Location of the Major ∊-(γ-Glutamyl)lysyl Cross-linking Site in Transglutaminase-modified Human Plasminogen
1995; Elsevier BV; Volume: 270; Issue: 30 Linguagem: Inglês
10.1074/jbc.270.30.17929
ISSN1083-351X
AutoresEm Bendixen, Peter C. Harpel, Lars Sottrup‐Jensen,
Tópico(s)Protease and Inhibitor Mechanisms
ResumoTissue and plasma transglutaminases cross-link human plasminogen into high molecular weight complexes (Bendixen, E., Borth, W., and Harpel, P. C.(1993) J. Biol. Chem. 268, 21962-21967). A major cross-linking site in plasminogen involved in the tissue transglutaminase-mediated polymerization process has been identified. The ∊-(γ-glutamyl)lysyl bridges of the polymer are formed between Lys-298 and Gln-322. Both the acyl donor Gln residue and the acyl acceptor Lys residue are located in the kringle 3 domain of plasminogen, i.e. cross-linking of plasminogen by tissue transglutaminase involves neither the catalytic domain nor the lysine-dependent binding sites of plasminogen. This study documents that kringle 3 contains a novel functional site with the potential to participate in transglutaminase-mediated cross-linking interactions with plasma, cell-surface, and extracellular proteins. Tissue and plasma transglutaminases cross-link human plasminogen into high molecular weight complexes (Bendixen, E., Borth, W., and Harpel, P. C.(1993) J. Biol. Chem. 268, 21962-21967). A major cross-linking site in plasminogen involved in the tissue transglutaminase-mediated polymerization process has been identified. The ∊-(γ-glutamyl)lysyl bridges of the polymer are formed between Lys-298 and Gln-322. Both the acyl donor Gln residue and the acyl acceptor Lys residue are located in the kringle 3 domain of plasminogen, i.e. cross-linking of plasminogen by tissue transglutaminase involves neither the catalytic domain nor the lysine-dependent binding sites of plasminogen. This study documents that kringle 3 contains a novel functional site with the potential to participate in transglutaminase-mediated cross-linking interactions with plasma, cell-surface, and extracellular proteins. Plasminogen is the circulating 90-kDa precursor of the serine protease plasmin. Although best known as the primary fibrinolytic enzyme in plasma, plasmin also degrades a broad spectrum of other plasma proteins as well as proteins in the extracellular matrix of vascular tissues (for a review, see 29Mayer M. Clin. Biochem. 1990; 23: 197-211Google Scholar). Plasmin also activates procollagenases (38Reich R. Thompson E.W. Iwamoto Y. Martin G.R. Deason J.R. Fuller G.C. Miskin R. Cancer Res. 1988; 48: 3307-3312Google Scholar) and converts latent transforming growth factor-β to its active form (20Kojima S. Harpel P.C. Rifkin D.B. J. Cell Biol. 1991; 113: 1439-1445Google Scholar) and is thereby actively involved in controlling cell migration and tissue remodeling (34Ossowski L. Reich E. Cell. 1983; 35: 611-619Google Scholar). Plasminogen consists of an NH2-terminal 77-residue domain, five internally homologous kringle domains, and a COOH-terminally located serine protease domain (46Sottrup-Jensen L. Claeys H. Zajdel M. Petersen T.E. Magnusson S. Prog. Chem. Fibrinolysis Thrombolysis. 1978; 3: 191-209Google Scholar). Some of the kringle domains are able to bind COOH-terminal Lys residues found in several plasma and cell-surface proteins (15Harpel P.C. Chang T.S. Verderber E. J. Biol. Chem. 1985; 260: 4432-4440Google Scholar). This is likely to be the basis for the specific interaction of plasminogen with fibrin (49Thorsen S. Clemmensen I. Sottrup-Jensen L. Magnusson S. Biochim. Biophys. Acta. 1981; 668: 377-387Google Scholar; 48Suenson E. Lutzen O. Thorsen S. Eur. J. Biochem. 1984; 140: 513-522Google Scholar; 15Harpel P.C. Chang T.S. Verderber E. J. Biol. Chem. 1985; 260: 4432-4440Google Scholar), fibronectin (40Salonen E.M. Saksela T. Vartio T. Vaheri A. Nielsen L.S. Zeuthen J. J. Biol. Chem. 1985; 260: 12302-12307Google Scholar), thrombospondin (44Silverstein R.L. Leung L.L.K. Harpel P.C. Nachman R.L. J. Clin. Invest. 