Inr. J. Biochem.Vol. 24, No. 1, pp. 91-97, 1992
M)20-711X/92$5.00+ 0.00 Copyright 0 1991 Pergamon Press plc
Printed in Great Britain. All rights reserved
PURIFICATION AND PROPERTIES OF FACTOR XIII FROM HUMAN PLACENTA CORINNEDE BACKER-R•YER,FATOUMATA TRAORRand JEAN-CLAUDEMJWNIER* Laboratoire de Chimie Biologique, Institut National Agronomique Paris-Grignon, Centre de Grignon, 78850 Thiverval-G&on,
France [Tel. (1) 30-54-45-101
(Received 21 February 1991) Abstract-l. FXIII was isolated and purified over 4000 fold from human placenta to apparent electrophoretic homogeneity by a new procedure including ethanol precipitation. DEAE-Cellulose, molecular sieving on Sephacryl S-300 and Phenyl-Sepharose chromatography. 2. Its p1 was about 5.1. Under appropriate conditions, the incubation of FXIII in the presence of thrombin did not lead to inactivation cut in the polypeptidic chain. 3. FXIII was also activated by CaCI, and, in a lesser extent, by other divalent cations like SrCl,, BaCl, or MgCl,. 4. The binding of calcium to FXIII exhibited a negative cooperativity. 5. The activity-pH curve of the calcium-activated enzyme did not appear very different from that of the thrombin-activated enzyme.
NADH were obtained from Boeringer, casein and bovine plasma thrombin from Sigma chemical corporation. All other reagents were of analytical grade.
INTRODUCTION Blood coagulation FXIII is a transglutaminase that occurs as a zymogen in plasma, platelets, and placenta (Bohn, 1972). Placental and platelets enzymes are identical and composed of two identical 80 kD polypeptides designed a-chains (Bohn, 1978). The activation of enzyme to FXIIIa occurs through a thrombin cleavage of an Arg,,-Gly,, peptide near the amino-terminus of the a-chains (Takagi et al., 1974; Takahashi et al., 1986; Ichinose et al., 1988).
Measurement of protein concentration Protein concentration was determined by the method of Bradford (1976), with bovine serum albumin (Sigma highest grade) as standard. FXIII activation Activation of FXIII was routinely carried out by incubating enzyme (260 nM) with 160 NIH U/ml of thrombin in a 100 mM Tris-HCl buffer pH 7.5 at 37°C for 10 min (see results). Under these conditions, thrombin hydrolyses the pro-enzyme at the level of the activation site. No cut appeared elsewhere in the polypeptidic chain by electrophoresis (Fig. 2). Activation of FXIII by divalent cations was performed by incubating enzyme (260 nM) with various concentrations of cation for 10 min in a 100 mM Tris-HCl buffer, pH 7.5 at 37°C.
Like all the transglutaminases, FXIIIa catalyses the formation of c (y -glutamyl)Iysine bonds between glutaminyl and lysyl residues (Pisano et al., 1968; Lorand et al., 1968). In viva, plasma FXIIIa crosslinks the fibrin monomers leading to the formation of the blood clot (Lorand et al., 1981) and placental FXIIIa is responsible for the retention of placenta (Bohn, 1978). To our knowledge, one method only has been published concerning the purification of placental FXIII (Bohn et al., 1971). The goal of this paper is to present another method of purification of FXIII from human placenta, giving more reproducible results. This procedure is also more convenient to perform. We report also on the optimum conditions in vitro of activation of FXIII by thrombin and
by divalent cations. Some properties activated enzyme are also described.
Kinetic measurement The transfer activity of FXIIIa was measured by the formation of ammonia. The measurement was carried out by coupling the FXIIIa reaction to the reaction catalysed by glutamate dehydrogenase. The assay was performed in a 0.1 M Tris-HCl buffer DH 7.8 containina 7 mM a-Ketoglutarat, 0.4mM NADH, 3mM CaCl, 2mM DTT, 10 mg/ml casein, 0.2mg/ml glutamate dehydrogenase. The presence of DTT in the assay medium was not obligatory. It can be replaced by ascorbic acid, glutathione or cysteine. The reaction rates were respectively equal to 91, 95, 90% with respect to that in the presence of DTT. Without any reductant the reaction rate was still 85% with respect to DTT. A 1-5 Perkin-Elmer spectrophotometer was used to monitor the reaction which was linear for 20min at least with 6 nM FXIIIa. One unit of enzymatic activity was defined as the amount of enzyme which catalyzed the formation of 1 nmol product/min at 30°C.
of the Ca2+-
MATERIALSAND METHOLB DEAE-cellulose (23s) was purchased from Serva, Phenylsepharose CL-4B and Sephacryl S-300 superfine from Pharmacia. Glutamate dehydrogenase, a-Ketoglutarat, *To whom all correspondence should be addressed. Abbreviations: IEF, isoelectric focusing; CPC, N-cetylpyridiniumchlorid; FXIII, Factor XIII zymogen; FXIIIa, activated FXIII. 91
C~RINNE DE BACKER-R•
92
no
YER et al.
