INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Tissue plasminogen activator (t-PA),¹ a M_r 68,000 glycoprotein synthesized and secreted by endothelial cells, activates plasminogen (PLG) to form the principal fibrinolytic protease, plasmin (1). Both PLG and t-PA bind to the surface of endothelial cells, a process that is mediated by the calcium-dependent, phospholipid-binding protein annexin II (Ann-II) (2). Assembly of a functional PLG-t-PA·Ann-II complex, which is associated with a 60-fold increase in catalytic efficiency of plasmin generation (3, 4), appears to require at least three independent binding domains of Ann-II. First, interaction of Ann-II with endothelial cell membrane phospholipid is mediated by the calcium-regulated juxtaposition of Lys¹³⁰-Thr¹³⁴ with Asp¹⁶¹ within endonexin repeat II (5). Second, PLG binding requires the carboxyl-terminal Lys³⁰⁷ of processed Ann-II, which interacts specifically with "kringle"-associated lysine binding sites of PLG (2). Third, an independent domain of Ann-II binds t-PA, increasing its affinity for PLG and subsequent generation of plasmin (3, 4).

Cell surface plasmin generation is thought to contribute to the thromboresistance of endothelial cells (6). The specific interaction of t-PA with endothelial cells is inhibited by the thiol-containing amino acid, homocysteine (HC) (7), which is formed upon demethylation of dietary methionine. HC accumulates in several inborn errors of metabolism and in some nutritional deficiencies, and homocysteinemia has been identified as an independent risk factor for atherothrombotic vascular disease (8-10). Here, we postulated that HC selectively impairs the t-PA binding domain of Ann-II, thereby inhibiting vessel wall thromboresistance and predisposing to blood vessel occlusion (7). We sought to identify the t-PA binding domain of annexin II and to determine the molecular mechanism by which HC impairs its interaction with t-PA.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- Bovine serum albumin, DL-homocysteine, L-cysteine, and -chymotrypsin were purchased from Sigma. Peptides were synthesized and purified by reverse phase HPLC by Quality Control Biochemicals, Inc. (Hopkinton, MA).

Purified Proteins-- Native Ann-II was purified from human placental membranes and characterized as described previously (4). Recombinant Ann-II was prepared from Escherichia coli transformed with the pET21b(+) vector containing the Ann-II cDNA as described previously (5). This construct lacked the codon for the initial methionine of Ann-II but added sequences encoding both an amino-terminal tag (MASMTGGQQMGRDP, designated 14 to 1) and a carboxyl-terminal tag (LEHHHHHH, designated +340 to +347). The integrity of the initial amino-terminal domain (MASMTGGQQMGRDPSTVHEILCK) was verified by automated amino acid sequencing at the Protein/DNA Technology Center at Rockefeller University, and the recombinant protein was characterized as described previously (5). Recombinant human t-PA was generously provided by Genentech. Amino-terminal lysine plasminogen was provided by Immuno (Vienna, Austria).

Preparation of the "Core" Fragment of Annexin II-- Purified recombinant Ann-II (1 mg) was subjected to controlled proteolysis with -chymotrypsin-agarose beads (30 units, 250 µg/ml) which had been washed three times in 10 ml of phosphate-buffered saline/5 mM dithiothreitol and resuspended as a 1:2 (v/v) slurry (11, 12). Proteolytic products were analyzed on Coomassie Blue-stained 15% SDS-PAGE and by Western blotting using polyclonal rabbit IgG directed against human placental Ann-II (4).