1984; 74: 1625-1633Google Scholar), and the extracellular matrix of vascular tissues (19Knudsen B.S. Silverstein R.L. Leung L.L.K. Harpel P.C. Nachman R.L. J. Biol. Chem. 1986; 261: 10765-10771Google Scholar) as well as a number of vascular cell surfaces (31Miles L.A. Levin E.G. Plescia J. Collen D. Plow E.F. Blood. 1988; 72: 628-635Google Scholar; 14Hajjar K.A. Harpel P.C. Jaffe E.A. Nachman R.L. J. Biol. Chem. 1986; 261: 11656-11662Google Scholar). It has become a basic concept that the interaction of plasminogen with protein matrices and cell surfaces plays a key role in the regulation of plasminogen activation, so plasmin activity is restricted to sites where its proteolytic activity is required (33Nachman R.L. Blood. 1992; 79: 1897-1906Google Scholar). We have recently found that plasminogen is a substrate for transglutaminases from plasma as well as tissues. These enzymes cross-link plasminogen into high molecular weight homopolymers and also cross-link plasminogen to fibronectin, a major constituent of the extracellular matrix of vascular tissues (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar). Transglutaminases are Ca2+-dependent enzymes that catalyze the cross-linking of certain Gln and Lys side chains in proteins, resulting in the formation of ∊-(γ-glutamyl)lysyl bridges. Transglutaminases are ubiquitously found in tissues and body fluids (for a review, see 22Lorand L. Conrad S.M. Mol. Cell. Biochem. 1984; 58: 9-35Google Scholar). The best characterized transglutaminase is plasma transglutaminase Factor XIIIa, formed from its precursor, Factor XIII, by activation with thrombin. It cross-links polymerized fibrin into a tight matrix during the final steps of blood coagulation (23Lorand L. Losowsky M.S. Miloszewski K.J.M. Prog. Hemostasis Thromb. 1980; 52: 245-290Google Scholar) and also cross-links α2-antiplasmin to the fibrin clot (39Sakata Y. Aoki N. J. Clin. Invest. 1980; 65: 290-297Google Scholar). The physiological roles of tissue transglutaminases are not understood. However, since several of their substrates are constituents of the extracellular matrix and basement membrane of vascular tissues, transglutaminases are believed to play a major role in the formation and stabilization of these matrices (27Martinez J. Rich E. Barsigian C. J. Biol. Chem. 1989; 264: 20502-20508Google Scholar; 17Kinsella M.G. Wright T.N. J. Biol. Chem. 1990; 265: 17891-17898Google Scholar; 41Sane D.C. Moser T.L. Greenberg C.S. Biochim. Biophys. Res. Commun. 1991; 174: 465-469Google Scholar; 12Greenberg C.S. Birckbichler P.J. Rice R.H. FASEB J. 1991; 5: 3071-3077Google Scholar; 1Aeschliemann D. Paulsson M. Mann K. J. Biol. Chem. 1992; 267: 11316-11321Google Scholar). It has previously been suggested that plasminogen cross-linked into high molecular weight homopolymers may also be cross-linked to the extracellular matrix as well and that this cross-linking of plasminogen may represent a novel mechanism by which plasmin activity is targeted to the extracellular matrix of vascular tissues (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar). We report here the identification of the major cross-linking site in plasminogen involved in its transglutaminase-catalyzed polymerization. The site has been identified by comparison of HPLC1 1The abbreviations used are: HPLChigh performance liquid chromatographyDTTdithiothreitolPAGEpolyacrylamide gel electrophoresisPTHphenylthiohydantoindansyl5-dimethylaminonaphthalene-1-sulfonyl. profiles of tryptic peptides of monomeric and cross-linked plasminogens, and from sequence analysis, it is found to be constituted by Lys-298 and Gln-322 located in the kringle 3 domain. high performance liquid chromatography dithiothreitol polyacrylamide gel electrophoresis phenylthiohydantoin 5-dimethylaminonaphthalene-1-sulfonyl. Human plasminogen (90 kDa) was purified as described earlier (46Sottrup-Jensen L. Claeys H. Zajdel M. Petersen T.E. Magnusson S. Prog. Chem. Fibrinolysis Thrombolysis. 