Factor XIIIa-derived Peptides Inhibit Transglutaminase Activity

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 28, Issue of October 5, pp. 21284-21292,1993 Printed in U.S.A.

Factor XIIIa-derived Peptides Inhibit TransglutaminaseActivity LOCALIZATION OF SUBSTRATERECOGNITIONSITES* (Received for publication, May 10, 1993)

Komandoor E. AchyuthanS, Thomas F. SlaughterQll,Manuel A. SantiagoSII, Jan J. Enghild**, and Charles S . Greenberg$**$$ From the Departments of $Medicine, **Pathologyand §Anesthesiology,Duke University Medical Center, Durham. North Carolina 27710

Factor XIIIa is a transglutaminase that catalyzes intermolecular y-glutamyl-e-lysyl bonds between fibrin and other proteins involved in hemostasis. We of factor synthesized 25 peptides from various regions XIIIa and studied their effects on cross-linking fibrin, N,N’-dimethylcasein, or fibronectin. We found that Asp1gotwo peptides, A ~ n ~ ~ - A(peptide-4) s p ~ ~ and Pheaso (peptide-7), inhibited factor XIIIa cross-linking of these substrates. The other peptides did not inhibit factor XIIIa activity. The inhibition of cross-linking was reversed by excess Substrate, indicating that the peptides were interacting with fibrin and not factor XIIIa. The peptides were not pseudosubstrates since they were not cross-linked to fibrin. Thepeptides did not modify the primaryamine bindingsite as increasing the primary amine concentration did not reverse inhibition. Peptides-4 and -7 also had no effect on exposure of the active siteof factor XIIIa and no synergistic inhibitory effects were detected. Peptides-4 and -7 had no effect on factor XIIIa binding to fibrin suggesting that the binding sites and the substrate recognition sites were distinct. Synthetic peptides containing shorteramino acid sequences of peptide-4 were (Asp”oinactive. Incontrast,theamino-terminal L Y S ’ ~ ~Tyr’g4-Tyr204) , and the carboxyl-terminal (LysZa1-Pheaso) portions of peptide-7 were 20-60-fold less inhibitory compared to intactpeptide-7. Peptides4 and-7 also inhibited guineapig liver tissue transglutaminase from cross-linking fibrinogen, N,N’-dimethylcasein, and fibronectin. In conclusion, we have identified two regions outside the active site pocket which are important for substrate recognition in factor XIIIa and tissue transglutaminase.

Folk and Finlayson, 1977; Greenberg et al., 1991; McDonagh, 1987). The keratinocyte transglutaminase, a membrane anchored protein, functions during keratinocyte differentiation to promote formation of the epidermis (Phillips et al., 1990). The tissue transglutaminase (Greenberg et al., 1991), present in a wide variety of tissues cross-linksfibrin(ogen) and many plasma, cellular, and extracelluar matrix adhesive molecules (Gentile et al., 1991; Martinez et al., 1989; Lee et al., 1992; Achyuthan et al., 1988; Greenberg et al., 1987a; Aeschlimann and Paulsson, 1991). During hemostasis plasma factor XI11 is converted by thrombin to factor XIIIa which has fibrin stabilizing activity (Lorand andKonishi, 1964). Factor XIIIa acts within the fibrin clot to increase the mechanical strength and tolimit the susceptibility of the fibrin network to plasmin degradation (Lorand et al., 1980; McDonagh, 1987). Factor XIIIa catalyzes the formation of covalent isopeptide bonds between fibrin molecules within the clot (Pisano et al., 1968) and also cross-links fibronectin (Mosher, 1975), a*-antiplasmin (Sakata and Aoki, 1980))vitronectin (Sane et al., 1988), von Willebrand factor (Bockenstedt et al., 1986), thrombospondin(Baleand Mosher, 1986), factor V (Francis et al., 1986), and collagen (Mosher et al., 1979). The interaction of factor XIIIa with these proteins at thesite of vascular injury contributes to hemostasis and wound healing. In the first stepof transglutaminase-mediatedcatalysis, the enzyme forms a calcium-dependent thioesterintermediate with a glutamine residue in the protein substrate and in the process ammonia is released from glutamine (Folk and Finlayson, 1977). Then the enzyme-substrate complex interacts with either a primary amine or a peptide-bound lysine residue to form an isopeptide bond. The transglutaminases do not display significant substrate specificity with regard to the primary amine (Folk and Finlayson, 1977). The amino acid sequence of factor XIIIa regulating the recognition of protein Blood coagulation factor XIIIa is amember of the recently glutamine residues is unknown. identified transglutaminase gene family (Ichinose et al., 1986, We postulated that the substraterecognition domain must 1990; Phillips et al., 1990; Korsgren and Cohen, 1991; Gentile reside between amino acids Gly38and Lys613in factor XIIIa, et al., 1991). A major function of factor XIIIa and the other based on the following data. 1) The purified fragment Gly38transglutaminases is to stabilize tissues by catalyzing inter- Lys613cross-links fibrin (Greenberg et al., 1988a), 2) a similar molecular isopeptide bonds between proteins (Chung, 1972; domain (Metl-AspS1’) expressed in Escherichia coli and activated by thrombin cross-linked fibrin (Lai et al., 1991), and *This study was supported in part by Grants HL-38245, HL- 3) Fukue et al. (1992) described an anti-factor XI11 antibody 26309, HL-28391, and AR39162. The costs of publication of this that cross-reacted with the peptide and inhibited article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accord- fibrin cross-linking. In this study, synthetic factor XIIIa peptides were used to ance with 18 U.S.C. Section 1734 solely to indicate this fact. 11 Supported by a Foundation for Anesthesia Education and Re- localize substrate recognition sites in factor XIIIa. The pepsearch Fellowship Award. tides were tested for their ability to inhibit cross-linking of 11 Supported by an American Heart Association Minority Fellow- protein substrates. Twenty-five different synthetic peptides ship. $$To whom all correspondence should be addressed Dept. of were evaluated for their ability to inhibit cross-linking (Fig. Medicine, Box 3031, Duke University Medical Center, Durham, NC 1 and Table I). Only two sequences affected substrate recognition.Peptides Asn7’-Aspg7 (peptide-4) and A ~ p ’ ~ - P h e ~ ~ ~ 27710. Tel.: 919-684-5823; Fax: 919-681-6160.

