Báo cáo khóa học: Nonlysine-analog plasminogen modulators promote autoproteolytic generation of plasmin(ogen) fragments with angiostatin-like activity

doi:10.1111/j.1432-1033.2004.03985.x

Eur. J. Biochem. 271, 809–820 (2004) (cid:1) FEBS 2004

Nonlysine-analog plasminogen modulators promote autoproteolytic generation of plasmin(ogen) fragments with angiostatin-like activity

Shigeki Ohyama, Tomotaka Harada, Toshihiro Chikanishi, Yutaka Miura and Keiji Hasumi

Department of Applied Biological Science, Tokyo Noko University, Saiwaicho, Fuchu-shi, Tokyo, Japan

mechanism in PMD generation. Thioplabin B and com- plestatin, two other nonlysine-analog modulators, were also active in producing similar PMDs, whereas the lysine analog 6-aminohexanoic acid was inactive while it enhanced plasminogen activation. Peptide sequencing and mass spectrometric analyses suggested that plasmin fragmenta- tion was due to cleavage at Lys615-Val616, Lys651-Leu652, Lys661-Val662, Lys698-Glu699, Lys708-Val709 and several other sites mostly in the catalytic domain. PMD was inhi- bitory to proliferation, migration and tube formation of endothelial cells at concentrations of 0.3–10 lgÆmL)1. These results suggest a possible application of nonlysine-analog modulators in the treatment of cancer through the enhancement of endogenous plasmin(ogen) fragment for- mation.

Keywords: angiostatin; plasminogen; plasmin; autoproteo- lysis; angiogenesis.

We recently discovered several nonlysine-analog conforma- tional modulators for plasminogen. These include SMTP-6, thioplabin B and complestatin that are low molecular mass compounds of microbial origin. Unlike lysine-analog mod- ulators, which increase plasminogen activation but inhibit its binding to fibrin, the nonlysine-analog modulators enhance both activation and fibrin binding of plasminogen. Here we show that some nonlysine-analog modulators promote autoproteolytic generation of plasmin(ogen) derivatives with its catalytic domain undergoing extensive fragmentation (PMDs), which have angiostatin-like anti-endothelial activ- ity. The enhancement of urokinase-catalyzed plasminogen activation by SMTP-6 was followed by rapid inactivation of plasmin due to its degradation mainly in the catalytic domain, yielding PMD with a molecular mass ranging from 68 to 77 kDa. PMD generation was observed when plasmin alone was treated with SMTP-6 and was inhibited by the plasmin inhibitor aprotinin, indicating an autoproteolytic

The plasminogen/plasmin system plays crucial roles not only in blood clot lysis but also in various physiological and pathological events including inflammation, tissue remode- ling, tumor metastasis and angiogenesis, where localized proteolysis is required [1–3]. Plasminogen consists of an N-terminal peptide, five kringle domains and a serine protease domain [4]. Two serine proteases, tissue-type plasminogen activator and urokinase-type plasminogen activator (u-PA), catalyze the activation of plasminogen by cleaving the Arg561-Val562 bond. The resulting plasmin in turn cleaves the Lys77-Lys78 and/or Lys78-Val79 bonds

resulting in conformational

to yield mature plasmin, which consists of two polypeptide chains that are held together by disulfide bridges. The heavy chain (A-chain; Lys78- or Val79-Arg561) contains five krin- gle domains and the light chain (B-chain; Val562-Asn791) has the serine protease domain. Native plasminogen resists activation, as it adpots a tight, spiral conformation [5] due to intramolecular binding of Lys50 and/or Lys62 in the N-terminal peptide to a lysine binding site (aminohexyl site) in the fifth kringle domain (K5) [6,7]. Fibrin and cellular receptors bind plasminogen and relax its conformation to be highly activatable, promoting efficient localized proteo- lysis [8–10]. Similarly, lysine analogs bind to plasminogen kringles, relaxation and enhancement of activation to plasmin [11]. Thus, conform- ational modulation of plasminogen is important for its activation.

Under certain circumstances, including tumor progres- sion and inflammation, plasmin(ogen) undergoes proteo- lysis to form various kringle-containing A-chain fragments, collectively called angiostatins [12–17]. It is postulated that physiological formation of angiostatin involves partial disulfide reduction in plasminogen and cleavages of plasmi- nogen by plasmin itself and/or other proteinases such as matrix metalloproteinases, metalloelastase and cathepsin D that are derived from tumor cells or infiltrating macro- phages [15,18–21]. Typical angiostatin consists of the first four kringle domains (K1–4) of plasmin [12,22,23]. Angio- statin and its relatives, including K1, K2, K3, K5, K1–3, K1–4½ and A61, inhibit the proliferation of vascular

Correspondence to K. Hasumi, Department of Applied Biological Science, Tokyo Noko University, 3-5-8 Saiwaicho, Fuchu-shi, Tokyo, 183–8509 Japan. Fax: + 81 42 3344661; Tel.: + 81 42 3675710; E-mail: hasumi@cc.tuat.ac.jp Abbreviations: u-PA, urokinase-type plasminogen activator; PMD, plasmin(ogen) derivative with its catalytic domain undergoing exten- sive fragmentation; 6-AHA, 6-aminohexanoic acid; VLK-pNA, H-Val-Leu-Lys-p-nitroanilide; MPB, 3-(N-maleimidylpropionyl)bio- cytin; HUVEC, human umbilical endothelial cells; FBS, fetal bovine serum; EGM, endothelial cell growth medium; bFGF, basic fibroblast growth factor; EGF, epidermal growth factor. Enzymes: human plasminogen (Swiss-Prot accession No. P00747; EC 3.4.21.7); human urokinase-type plasminogen activator (Swiss-Prot accession No. P00749; EC 3.4.21.73). (Received 20 October 2003, revised 30 December 2003, accepted 7 January 2004)

810 S. Ohyama et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

endothelial cells, which is a fundamental process in angio- genesis [12–16]. By inhibiting angiogenesis, angiostatin suppresses both in situ and metastatic tumor growths in animal models [12,15,19,22,23].

trichloroacetic acid, and the resulting precipitates were washed with acetone and dissolved in 11 lL of SDS-sample buffer [2% (w/v) SDS, 62.5 mM Tris/HCl, pH 6.8, 10% (w/v) sucrose and 0.02% (w/v) bromophenol blue] containing 5% (v/v) 2-mercaptoethanol. A portion (10 lL) of the mixture was subjected to SDS/PAGE on a 10% (w/v) gel, and the gel was stained with Coomassie Brilliant Blue R250.