1978; 3: 191-209Google Scholar). Guinea pig liver transglutaminase (75 kDa) was from Sigma. 1-Chloro-3-(4-tosylamido)-4-phenyl-2-butanone-treated trypsin was from Worthington. Standard chemicals were analytical grade and were obtained from Merck, Fluka, and Sigma. ∊-(γ-Glutamyl)lysine was from Sigma. Packing materials for reverse-phase and ion-exchange HPLC were from Machery Nagel, Phase Separations, and PolyLC, Inc. A Mono S HR 5/5 column was from Pharmacia Biotech Inc. Reagents and solvents for sequence analysis were from Applied Biosystems, Inc. 550 nmol (50 mg) of human plasminogen was incubated with 2.75 nmol (0.208 mg) of guinea pig liver tissue transglutaminase in the presence of 10 mM Ca2+, 0.5 mM DTT, and 50 mM Tris-buffered saline at pH 7.5. The total volume was 50 ml. The reaction was started by the addition of transglutaminase, proceeded for 1 h at 20°C with end-over-end mixing, and was stopped by the addition of 20 mM EDTA. For SDS-PAGE, the Tris/glycine system of 21Laemmli U.K. Nature. 1970; 227: 680-685Google Scholar was used with 0.5 × 80 × 100-mm 10-20% gradient gels. ~200-μg aliquots of cross-linked plasminogen or untreated plasminogen were digested with a mixture of endo- and exopeptidases using the procedure of 13Griffin M. Wilson J. Lorand L. Anal. Biochem. 1982; 124: 406-413Google Scholar. Because ∊-(γ-glutamyl)lysine elutes close to methionine in the amino acid analysis system used (45Sottrup-Jensen L. Biochem. Mol. Biol. Int. 1993; 30: 789-794Google Scholar), samples were oxidized with performic acid prior to analysis. The concentrations of the plasminogen solutions were determined by amino acid analysis after hydrolysis with 6 M HCl for 18 h at 110°C. Plasminogen cross-linked as described above or native plasminogen was reduced by adding 5 mM DTT in the presence of 6 M guanidinium chloride at pH 9.0. Reduction proceeded at 20°C for 30 min, where after all free sulfhydryl groups were blocked by reaction with 20 mM iodoacetic acid for 30 min at pH 8.0. The reduced and alkylated samples were desalted by repeated dialysis against 50 mM NH4HCO3, and digestion of plasminogen (37°C, 12 h) was performed by adding trypsin (1:100, w/w) in 50 mM NH4HCO3 (total volume of 150 ml). The peptide material was freeze-dried, redissolved in 25 ml of 0.1% trifluoroacetic acid, and stored as aliquots at −20°C. Separations of tryptic peptides derived from either cross-linked or monomeric plasminogen were performed by reverse-phase HPLC on a Hewlett-Packard 1084 B instrument. Columns (4 × 250 mm) were packed with 5-μm Nucleosil C-18 or 10-μm Vydac C-18 material and eluted with gradients of acetonitrile in 0.1% trifluoroacetic acid at 50°C at a flow rate of 1 ml/min. The elution profiles were monitored at 220 nm. For cation-exchange chromatography, an instrument assembled from Pharmacia Biotech HPLC modules was used. Separations were performed on a 4 × 120-mm polysulfoethyl-A column and on a Mono S column using a gradient of NaCl in 25% (v/v) acetonitrile, 5 mM phosphoric acid (2Alpert A.J. J. Chromatogr. 1988; 443: 819-825Google Scholar; 8Crimmins D.L. Thoma R.S. McCourt D.W. Schwartz B.D. Anal. Biochem. 