21284

Inhibitors Peptide

of Transglutaminases

21285

terns 120A HPLC system. The samples were spotted on Porton Instruments sample support discs and sequenced using the modified cycles PI-BGN and PI-1 recommended by Porton Instruments. The HCl samples (approximately 500 pmol) were hydrolyzed in 6 containing 0.1% phenol for 24 h. The tubes were evacuated and flushed with nitrogen several times before they were sealed under vacuum. The hydrolysates were analyzed in a Beckman Model 6300 amino acid analyzer using the sodium citrate buffer system provided by the manufacturer. Based upon these analyses, authentic, fulllength peptides-4 and -7 were identified (Table I) for use in crosslinking studies. Fibrin Cross-linking by Plasma FactorXZIIa-Cross-linking experEXPERIMENTAL PROCEDURES iments, unless otherwise stated, were conducted in a total volume of 20 pl of 0.1 M Tris-C1, pH 7.4, buffer at 25 “C, for 10 min in the Materials presence of 10 mM CaC12, with 2 p M fibrinogen and 0.02-0.04 p M Factor XI11 A-chain peptides were synthesized either at the Bio- plasma factor XIIIa. Plasma factor XI11 was activated during the polymer Resource Facility, Howard Hughes Medical Institute, Duke cross-linking reaction with 10 units/ml a-thrombin. When present, University or at Multiple Peptide Systems, San Diego, CA. Peptides- the peptides were allowedto preincubate with fibrinogen for 5 min a t 4 and -7 were further purified by reverse-phase high performance ambient temperature, before other components of the reaction mixliquid chromatography (RP-HPLC).’ In several experiments, unpu- ture were added. In preliminary experiments, conditions were estabrified peptide-4 and peptide-7 were also utilized. Unpurified peptide- lished to quantitate fibrin cross-linking by plasma factor XIIIa. A 4 was only partially soluble at high concentrations. The remaining time course of fibrin cross-linking with plasma factor XIIIa conducted peptides were used without further purification. Reverse-phase HPLC at 25 “C revealed that by 10 min, a y-dimer band was clearly and chromatograms of the remaining unpurified peptides indicated they reproducibly visible with as little as 0.25 p M fibrin. Fibrinogen Cross-linking by Guinea Pig Liuer Tissue Transglutawere between 52.9 and 100% pure. The peptides had a wide range of isoelectric points (3.2-12.5) based upon predicted values obtained mime-Cross-linking by tissue transglutaminase was carried out in using the MacVector program (IBI, New Haven, CT). Partially pu- a total volume of20p1 as described above with the following differrified human fibrinogen was purchased from Kabi Vitrum, Stock- ences. Plasma factor XIIIa was replaced with guinea pig liver tissue ) a-thrombin was omitted from holm, Sweden. Purified human plasma factor XI11 and purified factor transglutaminase (1.25-2.5 p ~ and XIII-free humanfibrinogen were purchased from American Diagnos- the reaction mixture. The reaction was terminated and samples tics, Greenwich, CT. Purified guinea pig liver tissue transglutaminase, prepared for sodium dodecyl sulfate-polyacrylamide gel electrophop-nitrophenyl phosphate, iodoacetamide, human serum albumin, and resis (SDS-PAGE) as described below. The following proteins were ovalbumin were purchased from Sigma. Bovine serum albumin (frac- studied for cross-linking: fibrinogen, N,N”dimethylcasein, and fibrotion V, Heat Shock) was purchased from Boehringer Mannheim. nectin. Solid Phose Microtiter Plate Assay for Transglutaminase ActiuityHuman a-thrombin was a gift from Dr. J. W. Fenton, 11, New York State Department of Health, Albany, NY. Flat bottom, high binding, An assay for measuring plasma factor XIIIa activity was used based upon the incorporation of 5‘-(biotinamido)pentylamine into N,N” 96-well enzyme immunoassay/radioimmunoassay microtiter plates were from Costar. Polyclonal rabbit anti-human factor XI11 A-chain dimethylcasein or fibronectin-coated microtiter plates (Slaughter et antiserum, N,N”dimethylcasein, purified human fibronectin, puri- al., 1992). To analyze whether the peptide inhibitorsaltered the fied bovine plasma fibronectin, and D-phenylalanyl-L-prolyl-L-argi-primary amine binding site or the glutamine substrate site, factor nine-chloromethyl ketone, were purchased from Calbiochem. Strep- XI11 (1.56 Fg/ml) and 5’-(biotinamido)pentylamineconcentrations tavidin-conjugated alkaline phosphatase was purchased from The from 0.04 to 5 mM were reacted in the presence of inhibitory concentrations of either peptide-4 orpeptide-7. The transglutaminase activBinding Site, Inc., Birmingham, United Kingdom. NalZ61(100 mCi/ ml) was from Amersham. Sephadex PD-10 gel filtration columns were ity was measured as described previously (Slaughter et al., 1992). SDS-PAGE-The cross-linking reaction was terminated by the from Pharmacia LKB Biotechnology Inc. All other chemicals were addition of15-30 pl of reducing SDS-PAGE solution and boiling reagent grade. (Mary et al., 1988; Greenberg et al., 1988a). Reaction mixtures were separated by SDS-PAGE in gels composed of 8, 10, or 4-10% linear Methods gradient separating gels with a 4% stacking gel. After SDS-PAGE, Peptide Purification and Sequencing-Synthetic peptides-4 and -7 the gels were stained using Coomassie Brilliant Blue and destained (Table I) were purified using either analytical (0.46 X 25 cm) or (Mary et al., 1988; Greenberg et al., 1988a, 198813). preparative (2.2 X 25 cm) Vydac C-18 RP-HPLC columns (The Quantitation of Protein Cross-linking-The destained gels containSeparations Group, Hesperia, CA). Peptide-4 was dissolved to a final ing the separated, cross-linked products, were dried on Whatman No. concentration of4.5 mg/ml in a mixture composed of4.5% (v/v) 1 filter paper. Transparencies of the dried gels were made, and the dimethylformamide, 13.6% (v/v) of a mixture of 0.1% trifluoroacetic protein bands were scanned using a GS300 Transmittance/Reflecacid in a 7030 mixture of acetonitrile and water and 81.8% of 0.1% tance Scanning Densitometer (Hoefer Scientific Instruments, San trifluoroacetic acid. Peptide-7 was dissolved in 0.1 M Tris-HC1, pH Francisco, CA). Densitometric tracings were quantitated using a 7.4, a t a concentration of 5 mg/ml and centrifuged. Both peptide GS370 1-D Electrophoresis Data System (Version 2.0) installed on a solutions were further diluted (typically 10-foldin 0.1% trifluoroacetic Macintosh SE Computer. Peak areas were calculated using the manacid in water) before injecting onto the RP-HPLC column. Solvents ual integrationmode. The distinct y-dimerformation was quantitated used for elution were 0.1% trifluoroacetic acid in water (solvent A) and expressed as the degree of cross-linking. Guinea pig liver tissue and 0.1% trifluoroacetic acid in a mixture of 7030 acetonitri1e:water transglutaminase cross-linking of fibrinogen was quantitated by (solvent B). A gradient of 0-100% solvent B was achieved in 30 min measuring the extent of formation of high molecular weight crossa t a flow rate of either 3.0 ml/min (preparative RP-HPLC) or 1.0 ml/ linked multimers ( d i m e r s , a-y-multimers, and a-multimers) and min (analytical rp-hplc). The column eluate was monitored at 280 the disappearance of the Aa- andy-chains of fibrinogen. nm. When the purified peptide was rechromatographed on the RPZmmunoblotting of Factor XIII A-chains-Samples containing HPLC column, it eluted with the same retention time with no traces plasma factor XI11 were subjected to SDS-PAGE and immunoblotting of contaminating materials. using polyclonal rabbit anti-factor XI11 A-chain antisera as described The major peaks absorbing a t 280 nm following RP-HPLC of previously (Greenberg et al., 1988b; Mary et al., 1988). peptides4 and -7 were collected as they emerged from the column a-Thrombin Cleavage of Plasma Factor XIZI-Purified plasma and dried. The dried materials were subjected to amino acid compo- factor XI11 (10 pg/ml) was incubated in 0.1 M Tris-C1, pH 7.4, with sition and sequence analyses. Automated Edman degradation was fibrinogen (680 pg/ml), CaC12 (10mM), in the presence or absence of performed in an Applied Biosystem 477A protein sequenator with peptide4 or peptide-7. Factor XI11 was activated by adding a-throm“on-line” analysis of the phenylthiohydantoins in anApplied Biosys- bin (10 NIH units/ml) and incubating at ambient temperature for 10 min. The reaction was stopped by the addition of reducing SDSThe abbreviations used are: RP-HPLC, reverse-phase high per- PAGE stop solution. The proteins were resolved by SDS-PAGE and formance liquid chromatography; PAGE, polyacrylamide gel electro- visualized by staining with Coomassie Brilliant Blue or by immunophoresis; TBS, Tris-buffered saline. blotting to detect factor XI11 A-chain antigen (Mary et al., 1988;

(peptide-7) reversibly inhibited factor XIIIa mediated crosslinking of fibrin, N,N’-dimethylcasein, and fibronectin. Peptides-4 and -7 had no effect on thebinding of factor XIIIa to fibrin demonstrating that the fibrin binding sites and the substrate recognition sites are distinct. Peptides4 and -7 also inhibited tissue transglutaminase cross-linkingof fibrinogen, N,N‘-dimethylcasein, and fibronectin, suggesting that these sequences play an important role in substrate recognition for other members of the transglutaminase gene family.