Cysteine Mutant Constructs-- Using site-directed mutagenesis by overlap extension as described previously (2, 13), the four cysteine residues of Ann-II (Cys⁹, Cys¹³³, Cys²⁶², and Cys³³⁵) were mutated individually to glycine residues. The pCMV5 expression vector containing the wild type cDNA for human Ann-II served as template. For the C9G mutation, initial primer pairs consisted of 5'-CCGCGTTACATAACTTA-3' (sense) with 5'-GAGCTTGCCCAGGATTT-3' (antisense) and 5'-AAATCCTGGGCAAGCTC-3' (sense) with 5'-AAAGTCGACATTTCTGGACGCTCA-3' (antisense). For C133G, initial primer pairs consisted of 5'-AAAAGATCTCCAGCTTCCTTCAAA-3' (sense) with 5'-TCTGGAGCCGATGATCT-3' (antisense) and 5'-AGATCATCGGCTCCAGA-3' (sense) with 5'-AAAGTCGACATTTCTGGACGCTCA-3' (antisense). For the C262G mutant, initial primer pairs consisted of 5'-AAAAGATCTCCAGCTTCCTTCAAA-3' (sense) with 5'-CTGAATGCCCTGAACCA-3' (antisense) and 5'-TGGTTCAGGGCATTCAG-3' (sense) with 5'-AAAGTCGACATTTCTGGACGCTCA-3' (antisense). For the C335G mutant, initial primer pairs consisted of 5'-AAAAGATCTCCAGCTTCCTTCAAA-3' (sense) with 5'-TCCACCACCCAGGTACA-3' (antisense) and 5'-TGTACCTGGGTGGTGGA-3' (sense) with 5'-CCCGTCTCTACCAAAAA-3' (antisense). Secondary reactions were carried out with primers 5'-AAAAGATCTCCAGCTTCCTTCAAA-3' (sense) and 5'-AAAGTCGACATTTCTGGACGCTCA-3' (antisense). All polymerase chain reactions were performed by preincubating the reactants at 94 °C × 90 s and consisted of 30 cycles of 94 °C × 30 s, 51-60 °C × 60 s, and 72 °C × 90 s and were followed by a 5-min incubation at 72 °C. Polymerase chain reaction products were ligated directionally into pCMV5 using BglII and SalI restriction sites. All constructs were sequenced completely to verify their integrity. Plasmids were propagated in HB101 E. coli.

Transient Transfection Assays-- Renal epithelial 293 cells were propagated in RPMI 1640 medium containing 20% fetal bovine serum in 24-well plates to a density of 1.0-1.5 × 10⁵ cells/well. Cells were transferred to serum- and antibiotic-free (basal) medium for 24 h. Each well was rinsed once and then treated with 0.4 µg of DNA/2.4 µl of lipofectamine (Life Technologies, Inc.) in 0.2 ml of basal medium. After 24 h, the transfection medium was replaced with RPMI 1640, 20% fetal bovine serum. t-PA binding capacity was studied after an additional 24 h of incubation. Transfection efficiency was optimized using the -galactosidase expression construct (pSV-gal) and assay system according to the manufacturer's instructions (Promega). Ann-II mRNA levels in transfected cells were estimated by a ribonuclease protection assay.

Mass Spectrometry-- Recombinant human Ann-II (2.8 mg/ml) was dialyzed (18 h, 4 °C, 1:1000 (v/v)) against phosphate-buffered saline, divided into two equal aliquots, and treated with HC (5 mM) or buffer. Following adjustment of the pH to 7.4 with NH₄HCO₃, the tubes were incubated at 37 °C, 3 h and stored at 4 °C until analyzed. The integrity of HC-treated recombinant Ann-II (rAnn-II) was evaluated by SDS-polyacrylamide gel electrophoresis, which showed no alteration in electrophoretic mobility. Using electrospray ionization (ESI), molecular mass spectra were derived from multiply charged ions observed for both HC-treated and untreated Ann-II as described previously (14). The purified protein was diluted 1:10 with 50% methanol, 50% water (v/v), pH 3.0, and 5 µl of analyte solution injected into a Finnigan-MAT TSQ-700 triple quadrupole instrument. In some experiments, both treated and untreated rAnn-II (40 pmol) were digested with trypsin, and the resulting peptides were analyzed by on-line liquid chromatography-ESI-mass spectrometry and by liquid chromatography-ESI-tandem mass spectrometry, using a Finnigan-MAT LCQ ion trap mass spectrometer. Liquid chromatography was performed with an HPLC system (Ultrafast Microprotein Analyzer, Microchrom BioResources, Inc.), using a reverse phase (C18) column (5 µm, 1 × 150 mm). Chromatographic separation was performed using a linear gradient from 5 to 65% of buffer B (5% water containing 0.09% trifluoroacetic acid, 95% acetonitrile) in buffer A (5% acetonitrile in 0.1% trifluoroacetic acid) over 35 min at a flow rate of 50 µl/min.