Detection of kringle 1–3

We recently discovered novel nonlysine-analog modula- tors of plasminogen activation [24–29]. These include the microbial products staplabin, SMTPs and thioplabins. Such nonlysine-analog modulators enhance both plasminogen activation and plasminogen-fibrin binding, resulting in augmentation of fibrinolysis [24–26]. These effects are strikingly different from the action of lysine-analog modu- lators, which enhance plasminogen activation but inhibit plasminogen-fibrin binding and fibrinolysis [9]. Therefore, conformational change induced by such nonlysine-analog modulators may be distinct from changes induced by lysine analogs [24,26]. We thought that conformational modula- tion of plasminogen by nonlysine-analog modulators not only enhances plasminogen activation, but also confers unusual proteolytic susceptibility on plasmin(ogen). Here we show that one such compound (SMTP-6) promotes the autoproteolytic generation of plasmin(ogen) derivative with its catalytic domain (B-chain) undergoing extensive frag- mentation (designated PMD), which inhibits the prolifer- ation, migration and tube formation of vascular endothelial cells. Similar PMD generation is induced by two other structurally distinct nonlysine-analog modulators but not by a lysine analog modulator, 6-aminohexanoic acid (6-AHA).

peroxidase-conjugated

rabbit

Materials and methods

Materials

Plasmin(ogen) fragments containing K1–3 were detected by immunoblotting. Plasminogen (200 nM) and u-PA (100 UÆmL)1) were incubated in 60 lL of TBS/T at 37 (cid:2)C for 60 min and the mixture was treated with 10% (v/v) trichloroacetic acid to precipitate proteins, which were then washed with acetone and dissolved in 12 lL of SDS sample buffer. A portion (10 lL) of the mixture was subjected to SDS/PAGE on a 7.5% (w/v) gel, and proteins were transferred to poly(vinylidene difluoride) membrane. The membrane was incubated in NaCl/Pi (20 mM sodium phosphate and 150 mM NaCl, pH 7.4) containing 0.5% (v/v) Tween 20 and 5% (w/v) BSA at 4 (cid:2)C overnight, washed twice with NaCl/Pi containing 0.5% (v/v) Tween 20 and incubated in NaCl/Pi containing mouse anti-(human plasminogen K1-3) IgG (1 : 500 dilution), 0.5% (v/v) Tween 20 and 1% (w/v) BSA at room temperature for 1 h. After washing five times with NaCl/Pi containing 0.5% (v/v) Tween 20, the membrane was incubated in NaCl/Pi anti-(mouse containing IgG + A + M) (H + L) antibody (1 : 2000 dilution), 0.5% (v/v) Tween 20 and 1% (w/v) BSA at room temperature for 1 h and washed five times with NaCl/Pi containing 0.5% (v/v) Tween 20. Finally, the membrane was incubated in 0.1 M sodium phosphate, pH 6.4 contain- ing 0.2 mgÆmL)1 diaminobenzidine and 0.06% (v/v) H2O2 at room temperature for 10 min.

(H + L)

Human plasmin, bovine aprotinin and bovine insulin were from Wako, Osaka, Japan; u-PA from JCR Pharmaceuti- cals, Kobe, Japan; H-Val-Leu-Lys-p-nitroanilide (VLK- pNA) from Bachem, Bubendorf, Switzerland; monoclonal anti-(human plasminogen K1–3) IgG (2PG) from Tech- noclone, Vienna, Austria; peroxidase-conjugated rabbit anti-(mouse IgG + A + M) from ZYMED Laboratories, San Francisco, CA, USA. SMTP-6, thiopla- bin B (antibiotic A10255B) and complestatin were isolated from cultures of Stachybotrys microspora IFO30018, Strep- tomyces sp. R1401 and Streptomyces sp. A1631, respectively [25–27]. Human plasminogen was affinity-purified from human plasma [30]. Angiostatin K1–3 was purified by lysine-Sepharose affinity chromatography from porcine elastase-digested plasminogen as described previously [22].

A time course of plasmin generation

Detection of free thiol Plasminogen (100 nM) and u-PA (50 UÆmL)1) were incu- bated in the absence or presence of 200 lM SMTP-6 in 150 lL of TBS/T at 37 (cid:2)C for 60 min. The resulting mixture was treated with 3-(N-maleimidylpropionyl)biocytin (MPB; Molecular Probes, Eugene, OR, USA) to detect free thiol groups in proteins as described previously [31,32] with slight modification. Briefly, reactions were labeled with MPB (100 lM) for 30 min at room temperature, followed by quenching of the unreacted MPB with reduced glutathione (200 lM) for 10 min at room temperature. Unreacted glutathione and other free sulfhydryls in the system were blocked with N-ethylmaleimide (400 lM) for 10 min at room temperature. The resulting mixtures were treated with 10% (v/v) trichloroacetic acid to precipitate proteins, which were then washed with acetone. Proteins were resolved on nonreduced 10% (v/v) SDS/PAGE, transferred to nitrocel- lulose membrane, and blotted with streptavidin-peroxidase (Sigma) to detect the MPB-labeled proteins.

Identification of plasmin fragments

u-PA-catalyzed activation of plasminogen was assayed by VLK-pNA hydrolysis and SDS/PAGE. Plasminogen (100 nM) was incubated at 37 (cid:2)C for up to 60 min with 50 UÆmL)1 (6.3 nM) u-PA in TBS/T [50 mM Tris/HCl, 100 mM NaCl and 0.01% (w/v) Tween 80, pH 7.4]. After incubation, the mixture received 1/9 volume of 1 mM VLK- pNA in TBS/T and was further incubated at 37 (cid:2)C to measure the absorbance at 405 nm with 2 min intervals. From the initial rate of the release of pNA, the plasmin concentration was determined. For determination of mole- cular species on SDS/PAGE, the above-mentioned reaction mixture (incubated for 0–60 min) was treated with 10% (v/v)

Both MALDI-TOF/MS and automated Edman degrada- tion were used to identify plasmin fragments. Plasmin

Nonlysine-analogs promote plasmin autoproteolysis (Eur. J. Biochem. 271) 811

(cid:1) FEBS 2004

Cell culture

(100 nM) was incubated in 12 mL of TBS/T at 37 (cid:2)C for 60 min in the presence of SMTP-6 (250 lM), thioplabin B (10 lM) or complestatin (0.3 lM), and proteins were preci- pitated by treatment with 10% (v/v) trichloroacetic acid.

Human umbilical vein endothelial cells (HUVEC; obtained from Clonetics, San Diego, CA, USA) were grown in endothelial cell growth medium (EGM; Clonetics) contain- ing 2% (v/v) fetal bovine serum. HT1080 cells and rat primary fibroblasts were maintained in DMEM containing 10% (v/v) fetal bovine serum, and Lewis lung carcinoma cells and CHO-K1 cells were grown in RPMI 1640 supplemented with 10% (v/v) fetal bovine serum.

Cell proliferation assay

For MALDI-TOF/MS analyses, proteins (50 lg) were dissolved in a mixture consisting of 41 lL of 8 M urea, 3 lL of 5% (v/v) EDTA, 30 lL of 1.44 M Tris/HCl, pH 8.6, and 1 lL of 14.4 M 2-mercaptoethanol. After incubation at room temperature for 4 h under N2 gas, the mixture was incubated with 10 lL of 1.44 M iodoacetic acid and 1.0 M NaOH at room temperature in the dark for 15 min. The resulting reduced, S-carboxymethylated peptides were applied to a silica-C18 spin column equilibrated with 0.1% (v/v) trifluoroacetic acid. The column was washed with 0.1% (v/v) trifluoroacetic acid and peptides were eluted with 0.1% (v/v) trifluoroacetic acid and 50% (v/v) CH3CN. The eluate was treated with ZipTip C18 (Millipore) and peptides were recovered in 4 lL of 0.1% (v/v) trifluoroacetic acid and 50% (v/v) CH3CN. A portion (0.5 lL) was analyzed by MALDI-TOF/MS on a Voyager DE-STR (Applied Biosystems) employing a linear mode positive ionization. Sinapinic acid was used as a matrix and bovine insulin as an internal standard for calibration. Results obtained were compared with a panel of theoretical peptide masses calculated for trypsin-cleaved, S-carboxymethylated plas- min fragments.