21286

Peptide Inhibitors of Transglutaminases

Greenberg et al., 1988b). Thrombin cleaved and uncleaved plasma factor XI11 A-chain antigen bands were quantitated by scanning densitometry. Effect of Peptides on the Binding of Factor XIIIa to Cross-linked Fibrin-Factor XIIIa binding to fibrin was measured by quantitating: ( a ) the factor XIIIa antigen bound to the fibrin pellet, or (6) the residual factor XIIIa activity in the supernatant. Withpeptides 5-7, 9-10, and 13-16, we measured factor XIIIa activity in the supernatant. The effect of peptide-4 on factor XIIIa binding to fibrin was analyzed by quantitating the factor XIIIa antigen bound to fibrin in the presence or absence of the peptide. ( a ) Plasma factor XI11 (20 pg/mL) was mixed with fibrinogen (0.4 mg/ml), 10 mM CaC12,2.4 mg/ml bovine serum albumin, in the ) buffered using 20 mM presence or absence of peptide-4 (259 p ~ and Tris-C1, 130 mM NaCl, pH 7.4 (Tris-buffered saline, TBS). Fibrin was clotted by the addition of a-thrombin (20 units/ml) and incubated at 37 “C for 10 min. Fibrin was centrifuged (15,000 rpm, 10 min) in a Microspin tabletop microcentrifuge (Sorvall Instruments, Du Pont Co.) and washed twice with TBS. The fibrin pellet and supernatant were then solubilized in SDS-PAGE reducing solution and subjected to SDS-PAGE and immunoblotting (Greenberg et al., 1988a, 198813; Mary et al., 1988). Fibrin-bound and unbound factor XI11 A-chain antigen bandsappearing in the immunoblots (Greenberg et al., 1988a; Achyuthan et al., 1988) were quantitated by scanning densitometry. b) Plasma factor XI11 (30 pg/ml) was mixed with purified human fibrinogen (1.0 mg/ml) in the presence of 1%(w/v) bovine serum albumin, 10 mM CaC12, and TBS in the presence or absence of the synthetic peptides as described above. Each peptide was tested at a final concentration of 2.0 mM except for peptide-7 which was tested at a concentration of0.5mM. Peptide concentrations were chosen such that they were either inhibitory (peptide-7) or noninhibitory (peptides-5,6,8-10, and 13-16) in the fibrin cross-linking assays (see “Results”). Control incubations included either omitting the peptide or omitting the fibrinogen from the binding mixture. Clotting was initiated by the addition of a-thrombin (20 units/ml) and incubation at 37 “C for 15 min. The reaction mixtures were then centrifuged in an Air Driven Ultracentrifuge (Beckman Instruments, Inc.) at ambient temperaturea t approximately 150,000 X g for 15 min. The clear supernatant was aspirated and the plasma factor XIIIa activity was estimated by measuring the incorporation of [3H]putrescine into N,N”dimethylcasein as described previously (Miraglia and Greenberg, 1985). Analysis of Peptides (IS Glutamine Donors or Lysine AcceptorsThe possibility that the peptides functioned as pseudosubstrates in the cross-linking reactions was tested by attempting tocross-link ‘“Ilabeled peptide-7 to fibrin. Peptide-7 was radiolabeled with Na’T using IODO-BEADS and purified by Sephadex gel filtration (Greenberg et al., 1988a). ‘261-Peptide-7(56-45,000 cpm) was mixed with Kabi fibrinogen (1.0 mg/ml) containing factor XIII, bovine serum albumin (O.l%, w/v), 0.1% (v/v) Tween 20, and 0.15 M NaC1. The reaction was buffered with 50 mM Tris-C1, pH 7.4. The mixture was preincubated at 37 “C for 10 min. Then 10 mM CaClz and 5 units/ml a-thrombin were added. EDTA (10 mM) controls were included to inhibit cross-linking. The reaction was incubated for 10 min at 37 “C. The fibrin pellet was compressed by a wooden stick to extrude all liquid. The fibrin was washed twice with 50 mM Tris-C1, pH 7.4, and the radioactivity measured in a y-counter. A gel shift assay was also used to determine if the peptides were cross-linked to fibrin by factor XIIIa. In this assay, Kabi fibrinogen (0.272 mg/ml), containing factor XIII, was clotted with a-thrombin (10 units/ml) in the presence of CaC12 (10 mM) and buffered with 0.02 M Tris, 0.13 M NaCl, pH 7.4 (TBS). EDTA was included in some experiments to inhibit cross-linking. Either peptide-4 (137 pM) or ) also added to some reactions. Fibrinogen and peptide-7 (114 p ~was either peptide-4 or -7 were allowed to preincubate at 37 “C for 15 min in TBS. Then the cross-linking reaction was initiated by adding calcium chloride anda-thrombin. After allowing the reaction to proceed at 37 “C for 15 min, the reaction was stopped and themixture separated by either a 10 or 4-15% linear gradient SDS-PAGE. The gel was stained with Coomassie Brilliant Blue and destained. Effects of Peptides-4 and -7 in Combination on 5’4Biotinamido)pentylamine Incorporation into N,N”Dimethylcasein by Factor XZIZa-To determine whether the peptides acted in an additive or synergistic pattern to inhibit factor XIIIa, purified peptide4 or -7 was added at increasing concentrations (1.5-100 p ~ to) factor XIIIa (1.56 pglml)) and the ICW concentration of the other peptide. A control assay was performed in parallel using either peptide-4 or -7 alone at concentrations from 1.5 to 100 p ~ The . microtiter plate

assay (Slaughteret al., 1992)was performed and thetransglutaminase activity measured. The transglutaminase activity recovered was determined by measuring activity in the presence of both peptides and comparing to the measurements with only one peptide present. All assays were performed in triplicate. Effect of Peptides-4 and -7on the Active Site of Factor XIIIaIncreasing concentrations of iodoacetamide (10-9-10-3M) were incubated with factor XIIIa in the presence of a concentration 10-fold in excess of the IC,of either peptide4 or -7. Factor XI11 (1 pg) was activated with a-thrombin (50 units/ml) and CaClz (10 mM) in a total volume of 50 p1 of 0.1 M Tris-C1,0.15 M NaCl, pH 7.4, in the presence of iodoacetamide and either peptide-4 or -7.Following a 20-min incubation at 37 “C, a sample from the reactant vials was added to the microtiter plate resulting in a 20-fold dilution. Recovery of factor XIIIa activity was determined by measuring the incorporation of 5’(biotinamid0)pentylamine into N,N”dimethylcasein as previously described (Slaughter et al., 1992). Control experiments were performed in parallel omitting either the iodoacetamide or the peptide during activationof factor XIII. RESULTS AND DISCUSSION