fraction
Fig. 1. ARinity chromatography of FXIII. Phenyl-Sepharose column (1 x 20 cm) was equilibrated in buffer 1. Elution was performed with a linear (&SO% (v/v) ethylene-glycol in buffer 2, at a flow rate of 60 ml/hr. The volume of fractions was I ml. FXIII activity was monitored as described in text. Polyacryl~jde gel eleccropho?esis The purity of enzyme was checked by PhastSystem electrophoresis (Pharmacia) with PhastGel Gradient 8-25 (continuous 8-25% gradient polyacrylamide gel) under denaturing conditions, and by PhastSystem isoelectric focusing with PhastGel IEF 3-9 (gel covering the pH range 3-P). pH study To determine the conditions of the maximum rate I’,,, at each pH, a kinetic study was carried out with casein as acyl donor and acceptor at a fixed CaCl, concentration (3 mM). The effect of pH on the stability of the enzyme was studied by incubating the enzyme at various pH for 20 min at 30°C. The enzyme activity was determined at the pH optimum OtH 8). Titrationof FXIII by CaCl, Calcium binding to FXIII decreased the intrinsic fluorescence of the protein. Titration of the protein by the ligand was constructed from fluorescence decrease upon addition of CaCl,. The fluorescence light was monitored at 338 nm (excitation at 290 nm). FXIII concentration was 0.4 PM in a 100 mM TEA buffer, pH 7.5 at 25°C. We used a Jobin Yvon JY3-D spectrofluorometer. RESULTS
PuriJcation of FXIII
All the operations were performed at 4°C. Four thawed human placentas were crushed in a Warring-Blendor in the presence of I liter per kg of placenta of 20 mM Tris-HCl buffer pH 7.1 containing 2 mM EDTA, 2 mM benzamidine, 0.1 mM DTT, 6 g/l NaCl, 0.002% (w/v) chlorexidine, and 12% (v/v) ethanol. Cellular debris were discarded by filtration and cent~fugation (27,O~g, 30min). The crude extract obtained in that way was fractionated by ethanol (25%, v/v). After centrifugation (27,000 g, 30min), the pellet was dissolved in the extraction buffer at pH 7.5 without ethanol (buffer 1) and dialysed against this buffer without NaCl (buffer 2). The pellet was then passed over a DEAE-cellulose
column. The column was washed with buffer 2. FXIII was eluted by a linear gradient of NaCl, from 0 to 3.50mM. One major peak of activity was observed, eluted between 100 and 200mM NaCl. Fractions corresponding to this peak were collected, concentrated by addition of 40% ammonium sulphate. After cent~fugation (2O,OOOg,30 min), the precipitate was dissolved in buffer 2 and dialysed against the same buffer. The solution was concentrated to 5 ml by ultrafiltration with a diaflo apparatus and submitted to an exclusion chromatography over a S-300 Sephacryl column. The active fractions were collected, concentrated by ultrafiltration and passed over a Fhenyl-~pharose CL-4B column equilibrated with buffer 1 (Fig. 1). FXIII was eluted by a linear gradient of ethylene-glycol between 30 and 45%. The active fractions were collected, dialysed against buffer 2, concentrated by ultrafiltration and stored at - 20°C in the presence of 30% (v/v) glycerol, without loss of activity for several months. The protein preparation thus obtained is homogeneous to SDSPAGE with a M, equal to 80,~O (Fig. 2). In isoelectric focusing, enzyme appears also homogeneous, but it separates into subbands (Fig. 3). The p1 value is about 5.1. The yield of the purification procedure is shown in Table 1. The results shown in Fig. 4 indicate that an incubation for 10 min with 160 NIH units/ml of thrombin was sufficient to obtain the maximum activity of FXIIIa between 60 and more than 600 nM enzyme concentration. Activation by divalent cations It is known that the plasma1 FXIII is activated at non-physiolo~~ally high con~ntration (> 100 mM) of CaCl, (Credo et al., 1978). In Fig. 5, we show that other divalent cations (SrCl,, BaCl,, M&l,) activated placental FXIII, but to a lesser extent than CaCl,. Results shown in this figure indicate a sigmoidal dependence of activity on cations concentration. For CaCl,, the maximum activity was
Factor XIII from human placenta
Fig. 2. M, and apparent electrophoretic homogeneity of FXIII and FXIIIa. Purified FXIII (30 ng) was submitted to SDS-PAGE (PhastGel 8-Z), before (lane 3) or after (lane 4) activation by 160 NIH units/ml of thrombin, during 10min at 37”C, pH 7.5 (FXIII concentration was 260nM). M, calibration kit [lane 1) and thrombin alone (lane 2) were also subjected to electrophoresis.
93
94
CORINNEDE BACKER-R•YER et al.
Fig. 3. Isoelectric focusing of FXIII. Purified FXIII (170 ng) was submitted to PhastGel IEF 3--9 (lane 2) Lane 1: pl calibration kit.