The proliferation of endothelial cell was assayed as described [13,15] with slight modification. HUVEC were dispersed in trypsin/EDTA solution (Clonetics) and resus- pended with EGM defective in human epidermal growth factor (EGF). Cells (2500 cells in 100 lL) were seeded into 96-well plates and incubated at 37 (cid:2)C overnight. The medium was replaced by 100 lL of fresh medium contain- ing PMD or angiostatin. After incubation for 30 min, 100 lL of EGM containing 20 ngÆmL)1 EGF was added, and cells were incubated for 3 days. After discarding the medium, fresh medium (100 lL) containing 0.5 mM 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)- 2H-tetrazolium (WST-1 reagent; Dojindo Laboratories, Kumamoto, Japan), a water-soluble formazan-forming reagent, was added at 37 (cid:2)C for 4 h. Absorbance at 405 nm was measured with a reference at 655 nm as an index of cell proliferation.

The growth of nonendothelial cells was assayed by a similar method employed for the assay for HUVEC using respective growth medium shown above.

Endothelial cell migration assay

For peptide sequencing, proteins (50 lg) were dissolved in 20 lL of SDS sample buffer containing 5% (v/v) 2-mercaptoethanol. A portion (10 lL) of the mixture was subjected to SDS/PAGE on a 20% (w/v) gel (Tris/ Tricine buffer system) and transferred to poly(vinylidene difluoride) membrane. The membrane was stained with Coomassie Brilliant Blue R250 and a section of mem- brane was subjected to N-terminal amino acid sequence analysis on a model 476A protein sequencer (Applied Biosystems).

Preparation and purification of PMDs for biological assays Plasminogen (100 nM) and u-PA (50 UÆmL)1) were incu- bated with SMTP-6 (250 lM), thioplabin (20 lM) or complestatin (0.3 lM) in TBS/T at 37 (cid:2)C for 60 min. After adding phenylmethanesulfonyl fluoride to a final concen- tration of 1 mM, the mixture was dialyzed against water at 4 (cid:2)C and lyophilized. The resulting sample was used directly for assays after filter-sterilization. The endotoxin levels in the PMD-S preparation, as determined using a Kinetic- QCL kit (BioWhittaker; Walkersville, MD, USA), were in a range of 0.082–0.21 ngÆmg)1 protein.

A modified Boyden chamber method was used [33]. Polycarbonate membrane with 8 lm pores was coated with 15 lg of fibronectin and placed over a 24-well culture plate containing 500 lL of EGM supplemented with 0.1% (v/v) BSA and 10 ngÆmL)1 bFGF. After 30 min, HUVEC that had been starved overnight in serum- and growth factor-free EGM containing 0.1% (v/v) BSA were seeded into the upper chamber (1.1 · 105 cells in 500 lL of EGM contain- ing 20 lgÆmL)1 of PMD or angiostatin) and incubated for 5 h. The upper chamber was removed and nonmigrated cells were scraped off with cotton swabs. The migrated cells were fixed with 70% (v/v) ethanol for 30 min, stained with Giemsa for 30 min and counted under a light microscope at ·100 magnification.

Endothelial cell tube formation assay Approximately 4.0 · 104 cells in 100 lL of MCDB 131 medium were seeded into a 24-well culture plate containing 200 lL of collagen gel. After incubation at 37 (cid:2)C for 2 h, medium was discarded, and cells were overlaid with 200 lL of collagen gel. After gel formation for 30 min, each well received 200 lL of MCDB 131 medium. Both medium and collagen gel contained 10 lgÆmL)1 of PMD or angiostatin. After 24 h of incubation, cultures were examined for tube formation under a light microscope at ·40 magnification.

For HPLC purification, the lyophilized sample was dissolved in 0.1% (v/v) trifluoroacetic acid and subjected to preparative HPLC on a lBONDASPHERE 5 l C8– 300 A˚ (19 · 150 mm; Waters), which was developed with a linear gradient of 2-propanol in water containing 0.1% (v/v) trifluoroacetic acid at a rate of 5 mLÆmin)1. The concen- tration of 2-propanol was 0% for the initial 20 min and increased to 100% at a rate of 0.8%Æmin)1 thereafter. The elution was monitored by UV absorption at 210 nm, and a peak at a retention time of 60–62.5 min was pooled and lyophilized. The lyophilized sample was dissolved in NaCl/ Pi and filter-sterilized before assay.

812 S. Ohyama et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

The length of tubes in 8–9 fields were measured and expressed as percent of control.

Results

Nonlysine-analog modulators induce autoproteolytic plasmin(ogen) fragmentation

these incubations (Fig. 2E,F). In the absence of u-PA, the three nonlysine-analogs induced neither plasminogen acti- vation nor its fragmentation (data not shown). These results suggested that these nonlysine-analog modulators enhanced not only plasminogen activation (cleavage at Arg561- Val562 bond by u-PA) but also its fragmentation mainly in the B-chain, which harbors the catalytic domain. This may account for the transient increase in plasmin activity. The apparent molecular masses of the plasmin(ogen) derivative PMD produced in the presence of the nonlysine- analog modulators were estimated to range from 68 to 77 kDa on SDS/PAGE under nonreducing conditions (Fig. 3A). In immunoblot analysis, the PMD reacted with anti-(K1-3) IgG (Fig. 3B). The molecular size of plasmino- gen was estimated to be rather smaller in nonreduced SDS/ PAGE: 84–92 kDa (Fig. 3A) vs. 100–110 kDa in reduced SDS/PAGE (Fig. 2). Other investigators have reported similar observations with A-chain fragments. For example, the apparent sizes observed in nonreduced and reduced SDS/PAGE, respectively, are reported to be 45 kDa and 66 kDa for (cid:2)K1–4 + 69 residues of K5(cid:3) [32], and 50 kDa and 61 kDa for (cid:2)K1-4 + 7 residues of K5(cid:3) [16]. Thus, the molecular mass of PMDs at 68–77 kDa in nonreduced SDS/PAGE was higher than that of such A-chain fragments by 18–32 kDa. Therefore, the data in Figs 2 and 3 suggested that PMDs consist of not only A-chain but also fragment(s) of B-chain that are connected by disulfide bridges.

To test whether the nonlysine-analog-mediated plas- min(ogen) degradation was due to cleavage by u-PA or plasmin, we incubated plasminogen and u-PA in the presence of aprotinin at a concentration of 1000 kallikrein inhibitor units per mL, which inhibits plasmin but not u-PA. As shown in Fig. 3C, the nonlysine-analog-mediated

SMTP-6 (Fig. 1A), one of the nonlysine-analog plasmino- gen modulators, enhances the initial rate of u-PA-catalyzed plasminogen activation [25]. We found that the SMTP-6 enhancement of plasminogen activation was transient and that a rapid decrease in plasmin activity followed this enhancement (Fig. 2A). On the other hand, u-PA-catalyzed plasminogen activation in control incubation proceeded constantly for up to 60 min. The lysine analog 6-AHA markedly enhanced plasmin generation throughout the incubation period. The time-dependent increase in plasmin activity in the control and 6-AHA incubations accompanied an increase in the intensity of protein bands both at 68/70 kDa doublet and 27.2 kDa in reduced SDS/PAGE analysis, which corresponded to A- and B-chains of plasmin, respectively (Fig. 2B,C). In incubation in the presence of SMTP-6, A-chain band appeared to be smeared while its intensity increased time-dependently, whereas the intensity of B-chain band increased transiently and decreased after 10 min of incubation (Fig. 2D). Thiopla- bin B and complestatin, two other structurally distinct nonlysine-analog modulators (Fig. 1B,C), also caused rapid increases in plasmin generation, which was halted after 20–30 min of incubation (Fig. 2A). The level of plasmin B-chain appeared to be consistent with plasmin activity, while A-chain production increased time-dependently in

Fig. 1. Structures of SMTP-6 (A), thioplabin B (B) and complestatin (C).