1989; 176: 255-260Google Scholar). Peptides were collected manually and dried in a Speed-Vac centrifuge. Edman degradation of selected peptides was performed using an Applied Biosystems 477A sequencer with an on-line 120A HPLC apparatus. The Normal-1 reaction and conversion cycles were used, and initial yields of PTH-derivatives were 100-1000 pmol. A sample of plasminogen incubated without transglutaminase under the conditions used for cross-linking was desalted on Sephadex G-25 and immediately treated with 20 mM iodoacetamide at pH 8.0 for 1 h to block free sulfhydryl groups. The content of S-carboxymethylcysteine was determined after acid hydrolysis (45Sottrup-Jensen L. Biochem. Mol. Biol. Int. 1993; 30: 789-794Google Scholar). Plasminogen incubated as described above was freed from reductant and treated with 0.5 mM 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole to introduce a fluorescent label on the free sulfhydryl groups (51Toyo'oka T. Imai K. Anal. Chem. 1984; 56: 2461-2464Google Scholar). The remaining disulfide bridges were then reduced with 10 mM DTT and alkylated with iodoacetamide in 6 M guanidine HCl, 0.1 M Tris-HCl at pH 8.5. After tryptic digestion, the peptides were separated by reverse-phase HPLC as described above. The column effluents were monitored by recording the absorbance at 220 nm and by recording the fluorescence (excitation at 374 nm and emission at 500 nm (51Toyo'oka T. Imai K. Anal. Chem. 1984; 56: 2461-2464Google Scholar)) using a Hitachi F-1000 fluorescence spectrophotometer equipped with a 12-μl flow cell. Major fluorescent pools identified from the fluorescence profile were subjected to sequence analysis. The fluorescent Cys residues were identified by HPLC as described by 6Chin C.C.Q. Wold F. Anal. Biochem. 1993; 214: 128-134Google Scholar. RasMol version 2.4 (Roger Sayle, Glaxo Research and Development) was run under Microsoft Windows and used to display the three-dimensional structure of human kringle 4 (32Mulichak A.M. Tulinsky A. Ravichandran K.G. Biochemistry. 1991; 30: 10576-10588Google Scholar). The program was obtained by anonymous FTP (ftp.dcs.ed.ac.uk), and the coordinate file was from Brookhaven Protein Data Bank. From previous studies, we know that plasminogen is readily cross-linked into high molecular weight complexes by several transglutaminases (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar). The ∊-(γ-glutamyl)lysyl cross-links are stable in the currently used methods for reduction, digestion, and peptide separation. Therefore, to locate the sites of cross-linking, tryptic HPLC profiles of plasminogen cross-linked with guinea pig liver tissue transglutaminase were compared with those of untreated plasminogen. The novel peptides found in the digest of cross-linked plasminogen must contain the sites of cross-linking. Consequently, the yields of peptides that harbor the reactive Gln and Lys residues must be reduced when compared with those of untreated plasminogen. When sequencing a peptide containing an ∊-(γ-glutamyl)lysyl bridge, two NH2-terminal sequences of equimolar ratio are expected. Because the isopeptide bond is not cleaved during Edman degradation, the cross-linked pair of residues is released as bis-PTH-∊-(γ-glutamyl)lysine, which is easily distinguished from the normal PTH-derivatives in the standard HPLC separation (47Sottrup-Jensen L. Hansen H.F. Pedersen H.S. Kristensen L. J. Biol. Chem. 1990; 265: 17727-17737Google Scholar). To minimize the differences in the peptide profiles of cross-linked and native plasminogens, a batch of 100 mg of plasminogen was divided in two equal portions. One batch was incubated with guinea pig liver transglutaminase in the presence of Ca2+, while the other was incubated with Ca2+ only. The molar ratio of transglutaminase to plasminogen was 1:200, so peptides of transglutaminase origin would not be expected to interfere with the interpretation of the peptide profiles of tryptic plasminogen digests. Fig. 1 shows a reducing SDS-PAGE analysis of plasminogen incubated with or without transglutaminase. Plasminogen incubated with transglutaminase and Ca2+ was essentially completely cross-linked after 1 h of incubation and was retained at the top of the stacking gel (lane 3), while untreated plasminogen was still in the 90-kDa form (lane 2), identical to the starting material (lane 1). DTT was used in the cross-linking reaction to preserve the activity of guinea pig liver transglutaminase. Plasminogen incubated with 0.5 mM DTT as