The mechanism by which factor XIIIa recognizes proteins with reactive glutamine residues is not well defined. The first step in the catalytic mechanism for transglutaminases is the formation of a thioester bond between the active site of the enzyme and thereactive glutamine residue (reviewedby Folk and Finlayson (1977)). In this study we examined the possibility that common substrate recognition sequences for reactive glutamines resided outside of the active site pocket of the transglutaminase.Many enzymes have domains mediating substrate recognition that exist outside the active site pocket. The use of synthetic peptides derived from enzymes to map substrate recognition sites was used to study blood coagulation factors Xa, XIa, activated human protein C, and thrombin (Chattopadhyay and Fair, 1989; Chattopadhyay et al., 1992; Altieri et al., 1991; Mesters et al., 1991; Baglia et al., 1991; Binnie and Lord, 1991). We used a similar approach to define the substrate recognition domains in factor XIIIa. In earlier studies we localized the substrate recognition sites to amino acid sequences G l ~ ~ ~ - L(Greenberg y s ‘ ~ ~ et al., 1988a; Lai et al., 1991). Since guinea pig liver tissue transglutaminase (Greenberg et al., 1991) exhibited significant cross-species amino acid homology within the Gly38-Lys613 sequence of factor XIIIa (Gentile et al., 1991) (Fig. l), we examined whether the substrate recognition sites were shared. Synthetic peptides with amino acid sequences spanning different regions of the factor XI11 A-chain (Fig. 1, Table I) were tested for their effects on factor XIIIa activity. These peptides had a wide range of molecular masses (748-4891 Da) and predicted isoelectric points (PI 3.2-12.5). Several peptides had significant homology with guinea pig liver tissue transglutaminase (Fig. 1).The peptides (Table I) represented approximately 30% of the total factor XI11 A-chain amino acid sequence, 36% of the sequence of the Gly38-Lys613fragment, and 51% of the regions of homology with guinea pig liver tissue transglutaminase (Fig. 1).Peptide-4 represented about 40% of the exon I11 sequence and was 70% homologous to a similar region of the guinea pig liver tissue transglutaminase (Ikura et al., 1988). Peptide-7 represented the entire exon V sequence andexhibitedabout 61% homology to a similar region on guinea pig liver tissue transglutaminase (Ikura et al., 1988). The remaining factor XI11 A-chain sequences were predicted to be very hydrophobic, based upon Kyte-Doolittle analyses and were not synthesized (Fig. 1). Effect of Factor XIII-derived Peptides on Plasma Factor XIIIa Cross-linking of Fibrin-Fibrin y-chain dimers are formed by the reciprocal cross-linking of the y-chains between adjacent fibrinmolecules and involves Gln3”.Gln399f and Lysm on the y-chains resulting in a y-dimer that is oriented

21287

Peptide Inhibitors of Transglutaminases Thrombin

Active site

4

4

Thrombln 4

.. .............................................................................................................................................................................................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . :::::: . .. .. .. .. .. .. .. .. ............... .. .. .. .. .. .. .. .. .. .. .. .. ............... .. .. .. .. .. .. .. .. .. .. .. .. ............... .. .. .. .. .. .. .. .. .. .. .. .. ............... .. .. .. .. .. .. .. .. .. .. .. .. ............... .. .. .. .. .. .. .. .. .. .. .. .. .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Factor Xlll A-chain t si Fibrin Binding Domain

..........................................................................

t M73i

.................................... t

;0 EP

Q38

Peptides

& hi

Homologous Domains ExonNumber

H

111

111

I

SH d i

id

a

- 50 koa

H

H H , ” W

IV

t

%I3

I v l v 1 1 VII ( V I I I I I X I X I

XI

1

XII

IXIIIIXIV~XV

F I G . 1. Schematic diagram of factor XI11 A-chain. Schematic representation of the A-chain of factor XI11 (stippled bar) beginning with N-acetyl-serine 1 and ending with methionine 731 and showing the following: horizontal lines, the regions of amino acid sequence homology to guinea pig liver tissue transglutaminase; roman numerals, the location of the exons; hatched bar, the location of the 50-kDa fibrin-binding domain of factor XIII; open bars, the location of peptides synthesized from factor XI11 A-chain sequence; GB and &13, major thrombin cleavage sites; and SH, the location of the active site cysteine 314 residue.

TABLEI Synthetic peptides derived from factor XIII A-chain sequence The amino acid sequences were as reported by Ichinose et al. (1986). PeDtide No. Amino acid sequence 1 2 3 4 4a

4b” 4c 5 6 7 7d” 7e” 7c 7d 7e 7f 8 9 10 11 12 13 14 15 16

M,

1324 S-E-T-S-R-T-A-F-G-G-R-R 1124 R-R-A-V-P-P-N-N-S-N 1143 P-P-N-N-S-N-A-A-E-D-D 3226 N-K-L-I-V-R-R-G-Q-S-F-Y-V-Q-I-D-F-S-R-P-Y-D-P-R-RD 1170 N-K-L-I-V-R-R-G-Q-S 1212 R-G-Q-S-F-Y-V-Q-I-D 1423 D-F-S-R-P-Y-D-P-R-R-D 1149 G-R-Y-P-Q-E-N-K-G-T 1176 R-T-S-R-N-P-E-T-D-T 4891 D-D-A-V-Y-L-D-N-E-K-E-R-E-E-Y-V-L-ND-I-G-V-I-F-Y-G-E-V-N-D-I-K-T-R-S-W-S-Y-G-Q-F 1181 D-D-A-V-Y-L-D-N-E-K 1487 Y-L-D-N-E-K-E-R-E-E-Y 1279 E-R-E-E-Y-V-L-N-D-I 1315 Y-V-L-N-D-I-G-V-I-F-Y 1112 G-V-I-F-Y-G-E-V-N-D 1259 K-T-R-S-W-S-Y-G-Q-F 1056 D-L-S-G-R-G-N-P-I-K 748 N-A-K-D-D-E-G G-S-W-D-N-I-Y-A-Y-G-V-P-P-S-A-W-T-G-S 2028 R-Y-G-Q-C-W 812 1172 H-D-N-D-A-N-L-Q-M-D 1008 E-E-D-G-N-V-M-S-K 1049 D-S-T-P-Q-E-N-S-D-G D-L-I-Y-I-T-A-K-K-D-G-T-H-V-V-V-E-N-V-D-A-T-H-I

K-F-0-E-G-0-E-E-E-R

2553 1279

These peptides were “Insoluble” and could not be tested.