Nonlysine-analogs promote plasmin autoproteolysis (Eur. J. Biochem. 271) 813

(cid:1) FEBS 2004

production of PMDs was completely inhibited by aprotinin. Furthermore, the incubation of plasmin alone in the presence of SMTP-6 produced extensive B-chain degrada- tion, whereas the A-chain appeared to remain intact (Fig. 3D). These results demonstrated an involvement of an autoproteolytic mechanism in the nonlysine-analog- mediated generation of PMDs.

Structural characterization of PMD

S-carboxymethylation of plasmin derivatives. Because plasmin cleaves lysyl and arginyl bonds, results obtained were compared with a panel of theoretical peptide masses calculated for S-carboxymethylated plasmin cleaved at lysyl and arginyl bonds. Seven to 13 peptide fragments were identified within a difference of 1.2 mass units in compar- ison with the theoretical peptide masses calculated for each fragments (Table 2). All but one fragment species identified were assigned to be from the B-chain. Taking the data obtained by the two methods together, autoproteolytic cleavage in plasmin incubated in the presence of the three nonlysine-analog modulators was assigned to occur at least at the positions shown in Table 3.

Figure 3F shows a time course for B-chain fragmentation in the presence of SMTP-6. The results demonstrated that peptide bands at 9.9 and 8.1 kDa appeared after 5–10 min, and their intensity remained constant after 20–60 min. This observation suggested that the two peptides at 9.9 kDa (with N-terminal Glu699) and 8.1 kDa (with N-terminal Val709) (Table 1) have been early intermediates in a sequence of fragmentation events. Therefore, it is likely that Lys698-Glu699 and/or Lys708-Val709 are most sus- ceptible cleavage sites in the B-chain. The cleavage at such sites may enable subsequent cleavage at many other sites.

It has been reported that autoproteolytic generation of is

the A-chain fragment (cid:2)K1–4 + 69 residues of K5(cid:3)

To determine the cleavage sites in PMD, we employed both N-terminal amino acid sequencing and MALDI-TOF/MS analyses. Plasmin was incubated with the nonlysine-analog modulators to allow autoproteolytic formation of PMDs, which were then reduced and fractionated by SDS/PAGE on a 20% (w/v) gel for sequence analysis by automated Edman degradation. The amino acid sequences of bands numbered in Fig. 3E were reasonably assigned as shown in Table 1. We could not determine sequences of fragment bands at 17–22 kDa because of the appearance of multiple amino acid peaks in the early part of several cycles in their analyses. The assignments demonstrated that, although one fragment at 14.9 kDa (Fig. 3E, band 1) was derived from the A-chain (starting with Val79), all the remaining fragments detected were from B-chain. MALDI-TOF/ MS analyses were performed after the reduction and

Fig. 2. Nonlysine-analog modulators enhance both initial plasminogen activation and subsequent fragmentation of plasmin(ogen). (A) Plasminogen (100 nM) was incubated with u-PA (50 UÆmL)1) in the absence (d) or presence of 20 mM 6-AHA (s), 250 lM SMTP-6 (n), 30 lM thioplabin B (h) or 0.3 lM complestatin (e). After incubation for the indicated time, plasmin activity was measured by adding VLK-pNA. Each value represents the mean ± SD from triplicate determinations. (B–F), SDS/PAGE analyses of plasmin(ogen) in the first incubation in (A). Aliquot of the incubation (2 lg protein) in the absence (B) or presence of 6-AHA (C), SMTP-6 (D), thioplabin B (E) or complestatin (F) was reduced and subjected to SDS/ PAGE on a 10% gel. The positions of molecular mass markers (left) as well as of plasminogen (Glu-Plg) and A- and B-chains of plasmin (right) are shown.

814 S. Ohyama et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

facilitated after cleavage of a disulfide bond in K5 under alkaline pH conditions [15] or through phosphoglycerate kinase-mediated conformational change [31]. Such disulfide bond cleavage accompanies the formation of a free thiol in K5. To test for the possibility that similar disulfide bond cleavage had occurred during PMD formation, PMD generated in the presence of SMTP-6 was subjected to labeling with the thiol-specific reagent MPB, followed by probing with streptavidin-peroxidase blotting. As shown in Fig. 3G, MPB labeling was not detected in PMD, suggest- ing the presence of no free thiol in PMD.

60–62.5 min was pooled (Fig. 5). PMD partially purified from plasminogen that had been incubated with u-PA in the presence of SMTP-6 (PMD-S) appeared as a diffuse band at 68–77 kDa in SDS/PAGE analysis under nonreducing conditions (Fig. 5, inset). Unlike angiostatin, the HPLC- purified PMD-S failed to bind lysine-Sepharose (data not shown). This may be due to the denaturation of kringles by exposure to 0.1% (v/v) trifluoroacetic acid and 2-propanol during the HPLC purification. PMDs with similar mole- cular mass were obtained from incubations in the presence of thioplabin B and complestatin (PMD-T and PMD-C, respectively).

Although PMDs are a heterogeneous mixture, we propose some possible structures from these observations, shown schematically in Fig. 4.

Inhibition of proliferation, migration and tube formation of HUVEC by PMD-S

Purification of PMDs

As angiostatins inhibit the proliferation of vascular endo- thelial cells in culture, we examined the activity of PMD using HUVEC. In these experiments, we used nonpurified

PMDs were fractionated by reverse-phase HPLC using a silica-C8 column, and a major peak at a retention time of