Preparation of L-[³⁵S]Homocysteine-- L-[³⁵S]Homocysteine was prepared from L-[³⁵S]methionine as described by Ewadh et al. and Mudd et al. (15, 16). Briefly, L-[³⁵S]methionine (7.6 µmol; Amersham Pharmacia Biotech catalog number SJ123) was refluxed with 100 µl of HI and 2 µl of H₃PO₃ under N₂ (100 °C, 8 h) (16). The resulting yellow solution (homocysteine thiolactone) was cooled to 21 °C (30 min) and then dried under N₂. The amber residue was dissolved in 30 µl of H₂O and incubated at 21 °C (3 min) following the addition of 20 µl of 2.5 N NaOH to generate the free sulfhydryl (16, 17). The pH was adjusted to 7.4 with 5× phosphate-buffered saline (pH 5.0), and the molar quantity of free sulfhydryl was estimated using 5,5'-dithiobis-(2-nitrobenzoic acid) (Ellman's reagent; Pierce catalog number 22582) according to the manufacturer's instructions. Free sulfhydryl content was calculated based upon the molar absorptivity of Ellman's reagent using the formula E = A/(b)(c), where E is the molar absorptivity of 5,5'-dithiobis-(2-nitrobenzoic acid) (14,150), A is the observed absorbance at 412 nm, b is the spectrophotometric cell path length, and c is the sulfhydryl concentration in mol/liter. Values for c were determined using a homocysteine standard curve. On a molar basis, mean recovery of free sulfhydryl was 105.4 ± 9.8% (S.E., n = 3) of the starting methionine.

Metabolic Labeling of HUVEC-- Confluent HUVEC (~5 × 10⁶ cells) were pretreated with cycloheximide (100 µM, 2 h, 37 °C) to inhibit de novo protein synthesis and then treated with L-[³⁵S]HC (~6 µmol) for 18 h. Cells were harvested by gentle scraping and lysed by three cycles of freeze-thaw in the presence of protease inhibitors (21 µM leupeptin, 15 µM pepstatin A, and 1 mM phenylmethylsulfonyl fluoride), and the 25,000 × g supernatant was immunoprecipitated with monoclonal anti-Ann-II IgG (2.5 µg/ml, 4 °C, 2 h) followed by Protein G-Sepharose beads (4 °C, 2 h). The beads were washed five times with 100 mM Tris base (pH 8.0) and resuspended in one-fifth volume of 1.25% SDS, 12.5 mM Tris, 1.25 mM EDTA, 6.25% sucrose (w/v), and 12.5 µg/ml bromphenol blue with or without 40 mM dithiothreitol (100 °C, 5 min). Final immunoprecipitates were collected as the 500 × g supernatant and resolved by 12% SDS-polyacrylamide gel electrophoresis. Gel fluorography was performed as described previously (5).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