in an antiparallel fashion (Hoeprichand Doolittle, 1983; Purves et al., 1987). Peptides-4 and -7 individually produced a concentration-dependent inhibition of y-chain cross-linking of fibrin (Fig. 2 A ) . A visible clot was observed in the tube after cross-linking in the presence of these peptides indicating that they were not affecting fibrin polymerization and gelation. Peptide-4 was 4-5-fold more potent an inhibitor than peptide-7 (Table 11) with ICbovalues for peptides-4 and -7, 12 and 52 p ~ respectively , (Table 11). The other peptides at 1 mM had no effect on cross-linking. Peptides4 and-7 also inhibited preactivatedplasma factor XI11 or placental factor XI11 suggesting that thepeptides were not inhibiting thrombincleavage of factor XI11 or altering the expression of the active site cysteine. Neither peptide-4 (324 p M ) nor peptide-7 (212 p ~ had ) any effect on thrombin cleavage of plasma factor XI11 when examined by SDS-PAGE (data not shown). The CaClz concentration (10 mM) used in the fibrin cross-linking assays was 50-100-fold higher than the peptide concentration (Fig. 2 A ) ensuring that thepeptides were not interacting with calcium ions. The concentration

(molar ratio) of peptide-4 in excess of fibrin resulting in 50 and >80% inhibition were 6 and 20:1, respectively. In contrast, peptide-7 required a 26- and 69-fold molar excess to produce 50 and >80% inhibition, respectively. Binding of Peptides-4 or -7 to Fibrin: Potential Role as Pseudosubstrates-We successfully radioiodinated and purified peptide-7 and found that less than 10%of the 9-peptide7 remained bound to washed fibrin clots. When solubilized fibrin from these experiments was separated by SDS-PAGE and analyzed by autoradiography, there was no evidence of cross-linking of lZ5I-peptide-7to any of the fibrin chains. Since peptide-4 could not be readily radiolabeled and purified, we attempted to cross-link unlabeled peptide-4 and peptide-7 to fibrin. If the peptides were acting as apseudosubstrates in the cross-linking reaction, an increase between 3.2 and 4.9 kDa in themolecular mass of the Aa- and/or y-chainsshould be detected by SDS-PAGE. As shown in Fig. 3, no change in molecular weight of Aa- or y-chains was detected. These studies directly demonstrate that thepeptides were not acting as pseudosubstrates.

2 1288

Peptide Inhibitors of Transglutaminases 120 -.

JI

100

-

Y

-E -ii

80-

“Q- Peptide4

+ Peptide-7

Y C

a

60-

L

0

-c

40

-

L

f

200 0

I

I

20

40

I

FIG. 2. Peptides-4 and -7 inhibit fibrin cross-linking ( A )and also inhibit the incorporation of 6‘-(biotinamido)pentylamine into N,N‘-dimethylcasein ( B )or fibronectin (C). Plasma factor XIIIa catalyzed crosslinking of fibrin y-chains (A) or the incorporation of 5’-(biotinamido)pentylamine into either N,N”dimethylcasein ( B ) or into fibronectin (C) in the presence or absence of peptides-4 or -7 were performed as described under “Experimental Procedures.”

I

i

I

I

6 0 140 8 0120 Peptldo Y M )

100

+ Peptide4 + Peptide7 a

20

-

Ln

0 40 0

I

I

20 Peptides

1

I

60

80

a I

100

(JJM)

U Peptide-4

+ Peptide-7

L

0

0

25-

a

0

I

0

Fibrin Specific Reversal of the Inhibitory Effectsof Peptides tested whether the peptides were interacting with the protein substrates by examining the effects of increasing fibrinconcentrationson cross-linking. Fibrin reversed the inhibition by both peptides-4 and -7 (Fig. 4). More fibrin was required to reverse peptide4 inhibition compared to peptide-7 (Fig. 4) consistent with the potent inhibition of peptide-4. Fibrin concentrations to reverse inhibition with peptides-4 and -7by 50% were 8 and 3.2 PM,respectively (Fig. -4 and -7-We

20

I

40 Peptide

I

1

I

60

80

100

(yM)

4). Inclusion of 15 p~ bovine serum albumin, human serum albumin, or ovalbumin had no effect on the inhibition of fibrin cross-linking by either peptide-4 or peptide-7. These results suggest that peptides-4 and -7 reversibly inhibit fibrin cross-linking by interacting with the substrate rather than the enzyme. Effect of Peptides on5’-(Bwtinamido)pentyylanine Incorporation Into N,N”Dimethylcasein by Plasma Factor XIIIaSince fibrin contains bothglutamine and lysine cross-linking

Peptide Inhibitors of Tramglutaminases

21289

TABLE I1 Inhibition concentrations ofpeptides for foctor XZZZa The remaining peptideswere non-inhibitory when tested to 1 mM. 5’-(Biotinamido)pentylamine incorporation into

Fibrin crosslinking ICw

Peptide

N,W-Dimethylcasein

IC,

Fibronectin 1C.m ( W ) O

(pM)

PM

5 25 12 Peptide-4 5 52 10 Peptide-7 >loo0 1935 375 Peptide-7a 1050 NIb NIb Peptide-7b 2265 678 4000 Peptide-7c 1276 225 4000 Peptide-7f a IC,, peptide concentration required to produce 50% inhibition of factor XIIIa activity. NI, noninhibitory to atleast 1 mM concentration.

kDa

A

-180 -116

w-

-84 58

-

Aa -

BP -

-48.5

Y-

-36.5 “26.3

Peplide-4-

1

2

3

4

-

B

a

kDa -180

-36.5

-26.6

5

6

7

8

FIG.3. Peptides-4 or -7 do not alter the mobility of the ychains while inhibiting fibrin cross-linking. Fibrin was crosslinked in the presence or absence of either peptide-4 (panel A ) or peptide-7 (panel B ) . Cross-linked products were then analyzed by either a 4-15% linear gradient (panel A ) or a 10% (panel B ) SDSPAGE. Lanes 1, 2, 5, and 6 show cross-linked fibrin; lunes 3 and 7 show non-cross-linked fibrin; and lunes 4 and 8 show the mobilities of molecular weight standards. Lane 1, no peptide; lune 2, with 137 p~ peptide-4; lune 5,with 114 p~ peptide-7; lune 6, no peptide.

V

0

3

6

9

Peptide-4 Peptide7 12

15

18

Fibrin

(YM) FIG. 4. Fibrin reverses the inhibitionof y-chain cross-linking produced by either peptide-4 or peptide-7. ”Chain crosslinking by plasma factor XIIIa was carried out in the presence of either peptide-4 (324 PM) or peptide-7 (213 or 340 p M ) and increasing . other conditions are as deconcentrations of fibrin (0-15 p ~ ) All scribed in the text. Datapresented are theaverage of either duplicate (peptide-4) or triplicate (peptide-7) measurements.

sites, the peptides could be modifying either site. Therefore, we tested whether the peptides could inhibit factor XIIIa using a protein that could only serve as a glutamine donor. N,N’-Dimethylcasein was selected because the lysine residues are chemically blocked producing a specific glutamine donor for the transglutaminase reaction. The effects of the peptides on the incorporation of 5’-(biotinamido)pentylamine into N,N”dimethylcasein by factor XIIIa (Slaughter et al., 1992) were remarkably similar to those obtained for fibrin crosslinking. Only peptides-4 and -7 were inhibitory (Fig. 2B) with ICSo values of 5and 10 PM, respectively (Table 11). The inhibitory peptides could beeither directly altering the active site of factor XIIIa, interfering with the formation of thioester bond between the active site cysteine and the reactive glutamine, and/or altering the primary amine binding site. We performed experiments described in the following section to distinguish between these possibilities. Effect of Peptides-4 and -7 on the Active Site of Factor XIIIa and Primary Amine Binding Site-The alkylation of factor XIIIa requires that the active site Cys314residue be freely exposed. Increasing concentrations of iodoacetamide alkylated theactive site and inhibitedfactor XIIIa activity. Neither peptide interfered with inactivation of the active site demonstrating that the peptides did not interact directly with the enzyme or interfere with reactivity of Cys314. Since the peptides do not modify exposure of the active site, they must interfere with either the reactivity of the glutamine or primary amine substraterecognition site(s). Thepeptides could either cause a conformational change in the substrates or produce steric hindrance limiting access to the glutamine residue or primary amine. T o further define whether the peptides altered the primary amine or glutamine cross-linking sites, factor XIIIa was incubated with increasing concentrations of 5’-(biotinamido)pentylamine and a fixed concentration of the glutamine substrate, N,N”dimethylcasein, and the peptide inhibitors. Increasing concentrations of the primary amine 5’4biotinamido)pentylamine failed to reverse the inhibition of factor XIIIa by peptide-4 and -7 (Fig. 5). This is in contrast to reversing the reaction with increasing fibrin concentrations. The maximal velocity of the reaction was reduced with depletion of the glutamine substrate (Fig. 5). Attempts to measure ammonia release using the coupled reaction with glutamate