described above but in the absence of transglutaminase was found to contain 4.9 mol of S-carboxymethylcysteine/mol of protein (~10% of the content of halfcystine). To investigate whether this level of S-carboxymethylcysteine represented nearly full reduction of a limited set of disulfide bridges or whether a large number of bridges had randomly been reduced to a low extent, a sample of plasminogen was incubated with 0.5 mM DTT. Free sulfhydryl groups were tagged with a fluorescent label by reaction with 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (51Toyo'oka T. Imai K. Anal. Chem. 1984; 56: 2461-2464Google Scholar), and after complete reduction and carboxamidomethylation, tryptic peptides were separated by the procedures used for the main digests (see below). It was found that the extent of reduction described above resulted from a low level of reduction of randomly distributed disulfide bridges in plasminogen (data not shown). To estimate the amount of cross-links, aliquots of both cross-linked and monomeric plasminogens were exhaustively digested with a mixture of endo- and exopeptidases (13Griffin M. Wilson J. Lorand L. Anal. Biochem. 1982; 124: 406-413Google Scholar). It was found by amino acid analysis that 0.6-0.8 mol of ∊-(γ-glutamyl)lysine was present per mol of cross-linked plasminogen, while untreated plasminogen, as expected, did not contain the isopeptide. The estimation of the amount of cross-link in plasminogen is not accurate because the small peak corresponding to ∊-(γ-glutamyl)lysine is located in the valley between the huge peaks corresponding to valine and isoleucine (45Sottrup-Jensen L. Biochem. Mol. Biol. Int. 1993; 30: 789-794Google Scholar), but the data suggest that cross-linking results in the introduction of only 1 mol of isopeptide bond/mol of plasminogen. Both monomeric and cross-linked plasminogens were digested with trypsin. To ensure complete digestion, the two batches were first reduced with DTT and alkylated with iodoacetic acid in the presence of guanidinium chloride. To investigate whether the cross-links in transglutaminase-treated plasminogen could be identified by comparison of the tryptic peptide profiles, identical aliquots of the digests were separated on Nucleosil C-18 and Vydac C-18 columns and on Mono S and polysulfoethyl-A columns using a variety of conditions for gradient elution (data not shown). With the exception of a few peaks, the elution profiles of the peptides obtained from plasminogen and cross-linked plasminogen were similar. Representative peptide profiles obtained on Nucleosil C-18 are shown in Fig. 2(A (plasminogen) and B (cross-linked plasminogen)). The heights of peaks P-1, P-2, and P-3 in Fig. 2A were markedly reduced in the peptide profile obtained from cross-linked plasminogen. In Fig. 2B, the corresponding peaks are X-1, X-2, and X-3. Furthermore, a single peak (X-33) seen in the profile obtained from cross-linked plasminogen had no counterpart in the profile obtained from monomeric plasminogen. Therefore, this peak was expected to represent two peptides cross-linked by an ∊-(γ-glutamyl)lysyl isopeptide bond. Peak X-33 was further purified by cation-exchange chromatography (Fig. 3).Figure 3:Purification of the material in peak X-33. Peak X-33, harboring an ∊-(γ-glutamyl)lysyl bond, was separated by cation-exchange chromatography. A polysulfoethyl-A column was eluted with a linear gradient of 5 mM H3PO4, 25% (v/v) acetonitrile (solvent A) and of 5 mM H3PO4, 25% (v/v) acetonitrile, 1 M NaCl (solvent B) at a flow rate of 1 ml/min.View Large Image Figure ViewerDownload (PPT) Sequence analysis of the material in peaks P-1, P-2, and P-3 showed that each peak contained a main sequence, summarized in Table 1. A background of several minor peptides that could not be clearly identified was present in these peaks as well. As determined from the initial yields of PTH-derivatives, the levels of