Fig. 3. SDS/PAGE analyses of PMDs. (A) Plasminogen (100 nM; lane 1) was incubated with u-PA (50 UÆmL)1) for 60 min in the absence (lane 2) or presence of 20 mM 6-AHA (lane 3), 250 lM SMTP-6 (lane 4), 10 lM thioplabin B (lane 5) or 0.3 lM complestatin (lane 6), and aliquot of the mixture (2 lg protein) was analysed by nonreduced SDS/PAGE on a 7.5% gel. (B) Plasminogen (200 nM) and u-PA (100 UÆmL)1) were incubated for 60 min in the absence (lane 1) or presence of 250 lM SMTP-6 (lane 2). An aliquot (2 lg protein) was subjected to nonreduced SDS/PAGE, followed by immunoblot analysis using anti-(kringle 1–3) IgG. (C) Plasminogen was treated as in (A), except that aprotinin (1000 KIUÆmL)1) was included in the mixture. Lane numbers are the same as those in (A). (D) Plasmin (100 nM) was incubated for 60 min in the absence (lane 2) or presence of 250 lM SMTP-6 (lane 3). An aliquot (2 lg protein) was analysed by reduced SDS/PAGE on a 12.5% gel. Lane 1 represents nontreated plasmin (2 lg). (E) Plasmin (100 nM) was incubated for 60 min in the presence of 250 lM SMTP-6 (lane 1), 30 lM thioplabin B (lane 2) or 0.3 lM complestatin (lane 3) to allow autoproteolytic fragmentation. An aliquot (10 lg protein) was subjected to reduced SDS/PAGE on a 20% gel using a Tris/Tricine buffer system to analyse low molecular mass fragments. N-terminal amino acid sequences of the fragments with numbered arrows were determined by automated Edman degradation. (F) Plasminogen (100 nM) and u-PA (50 UÆmL)1) were incubated in the presence of 200 lM SMTP-6 for 0–60 min. After incubation, an aliquot (5 lg protein) was resolved on 20% SDS/PAGE under reducing conditions as described above. Arrows denote key fragments. (G) Plasminogen (100 nM) and u-PA (50 UÆmL)1) were incubated in the absence (lane 3) or presence of 200 lM SMTP-6 (lane 4) for 60 min. After incubation, proteins (1.4 lg) were labeled with MPB and resolved on 10% SDS/PAGE. As controls, reduced plasminogen (1.4 lg; lane 1) and native plasminogen (1.4 lg; lane 2) were also treated with MPB, followed by SDS/PAGE. The gel was processed for Coomassie blue staining (left) or blotting with streptavidin–peroxidase to detect MPB-labeled proteins (right). The positions of molecular mass markers as well as of A- and/or B-chains of plasmin are shown.

Nonlysine-analogs promote plasmin autoproteolysis (Eur. J. Biochem. 271) 815

(cid:1) FEBS 2004

Table 1. N-terminal sequence analysis of PMDs generated in the presence of SMTP-6, thioplabin B and complestatin. Plasmin (100 nM) was incubated for 60 min in the presence of SMTP-6 (250 lM), thioplabin B (30 lM) or complestatin (0.3 lM), and aliquot of the mixture (25 lg protein) was subjected to SDS/PAGE on a 20% gel and transferred to membrane. After staining, peptide bands were excised from membrane and subjected to N-terminal amino acid sequence analysis. The fragment number corresponds to that in Fig. 3E. X represents no significant recovery.

Amino acid in cycle (recovery in pmol)

Assignment Treatment No. kDa 1 2 3 4 5

SMTP-6 1 14.9 V (0.9)

E (1.8) P (2.4) Y (5.1) 2 3 4 9.9 8.1 5.7 V (0.1) Y (0.7) A (4.0) X V (15.3) G (0.2) L (0.5) Q (5.4) N (5.9) G (15.2) E (4.2) V (10.7) V (15.0) L (4.0) G (0.2) S (0.4) L (4.6) R (3.1) G (17.1) P (2.2) A (2.6) 5 4.7 V (8.2) G (6.1) V (5.5) I (4.2) G (4.2) L (3.6) A (5.6) L (4.5) P (2.6) V562– V79– E699– V709– V562– L652– V709– V616– L652–

Thioplabin B 1 14.9 G (7.2) V (12.1)

2 3 9.9 8.1 E (3.0) V (9.8) V (10.1) Y (3.0) A (4.7) X G (6.8) L (2.7) Q (4.7) N (3.0) L (4.5) R (4.2) E (1.8) X X V562– V79– E699– V709–

Complestatin 1 14.9 G (5.0) V (6.2)

E (7.5) 2 3 9.9 8.1 V (6.2) Y (1.3) A (10.3) I (4.2) V (10.1) G (4.6) L (1.4) Q (9.5) L (4.1) N (3.0) L (8.6) G (4.6) R (4.2) E (0.9) P (4.3) A (4.3) Y (3.4) V562– V79– E699– V616– V709–

Discussion

SMTP-6, thioplabin B and complestatin are modulators of plasminogen activation. These low molecular mass com- pounds are not structurally related to lysine or its analogs that bind to plasminogen kringles. Smilarly to lysine analog modulators, such nonlysine-analogs enhance plasminogen activator-catalyzed plasminogen activation by modulating plasminogen conformation. The difference between the two groups of plasminogen modulators is that the nonlysine- analogs enhance plasminogen-fibrin binding and fibrinolysis while lysine analogs inhibit these activities. We show in the present study that the nonlysine-analog modulators, but not lysine analogs, enhance u-PA-catalyzed activation of plasminogen (cleavage at Arg561-Val562) and subsequent fragmentation of plasmin(ogen) mainly in the catalytic domain. This fragmentation causes inactivation of plasmin and formation of PMD. The finding that the lysine analog 6-AHA protects plasmin(ogen) from degradation (Fig. 2C) is consistent with the observation by Ueshima et al. [35] and supports the idea that conformational change induced by the nonlysine-analogs is distinct from that induced by lysine analogs. The extensive B-chain degradation in the presence of the nonlysine-analogs suggests that these agents bind to a region of plasmin(ogen) that affects the topology or conformation of the catalytic domain.

Plasmin(ogen) fragmentation in the presence of the nonlysine-analogs is due to an autoproteolytic mechanism. This conclusion is supported by the observations that aprotinin inhibits the PMD generation at a concentration that enables u-PA cleavage and that the incubation of plasmin alone in the presence of SMTP-6 results in extensive

PMD preparation produced in the presence of SMTP-6 (PMD-S), as the preparation consisted mainly of 68–77 kDa molecular species even before purification (Fig. 5). As shown in Fig. 6A, PMD-S inhibited the pro- liferation of HUVEC at concentrations of 0.3–3 lgÆmL)1. In contrast to the inhibitory effect on vascular endothelial cells, PMD-S did not affect the growth of Lewis lung carcinoma, HT-1080 human fibrosarcoma, rat primary fibroblasts and CHO-K1 cells (Fig. 6B). These activities of PMD-S were similar to those of angiostatin K1–3 obtained from elastase-digested plasminogen. We next asked whe- ther PMD-S is inhibitory to migration and tube formation of endothelial cells, as angiostatins are also inhibitory to such activities characteristic of endothelial cells [13–15,33]. PMD-S inhibited migration of HUVEC by 17 ± 8% at 10 lgÆmL)1, while the change was not statistically signifi- cant. PMD-S (10 lgÆmL)1) also suppressed tube formation of HUVEC by 43 ± 21% (P < 0.01 relative to control by Student’s t-test), which was similar to the inhibition by K1–3 at 10 lgÆmL)1 (44 ± 22%, P < 0.01). The concen- tration of PMD-S at 0.3–10 lgÆmL)1 corresponds to 4.1–135 nM when taken 73 kDa as a mean molecular mass. These values were comparable to previously reported effective concentrations for angiostatins including K1–3, K1–4 and (cid:2)K1–4 + 7 residues of K5(cid:3), which exert anti- endothelial activity at 10–300 nM [12,13,16]. The levels of endotoxin in PMD-S preparation were in a range of 0.082– 0.21 ng per mg protein. At the highest concentration of PMD-S used (10 lgÆmL)1), the endotoxin level reached 2.1 pgÆmL)1, which was far below the lowest cytotoxic concentration of endotoxin (1 lgÆmL)1) reported by Lucas et al. [34].

8 1 6

S

.

O h y a m a

e t

a l .

( E u r .

J .