21290


-

I

I

Control Peptide 4; 72.5 uM Peptide 7; 20 uM

200

100

0

I I

I

I

I

I

I

I

0

1

2

3

4

5

6

Substrate (mM) FIG. 5. Effect of increasing 5‘-(biotinamido)pentylamineconcentration on the inhibition of N,N’-dimethylcasein crosslinking by peptides-4 and -7.Microtiter plates coated with N,N’-dimethylcasein were incubated with increasing concentrations of 5’(biotinamidobentvlamine, factor XIIIa. and inhibitorv concentrations of peptides4 and -7. Transglutaminase activity was measured as described und‘er “Experimental Procedures.”

dehydrogenase described by Fickenscher et al. (1991) and Muszbek et al. (1985) were unsucessful due to interference by the peptides with the assay. Therefore, the peptides produce an alteration in factor XIIIa’s recognition of the glutamine substrate. Effect of Factor XIII-derived Peptides on Plasma Factor XIIIa Binding to Fibrin-Since the peptides were inhibiting catalysis by altering glutamine substrate recognition, we were interested in establishing whether they also inhibited factor XIIIa localization to thefibrin clot. Factor XIIIa is known to remain noncovalently bound to insoluble fibrin clots (Greenberg et al., 198713; Hornyak and Shafer, 1992). Although peptides-4 and -7 inhibited fibrin cross-linking (Fig. 2 A ) , they did not inhibit factor XIIIa binding to fibrin under conditions where cross-linking was completely inhibited. Furthermore, peptides-5, -6, -9-10, and -13-16 at 2 mM had no effect on factor XIIIa binding to fibrin. In control experiments, the peptides did not affect the solubility of factor XIIIa. The peptides apparently interact with fibrin at a site (or multiple sites)that is distinct from that involved in factor XIIIa binding. Recent studies suggest that factor XI11 and factor XIIIa interact at distinct sites on fibrinogen and fibrin (Hornyak and Shafer, 1992). This suggests there may be multiple sites involved in factor XI11 and factor XIIIa binding to fibrin. Results from our studies suggest that thefibrin binding sites on factor XIIIa can be dissociated from those involved in substrate recognition. Factor XIIIa is a dimeric enzyme that displays a complex set of interactions between each A-chain during catalysis of covalent cross-links (Hornyak and Shafer, 1992). Multiple domains of this complex tertiary structure may localize factor XIIIa to the fibrin surface. We cannot exclude the possibility that one or more of these peptide sequences could affect the binding of factor XI11 or factor XIIIa to fibrin(ogen) in concert with other peptide sequences using different assay methods. Studiesarein progress to characterize the fibrin binding sites on factor XI11 and factor XIIIa. Effect of Peptides withSequences Derived from within Peptides-4 and -7 on Plasma Factor XIIIa Cross-linking of Fibrin and N,N’-Dimethylcasein-Since peptides4 and -7 were the most potent inhibitorsof fibrin cross-linking by plasma factor

XIIIa, we synthesized shorter peptides (10-11-mers) (Table I) to further define the amino acid sequence(s) responsible for their inhibitoryeffects. Peptides-4a and -4c had no effect on fibrin cross-linking by plasma factor XIIIa when tested at 4 mM. Peptides-7b and -7f had IC, values of 1050 and 1276 p~ and inhibited cross-linking more strongly than peptides7a and -7c with ICs0values of 1935 and 2265 p ~respectively , (Table 11).However, peptides-7b and -7f wereat least 20-fold less potent compared to the intact peptide-7 (Table 11).Therefore all portions of peptide-7 had some inhibitory activity although the amino- and carboxyl-termini of peptide-7 were more potent a t inhibiting fibrin cross-linking. In contrast, peptides-4a-4c (Table I), did not inhibit cross-linking suggesting the entire peptide-4 sequence was necessary to function as an inhibitor. We found a similar effect when these shorterpeptides were used to inhibit N,W-dimethylcasein cross-linking. Peptides7a and -7f inhibited at lower concentrations (ICs0,375 and 225 p ~than ) peptide-7c (ICIo,678 p ~ (Table ) 11).The shorter peptides required 20-60-fold higher concentrations compared to thefull-length peptide-7 to inhibit factor XIIIa (Table 11). Peptides-4a, -4c, and -7b were noninhibitory up to 1.0 mM concentrations. Therefore, factor XIIIa-derived peptides were inhibiting factor XIIIa’s interaction with two different glutamine containing protein substrates. Effects of Using Peptides-4 and -7 in Synergy on 5’4Biotinamid0)pentylamineIncorporationinto N,N‘-Dimethylcasein by Factor XIIIa-To establish whether the two inhibitory peptides were synergistic intheirinhibition, we tested whether one peptide would potentiate the inhibition of the other. The presence of increasing concentrations of one peptide and a fixed concentration of the other peptide had no synergistic effect since the factor XIIIa activity recovered was within 10% of the predicted value (data not shown). The failure to detect anysynergistic inhibitory effects between the peptides over a wide concentration range suggests that they interact with the glutamine substrates at noninteracting sites. These sites may be similar or in close proximity resulting in additive rather than synergistic effects. Effects of Peptides on 5’-(Biotinamido)pentylmine Incorporation into Fibronectin by Plasma Factor XIIIa-The pep-