t-PA Binds to the Amino Terminal Domain of Annexin II-- To determine the ligand binding properties of rAnn-II, microtiter plate wells were coated with either rAnn-II or native Ann-II (nAnn-II) and probed with either ¹²⁵I-t-PA or ¹²⁵I-PLG in a range of concentrations (Fig. 1). Parameters for ¹²⁵I-t-PA binding to nAnn-II and rAnn-II were essentially equivalent (K_d of 47 and 42 nM; B_max of 300 and 272 fmol/well, respectively) (Fig. 1A), suggesting that t-PA binding of Ann-II does not require post-translational modification of the parent protein. Binding capacity of rAnn-II for Lys-plasminogen, on the other hand, was significantly less than that of nAnn-II (K_d of 117 and 150 nM; B_max of 85 and 271 fmol/well) (Fig. 1B) possibly because plasminogen binding is thought to require proteolytic processing for generation of a carboxyl-terminal lysine residue (Lys³⁰⁸) in mammalian cells (18). Interaction of rAnn-II with the endothelial cell surface was equivalent to that of nAnn-II, as described previously (5). These experiments justified the use of rAnn-II to investigate the nature of the t-PA binding domain.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Ligand binding properties of native and recombinant Ann-II. Wells of a 96-well Nunc microtiter plate were coated with either nAnn-II (, ) or rAnn-II (, ; 5 µg/ml; 4 °C, 18 h) and washed three times with Tris-buffered saline containing 0.02% Tween 20 (TBS/Tw). Following incubation with either 125I-t-PA (, ; 0-138 nM; 1 h, 37 °C) (A) or 125I-PLG (, ; 0-525 nM; 1 h, 37 °C) (B), free radioactivity was sampled. The wells were washed rapidly three times, and bound ligand was solubilized in 1% SDS, 0.5 M NaOH, 0.1 M EDTA (1 h, 58 °C) (4, 26).

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Role of the core domain of Ann-II in t-PA binding. A, isolation of core fragment of Ann-II. rAnn-II (1 mg, lane 2) was incubated with 250 µg/ml -chymotrypsin beads, and cleavage products were evaluated by 15% SDS-polyacrylamide gel electrophoresis. The 33-kDa chymotryptic core fragment was recovered in the flow-through of the digest applied to a DE52 ion exchange column (lanes 3-5). Molecular mass markers are shown in lane 1. B, effect of core fragment on binding. Nunc Maxisorp plates were coated with nAnn-II as described in the legend to Fig. 1, washed, and incubated with 125I-t-PA (10 nM), either alone (CON) or in the presence of a 50-fold excess of unlabeled t-PA, intact rAnn-II, or core fragment of rAnn-II (1 h, 37 °C). Bound and free radioactivity were quantified as described in Fig. 1. Shown are mean values for one experiment representative of three. Error bars indicate standard deviation for three determinations for the experiment shown.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effect of Ann-II tail peptides on binding of t-PA to Ann-II and HUVEC. For A and B, microtiter plate wells were coated with rAnn-II (10 µg/ml; 37 °C; 2 h), washed three times, passivated with IB(5) (37 °C; 2 h), and exposed to 125I-t-PA (100 nM; 37 °C; 2 h) in the presence or absence of acetyl-STVHEILCKLSL, acetyl-LCKLSLEGDHST, or acetyl-EGDHSTPPSAYG (A), representing residues 2-13, 8-19, and 14-25 or LCKLSL or LGKLSL (B), representing residues 8-13 of Ann-II. Panels A and B show one experiment representative of three and four, respectively. C, confluent fourth-passage HUVEC, cultured as described previously (3), were washed three times, passivated with IB(5) (37 °C; 2 h), washed, and exposed to 125I-t-PA (20 nM) in the absence or presence of hexapeptide LCKLSL or LGKLSL. Panel C shows one experiment representative of two.

Cysteine 9 Is Essential for t-PA Binding to Annexin II-- To examine further the potential role of Cys⁹ in t-PA binding to Ann-II on cell surfaces, transient transfection of mutant Ann-II constructs were carried out in 293 cells (Fig. 4). Mammalian expression plasmids (pCMV5), containing either the wild type Ann-II cDNA or cDNAs in which Cys⁹, Cys¹³³, Cys²⁶², or Cys³³⁵ were individually mutated to Gly residues, were created. By ribonuclease protection assay, all four mutant mRNAs, as well as the wild type construct, were found to be expressed at 8-12 times the nontransfected control (Fig. 4A). In radioligand assays, specific binding of ¹²⁵I-t-PA to 293 cells transfected with the C133G, C262G, or C335G constructs did not differ significantly from that observed with the wild type expression construct (Fig. 4B). However, binding of ¹²⁵I-t-PA to the C9G-transfected cells, was dramatically reduced almost to the level of nontransfected or mock-transfected cells. These data indicate that Cys⁹, which resides within the putative t-PA binding domain (LCKLSL), contributes to the ligand-receptor binding interaction.

