds between Thr44 and Arg53 in the molecular of eglin C, which are critical for stablizing the reactive site loop of eglin C. Furthermore, some other basic residues in eglin C (including Arg48, Arg51, Arg53) provide unfavorable electrostatic interactions with basic P2 residue [Figure 6(B)].

 

Fig.6       Ribbon-plot representation of the modeled complex structure of furin and eglin C RVKR45, made with MOLSCRIPT

(A) P1, P2, P3 and P4 of eglin C (partially shown) are highlighted in ball-and-stick in dark gray. The residues in furin predicted to be important for binding are shown in ball-and-stick in light gray, labeled with sequence and residue type (one character symbol). (B) Focusing on P2 and its interacting residues (residues in eglin C were labeled in box). (C) Focusing on P3 and its interacting residues (See text for details).

 

Look at the P3 in the binding interface [Figure 6(C)]. The residues Leu37 and Phe55 of eglin C, constitute a hydrophobic environment with the S3 residues Trp147 and Val124 of furin. Therefore, a hydrophobic residue like valine at this subsite is favorable for furin inhibitor. Mutation with polar residue like V43T led to an apparent decrease in the inhibitor activity. Particularly, the mutant RDKR45 became a weak inhibitor. Previous observations on furin substrates showed that in its optimal consensus recognition sequence Arg-Xaa-Lys-Arg, the P3 could be any amino acid except Cys, which meant furin displayed no obvious preference on P3 while recognizing and processing a substrate[27]. However, in the case of enzyme-inhibitor interaction, the P3 could have more important contributions not only to the conformation or stability of inhibitor itself, but also to the interaction between enzyme and inhibitor.

 

References

1     Steiner DF. The proprotein convertases. Curr Opin Chem Biol, 1998, 2(1): 31-39

2     Julius D, Brake A, Blair L, Kunisawa R, Thorner J. Isolation of the putative structural gene for the lysine-arginine-cleaving endopeptidase required for processing of yeast prepro-alpha-factor. Cell, 1984, 37(3): 1075-1089

3     Fuller RS, Brake AJ, Thorner J. Intracellular targeting and structural conservation of a prohormone-processing endoprotease. Science, 1989, 246(4929): 482-486

4     Nakayama K. Furin: A mammalian subtilisin/Kex2p-like endoprotease involved in processing of a wide variety of precursor proteins. Biochem J, 1997, 327(Pt 3): 625-635

5     Rouille Y, Duguay SJ, Lund K, Furuta M, Gong Q, Lipkind G, Oliva A Jr et al. Proteolytic processing mechanisms in the biosynthesis of neuroendocrine peptides: the subtilisin-like proprotein convertases. Frontiers in Neuroendocrinology, 1995, 16: 322-361

6     Bresnahan PA, Leduc R, Thomas L, Thorner J, Gibson H L, Brake A J, Barr P J, Thomas G. Human fur gene encodes a yeast KEX2-like endoprotease that cleaves pro-beta-NGF in vivo. J Cell Biol, 1990, 111(6 Pt 2): 2851-2859

7     Mondino A, Giordano S, Comoglio P M. Defective posttranslational processing activates the tyrosine kinase encoded by the MET proto-oncogene (hepatocyte growth factor receptor). Mol Cell Biol, 1991, 11(12): 6084-6092

8     Brennan SO, Nakayama K. Furin has the proalbumin substrate specificity and serpin inhibitory properties of an in situ hepatic convertase. FEBS Lett, 1994, 338(2): 147-151

9     Wasley LC, Rehemtulla A, Bristol JA, Kaufman RJ. PACE/furin can process the vitamin K-dependent pro-factor IX precursor within the secretory pathway. J Biol Chem, 1993, 268(12): 8458-8465

10    Pei D, Weiss SJ. Furin-dependent intracellular activation of the human stromelysin-3 zymogen. Nature, 1995, 375(6528): 244-247

11    Hallenberger S, Bosch V, Angliker H, Shaw E, Klenk HD, Garten W. Inhibition of furin-mediated cleavage activation of HIV-1 glycoprotein gp160. Nature, 1992, 360(6402): 358-361

