Sulfation of a tyrosine residue in the plasmin-binding domain of alpha 2-antiplasmin.

Sulfation of human alpha 2-antiplasmin, the major plasma inhibitor of fibrinolysis, was examined using both protein isolated from human plasma and protein synthesized and biosynthetically labeled with [35S]sulfate by a human hepatoma-derived cell line. Linkage of sulfate to tyrosine was demonstrated by recovery of labeled tyrosine sulfate after base hydrolysis of sulfate-labeled alpha 2-antiplasmin. Analysis by reverse-phase high performance liquid chromatography of peptides released from alpha 2-antiplasmin by cleavage with trypsin or cyanogen bromide indicated that sulfate is linked to a single segment of the protein. A cyanogen bromide peptide corresponding to the sulfate-labeled peptide was prepared from alpha 2-antiplasmin isolated from human plasma. Consistent with the presence of tyrosine sulfate in this peptide, its chromatographic elution was altered by treatment with acid under conditions which release sulfate from a tyrosine residue. No peptide in the total digest of alpha 2-antiplasmin by cyanogen bromide eluted at the position of the peptide following desulfation, suggesting that all of the protein is in a sulfated form. The sequence of the sulfate-containing cyanogen bromide peptide as determined by sequential Edman degradation, amino acid composition, and fast atom-bombardment-mass spectrometry was: Glu-Glu-Asp-Tyr(SO4)-Pro-Gln-Phe-Gly-Ser-Pro-Lys-COOH. This peptide is a segment of the previously identified plasmin-binding domain of alpha 2-antiplasmin.


Tyr(S0,)-Pro-Gln-Phe-Gly-Ser-Pro-Lys-COOH. This peptide is a segment of the previously identified plasmin-binding domain of a2-antiplasmin.
Occurrence of tyrosine sulfate as an amino acid residue in proteins was considered to be vary rare, limited to fibrinogen, until 1982, when Huttner reported evidence that every tissue in rats contains a number of proteins with tyrosine sulfate (1). Subsequent studies by Huttner (2) demonstrated, further, that sulfation of tyrosine residues in proteins occurs in a diverse range of organisms. These findings stimulated efforts to identify proteins that contain tyrosine sulfate. Ensuing studies have identified a substantial number of proteins that contain tyrosine sulfate. Examples of human proteins noted to contain this amino acid are: the fourth component of complement (3, 4), fibrinogen (5), fibronectin (6), a-fetoprotein ( 5 ) , and heparin cofactor I1 (7). Although a considerable * This work was supported by grants from the Monsanto Chemical Co. and from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
7 Supported by an Individual National Research Service Award from the National Institutes of Health. To whom correspondence should be addressed. number of proteins containing tyrosine sulfate have been identified, in most cases, the sites and stoichiometry of sulfation have not been determined. Structural characterization of sulfation sites has been reported for only three proteins, fibrinogen (in a number of animal species) (8), the fourth component of human complement (3), and bovine coagulation factor X (9). Studies by our laboratory (3,7,10) have been directed at expanding the limited structural data on sites of sulfation by identifying human plasma proteins that contain sulfate and by performing detailed structural analysis of their sulfation sites. Major objectives are to clarify the structural specificity of the sulfation of proteins and to determine the effect of sulfation on protein function. A continuing effort has been made to identify a 75,000-Da protein which was previously noted to be one of the major sulfate-containing proteins secreted by a human hepatoma-derived cell line (10). Possible identity of the product was suggested by a recent report by Lijnen and co-workers (ll), who detected sulfate linked to a,-antiplasmin. The present paper describes a detailed analysis of the sulfation of this protein, which functions as the major physiological inhibitor of plasmin circulating in blood (12)(13)(14). Structural analysis of a,-antiplasmin indicated that sulfate is linked to a single site, to the tyrosine residue nearest the COOH terminus of the protein. The site of sulfation of a*-antiplasmin is within a 26-residue segment of the protein that is of particular functional significance because it comprises a plasmin-binding domain (15).