the minor peptides in peaks P-1, P-2, and P-3 were ~20% of that of each major sequence. As shown in Table 1, peak X-33 was found to consist of the stretches Thr-291-Arg-306 and Arg-312-Arg-324, cross-linked by an isopeptide bond involving Lys-298 and Gln-322. As expected, upon sequence analysis, the yields of the cross-linked mates were similar, and the cross-linking prevented tryptic cleavage at Lys-298. In the peptide profile from plasminogen peptides, residues 291-306 were recovered as the major peptides in peaks P-2 and P-1, while residues 312-324 were recovered as the major peptide in peak P-3.TABLE I Open table in a new tab As shown in Fig. 1, virtually all plasminogen had been cross-linked because only trace amounts of monomeric plasminogen were present. If peak X-33 represents the major site of cross-linking, peaks X-1, X-2, and X-3 would be expected to disappear nearly completely. This was evidently not the case (Fig. 2B), but upon sequence analysis of peaks X-1, X-2, and X-3, a set of NH2 termini similar to those representing the minor components in peaks P-1, P-2, and P-3 were revealed. In peaks X-1, X-2, and X-3, the major sequences of peaks P-1, P-2, and P-3 were recognized, respectively, but their levels were only ~20% of the average level of peptides present. Hence, we estimate that the pair Lys-298 and Gln-322 has been utilized in the cross-linking process to an extent of ~95%. Apart from the differences in the peptide profiles described above, it can be seen that the heights of several peaks with identical retention times in the separations shown in Fig. 2(A and B) are not the same. This could indicate different yields of certain noncross-linked peptides in the two digests or the presence of additional cross-linked peptides. To investigate this, we analyzed the peptide profiles of both monomeric and cross-linked plasminogens by dividing the eluates into 13 pools, as indicated in Fig. 2(A and B). Each of these pools was subjected to reverse-phase HPLC on a Vydac C-18 column using shallow gradients of acetonitrile. Differences between the profiles obtained from plasminogen and cross-linked plasminogen were observed only in the material from pools 2, 3, and 5 involving peaks P-1, P-2, P-3, X-1, X-2, X-3, and X-33 discussed above. Furthermore, sequence analysis of the relevant pools from Fig. 2B did not result in the appearance of bis-PTH-∊-(γ-glutamyl)lysine. Hence, the cross-link between Lys-298 and Gln-322 identified above is likely to be the only major cross-link in transglutaminase-polymerized plasminogen. Transglutaminase-catalyzed reactions result in the formation of isopeptide bonds between certain Gln residues (acyl donor sites) and Lys residues (acyl acceptor sites) in proteins. Potential acyl donor sites in proteins have generally been identified by transglutaminase-catalyzed incorporation of radioactive or fluorescent amines followed by characterization of the relevant peptide (see 30McDonagh R.P. McDonagh J. Petersen T.E. Thogersen H.C. Skorstensgaard K. Sottrup-Jensen L. Magnusson S. FEBS Lett. 1981; 127: 174-178Google Scholar, 1Aeschliemann D. Paulsson M. Mann K. J. Biol. Chem. 1992; 267: 11316-11321Google Scholar, and 11Gorman J.J. Folk J.E. J. Biol. Chem. 1984; 259: 9007-9010Google Scholar). It has recently been possible to characterize potential acyl acceptor sites by incorporating specifically designed Gln-containing fluorescent peptides into proteins (24Lorand L. Parameswaran K.N. Velasco P.T. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 82-83Google Scholar). However, the identification of ∊-(γ-glutamyl)lysyl cross-links in proteins is not straightforward, as documented by the variety of procedures employed (5Chen R. Doolittle R.F. Proc. Natl. Acad. Sci. U. S. A. 1970; 66: 472-479Google Scholar; 7Cottrell B.A. Strong D.D. Watt K.W.K. Doolittle R.F. Biochemistry. 