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Transfection of 293 cells with cysteine mutant Ann-II constructs. Renal epithelial 293 cells, propagated as described previously (2), were transfected with plasmids containing either wild type or mutant Ann-II cDNAs as described under "Experimental Procedures." A, ribonuclease protection assay. Expression of Ann-II mRNA was verified by ribonuclease protection assay of lysates from nontransfected cells, wild type transfectants, and C9G, C133G, C262G, and C335G transfectants (lanes 4-9, respectively). Glyceraldehyde-3-phosphate dehydrogenase and Ann-II riboprobes before (lanes 1 and 2) and after (lane 3) RNase treatment are shown. Numbers below each lane indicate the relative abundance of the 300-base pair Ann-II protected fragment after normalization to the 100-base pair glyceraldehyde-3-phosphate dehydrogenase protected fragment. B, binding of 125I-t-PA to transfected cells. Transfectants were equilibrated to 4 °C (5 min), washed three times in IB(5), and exposed to 125I-t-PA (20 nM, 4 °C, 30 min). Bound and free radioactivity were sampled as described in Fig. 1. Shown are mean values for three separate experiments for specific binding to wild type (WT); nontransfected (NT); mock-transfected (MT); and C9G, C133G, C262G, and C335G transfectants. Error bars indicate S.E. (n = 3).

Homocysteine Interacts Directly with the t-PA Binding Domain of Annexin II-- The thiol-containing amino acid HC has been previously shown to impair binding of t-PA to its endothelial cell receptor, Ann-II (7). Because the presence of Cys⁹ appears to be critical for t-PA binding, we elected to determine whether HC might directly modify the t-PA binding domain of Ann-II. Electrospray ionization mass spectrometry was employed to determine the molecular masses for untreated rAnn II (theoretical mass = 40,914 Da) and HC-treated rAnn II (Fig. 5). Spectra for the untreated protein revealed a major peak at 40,919 ± 4 Da (mean ± S.E., n = 4), while the HC-treated protein showed a mass of 41,054 ± 4 Da (Fig. 5). These data indicate an increment of 135 ± 4 mass units, suggesting the formation of an Ann-II-HC adduct with 1:1 stoichiometry (theoretical increment of 133.2 Da).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   ESI-mass spectrometry analysis of untreated and HC-treated Ann-II. rAnn-II (70 µM) was dialyzed against 100 volumes of phosphate-buffered saline (18 h, 4 °C) and either left untreated (thick line) or treated with 5 mM HC (thin line) (3 h, 37 °C, pH 8). Proteins (5 µl) were analyzed by ESI-mass spectrometry.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   ESI-mass spectrometry analysis of peptide 12-22 from untreated and HC-treated rAnn-II. Untreated (A) or HC-treated (B) Ann-II (40 pmol) was subjected to digestion with trypsin at a 5:1 molar ratio of Ann-II to trypsin (4 h, 37 °C) and then analyzed by on-line liquid chromatography-ESI-mass spectrometry.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Liquid chromatography-ESI-tandem mass spectrometry analysis of peptide 12-22 of Ann-II. Peptide 12-22 from either untreated (A) or HC-treated (B) Ann-II was analyzed by tandem mass spectrometry. The fragmentation pattern indicates that HC formed a disulfide bond with cysteine 9.