DISCUSSION
The present study performs a detailed characterization of the sulfation of human a,-antiplasmin, identifying stoichiometric sulfation of a single tyrosine residue. Obtaining structural data on proteins containing tyrosine sulfate, as presented here, is an important step in understanding the site specificity and the biological function of this widespread modification of proteins. Previously, the positions of sulfatecontaining tyrosine residues have been determined in only a few proteins (3,(7)(8)(9). The present study clearly demonstrates the high degree of site specificity of the sulfation of tyrosine residues. Only one of several tyrosine residues in a2-antiplas-' Portions of this paper (including "Materials and Methods," "Results," Figs. 2-6, and Table I)  min is sulfated. The issue of site specificity has been analyzed in detail for only two other proteins, the fourth component of complement (3) and heparin cofactor I1 (7), with similar results. In C4, sulfate is transferred only to 3 closely grouped tyrosine residues out of a total of more than 50 tyrosine residues in the protein (3), and, in heparin cofactor 11, sulfate is specifically added to 2 out of a total of 12 tyrosine residues (7). High specificity of the process of sulfation is also inferred from the experimental observation that no sulfate is transferred to some proteins such as albumin and the third component of complement (3, lo), which are synthesized by the same cells that stoichiometrically sulfate a,-antiplasmin and the fourth component of complement (3). The structural determinants that define sites of sulfation have not been identified. No consensus sequence is evident among the amino acid sequences of the few known sites of sulfation. The only obvious distinctive characteristic of amino acid sequences surrounding sulfation sites is the high abundance of acidic amino acid residues. az-Antiplasmin conforms to this pattern by having 3 consecutive acidic residues preceding the sulfation site. Further evidence for the importance of acidic residues in determining the specificity of sulfation results from experiments using synthetic peptides as substrates for sulfation. Acidic residues substantially increase the affinity of peptides for the sulfotransferase (16).
Identification of the site of sulfation of ctz-antiplasmin is of particular significance because the single tyrosine sulfate residue is located within a 26-residue segment of the protein which has high affinity (Kd = 5 X for plasmin (15). The location of the sulfation site relative to other sites in azantiplasmin is presented diagrammatically in Fig. 1. The protein consists of a single peptide chain with segments that are homologous to &,-antitrypsin and to other members of the serine proteinase (serpin) superfamily (17,18). Plasmin attacks a specific arginyl-methionine bond (18) and, in the process, forms a covalent bond with the carboxyl group of the arginine residue. The COOH-terminal segment of a,-antiplasmin, which contains the tyrosine sulfate residue, is not homologous to other serpins. This segment comprises a lysinebinding domain that mediates the rapid, reversible association of a,-antiplasmin with lysine-binding sites of plasmin (15). Fast association of these sites is an essential step for the efficient inactivation of plasmin by a,-antiplasmin (19,20). Complete determination of the structure of the plasminbinding domain of a,-antiplasmin, including any post-translational modifications, will contribute to identifying the structural basis for this physiologically important reaction. Lysinebinding sites of plasmin require ligands that contain a positive charge and a negative charge separated by an appropriate distance. This was deduced from the binding affinity of homologs of 6-aminohexanoic acid (21). The sulfate group could contribute a negative charge for interaction with plasmin. We intend to examine the contribution of sulfate to the interaction of a*-antiplasmin with plasmin by comparing the affinity for plasmin of synthetic peptides (corresponding to the plasmin-binding domain of a,-antiplasmin) with and without sulfate. Sulfate-labeled a*-antiplasmin synthesized by cultured hepatocytes as in the present study may serve as a useful reagent for investigating the function of this proteinase inhibitor. The unique site of sulfation is located within the 8000-Da peptide which is excised from az-antiplasmin when this protein forms a covalent complex with plasmin. Also, as a,-antiplasmin circulates in blood, it is converted to an inactive form lacking the plasmin-binding domain (22). The inactive form is about 8000 Da smaller than the active form; the precise COOH terminus of the inactive form has not been identified. Normally, about 70% of a,-antiplasmin circulates as the active form and 30% as the inactive form (22). Sulfate-labeled protein would serve as an ideal substrate for monitoring the conversion from the active to inactive form and for examining the peptide released by this process. The alotchiomerry of the rvlfslion o f the cyanogen bromide pepltde was investigated using revensphsw HPLC. Sulfated and nonsulfated forma of the peplndc wcm cl&y scpanted. Removd of IhC sulfate by heating the peptide briefly in 10% trifluomacstic acid increesed its reLenlion lime by 5 mi" (not shown) Significantly, we did not observe any p k in Ihe analysis of the 1omI cyanogen bmmide digesl of .zlmiplasmin which mrrespOnded IO the paition of Ihe nansulfated peptide. Only a single sharp peak.
comsponding to the position of the sulfaled peptide. eluted in this xgmml of the chromatogram (See Fig.   5 . Miniprint ) Similar m m l t~ were oblained using 02-anliplumin purified from B ringle individual Bnd from plume p m l e d from 20 donom. Thus. vinuaily all mOleFuies of mz-anlipi~min are lulfaled. and there IS probably lillle variation among normal mdividuals ~n the rulfmon of mpnupl.rmin.