B i o c h e m

Table 2. Mass spectrometric analysis of PMDs generated in the presence of SMTP-6, thioplabin B and complestatin. Plasmin (100 nM) was incubated for 60 min in the presence of SMTP-6 (250 lM), thioplabin B (30 lM) or complestatin (0.3 lM) to allow autoproteolytic generation of PMDs, which were then reduced and S-carboxymethylated. The samples were analyzed by MALDI-TOF/MS. Mcalc, calculated theoretical peptide mass; Mmeas, measured peptide mass; DM, Mcalc – Mmeas; accuracy, DM/Mcalc · 100 (%).

.

SMTP-6 Thioplabin B Complestatin

2 7 1 )

Assignment DM Accuracy DM Accuracy DMS Accuracy Mcalc Mmeas Mmeas Mmeas

1773.0405 ± 0.8748 2174.2918 ± 0.7226 0.0278 0.2090 0.0016 0.0096 1428.6779 ± 0.8308 1771.9942 ± 0.6562 2173.6048 ± 0.6312 2693.6986 ± 0.6484 0.9251 1.0741 0.8960 0.4434 0.0647 0.0606 0.0412 0.0165 1429.6506 ± 0.2611 1772.8040 ± 0.5744 2174.4891 ± 0.7819 2693.4624 ± 1.5563 3242.3389 ± 0.3630 0.0476 0.2643 0.0117 0.6796 0.1698 0.0033 0.0149 0.0005 0.0252 0.0052 3483.5837 ± 0.7011 0.5616 0.0161

3549.3702 ± 1.6531 4029.4426 ± 1.0531 0.3867 0.1530 0.0109 0.0038 4029.5173 ± 0.8347 4137.2671 ± 0.4110 0.0783 0.4222 0.0019 0.0102

3548.5076 ± 0.5684 4028.9634 ± 0.7941 4137.3315 ± 0.4839 4499.1628 ± 1.0945 5040.7673 ± 0.7484 0.4759 0.6322 0.4866 0.1188 1.0033 0.0134 0.0157 0.0118 0.0026 0.0199 5042.1589 ± 1.2970 0.3883 0.0077

5042.1731 ± 0.7885 6161.7105 ± 0.2553 6182.1405 ± 0.5588 0.4025 0.8195 1.0272 0.0080 0.0133 0.0166

(cid:1) F E B S

2 0 0 4

1429.6030 1773.0683 2174.5008 2694.1420 3242.5087 3483.0221 3548.9835 4029.5956 4136.8449 4499.2816 5041.7706 6160.8910 6181.1133 7722.7671 7950.9755 10255.6379 10483.8463 V709-R719 V662-R677 V562-R580 Y753-R776 V720-K750 V616-K645 L532-R561 V662-K698 V616-K651 D751-R789 L652-K698 V720-R776 V562-K615 V720-R789 V720-N791 E699-R789 E699-N791 6181.8638 ± 1.2217 7722.9855 ± 1.8527 7952.1167 ± 5.1018 10256.2055 ± 2.2843 10484.3362 ± 2.6192 0.7505 0.2184 1.1412 0.5676 0.4899 0.0121 0.0028 0.0144 0.0055 0.0047

Nonlysine-analogs promote plasmin autoproteolysis (Eur. J. Biochem. 271) 817

(cid:1) FEBS 2004

Table 3. Assignment of autoproteolytic cleavage site in plasmin treated with the three nonlysine-analog modulators.

Number of fragment species for assignment

Cleavage site SMTP-6 Thioplabin B Complestatin

1 (E)

1 (E)b 1 (M) 1 (M) 1 (E), 1 (M) 1 (E) 1 (M) 1 (M) 1 (E), 1 (M)

2 (E), 3 (M) 2 (M) 1 (M) 1 (E), 3 (M) 2 (E) 1 (M) 3 (M) 1 (M) 2 (M) 2 (M) 1 (M) 1 (E), 2 (M) 1 (E), 1 (M) 2 (M) 1 (M)

a Probable assignment based on the apparent molecular mass of the fragment initiating with V79. b E, Edman degradation; M, mass spectrometry.

1 (M) 2 (M) K204 to N213a K531–L532 R580–T581 K615–V616 K645–D646 K651–L652 K661–V662 R677–T678 K698–E699 K708–V709 R719–V720 K750–D751 K752–Y753 R776–V777 R789–N790 1 (M) 1 (M) 2 (M) 1 (M) 1 (E), 3 (M) 1 (E), 1 (M) 4 (M) 1 (M) 1 (M) 1 (M) 2 (M)

and degraded B-chain that are connected by disulfide bonds. N-terminal sequencing and mass spectrometric analysis of PMD support this idea. Unless disulfide cleavage occurs, the fragment containing N-terminal Val79 must be connected with the B-chain even when the A-chain is fully

B-chain degradation. Autoproteolytic degradation of plas- min has been investigated by several investigators [35–39]. Shi and Wu have observed two autoproteolytic processes for plasmin inactivation. In a slightly acidic solution (pH 6.5), the B-chain is preferentially cleaved, while the A-chain is cleaved in an alkaline solution of pH near 11. Both A-chain and B-chain are cleaved at pH levels between 6.5 and 11 [37]. Thus, our experimental conditions at pH 7.4 may allow degradation of both chains of plasmin. In short- term incubations (< 60 min), however, we find minimum degradation in both A-chain and B-chain in the absence of the nonlysine-analog modulators, whereas degradation mainly in the B-chain was observed in their presence (Figs 2 and 3D). Therefore, it is likely that the agents selectively affect an autoproteolytic process responsible for B-chain fragmentation. Waisman and coworkers have reported that the annexin II tetramer, a major Ca2+-binding protein of endothelial cell surfaces, leads plasmin to autoproteolytic inactivation [38,39]. Annexin II tetramer can induce conformational change in plasminogen and stimulate its activation by plasminogen activators [40], but it does not compete with plasminogen for fibrin binding [39]. In these respects, the nonlysine-analog modulators are similar to the annexin II tetramer, while the autoproteolysis stimulated by the protein is found both in the A-chain and B-chain [37]. The plasmin(ogen) fragment PMD produced in the presence of the nonlysine-analog modulators is a hetero- geneous mixture. This is true even for PMD purified by HPLC as shown in Fig. 5. The complexity of the molecular constitution of PMD has made it difficult to purify to homogeneity in sufficient quantities to elucidate the exact structure. We nevertheless believe that PMD of 68–77 kDa is a mixture of molecules, as shown in Fig. 4 and from the following observations. The pattern and apparent molecular masses in SDS/PAGE analyses are consistent with the idea that PMDs of 68–77 kDa consist of almost intact A-chain

Fig. 4. Schematic representation of some examples for possible struc- tures of PMDs generated in the presence of the three nonlysine-analog modulators. In the plasmin molecule (top), all of the possible auto- proteolytic cleavage sites are shown as circles. Sites at which cleavage has been found in all the three PMD preparations are filled black, and sites where cleavage has been found in one or two preparations are filled gray. The site with asterisk denotes tentative assignment (Table 3 for detail). Dashed lines represent disulfide bridges. In addition to an extensively cleaved example (bottom), two possible structures without cleavage in the A-chain are shown, because the A-chain appears to be far more stable than does the B-chain.