1979; 18: 5405-5410Google Scholar; 18Kimura S. Aoki N. J. Biol. Chem. 1986; 261: 15591-15595Google Scholar; 37Purves L. Purves M. Brandt W. Biochemistry. 1987; 26: 4640-4646Google Scholar; 36Pucci P. Malorni A. Marino G. Metafora S. Esposito C. Porta R. Biochem. Biophys. Res. Commun. 1988; 154: 735-740Google Scholar; 47Sottrup-Jensen L. Hansen H.F. Pedersen H.S. Kristensen L. J. Biol. Chem. 1990; 265: 17727-17737Google Scholar; 16Hohl D. Mehrel T. Lichti U. Turner M.L. Roop D.R. Steinert P.M. J. Biol. Chem. 1991; 266: 6626-6636Google Scholar). Plasminogen is a substrate for transglutaminases (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar), and incubation of native plasminogen with guinea pig liver tissue transglutaminase results in the formation of a polymer of high molecular weight (Fig. 1). The content of ∊-(γ-glutamyl)lysine was estimated at 0.6-0.8 mol/mol of plasminogen, which, in view of the near absence of monomeric plasminogen after treatment with transglutaminase, is likely to reflect the introduction of only 1 mol of cross-link/mol of plasminogen. We investigated whether the cross-linking site(s) could be identified by comparison of the peptide profiles of tryptic digests of cross-linked and monomeric plasminogens. Trypsin appeared to degrade cross-linked and monomeric plasminogens equally well, suggesting that cross-linked plasminogen retains a rather open structure, and the majority of the peaks of the two profiles were easily paired (Fig. 2, A and B). Peptides were selected that were unique for cross-linked plasminogen and that appeared in reduced quantities in the profile of cross-linked plasminogen. As such, we found only one tryptic peptide that was unique for cross-linked plasminogen, namely peak X-33 (Fig. 2B). From sequence analysis of the pure cross-linked peptide (Fig. 3 and Table 1), in which the cross-link was identified as bis-PTH-∊-(γ-glutamyl)lysine, we conclude that Lys-298 and Gln-322 are cross-linked by an ∊-(γ-glutamyl)lysyl isopeptide bond. From the digest of monomeric plasminogen, we have characterized the corresponding noncross-linked peptides and found that these peptides are present in the digest of cross-linked plasminogen in very low amounts. Because the peptide separations have been performed under a variety of conditions and because only one cross-linked peptide was identified in cross-linked plasminogen, we conclude that Lys-298 and Gln-322, both located in the kringle 3 domain, constitute the acyl acceptor and acyl donor sites in the cross-linking process, respectively, and that these residues have been utilized to an extent of ~95%. It was found earlier (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar) that fluorescent substrates reacting with donor and acceptor sites become incorporated into plasminogen. The initial incorporation took place in monomeric plasminogen, but upon prolonged incubation, the substrates were also found in polymerized plasminogen. We believe that these substrates react with the donor and acceptor sites identified here and may become released from plasminogen in the polymerization process. However, there might be other reactive sites because incubation with dansylcadaverine did not block the polymerization process and the polymeric product did contain some fluorescent label (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar). The conformation of Glu-plasminogen has been described as a compact ellipsoid, coiled up in a spiral in which the NH2-terminal portion is in close proximity to the serine protease portion (3Banyai L. Patthy L. J. Biol. Chem. 1984; 259: 6466-6471Google Scholar; 35Ponting C.P. Holland S.K. Cederholm-Williams S.A. Marshall J.M. Brown A.J. Spraggon G. Blake C.C.F. Biochim. Biophys. Acta. 1992; 1159: 155-161Google Scholar; 26Mangel W.F. Lin B. Ramakrishnan V. Science. 