Factor XIII from human placenta
95
Table 1. Purification of FXIII from human olacente SpCifiC
Step Crude extract Ethanol precipitation DEAE-
Activity (UilitS)
Protein (mg)
19,300
82,440
8280
4320
o‘r%&=, Precipitation Sephacryl s-300
Recovery W)
0.23
1.92
100
I
43
8.4
I5
15.5
65
1840
68
27
9.5
117
800
12
65.3
4.1
284
2.6
3913
PhenylScpharese 540 0.6 900 Activity of PXIIIa was monitored as described in text,
reached at about 1SOmM and for other cations at higher concentrations. The time dependence of activation of FXIII by this cations at their optimum concentration indicated that the activity reached a plateau within 10 mm and that the loss of activity at 37°C with respect to time was weak (data not shown). Calcium is also essential to activity of FXIIIa. But stronti~, barium and magnesium can replace calcium in this effect. The kinetics with respect to CaCl,, SrCl,, BaCl,, and MgCI, were michaelian (data not shown). The Km values were 0.067, 0.34, 0.215, and 0.21 mM for CaCI,, SrCl,, BaCl, and MgCI, respectively. pH-Dependence
Purification tkctor
200
3ooo
CdUIO~
activity (uni~/m6)
and stability
The maximum
cross-linking of casein by Gazeactivated FXIII occurred at pH 7.6 (Fig. 6). NO significant difference was observed in this profile when FXIII activated by thrombin was used. The results shown in Table 2 indicate that the differences between the observed pK values (defined as pH corresponding to apparent V,,,/2) were below 3. Therefore, we have calculated the real pK values according to Segel (1975). These values are listed in Table 2. The pH stability profile shows that the decrease of activity of enzyme was not due to a
pH denaturation: and 10.
FXIIIa
was stable between pH 5
Titration of FXHI by calcium
Figure 7(A) shows the variation of the saturation fraction against calcium concentration. The results with the Scatchard representation [Fig. 7(B)] present a negative cooperativity in the binding. DISCUSSION
FXIII has been purified to apparent homogeneity from human placenta according to a new procedure. This method involves classical steps as ethanol and ammonium sulphate precipitations, and chromatographic separations (three steps: DEAE-Cellulose, S-300 Sephacryl, and Phenyl-Sepharose). CPC and Rivanoi (Bohn ez al., 1971; Bohn, 1972), which were not easy to employ, were not used. This procedure gives reproducible results. In isoelectric focusing, placental FXIII separates into subbands. An explanation may come from the fact that the primary structure of placental FXIII by direct amino-acid analysis, as well as c-DNA sequencing, gave clear evidence for microhet~o~neity (Takahashi et al., 1986; Ichinose et al., 1986; Ichinose et al., 1988).
3
200 Thrombin (NIH
4OQ
600
concentration units/ml
20 Incubation
40 time
60 (minf
200 FXIJJ
400
600
concentration
(nM)
)
Fig. 4. FXIII activation by thrombin. (a) FXIII activation as a function of thrombin concentration: FXIII (260 nM) was preincubated in a constant volume at pH 7.5, 37’C.Gin 0.1 M Tris-HCl buffer containing increased thrombin concentrations. FXIfia activity was then tested after 10 min of incubation. (b) FXIII activation as a function of incubation time: FXIII was preincubated in the same conditions as (a), with 160 NIH units/ml (a) or 80 NIH units/ml (m) of thrombin. FXIIIa activity was tested at various times of incubation. (c) FXIII activation as a function of FXIII concentration: FXIII was incubated at various concentrations in the same conditions as (a) and (b), but thrombin concentration was 160 NIH units/ml. During the kinetics, FXIII ~n~ntration was equal to 6nM.
96
CORINNE
DE BACKER-R• YER et al. Table 2. pK and pH optimum values of FXIIIa activated by thrombin or C&I, Ca*+-activated FXIII
Thrombin-activated FXIII
Parameter
6.1 a.9 1.5 6.3 f 0.05 9.0 f 0.07 2.8 3.1 + 0.07
6.3 8.95 1.6 6.45 + 0.07 9.0 + 0.08 3.0 3.38 f 0.08
PK, obr. PK,,. PH., PK,, PKW Vrnqy. (S-I) Vmlhsor.W’)
The observed pK (pK, ), pH optimum (pH,,) and the optimum apparent rate values (V, sw ) were determined from experimental results given in Fig. 6. The values of theoretical I’, , pK,, , and pK,, were obtained by fitting the experimental results to equation: Vmspp= V, x [H+ I x &s, /((ks,
x [H+ I)
+ IB+12+0% x KY&) 0
200
100
300
CXClzl (mM I Fig. 5. FXIII activation by divalent cations. FXIII (260 nM) was preincubated in a constant volume at pH 7.5, 37”C, in a 100 mM Tris-HCl buffer containing increased CaCl, (a), SrCl, (O), BaCl, (0) or MgCl, (A) concentrations. FXIIIa activity was then tested after 10 min of incubation.
When thrombin concentration was increased (above 120 NIH units/ml) in the activation medium, the constancy of activity up to 630 NIH units/ml indicates that during the 10min of incubation no inactivation cut occurs in the peptidic chain. This finding was checked by electrophoretic analysis of FXIII treated by thrombin. The pattern shows clearly only one band at 76,000 corresponding to the FXIIIa. However, under different experimental conditions, small peptides appear leading to the inactivation of FXIIIa (Schwartz et al., 1973; Takahashi et al., 1986; Mary et al., 1988). Strontium, barium and magnesium (in decreasing order of action) were capable to activate FXIII at high
according to Segel(l975). pK,, and pK,, of the enzyme-substrate complex.
refer to the true pK
concentration (> 100 mM). Strontium was nearly as efficient as calcium in this non-proteolytic effect. The counter-ion (Cl-) did not play any role in this activation since this anion was used with all the cations, and since LiCl had no effect on the inactive enzyme (data not shown). The maximum activity of the enzyme activated by calcium was of the same order of magnitude as the enzyme activated by thrombin.
l.Or
0
(Al
I 0.1
I
I
I
0.2
0.3
0.4
kalciuml
(M-r)
60
PH
Fig. 6. pH dependence and stability of FXIIIa. The curves (0, n) are theoretical and were obtained after fitting the experimental data according to Segel (1975). The enzyme (260 nM) was activated 10 min at pH 7.5,37”C, by thrombin 160 NIH units/ml (m) or by 150mM CaCl, (0). Enzyme concentration was 6nM during the kinetics. Glutamate dehydrogenase and casein were saturating at each pH. Buffer was MES 0.05 M below pH 7, Tris-HCl 0.1 M from pH 7 to 8.8, and GlycineNaOH 0.1 M over pH 8.8. The enzyme stability (0) was also drawn after activation by Cat&.