818 S. Ohyama et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

at Cys536 [31]. The Cys536 thiol is then available to exchange with the Cys462-Cys541 disulfide bond, resulting in formation of a new disulfide bond at Cys536-Cys541 and a free thiol at Cys536. Although the nonlysine-analogs induce conformational change in the plasminogen molecule, it is unlikely that in PMD formation such conformational change facilitates a similar sequence of events that are achieved by phosphoglycerate kinase, as our experiments have failed to detect free thiols in the PMD preparation.

Similarly to angiostatin, PMD-S, which is generated in the presence of SMTP-6, inhibits the proliferation, migra- tion and tube formation of endothelial cells. The inhibition is observed at concentrations of 0.3–10 lgÆmL)1 (4.1– 135 nM), which are similar to the concentration required for the angiostatin action [12,13,16]. Thus, although the composition of fragments may differ slightly among the three PMD preparations (PMD-S, PMD-T and PMD-C; Tables 1–3), it is likely that PMD-T and PMD-C also have angiostatin-like activity. This observation may extend the previous understanding that several kringle combinations as well as isolated kringles of plasminogen have anti- endothelial activity [12,13,17].

Fig. 5. Purification of PMDs. (A) PMDs produced in the presence of 250 lM SMTP-6 (PMD-S), 30 lM thioplabin B (PMD-T) or 0.3 lM complestatin (PMD-C) were purified by reverse-phase HPLC using a silica-C8 col- umn. A chromatogram for the purification of PMD-S is shown. The bar above the major peak represents fractions pooled for SDS/ PAGE analysis. The inset shows the results of nonreduced SDS/PAGE analysis of PMDs (2 lg protein) before (lane M) and after (lane P) purification.

Systemic treatment of tumor-bearing mice with high doses of angiostatin K1–4 (40–50 mgÆkg)1, twice daily) has been shown not only to inhibit the growth of primary tumors [22,23] but also to induce and sustain their dormancy [22]. Thus, it is possible that pharmacological promotion of endogenous production of angiostatin-like plasmin(ogen) fragments by the nonlysine-analog plasmi- nogen modulators may benefit tumor therapy. Indeed, we have found that the administration of a derivative of SMTP-6 to tumor-bearing mice resulted in significant retardation of tumor growth without any symptoms of toxicity (W. Hu, S. Abe, K. Shibuya, & K. Hasumi, unpublished results). The exploitation of the host-mediated response using such low molecular mass compounds would cause in situ generation of PMD in tumors where the supply of plasminogen activators are abundant and therefore be advantageous with respect to efficacy and selectivity as well as tolerability, which is problematic for dosing with high levels of a proteinaceous agent [41].

degraded by autoproteolysis (Fig. 4). Although it has been reported that the B-chain is autoproteolytically released from the A-chain, even in the absence of reductants, through both a disulfide exchange reaction in K5 and plasmin cleavage at Lys468 and/or Arg471 [16] and Arg530 [15,36], such phenomena are only observed at pH levels at 9.0 [16] to 11.0 [36]. The scrambling of especially sensitive disulfide bonds in K5 is postulated to involve a reaction of OH– with Cys512, and this mechanism requires relatively strong alkaline conditions [36], while our experiments were performed at pH 7.4. Alternatively, recent observations by Lay et al. suggest a novel mechanism that phosphoglycerate kinase binds plasminogen and induces conformational change in K5 that facilitates the OH– attack on the Cys512-Cys536 disulfide bond at neutral pH, resulting in a formation of a sulfenic acid at position 512 and a free thiol

Fig. 6. Characterization of the anti-endothelial activities of PMD-S. The proliferation of HUVEC (A) and four nonendothelial cell lines (B) was assayed in the absence (control) or presence of the indicated concentrations of PMD-S or angiostatin K1–3. The PMD-S prepar- ation used was nonpurified one, and it may contain u-PA. Possible concentrations of u-PA were 0.26–8.6 nM for incubation with PMD-S at 0.3–10 lgÆmL)1, respectively. HT, HT1080 cell; LLC, Lewis lung carcinoma; Fibro, rat fibroblast; CHO, CHO-K1 cell. Results are the mean ± SD from triplicate determinations. *, P < 0.01 relative to control by Student’s t-test employing normally distributed variables.

Nonlysine-analogs promote plasmin autoproteolysis (Eur. J. Biochem. 271) 819

(cid:1) FEBS 2004

Acknowledgements

fragment produced by plasmin autodigestion in the absence of sulfhydryl donors. J. Biol. Chem. 276, 8924–8933.

The authors thank Dr Nobuhiro Takahashi for encouragement and Miss Ryo Akamine for technical assistance. This work was supported in part by a grant from the Takano Life Science Research Foundation. 17. Kwon, M., Yoon, C.S., Fitzpatrick, S., Kassam, G., Graham, K.S., Young, M.K. & Waisman, D.M. (2001) p22 is a novel plasminogen fragment with antiangiogenic activity. Biochemistry 40, 13246–13253.

References

18. Lijnen, H.R., Ugwu, F., Bini, A. & Collen, D. (1998) Generation of an angiostatin-like fragment from plasminogen by stromelysin- 1 (MMP-3). Biochemistry 37, 4699–4702.

1. Carmeliet, P. & Collen, D. (1995) Gene targeting and gene transfer studies of implications in the plasminogen/plasmin system: thrombosis, hemostasis, neointima formation, and atherosclerosis. FASEB J. 9, 934–938. 19. Cornelius, L.A., Nehring, L.C., Harding, E., Bolanowski, M., Welgus, H.G., Kobayashi, D.K., Pierce, R.A. & Shapiro, S.D. (1998) Matrix metalloproteinases generate angiostatin: effects on neovascularization. J. Immunol. 161, 6845–6852.

2. Tryggvason, K., Ho¨ yhtya¨ , M. & Salo, T. (1987) Proteolytic degradation of extracellular matrix in tumor invasion. Biochim. Biophys. Acta 907, 191–217. 3. Pepper, M.S. (2001) Extracellular proteolysis and angiogenesis. Thromb. Haemost. 86, 346–355. 20. Morikawa, W., Yamamoto, K., Ishikawa, S., Takemoto, S., Ono, M., Fukushi, J., Naito, S., Nozaki, C., Iwanaga, S. & Kuwano, M. (2000) Angiostatin generation by cathepsin D secreted by human prostate carcinoma cells. J. Biol. Chem. 275, 38912–38920.

4. Forsgren, M., Raden, B., Israelsson, M., Larsson, K. & Heden, L.O. (1987) Molecular cloning and characterization of a full- length cDNA clone for human plasminogen. FEBS Lett. 213, 254–260. 21. Lay, A.J., Jiang, X.M., Kisker, O., Flynn, E., Underwood, A., Condron, R. & Hogg, P.J. (2000) Phosphoglycerate kinase acts in tumour angiogenesis as a disulphide reductase. Nature 408, 869–873. 5. Mangel, W.F., Lin, B.H. & Ramakrishnan, V. (1990) Character- ligand-induced conformational ization of an extremely large, change in plasminogen. Science 248, 69–73. 22. O’Reilly, M.S., Holmgren, L., Chen, C. & Folkman, J. (1996) Angiostatin induces and sustains dormancy of human primary tumors in mice. Nat. Med. 2, 689–692.