1990; 248: 69-73Google Scholar). According to this model, the kringle 3 domain protrudes from the compact plasminogen molecule. Our results substantiate this proposal by showing that part of the surface of kringle 3 is exposed for transglutaminase-mediated polymerization. Furthermore, the lack of oligomers (Fig. 1) (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar) in the cross-linking of plasminogen points to highly specific protein-protein interactions in the system. No high resolution structure is available for kringle 3, but to illustrate the putative location of Lys-298 and Gln-322, Fig. 4 shows a model of the homologous kringle 4 of plasminogen (32Mulichak A.M. Tulinsky A. Ravichandran K.G. Biochemistry. 1991; 30: 10576-10588Google Scholar) in which Lys-298 and Gln-322 have been modeled. Clearly, both the donor and acceptor positions are located at the surface of the kringle, spaced apart by ~13 Å. Kringles 1, 4, and 5 mediate binding of human plasminogen to fibrin and cellular surfaces (49Thorsen S. Clemmensen I. Sottrup-Jensen L. Magnusson S. Biochim. Biophys. Acta. 1981; 668: 377-387Google Scholar; Lukas et al., 1983; 28Matsuka Y.V. Novokhatny V.V. Kudinov S.A. Eur. J. Biochem. 1990; 190: 93-97Google Scholar; 52Wu H.L. Chang B.I. Wu D.H. Chang L.C. Gong C.C. Lou K.L. Shi G.Y. J. Biol. Chem. 1990; 265: 19658-19664Google Scholar). The present finding that the cross-linking of plasminogen involves residues in kringle 3 is the first indication of a specific function for this domain. Lys-298 is conserved in plasminogen from all species sequenced so far (human (46Sottrup-Jensen L. Claeys H. Zajdel M. Petersen T.E. Magnusson S. Prog. Chem. Fibrinolysis Thrombolysis. 1978; 3: 191-209Google Scholar; 10Forsgren M.R.B. Israelson M. Larsson K. Heden L.-O. FEBS Lett. 1987; 213: 254-260Google Scholar), rhesus monkey (50Tomlinson J.E. McLean J.W. Lawn R.M. J. Biol. Chem. 1989; 264: 5957-5965Google Scholar), mouse (9Degen S.J.F. Bell S.M. Schaefer L.A. Elliot R.W. Genomics. 1990; 8: 49-61Google Scholar), cattle (42Schaller J. Moser P.W. Dannegger-Muller G.A.K. Roesselet S.J. Kampfer U. Rickli E.E. Eur. J. Biochem. 1985; 149: 267-278Google Scholar), and pig (43Schaller J. Marti T. Roesselet S.J. Kampfer U. Rickli E.E. Fibrinolysis. 1987; 1: 91-102Google Scholar)). Gln-322 is conserved in human, rhesus monkey, and mouse plasminogen, being replaced by a Glu residue in plasminogen from cattle and pig. Therefore, plasminogen from rhesus monkey and mouse is a potential substrate for transglutaminases, but if plasminogen from cattle and pig is a substrate for transglutaminases, the acyl donor residue must be different from that of human plasminogen. Polymeric plasminogen cross-linked by transglutaminases maintains the characteristic behavior of 90-kDa plasminogen in the sense that it binds to endothelial cells and is readily activable by physiological plasminogen activators, and when activated, it shows full plasmin activity (4Bendixen E. Borth W. Harpel P.C. J. Biol. Chem. 1993; 268: 21962-21967Google Scholar). We suggest that cross-linked plasminogen is physiologically important on cell surfaces and in the extracellular matrix. Besides being cross-linked to itself, plasminogen might become cross-linked to cell-surface and extracellular matrix proteins because the plasminogen polymers possess one free donor and acceptor site. The basement membrane proteins degraded by plasmin are also substrates for transglutaminases, and it is possible that plasminogen is cross-linked to some of these substrate proteins and thus is able to locally become activated when proteolysis of the extracellular matrix is required. Transglutaminase-mediated cross-linking of plasminogen may be a novel mechanism by which plasmin can interact with its specific target sites. These could include fibrin, fibrinogen, the extracellular matrix, cell-surface proteins, or circulating plasma proteins. We thank Lene Kristensen for excellent technical assistance and Jesper Haaning for help with display of protein structures.
Referência(s)