Homocysteine Specifically Impairs the Binding Capacity of Annexin II for t-PA-- To determine whether HC directly interferes with the ability of Ann-II to bind t-PA, a series of binding experiments was conducted (Fig. 8). ¹²⁵I-rAnn-II, untreated or treated with either HC (Fig. 8A) or L-cysteine (Fig. 8B) (5-50 µM) for 3, 6, or 12 h, was added to t-PA immobilized on microtiter plate wells. After incubation at 4 °C, bound and free radioactivity were quantified in the presence and absence of excess unlabeled ligand (rAnn-II). Exposure of ¹²⁵I-rAnn-II to HC (panel A), but not L-cysteine (panel B), was associated with a time- and dose-dependent decrease in specific binding. In three separate experiments, the maximal inhibitory effect of HC (60-66%) was observed after a 12-h incubation at 50 µM. Interestingly, the I₅₀ for this effect was ~11 ± 3 µM (S.E., n = 3), a value that approximates the upper limit of normal HC concentrations in plasma (8). Maximal inhibition of binding to t-PA after 3 and 6 h of HC preincubation was 20 and 30%, respectively. These data indicate that pathophysiologic concentrations of HC may, over time, significantly impair the t-PA binding capacity of rAnn-II.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Effect of HC on t-PA-binding capacity of Ann-II. Microtiter plate wells were coated with t-PA (100 µg/ml in carbonate buffer, 4 °C, 18 h) and washed three times. 125I-rAnn II (125,860 cpm/pmol) (5) was treated with HC (A) or L-cysteine (B) in a range of concentrations (5-50 µM) for 3 (, ), 6 (, ), or 12 h (, ). Treated or untreated 125I-rAnn II (150 nM) was added to washed t-PA-coated wells. Following incubation (4 °C, 30 min), unbound radioactivity was sampled, the wells were washed three times, and bound radioactivity was solubilized. Specific binding was calculated as that portion inhibited in the presence of a 100-fold molar excess of unlabeled Ann II. Panels A and B show one experiment representative of three.

[³⁵S]HC Metabolically Labels Annexin II in Cultured Endothelial Cells through a Disulfide Linkage-- To ascertain whether endothelial cell Ann-II can be metabolically labeled with [³⁵S]HC and whether such labeling results from disulfide bond formation, lysates from HUVEC, pretreated with cycloheximide and incubated overnight with [³⁵S]HC, were immunoprecipitated with a monoclonal anti-Ann-II IgG (Fig. 9). Washed precipitates were analyzed in the presence and absence of the reducing agent, dithiothreitol. In the absence of dithiothreitol, anti-Ann-II IgG precipitated a prominent band with an apparent molecular mass of ~36 kDa. This band disappeared completely when the sample was reduced in the presence of dithiothreitol. These data indicate that HC labeling of Ann-II occurs in cultured endothelial cells and that adduct formation requires disulfide bond formation.

View larger version (62K): [in this window] [in a new window]   Fig. 9.   Metabolic labeling of endothelial cell Ann-II by L-[35S]HC. HUVEC were pretreated with cycloheximide (100 µM, 2 h) and incubated with L-[35S]HC (345 µM, 18 h). Cells were washed and lysed, and the postnuclear supernatant was immunoprecipitated with monoclonal anti-Ann-II IgG as described previously (5). Whole cell lysates (lanes 2 and 4) as well as immunoprecipitates (lanes 1 and 3) were analyzed under nonreducing (A) or reducing (B) conditions by 12% SDS-polyacrylamide gel electrophoresis fluorography as described (5).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The present data identify a linear sequence of amino acids required for binding of t-PA to Ann-II, its major endothelial cell receptor. The t-PA binding domain of Ann-II was localized to the 23-amino acid amino-terminal tail region of the molecule and more specifically to the hexapeptide sequence LCKLSL, residues 8-13. Although the carboxyl-terminal core portions of the annexin family of proteins share roughly 40-50% homology at the amino acid level (19), the tail domains show limited homology and are thought to confer functional diversity (20). Indeed, our previous studies suggest that, of annexins I, II, IV, and VI, only Ann-II possesses t-PA binding capacity.² The present data, in addition, provide further evidence that the t-PA binding domain of Ann-II is independent of its PLG binding site, the latter having been localized to the extreme carboxyl terminus of the processed protein (2, 18).