0
0.2
0.4
0.6
0.8
1.0
Fig. 7. Calcium binding isotherm to FXIII. (A) Variation of the saturation fraction (I’) against calcium concentration (B). Scatchard plot of the data shown in (A). Intrinsic fluorescence was measured at 338 nm.
Factor XIII from human placenta
The binding isotherm of calcium showed a negative cooperativity. But since this isotherm has been drawn using a dependent function, fluorescence, and since the binding is weak, we have no way to check whether this negative cooperativity can be due to a difference in the quantum yield of the sites (two at least) or to interactions between them. For the same reasons, we cannot determine the number of calcium sites on the protein. In contrast, no kinetic cooperativity was observed with respect to calcium and other divalent cations. It means, if interactions between binding sites exist, that they are not transmitted to the catalytic sites. Since the differences between the observed pK values in the pH activity profile were below 3, we have calculated the real pK values in accordance to Segel(l975). However, since this difference was quite near 3, the apparent Vm was 90% of the theoretical value, and the pK theoretical values were not very different from the observed values. This pH behaviour is at first view similar to that of the thrombinactivated enzyme, casein being also the substrate. But a significant difference appeared between the pKEs, values. It means that the microenvironment of the amino acid residue is slightly different in the two activated forms. Although it is difficult to assign experimental pK value to the reactive group of an amino acid, one of the pK values certainly refers to cysteine residue, which has been described elsewhere as being at the active center of guinea pig liver transglutaminase (Folk et al., 1966), plasma1 FXIII (Curtis et al., 1973) and placental FXIII (Takahashi et al., 1986). REFERENCES
Bohn H. and Schwick H. G. (1971) Isolierung und characterisierung eines fibrinstabilisierenden factors aus menschlichen plazenten. Arzneim. Forsch. 21, 1432-1439. Bohn H. (1972) Comparative studies on the fibrin-stabilizing factors from human plasma, platelets and placenta. Ann. N.Y. Acad. Sci. 202, 256-272.
Bohn H. (1978) The human fibrin-stabilizing factors. Molec. cell. Biochem. 20, 67-75.
91
Bradford M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anafyt. Biochem. 72, 248-254.
Credo R. B., Curtis C. G. and Lorand L. (1978) Ca2+related regulatory function of fibrinogen. Proc. natn. Acad. Sci., U.S.A.
75, 42344237.
Curtis C. G., Srenberg P., Chou C.-H. J., Gray A., Brown K. L. and Lorand L. (1973) Titration and subunit localisation of active center cysteine in fibrinoligase. Biochem. biophys. Res. Cotnmun. 52, 51-56. Folk J. E. and Cole P. M. (1966) Identification of a functional cysteine essential for the activity of guinea-pig liver transglutaminase. J. biol. Chem. 241, 3238-3240. Ichinose A., Hendrickson L. E., Fujikawa K. and Davie E. W. (1986) Amino acid sequence of the a subunit of human factor XIII. Biochemistry 25, 6900-6906. Ichinose A. and Davie E. W. (1988) Characterization of the gene for the a subunit of human factor XIII (plasma transglutaminase), a blood coagulation factor. Proc. natn. Acad. Sci. U.S.A. 85, 5829-5833.
Lorand L., Downey J., Gotoh T., Jacobsen A. and Tokura S. (1968) The transpeptidase system which crosslinks fibrin by y-glutamyl-c-lysine bonds. Biochem. biophys. Res. Commun. 31, 222-230.
Lorand L., Credo R. B. and Janus T. J. (1981) Factor XIII (stabilizing factor). Meth. Enzymol. 80, 333-341. Mary A., Achyuthan K. E. and Greenberg C. S. (1988) The binding of divalent metal ions to platelet factor XIII modulates its proteolysis by trypsin and thrombin. Archs Biochem. Biophys. 261, 112-121. Pisano J. J., Finlayson J. S. & Peyton M. P. (1968) Cross-link in fibrin polymerized by factor XIII: 6-(y -glutamyl) lysine. Science, N. Y. 160, 892493. Schwartz M. L., Pizzo S. V., Hill R. L. and MacKee P. A. (1973) The subunit structures of human plasma and platelet factor XIII (fibrin-stabilizing factor). J. biol. Chem. 248, 13951407. &gel I. H. (1975) Enzyme Kinetics, pp. 914-915. Wiley, New York. Takagi T. and Doolitle R. F. (1974) Amino acid sequence studies on factor XIII and the peptide released during its activation by thrombin. Biochemistry 13, 750-756.
Takahashi N., Takahashi Y. and Putnam F. W. (1986) Primary structure of blood coagulation factor XIIIa (fibrinoligase, transglutaminase) from human placenta. Proc. natn. Acad. Sri. U.S.A. 83, 8019-8023.