6. An, S.S., Carreno, C., Marti, D.N., Schaller, J., Albericio, F. & Llinas, M. (1998) Lysine-50 is a likely site for anchoring the plasminogen N-terminal peptide to lysine-binding kringles. Protein Sci. 7, 1960–1969. 23. Sim, B.K., O’Reilly, M.S., Liang, H., Fortier, A.H., He, W., Madsen, J.W., Lapcevich, R. & Nacy, C.A. (1997) A recombinant human angiostatin protein inhibits experimental primary and metastatic cancer. Cancer Res. 57, 1329–1334. (1998) Evidence that

24. Takayasu, R., Hasumi, K., Shinohara, C. & Endo, A. (1997) Enhancement of fibrin binding and activation of plasminogen by staplabin through induction of a conformational change in plas- minogen. FEBS Lett. 418, 58–62. 7. Cockell, C.S., Marshall, J.M., Dawson, K.M., Cederholm- Williams, S.A. & Ponting, C.P. the conformation of unliganded human plasminogen is maintained via an intramolecular interaction between the lysine-binding site of kringle 5 and the N-terminal peptide. Biochem. J. 333, 99–105. 8. Hajjar, K.A. (1995) Cellular receptors in the regulation of plasmin generation. Thromb. Haemost. 74, 294–301.

25. Hasumi, K., Ohyama, S., Kohyama, T., Ohsaki, Y., Takayasu, R. & Endo, A. (1998) Isolation of SMTP-3-4-5 and – 6, novel analogs of staplabin, and their effects on plasminogen activation and fibrinolysis. J. Antibiot. (Tokyo) 51, 1059–1068. 9. Lucas, M.A., Fretto, L.J. & McKee, P.A. (1983) The binding of human plasminogen to fibrin and fibrinogen. J. Biol. Chem. 258, 4249–4256.

26. Ohyama, S., Wada, Y. & Hasumi, K. (2002) Antibiotic A10255 (thioplabin) enhances fibrin binding and activation of plasmino- gen. J. Antibiot. (Tokyo) 55, 83–91.

10. Gong, Y., Kim, S.O., Felez, J., Grella, D.K., Castellino, F.J. & Miles, L.A. (2001) Conversion of Glu-plasminogen to Lys-plas- minogen is necessary for optimal stimulation of plasminogen activation on the endothelial cell surface. J. Biol. Chem. 276, 19078–19083. 27. Tachikawa, K., Hasumi, K. & Endo, A. (1997) Enhancement of plasminogen binding to U937 cells and fibrin by complestatin. Thromb. Haemost. 77, 137–142.

11. Molgaard, L., Ponting, C.P. & Christensen, U. (1997) Glycosy- lation at Asn-289 facilitates the ligand-induced conformational changes of human Glu-plasminogen. FEBS Lett. 405, 363–368. 28. Kikuchi, T. & Hasumi, K. (2002) Enhancement of plasminogen activation by surfactin C: augmentation of fibrinolysis in vitro and in vivo. Biochim. Biophys. Acta 1596, 234–245.

12. O’Reilly, M.S., Holmgren, L., Shing, Y., Chen, C., Rosenthal, R.A., Moses, M., Lane, W.S., Cao, Y., Sage, E.H. & Folkman, J. (1994) Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79, 315–328. 29. Hu, W., Ohyama, S. & Hasumi, K. (2000) Activation of fibrino- lysis by SMTP-7 and – 8, novel staplabin analogs with a pseudosymmetric structure. J. Antibiot. (Tokyo) 53, 241–247. 30. Deutsch, D.G. & Mertz, E.T. (1970) Plasminogen: purification from human plasma by affinity chromatography. Science 170, 1095–1096.

31. Lay, A.J., Jiang, X.M., Daly, E., Sun, L. & Hogg, P.J. (2002) Plasmin reduction by phosphoglycerate kinase is a thiol- independent process. J. Biol. Chem. 277, 9062–9068. 13. Cao, Y., Ji, R.W., Davidson, D., Schaller, J., Marti, D., Sohndel, S., McCance, S.G., O’Reilly, M.S., Llinas, M. & Folkman, J. (1996) Kringle domains of human angiostatin. Characterization of the anti-proliferative activity on endothelial cells. J. Biol. Chem. 271, 29461–29467.

14. Cao, Y., Chen, A., An, S.S., Ji, R.W., Davidson, D. & Llinas, M. (1997) Kringle 5 of plasminogen is a novel inhibitor of endothelial cell growth. J. Biol. Chem. 272, 22924–22928. 32. Stathakis, P., Lay, A.J., Fitzgerald, M., Schlieker, C., Matthias, L.J. & Hogg, P.J. (1999) Angiostatin formation involves disulfide bond reduction and proteolysis in kringle 5 of plasmin. J. Biol. Chem. 274, 8910–8916.

15. Cao, R., Wu, H.L., Veitonma¨ ki, N., Linden, P., Farnebo, J., Shi, G.Y. & Cao, Y. (1999) Suppression of angiogenesis and tumor growth by the inhibitor K1–5 generated by plasmin-mediated proteolysis. Proc. Natl Acad. Sci. USA 96, 5728–5733. 33. Claesson-Welsh, L., Welsh, M., Ito, N., Anand-Apte, B., Soker, S., Zetter, B., O’Reilly, M. & Folkman, J. (1998) Angiostatin induces endothelial cell apoptosis and activation of focal adhesion kinase independently of the integrin-binding motif RGD. Proc. Natl Acad. Sci. USA 95, 5579–5583.

16. Kassam, G., Kwon, M., Yoon, C.S., Graham, K.S., Young, M.K., Gluck, S. & Waisman, D.M. (2001) Purification and characterization of A61. An angiostatin-like plasminogen 34. Lucas, R., Holmgren, L., Garcia, I., Jimenez, B., Mandriota, S.J., Borlat, F., Sim, B.K., Wu, Z., Grau, G.E., Shing, Y., Soff, G.A.,

820 S. Ohyama et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Bouck, N. & Pepper, M.S. (1998) Multiple forms of angiostatin induce apoptosis in endothelial cells. Blood 92, 4730–4741. 35. Ueshima, S., Okada, K. & Matsuo, O. (1996) Stabilization of 39. Choi, K.S., Fitzpatrick, S.L., Filipenko, N.R., Fogg, D.K., Kas- sam, G., Magliocco, A.M. & Waisman, D.M. (2001) Regulation of plasmin-dependent fibrin clot lysis by annexin II hetero- tetramer. J. Biol. Chem. 276, 25212–25221. plasmin by lysine derivatives. Clin. Chim. Acta 245, 7–18.

36. Wu, H.L., Shi, G.Y., Wohl, R.C. & Bender, M.L. (1987) Structure and formation of microplasmin. Proc. Natl Acad. Sci. USA 84, 8793–8795. 37. Shi, G.Y. & Wu, H.L. (1988) Differential autolysis of human 40. Kang, H.M., Choi, K.S., Kassam, G., Fitzpatrick, S.L., Kwon, M. & Waisman, D.M. (1999) Role of annexin II tetramer in plasminogen activation. Trends Cardiovasc. Med. 9, 92–102. 41. Cao, Y. (2001) Endogenous angiogenesis inhibitors and their therapeutic implications. Int. J. Biochem. Cell Biol. 33, 357–369. plasmin at various pH levels. Thromb. Res. 51, 355–364.

38. Fitzpatrick, S.L., Kassam, G., Choi, K.S., Kang, H.M., Fogg, D.K. & Waisman, D.M. (2000) Regulation of plasmin activity by annexin II tetramer. Biochemistry 39, 1021–1028.