The tail domain of Ann-II has previously been implicated in binding of the Ann-II light chain (p11) to form the Ann-II heterotetramer, a complex thought to be involved in intracellular membrane-bridging events (21). Interestingly, p11 is known to bind within the first 9 residues of the Ann-II tail (STVHEILCK), and Cys⁹ has been localized within the Ann-II-p11 interface by fluorescence quenching studies (22). The carboxyl-terminal region of p11 includes the sequence CRDGK (residues 61-65) (23), which closely mimics a portion of the "finger" domain of t-PA (CRDEK, residues 6-17 (RDEKTQMIYQQ)) (24), previously reported to mediate binding of t-PA to HUVEC (25). Thus, it is possible that t-PA and the light chain of Ann-II, p11, bind to closely related regions of Ann-II and interact with Ann-II in a mutually exclusive or co-competitive fashion.

Peptides mimicking this region of Ann-II effectively blocked binding of t-PA to the purified recombinant protein. There is, further, ample precedent for the observation that this effect was observed at relatively high peptide concentrations (I₅₀ = 200-1000 nM) compared with the apparent K_d for the intact protein (K_d = 10-20 nM) (5, 25). Potential explanations include the possibility that a 6-12-residue peptide might not assume a conformation identical to that of the parent protein or that additional vicinal residues might contribute to full binding capacity. In addition, peptides might interact nonspecifically with other cell-associated proteins or tissue culture surfaces. However, in the context of mutational experiments showing the requirement for Cys⁹ for t-PA binding to Ann-II and the direct blockade of this residue by homocysteine, it seems reasonable to conclude that the LCKLSL region of the Ann-II tail plays a role in t-PA binding.

Our data indicate that the thiol-containing amino acid, HC, can inhibit t-PA binding to annexin II by directly complexing with Cys⁹ within its putative hexapeptide binding domain (LCKLSL). This interaction with HC appears to be disulfide-mediated and unique among the four cysteine residues of Ann-II. Although blockade of Cys⁹ of Ann-II impairs its ability to bind t-PA, there is no evidence for a disulfide linkage between t-PA and annexin II within the time frame of this experiment (26). This is perhaps a reflection of the fact that the single unpaired cysteine of t-PA (Cys⁸³) is located not within the "finger"-based Ann-II binding module (residues 6-17) but rather some distance away within the epidermal growth factor domain (27).

Elevated plasma HC has been identified as an independent risk factor for atherosclerotic cardiovascular, cerebrovascular, and peripheral vascular disease and is further associated with deep vein thrombosis and thromboembolism (8, 9, 28, 29). Because it is not a dietary constituent, the sole source of HC in human tissues is methionine. Classical homocystinuria results from deficiency of the pyridoxal-5'-phosphate (vitamin B₆)-dependent enzyme, cystathionine -synthase, which condenses homocyst(e)ine with serine to form cystathionine, a precursor of cysteine (30, 31). However, genetic deficiencies of the enzymes 5-methyltetrahydrofolate-homocysteine methyltransferase and 5,10-methylenetetrahydrofolate reductase, which participate in the remethylation of HC and regeneration of methionine, are also causes of homocyst(e)inemia (32).

There is general agreement in the literature that elevated plasma HC levels may not only lead to endothelial cell dysfunction (7, 33-41) but also impose an independent risk of cardiovascular disease similar to that of smoking or hyperlipidemia (42). Normal adult males and females have HC levels of 6-10 and 8-12 µM, respectively (8), and levels exceeding 14 µM are associated with increased risk in nearly all studies (43). Untreated patients with homocystinuria may have plasma levels as high as 400 µM (10). One recent study found a strong, graded association between plasma HC levels from 9 to 20 µM and overall mortality in patients with angiographically confirmed coronary artery disease (44). Similarly, the risk of extracranial carotid artery stenosis for elderly men and women was 2-fold higher for individuals with plasma HC levels exceeding 14.4 µM compared with those with levels less than 9.1 µM (45). Our data suggest that HC at levels of ~ 11 µM can inhibit ~50% of t-PA binding to Ann-II and that higher levels can block as much as 66%. This loss of fibrinolytic potential could play an etiologic role in life-threatening vascular disorders.