M)20-711X/92$5.00+ 0.00 Copyright 0 1991 Pergamon Press plc
Printed in Great Britain. All rights reserved
PURIFICATION AND PROPERTIES OF FACTOR XIII FROM HUMAN PLACENTA CORINNEDE BACKER-R•YER,FATOUMATA TRAORRand JEAN-CLAUDEMJWNIER* Laboratoire de Chimie Biologique, Institut National Agronomique Paris-Grignon, Centre de Grignon, 78850 Thiverval-G&on,
France [Tel. (1) 30-54-45-101
(Received 21 February 1991) Abstract-l. FXIII was isolated and purified over 4000 fold from human placenta to apparent electrophoretic homogeneity by a new procedure including ethanol precipitation. DEAE-Cellulose, molecular sieving on Sephacryl S-300 and Phenyl-Sepharose chromatography. 2. Its p1 was about 5.1. Under appropriate conditions, the incubation of FXIII in the presence of thrombin did not lead to inactivation cut in the polypeptidic chain. 3. FXIII was also activated by CaCI, and, in a lesser extent, by other divalent cations like SrCl,, BaCl, or MgCl,. 4. The binding of calcium to FXIII exhibited a negative cooperativity. 5. The activity-pH curve of the calcium-activated enzyme did not appear very different from that of the thrombin-activated enzyme.
NADH were obtained from Boeringer, casein and bovine plasma thrombin from Sigma chemical corporation. All other reagents were of analytical grade.
INTRODUCTION Blood coagulation FXIII is a transglutaminase that occurs as a zymogen in plasma, platelets, and placenta (Bohn, 1972). Placental and platelets enzymes are identical and composed of two identical 80 kD polypeptides designed a-chains (Bohn, 1978). The activation of enzyme to FXIIIa occurs through a thrombin cleavage of an Arg,,-Gly,, peptide near the amino-terminus of the a-chains (Takagi et al., 1974; Takahashi et al., 1986; Ichinose et al., 1988).
Measurement of protein concentration Protein concentration was determined by the method of Bradford (1976), with bovine serum albumin (Sigma highest grade) as standard. FXIII activation Activation of FXIII was routinely carried out by incubating enzyme (260 nM) with 160 NIH U/ml of thrombin in a 100 mM Tris-HCl buffer pH 7.5 at 37°C for 10 min (see results). Under these conditions, thrombin hydrolyses the pro-enzyme at the level of the activation site. No cut appeared elsewhere in the polypeptidic chain by electrophoresis (Fig. 2). Activation of FXIII by divalent cations was performed by incubating enzyme (260 nM) with various concentrations of cation for 10 min in a 100 mM Tris-HCl buffer, pH 7.5 at 37°C.
Like all the transglutaminases, FXIIIa catalyses the formation of c (y -glutamyl)Iysine bonds between glutaminyl and lysyl residues (Pisano et al., 1968; Lorand et al., 1968). In viva, plasma FXIIIa crosslinks the fibrin monomers leading to the formation of the blood clot (Lorand et al., 1981) and placental FXIIIa is responsible for the retention of placenta (Bohn, 1978). To our knowledge, one method only has been published concerning the purification of placental FXIII (Bohn et al., 1971). The goal of this paper is to present another method of purification of FXIII from human placenta, giving more reproducible results. This procedure is also more convenient to perform. We report also on the optimum conditions in vitro of activation of FXIII by thrombin and
by divalent cations. Some properties activated enzyme are also described.
Kinetic measurement The transfer activity of FXIIIa was measured by the formation of ammonia. The measurement was carried out by coupling the FXIIIa reaction to the reaction catalysed by glutamate dehydrogenase. The assay was performed in a 0.1 M Tris-HCl buffer DH 7.8 containina 7 mM a-Ketoglutarat, 0.4mM NADH, 3mM CaCl, 2mM DTT, 10 mg/ml casein, 0.2mg/ml glutamate dehydrogenase. The presence of DTT in the assay medium was not obligatory. It can be replaced by ascorbic acid, glutathione or cysteine. The reaction rates were respectively equal to 91, 95, 90% with respect to that in the presence of DTT. Without any reductant the reaction rate was still 85% with respect to DTT. A 1-5 Perkin-Elmer spectrophotometer was used to monitor the reaction which was linear for 20min at least with 6 nM FXIIIa. One unit of enzymatic activity was defined as the amount of enzyme which catalyzed the formation of 1 nmol product/min at 30°C.
of the Ca2+-
MATERIALSAND METHOLB DEAE-cellulose (23s) was purchased from Serva, Phenylsepharose CL-4B and Sephacryl S-300 superfine from Pharmacia. Glutamate dehydrogenase, a-Ketoglutarat, *To whom all correspondence should be addressed. Abbreviations: IEF, isoelectric focusing; CPC, N-cetylpyridiniumchlorid; FXIII, Factor XIII zymogen; FXIIIa, activated FXIII. 91
C~RINNE DE BACKER-R•
92
no
YER et al.