21291

tides were also tested for their effect on factor XIIIa mediated ited fibrinogen a-chain cross-linking with an ICs0 value of incorporation of 5'-(biotinamido)pentylamine into fibronec- 1318 p~ (Table 111). The remaining peptides were either tin. Glutamine-3 isthe major cross-linking site on fibronectin noninhibitory or only weakly inhibitory (ICs0 values of X . 2 for factor XIIIa (McDonagh et al., 1981). Both peptides-4 and mM, Table 111). These data closely resembled that obtained -7 inhibited this reaction (IC, of 25 and 5 pM, respectively) with factor XIIIa cross-linking the y-chains. Tissue transglu(Fig. 2C, Table 11),whereas peptides 1-3, 5-6, and 8-16 were taminase is known to have a broader substrate specificity noninhibitory at 1.0 mM. We also tested the effect of short than factor XIIIa (Folk and Finlayson, 1977; Gorman and Folk, 1973) despite having a high degree of homology for the peptides derived from peptides4 and -7 on5'-(biotinamido)pentylamine incorporation intofibronectin. When 1.0 mM active site (Ikura et al., 1988). Since guinea pig liver tissue of peptides-7a, -712, and -7f were tested, they produced 39,59, transglutaminase is a monomer (Connellan et al., 1971) and and 90% inhibition, respectively (Table 11). Peptides-4a, -4c, factor XIIIa is a dimer this could allow different interactions and -7b were noninhibitory at 1.0 mM. Once again the amino- with the same substrate. The relative differences in the poin inhibiting factor XIIIa andtissue and carboxyl-terminal sequences of peptide-7 were inhibitory tency of peptides-4 and -7 while shorter peptides derived from peptide-4 were noninhi- transglutaminase could be explained by differences in either bitory. These data are consistent with our earlier results on the primary amino acid sequence of the enzymes, the fibrinthe cross-linking of fibrin by plasma factor XIIIa (Fig. 2 A ) . ogen Aa-chains and/or y-chain sequences. Despite these differences, factor XIII-derived peptides disrupted substraterecThe factor XIIIa-derived peptides inhibited three proteins having very different primary aminoacid sequences surround- ognition by the tissue transglutaminase (Table 111). The ining the reactive glutamine. This suggests that there arecom- hibitory peptides were 60-70% homologous to the sequences mon secondary or tertiary structures thatsurround the reac- inthe guinea pig liver transglutaminase, which probably tive glutamine allowing factor XIIIa to recognize them. Ad- account for their ability to act asinhibitors. Summary-Although amino acid sequences Asn12-Asp91 ditional studies are needed to define the three-dimensional were e ~ identified ~~ in these studies as potential structure of these sites.We knowthat in the case of fibrinogen and A ~ p ' ~ - P h the y-chain cross-linking site is conformationally dependent factor XIIIa substrate recognition sites, we cannot exclude since GPRP can bind to fibrinogen and inhibit cross-linking the possibility that other portions of the enzyme play a role (Achyuthan et al., 1986). In addition, fibrinogen Paris I, a in this process. The underlying mechanism for inhibition by congenital dysfibrinogenemia caused by a mutation in intron the peptides appears to be depletion of glutamine containing 8 results in the insertion of a 15-aminoacid sequence after y- protein substrateby the peptides. Since only excess glutamine chain amino acid 350 (Rosenberg et aL, 1993) and inhibits substrate reversed the inhibition, the peptide inhibitors appear to react with the substrate and decrease the effective recognition of the y-chain by factor XIIIa. Effect of Factor XIII-derived Peptides on Fibrinogen Cross- concentration in the reaction (Segel, 1975). Increasing the fibrin concentration increased the extentof y-y cross-linking linking by Guinea Pig Liver Tissue Transglutaminme-We next investigated whether the factor XIII-derived peptides and ultimately reversed the inhibition which is consistent with inhibition by substrate depletion (Segel, 1975). Increascould also inhibit guinea pig liver tissue transglutaminase cross-linking. The tissue enzyme is monomeric and does not ing the primary amine concentration in the reaction did not require thrombin for activation orexpression of its active site reverse the inhibition. We have excluded the possibility that the peptides were cysteine residue (Folk and Finlayson, 1977). Guinea pig liver transglutaminase preferentially cross-links the Aa-chains of acting as pseudosubstrates for the following reasons. 1) The fibrinogen, whereas factor XIIIa initially cross-links the y- presence of either peptide-4 (3.2 kDa) or peptide-7 (4.9 kDa) chains resulting in amarkedly different cross-linkingprocess did not result in any shift in the mobility of the individual (Shainoff et al., 1991). There is no sequence homology sur- fibrin chains when analyzed by SDS-PAGE (Fig. 3). 2) Several rounding the reactive glutamine residues in the Aa-chain or of the peptides had one or more glutamine and/or lysine y-chain of fibrinogen (Cottrell et al., 1979) and cross-linking residues in their sequence and were tested at 10-50-fold higher studies of fibrinogen and tissue transglutaminaseshould pro- concentrations than peptides-4 or -7and they were not inhibvide further insight into the molecular basis of substrate itors. 3) Peptide-7c lacking glutamine and lysine residues also recognition. Peptides-4 and -7 were potent inhibitors of fi- inhibited factor XIIIa. 4) Shorter peptides (peptides-4a-4c) brinogen a-chain cross-linking with ICs0values of 25 and 13 derived from peptide-4 failed to inhibit plasma factor XIIIa when tested with fibrin, fibronectin, or N,N'-dimethylcasein pM, respectively (Table 111). The other peptides were not in either cross-linking or amine incorporation assays. 5) Exinhibitory up to aconcentration of 4 mM. Whenshorter peptides (4a-4c and 7a-7f) were tested, only peptide-7a inhib- periments conducted with '251-peptide-7 directly demonstrated that the peptide was not a pseudosubstrate but was noncovalently bound to fibrin. TABLE111 Synthetic glutamine containing peptides were used in earInhibition of fibrinogen cross-linking by guinea pig liver tissue tramglutaminase lier studies to define the molecular basis of factor XIIIa and The remaining peptides were not inhibitory when tested up to 2 tissue transglutaminase substrate recognition (Gorman and mM. Folk, 1980). Results from these earlier investigations sugICw for gested that theprimary aminoacid sequence surrounding the Peptide fibrinogen glutamine containing substrates was not the only factor incross-linking volved in substrate recognition. Our results support this conWM clusion and provide further evidence for at least two areas on Peptide-4 24 the enzyme that interact with the glutamine substrate. CousPeptide-7 13 sons et al. (1992) reported that theglutamine side chain must Peptide-7a 1318 be solvent exposed or in a portion of the protein that is Peptide-7b 2471 Peptide-7c 2306 flexible. The presence of charged amino acid residues on either 'ICw, peptide concentration required to produce 50%inhibition of side of the reactive glutamine discouraged reactivity. Results guinea pig liver tissue transglutaminase activity. from our study suggest that there may be a more complex set

21292


of interactions between the transglutaminases and theirsubstrates. We interpret our studies tosuggest there are homologous sequences outside the active site pocket for the transglutaminases that bind to the substratefacilitating a productive catalytic eventin the active site pocket. We plan additional studies to define the biochemical nature of this substrate recognition domain using recombinant factor XIIIa. Since x-ray crystallography data to model the three-dimensional structure of factor XIIIa are notavailable, we can only speculate on the relationship between the substrate recognition domain and the active site pocket. The domains represented by peptides-4 and -7 may be located in close proximity to the active site pocket or may promote the proper orientation of the active site pocket with the substrate. In conclusion, using synthetic peptides derived from factor XIIIa, we have identified two regions (Asn''-Asps7 and AsplgOPhe230)of factor XIIIa that are involved in protein substrate recognition. Thesesubstrate recognition sitesaredistinct from those involved in factor XIIIa binding to fibrin. These inhibitors could prove useful in defining the action of transglutaminases in a wide variety of biochemical and cellular processes. Acknowledgments-We thank Drs. Yusuf A. Hannun, William H. Kane, Thomas L. Ortel, Jane Richardson, David Richardson, and Guy S. Salvesen for their helpful comments, Drs. Joanne M. Hettasch and Thung-Shenq Lai for carefully reviewing the manuscript, and Tammy L. Moser for her excellent technical assistance.