12    Volchkov VE, Feldmann H, Volchkova VA, Klenk HD. Processing of the Ebola virus glycoprotein by the proprotein convertase furin. Proc Natl Acad Sci USA, 1998, 95(10): 5762-5767

13    Spaete RR, Thayer RM, Probert WS, Masiarz FR, Chamberlain S H, Rasmussen L, Merigan T C et al. Human cytomegalovirus strain Towne glycoprotein B is processed by proteolytic cleavage. Virology, 1988, 167(1): 207-225

14    Klimpel KR, Molloy S S, Thomas G, Leppla S H. Anthrax toxin protective antigen is activated by a cell surface protease with the sequence specificity and catalytic properties of furin. Proc Natl Acad Sci USA, 1992, 89(21): 10277-10281

15    Tsuneoka M, Nakayama K, Hatsuzawa K, Komada M, Kitamura N, Mekada E. Evidence for involvement of furin in cleavage and activation of diphtheria toxin. J Biol Chem, 1993, 268(35): 26461-26465

16    Anderson ED, Thomas L, Hayflick JS, Thomas G. Inhibition of HIV-1 gp160-dependent membrane fusion by a furin-directed alpha 1-antitrypsin variant. J Biol Chem, 1993, 268(33): 24887-24891

17    Jean F, Stella K, Thomas L, Liu G, Xiang Y, Reason AJ, Thomas G. Alpha1-antitrypsin Portland, a bioengineered serpin highly selective for furin: application as an antipathogenic agent. Proc Natl Acad Sci USA, 1998, 95(13): 7293-7298

18    Dahlen JR, Jean F, Thomas G, Foster DC, Kisiel W. Inhibition of soluble recombinant furin by human proteinase inhibitor 8. J Biol Chem, 1998, 273: 1851-1854

19    Basak A, Ernst B, Brewer D, Seidah NG, Munzer J S, Lazure C, Lajoie GA. Histidine-rich human salivary peptides are inhibitors of proprotein convertases furin and PC7 but act as substrates for PC1. J Pept Res, 1997, 49(6): 596-603

20    Lu W, Zhang W, Molloy SS, Thomas G, Ryan K, Chiang Y, Anderson S et al. Arg15-Lys17-Arg18 turkey ovomucoid third domain inhibits human furin. J Biol Chem, 1993, 268(20): 14583-14585

21    Heinz DW, Hyberts SG, Peng JW, Priestle JP, Wagner G, Grutter MG. Changing the inhibitory specificity and function of the proteinase inhibitor eglin C by site-directed mutagenesis: Functional and structural investigation. Biochemistry, 1992, 31(37): 8755-8766

22  Rink H, Liersh M, Sieber P, Meyer F. A large fragment approach to DNA synthesis: Total synthesis of a gene for the protease inhibitor eglin C from the leech Hirudo medicinalis and its expression in E. coli. Nucleic Acids Res, 1984, 12(16): 6369-6387

23    Ding DF, Tang HX, Zhang BH. A knowledge-based protein modeling system and its evaluation. Biophysica Sinica, 1995, 11(3): 416-428

24    Komiyama T, Fuller R S. Engineered eglin C variants inhibit yeast and human proprotein processing proteases, Kex2 and furin. Biochemistry, 2000, 39: 15156-15165

25    Siezen RJ, Creemers JW, Van de Ven WJ. Homology modelling of the catalytic domain of human furin. A model for the eukaryotic subtilisin-like proprotein convertases. Eur J Biochem, 1994, 222(2): 255-266

26    Rockwell N C, Fuller R S. Interplay between S1 and S4 subsites in Kex2 protease: Kex2 exhibits dual specificity for the P4 side chain. Biochemistry, 1998, 37: 3386-3391

27    Seidah NG, Chretien M. Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides. Brain Res, 1999, 848: 45-62

Received: June 18, 2001        Accepted: July 11, 2001

This work was supported by the Special Funds for Major State Basic Research of China (No.G1998051121) and State 863 High Technology R&D Project of China (No.103-13-01-03)

*Corresponding author: Tel, 86-21-64337152; Fax, 86-21-64338357; e-mail, [email protected]