fraction
Fig. 1. ARinity chromatography of FXIII. Phenyl-Sepharose column (1 x 20 cm) was equilibrated in buffer 1. Elution was performed with a linear (&SO% (v/v) ethylene-glycol in buffer 2, at a flow rate of 60 ml/hr. The volume of fractions was I ml. FXIII activity was monitored as described in text. Polyacryl~jde gel eleccropho?esis The purity of enzyme was checked by PhastSystem electrophoresis (Pharmacia) with PhastGel Gradient 8-25 (continuous 8-25% gradient polyacrylamide gel) under denaturing conditions, and by PhastSystem isoelectric focusing with PhastGel IEF 3-9 (gel covering the pH range 3-P). pH study To determine the conditions of the maximum rate I’,,, at each pH, a kinetic study was carried out with casein as acyl donor and acceptor at a fixed CaCl, concentration (3 mM). The effect of pH on the stability of the enzyme was studied by incubating the enzyme at various pH for 20 min at 30°C. The enzyme activity was determined at the pH optimum OtH 8). Titrationof FXIII by CaCl, Calcium binding to FXIII decreased the intrinsic fluorescence of the protein. Titration of the protein by the ligand was constructed from fluorescence decrease upon addition of CaCl,. The fluorescence light was monitored at 338 nm (excitation at 290 nm). FXIII concentration was 0.4 PM in a 100 mM TEA buffer, pH 7.5 at 25°C. We used a Jobin Yvon JY3-D spectrofluorometer. RESULTS
PuriJcation of FXIII
All the operations were performed at 4°C. Four thawed human placentas were crushed in a Warring-Blendor in the presence of I liter per kg of placenta of 20 mM Tris-HCl buffer pH 7.1 containing 2 mM EDTA, 2 mM benzamidine, 0.1 mM DTT, 6 g/l NaCl, 0.002% (w/v) chlorexidine, and 12% (v/v) ethanol. Cellular debris were discarded by filtration and cent~fugation (27,O~g, 30min). The crude extract obtained in that way was fractionated by ethanol (25%, v/v). After centrifugation (27,000 g, 30min), the pellet was dissolved in the extraction buffer at pH 7.5 without ethanol (buffer 1) and dialysed against this buffer without NaCl (buffer 2). The pellet was then passed over a DEAE-cellulose
column. The column was washed with buffer 2. FXIII was eluted by a linear gradient of NaCl, from 0 to 3.50mM. One major peak of activity was observed, eluted between 100 and 200mM NaCl. Fractions corresponding to this peak were collected, concentrated by addition of 40% ammonium sulphate. After cent~fugation (2O,OOOg,30 min), the precipitate was dissolved in buffer 2 and dialysed against the same buffer. The solution was concentrated to 5 ml by ultrafiltration with a diaflo apparatus and submitted to an exclusion chromatography over a S-300 Sephacryl column. The active fractions were collected, concentrated by ultrafiltration and passed over a Fhenyl-~pharose CL-4B column equilibrated with buffer 1 (Fig. 1). FXIII was eluted by a linear gradient of ethylene-glycol between 30 and 45%. The active fractions were collected, dialysed against buffer 2, concentrated by ultrafiltration and stored at - 20°C in the presence of 30% (v/v) glycerol, without loss of activity for several months. The protein preparation thus obtained is homogeneous to SDSPAGE with a M, equal to 80,~O (Fig. 2). In isoelectric focusing, enzyme appears also homogeneous, but it separates into subbands (Fig. 3). The p1 value is about 5.1. The yield of the purification procedure is shown in Table 1. The results shown in Fig. 4 indicate that an incubation for 10 min with 160 NIH units/ml of thrombin was sufficient to obtain the maximum activity of FXIIIa between 60 and more than 600 nM enzyme concentration. Activation by divalent cations It is known that the plasma1 FXIII is activated at non-physiolo~~ally high con~ntration (> 100 mM) of CaCl, (Credo et al., 1978). In Fig. 5, we show that other divalent cations (SrCl,, BaCl,, M&l,) activated placental FXIII, but to a lesser extent than CaCl,. Results shown in this figure indicate a sigmoidal dependence of activity on cations concentration. For CaCl,, the maximum activity was
Factor XIII from human placenta
Fig. 2. M, and apparent electrophoretic homogeneity of FXIII and FXIIIa. Purified FXIII (30 ng) was submitted to SDS-PAGE (PhastGel 8-Z), before (lane 3) or after (lane 4) activation by 160 NIH units/ml of thrombin, during 10min at 37”C, pH 7.5 (FXIII concentration was 260nM). M, calibration kit [lane 1) and thrombin alone (lane 2) were also subjected to electrophoresis.
93
94
CORINNEDE BACKER-R•YER et al.
Fig. 3. Isoelectric focusing of FXIII. Purified FXIII (170 ng) was submitted to PhastGel IEF 3--9 (lane 2) Lane 1: pl calibration kit.