Fickenscher, K., Aab, A. & Stuber, W. (1991) Thromb. Haemostasis 66,535540 Folk, J. E. & Finlayson, J. S. (1977) Adu. Protein Chem. 3 1 , 1-133 Francis, R. R., McDonagh, J. & Mann, K. G. (1986) J. Bwl. Chem. 261,97879792 Fukue, H., Anderson, K., McPhedran, P., Clyne, L. & McDonagh, J. (1992) Blood 79.64-74 Gentile, V., Saydak, M., Chiocca, E. A,, Akande, N., Birckbichler, P. J., Lee, K. N., Stein, J. P. & Davies, P. J. A. (1991) J. BioL Chem. 266,478-483 Gorman, J. J. & Folk, J. E. (1973) J. Biol. Chem. 248,1301-1306 Gorman, J. J. & Folk, J. E. (1980) J. Biol. Chem. 266,419-427 Greenberg, C. S., Achyuthan, K. E., Borowitz, M. J. & Shuman, M. A. (1987a) Blood 70,702-709 Greenberg, C. S., Dobson, J. V. & Miraglia, C. C. (1987b) Blood 66,1028-1034 Greenberg, C. S., Enghild, J. J., Mary, A,, Dobson, J. V. & Achyuthan, K. E. (1988a) Biochem. J. 2 6 6 , 1013-1019 Greenberg, C. S., Achyuthan, K. E., Rajagopalan, S. & Pizzo, S. V. (1988b) Arch. Biochem. Biophys. 262,142-148 Greenberg, C. S., Birckbichler, P. J. & Rice, R. H. (1991) FASEB J. 6,30711n77

H g p n c h , P. D., Jr. & Doolittle, R. F. (1983) Biochemistry 2 2 , 2049-2055 Hornyak, T. J. & Shafer, J. A. (1992) Biochemistry 3 1 , 4 2 3 4 2 9 Ichinose, A., McMullen, B. A., Fujikawa, K. & Davie, E. W. (1986) Biochemistry 26,4633-4638 Ichinose, A,, Bottenus, R. E. & Davie, E. W. (1990) J. Biol. Chem. 266,1341113414

Ik&i,-K., Nasu, T., Yokota, H., Tsuchiya, Y., Sasaki, R. & Chiba, H. (1988) Biochemistry 27,2898-2905 Korsgren, C. & Cohen, C. M. (1991) Proc. Natl. Acad. Sci. U.S. A. 8 8 , 48404844 Lai, T-S., Santiago, M. A., Achyuthan, K. E. & Greenberg, C. S. (1991) Blood 7 8 , Suppl. 1,391a (Abstr. 1554) Lee, K. N., Maxwell, M. D., Patterson, M. K., Jr., Birckbichler, P. J. & Conway, E. (1992) Biochim. Biophys. Acta 1136.12-16 Lorand, L. & Konishi, K. (1964) Arch. Biochem. Biophys. 106,58-67 Lorand, L., Losowsky, M. S. & Miloszewski, K. J. M. (1980) Prog. Haemostasis Thromb. 6,245-290 Martinez, J., Rich, E. & Barsigian, C. (1989) J. Biol. Chem. 264,20502-20508 Marv. A.. Achvuthan. K. E. & Greenbere. C. S. (1988) . . Arch. Biochem. Bwuhvs. 2 6 1 , i12-151 ' McDonagh, J. A. (1987) in Hemostasis & Thrombosis (Colman, R. W., Hirsh, J.. Marder. V. J. & Salzman, E.W.. eds) PP. 289-300, J. B. Lippincott, .. Philadelphia McDonaeh. R. P.. McDonaeh. J.. & Petersen. T.E., Thoaersen. H. C.. Skorstengaird, K., Sottrup-Jeksen,'L., Magnusbn, S.,. Dell,-A. & M o m s , H. R. (1981) FEBS Lett. 127,174-178 Mesters. R. M., Houahten. R.A. & Griffin, J. H. (1991) J. Biol. Chem. 2 6 6 , . 24514124519. Miraglia, C. C. & Greenberg, C. S. (1985) Anal. Biochem. 144,165-171 Mosher, D. F.(1975) J.Biol. Chem. 260,6614-6621 Mosher, D. F., Schad, P. E. & Kleinman, H. K. (1979) J. Clin. Invest. 64,781-

_ _

I ,

"

REFERENCES Achyuthan, K. E., Dobson, J. V. & Greenberg, C. S. (1986) Biochim. Biophys. Acta 872,261-268 Achyuthan, K. E., Mary, A. & Greenberg, C. S. (1988) J. Biol. Chem. 2 6 3 , 1429614301 Aeschlimann, D. & Paulsson, M. (1991) J. Biol. Chem. 266,15308-15317 Altieri, D. C., Etingin, 0. R., Fair, D. S., Brunck, T. K., Geltosky, J. E., Hajjar, D. P. & Edgington, T. S. (1991) Science 264,1200-1202 Baglia, F. A,, Jameson, B. A. & Walsh, P.N. (1991) J. Biol. Chem. 266,2419024197 Bale, M. D. & Mosher, D. F. (1986) Biochemistry 26,5667-5673 Binnie, C. G. & Lord, S. T. (1991) Thromb. Haemostasis 66,165-168 Bockenstedt, P., McDonagh, J. & Handin, R. I. (1986) J. Clin. Invest. 78,551556 Chattopadhyay, A. & Fair, D. S. (1989) J. Bwl. Chem. 264,11035-11043 Chattopadhyay, A., James, H. L. & Fair, D. S. (1992) J. Biol. Chem. 267, 12323-12329 Chung, S. I. (1972) Ann. N. Y. Acad. Sci. 202,240-255 Connellan, J. M., Chung, S. I., Whetzel, N. K., Bradley, L. M. & Folk, J. E. (1971) J. Biol. Chern. 246,1093-1098 Cottrell, B. A., Strong, D.D., Watt, K. W. K. & Doolittle, R. F. (1979) Biochemistry 18,5405-5412 Coussons, P. J., Price, N.C., Kelly, S. M., Smith, B. & Sawyer, L. (1992) Biochem. J. 282,929-930

7Q7

M & e k L., Pol ar, J & Fesus, L. (1985) Clin. Chern. 31,35-40 Phillips M. A., &wart, B. E., gin, Q., Chakravarty, R., Floyd, E. E., Jetten, A. M.'& Rice, R. H. (1990) P m . Natl. Acad. Sci. U.S. A. 97,9333-9337 Pisano J. J., Finlayson, J. S. & Pe n, M. P. (1968),Science1 6 0 , 892-893 Purves: L., Purves, M. & Brandt, p ( 1 9 8 7 ) Biochemwtry 26,4640-4646 Rosenberg, J. B., Newman, P. J., Mosesson, M. W., Guillin, M-C. & Amrani, D. L. (1993) Thromb. Haemostasis 69,217-220 Sakata, Y. & Aoki, N. (1980) J. Clin. Invest. 66,290-297 Sane D. C. Moser, T. L., Pippen, A. M. M., Parker, C. J., Achyuthan, K. E. & Grkenbeig, C. S. (1988) Biochem. Biophys. Res. Commun. 1 6 7 , 115-120 Segel(l975) Enzyme Ktnetlcs, pp. 203-208, John Wlley & Sons, New York Shainoff, J. R., Urbanic, D. A. & Dibello, P. M. (1991) J. Biol. Chern. 2 6 6 , 6429437 Slaughter, T. F., Achyuthan, K. E.,Lai, T.3. & Greenberg, C. S. (1992) Anal. Bwchem. 2 0 6 , 166-171