Factor XIII from human placenta
95
Table 1. Purification of FXIII from human olacente SpCifiC
Step Crude extract Ethanol precipitation DEAE-
Activity (UilitS)
Protein (mg)
19,300
82,440
8280
4320
o‘r%&=, Precipitation Sephacryl s-300
Recovery W)
0.23
1.92
100
I
43
8.4
I5
15.5
65
1840
68
27
9.5
117
800
12
65.3
4.1
284
2.6
3913
PhenylScpharese 540 0.6 900 Activity of PXIIIa was monitored as described in text,
reached at about 1SOmM and for other cations at higher concentrations. The time dependence of activation of FXIII by this cations at their optimum concentration indicated that the activity reached a plateau within 10 mm and that the loss of activity at 37°C with respect to time was weak (data not shown). Calcium is also essential to activity of FXIIIa. But stronti~, barium and magnesium can replace calcium in this effect. The kinetics with respect to CaCl,, SrCl,, BaCl,, and MgCI, were michaelian (data not shown). The Km values were 0.067, 0.34, 0.215, and 0.21 mM for CaCI,, SrCl,, BaCl, and MgCI, respectively. pH-Dependence
Purification tkctor
200
3ooo
CdUIO~
activity (uni~/m6)
and stability
The maximum
cross-linking of casein by Gazeactivated FXIII occurred at pH 7.6 (Fig. 6). NO significant difference was observed in this profile when FXIII activated by thrombin was used. The results shown in Table 2 indicate that the differences between the observed pK values (defined as pH corresponding to apparent V,,,/2) were below 3. Therefore, we have calculated the real pK values according to Segel (1975). These values are listed in Table 2. The pH stability profile shows that the decrease of activity of enzyme was not due to a
pH denaturation: and 10.
FXIIIa
was stable between pH 5
Titration of FXHI by calcium
Figure 7(A) shows the variation of the saturation fraction against calcium concentration. The results with the Scatchard representation [Fig. 7(B)] present a negative cooperativity in the binding. DISCUSSION
FXIII has been purified to apparent homogeneity from human placenta according to a new procedure. This method involves classical steps as ethanol and ammonium sulphate precipitations, and chromatographic separations (three steps: DEAE-Cellulose, S-300 Sephacryl, and Phenyl-Sepharose). CPC and Rivanoi (Bohn ez al., 1971; Bohn, 1972), which were not easy to employ, were not used. This procedure gives reproducible results. In isoelectric focusing, placental FXIII separates into subbands. An explanation may come from the fact that the primary structure of placental FXIII by direct amino-acid analysis, as well as c-DNA sequencing, gave clear evidence for microhet~o~neity (Takahashi et al., 1986; Ichinose et al., 1986; Ichinose et al., 1988).
3
200 Thrombin (NIH
4OQ
600
concentration units/ml
20 Incubation
40 time
60 (minf
200 FXIJJ
400
600
concentration
(nM)
)
Fig. 4. FXIII activation by thrombin. (a) FXIII activation as a function of thrombin concentration: FXIII (260 nM) was preincubated in a constant volume at pH 7.5, 37’C.Gin 0.1 M Tris-HCl buffer containing increased thrombin concentrations. FXIfia activity was then tested after 10 min of incubation. (b) FXIII activation as a function of incubation time: FXIII was preincubated in the same conditions as (a), with 160 NIH units/ml (a) or 80 NIH units/ml (m) of thrombin. FXIIIa activity was tested at various times of incubation. (c) FXIII activation as a function of FXIII concentration: FXIII was incubated at various concentrations in the same conditions as (a) and (b), but thrombin concentration was 160 NIH units/ml. During the kinetics, FXIII ~n~ntration was equal to 6nM.
96
CORINNE
DE BACKER-R• YER et al. Table 2. pK and pH optimum values of FXIIIa activated by thrombin or C&I, Ca*+-activated FXIII
Thrombin-activated FXIII
Parameter
6.1 a.9 1.5 6.3 f 0.05 9.0 f 0.07 2.8 3.1 + 0.07
6.3 8.95 1.6 6.45 + 0.07 9.0 + 0.08 3.0 3.38 f 0.08
PK, obr. PK,,. PH., PK,, PKW Vrnqy. (S-I) Vmlhsor.W’)
The observed pK (pK, ), pH optimum (pH,,) and the optimum apparent rate values (V, sw ) were determined from experimental results given in Fig. 6. The values of theoretical I’, , pK,, , and pK,, were obtained by fitting the experimental results to equation: Vmspp= V, x [H+ I x &s, /((ks,
x [H+ I)
+ IB+12+0% x KY&) 0
200
100
300
CXClzl (mM I Fig. 5. FXIII activation by divalent cations. FXIII (260 nM) was preincubated in a constant volume at pH 7.5, 37”C, in a 100 mM Tris-HCl buffer containing increased CaCl, (a), SrCl, (O), BaCl, (0) or MgCl, (A) concentrations. FXIIIa activity was then tested after 10 min of incubation.
When thrombin concentration was increased (above 120 NIH units/ml) in the activation medium, the constancy of activity up to 630 NIH units/ml indicates that during the 10min of incubation no inactivation cut occurs in the peptidic chain. This finding was checked by electrophoretic analysis of FXIII treated by thrombin. The pattern shows clearly only one band at 76,000 corresponding to the FXIIIa. However, under different experimental conditions, small peptides appear leading to the inactivation of FXIIIa (Schwartz et al., 1973; Takahashi et al., 1986; Mary et al., 1988). Strontium, barium and magnesium (in decreasing order of action) were capable to activate FXIII at high
according to Segel(l975). pK,, and pK,, of the enzyme-substrate complex.
refer to the true pK
concentration (> 100 mM). Strontium was nearly as efficient as calcium in this non-proteolytic effect. The counter-ion (Cl-) did not play any role in this activation since this anion was used with all the cations, and since LiCl had no effect on the inactive enzyme (data not shown). The maximum activity of the enzyme activated by calcium was of the same order of magnitude as the enzyme activated by thrombin.
l.Or
0
(Al
I 0.1
I
I
I
0.2
0.3
0.4
kalciuml
(M-r)
60
PH
Fig. 6. pH dependence and stability of FXIIIa. The curves (0, n) are theoretical and were obtained after fitting the experimental data according to Segel (1975). The enzyme (260 nM) was activated 10 min at pH 7.5,37”C, by thrombin 160 NIH units/ml (m) or by 150mM CaCl, (0). Enzyme concentration was 6nM during the kinetics. Glutamate dehydrogenase and casein were saturating at each pH. Buffer was MES 0.05 M below pH 7, Tris-HCl 0.1 M from pH 7 to 8.8, and GlycineNaOH 0.1 M over pH 8.8. The enzyme stability (0) was also drawn after activation by Cat&.
0
0.2
0.4
0.6
0.8
1.0
Fig. 7. Calcium binding isotherm to FXIII. (A) Variation of the saturation fraction (I’) against calcium concentration (B). Scatchard plot of the data shown in (A). Intrinsic fluorescence was measured at 338 nm.
Factor XIII from human placenta
The binding isotherm of calcium showed a negative cooperativity. But since this isotherm has been drawn using a dependent function, fluorescence, and since the binding is weak, we have no way to check whether this negative cooperativity can be due to a difference in the quantum yield of the sites (two at least) or to interactions between them. For the same reasons, we cannot determine the number of calcium sites on the protein. In contrast, no kinetic cooperativity was observed with respect to calcium and other divalent cations. It means, if interactions between binding sites exist, that they are not transmitted to the catalytic sites. Since the differences between the observed pK values in the pH activity profile were below 3, we have calculated the real pK values in accordance to Segel(l975). However, since this difference was quite near 3, the apparent Vm was 90% of the theoretical value, and the pK theoretical values were not very different from the observed values. This pH behaviour is at first view similar to that of the thrombinactivated enzyme, casein being also the substrate. But a significant difference appeared between the pKEs, values. It means that the microenvironment of the amino acid residue is slightly different in the two activated forms. Although it is difficult to assign experimental pK value to the reactive group of an amino acid, one of the pK values certainly refers to cysteine residue, which has been described elsewhere as being at the active center of guinea pig liver transglutaminase (Folk et al., 1966), plasma1 FXIII (Curtis et al., 1973) and placental FXIII (Takahashi et al., 1986). REFERENCES
Bohn H. and Schwick H. G. (1971) Isolierung und characterisierung eines fibrinstabilisierenden factors aus menschlichen plazenten. Arzneim. Forsch. 21, 1432-1439. Bohn H. (1972) Comparative studies on the fibrin-stabilizing factors from human plasma, platelets and placenta. Ann. N.Y. Acad. Sci. 202, 256-272.
Bohn H. (1978) The human fibrin-stabilizing factors. Molec. cell. Biochem. 20, 67-75.
91
Bradford M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anafyt. Biochem. 72, 248-254.
Credo R. B., Curtis C. G. and Lorand L. (1978) Ca2+related regulatory function of fibrinogen. Proc. natn. Acad. Sci., U.S.A.
75, 42344237.
Curtis C. G., Srenberg P., Chou C.-H. J., Gray A., Brown K. L. and Lorand L. (1973) Titration and subunit localisation of active center cysteine in fibrinoligase. Biochem. biophys. Res. Cotnmun. 52, 51-56. Folk J. E. and Cole P. M. (1966) Identification of a functional cysteine essential for the activity of guinea-pig liver transglutaminase. J. biol. Chem. 241, 3238-3240. Ichinose A., Hendrickson L. E., Fujikawa K. and Davie E. W. (1986) Amino acid sequence of the a subunit of human factor XIII. Biochemistry 25, 6900-6906. Ichinose A. and Davie E. W. (1988) Characterization of the gene for the a subunit of human factor XIII (plasma transglutaminase), a blood coagulation factor. Proc. natn. Acad. Sci. U.S.A. 85, 5829-5833.
Lorand L., Downey J., Gotoh T., Jacobsen A. and Tokura S. (1968) The transpeptidase system which crosslinks fibrin by y-glutamyl-c-lysine bonds. Biochem. biophys. Res. Commun. 31, 222-230.
Lorand L., Credo R. B. and Janus T. J. (1981) Factor XIII (stabilizing factor). Meth. Enzymol. 80, 333-341. Mary A., Achyuthan K. E. and Greenberg C. S. (1988) The binding of divalent metal ions to platelet factor XIII modulates its proteolysis by trypsin and thrombin. Archs Biochem. Biophys. 261, 112-121. Pisano J. J., Finlayson J. S. & Peyton M. P. (1968) Cross-link in fibrin polymerized by factor XIII: 6-(y -glutamyl) lysine. Science, N. Y. 160, 892493. Schwartz M. L., Pizzo S. V., Hill R. L. and MacKee P. A. (1973) The subunit structures of human plasma and platelet factor XIII (fibrin-stabilizing factor). J. biol. Chem. 248, 13951407. &gel I. H. (1975) Enzyme Kinetics, pp. 914-915. Wiley, New York. Takagi T. and Doolitle R. F. (1974) Amino acid sequence studies on factor XIII and the peptide released during its activation by thrombin. Biochemistry 13, 750-756.
Takahashi N., Takahashi Y. and Putnam F. W. (1986) Primary structure of blood coagulation factor XIIIa (fibrinoligase, transglutaminase) from human placenta. Proc. natn. Acad. Sri. U.S.A. 83, 8019-8023.
