Biochimica et Biophysica Acta 965 (1988) 169-175 169 Elsevier

BBA22921

T h r o m b i n - i n d u c e d f i b r i n o p e p t i d e B r e l e a s e f r o m n o r m a l a n d v a r i a n t f i b r i n o g e n s :

i n f l u e n c e o f i n h i b i t o r s o f f i b r i n p o l y m e r i z a t i o n

W o l f r a m g u f a, A r m i n B e n d e r a, D a v i d A. L a n e b, K l a u s T. P r e i s s n e r a,

E b e r h a r d S e l m a y r a a n d G e r t M i ~ l l e r - B e r g h a u s a

a Clinical Research Unit for Blood Coagulation and Thrombosis of the Max-Planck-Gesellschaft at the Justus-Liebig- Universitiit, Giessen (F R. G. )

and b Department of Haematology, Charing Cross and Westminster Medical School, Hammersmith, London (U.K.)

(Received 17 December 1987)

Key words: Fibrinopeptide B release; Fibrin polymerization kinetics; Fragment D1; Fibrinogen variant

Thrombin preferentially cleaves fibrinopeptides A (FPA) from fibrinogen resulting in the formation of desAA-fibrin from which most of the fibrinopeptides B (FPB) are then released with an enhanced rate. Kinetics of fibrinopeptide release from normal and dysfunctional fibrin0gens were investigated in order to further characterize the mechanism of accelerated FPB release during desAA-fibrin polymerization. Dys- functional fibrinogens London I and Asldord, exhibiting primary polymerization abnormalities (i.e., an abnormality present when all fibrinopeptides have been cleaved), which in the case of fibrinogen London I is believed to be caused by a defect in the D-domain, were shown to exhibit a decreased rate of FPB release compared with normal fibrinogen. While Gly-Pro-Arg-Pro, an inhibitor of fibrin polymerization, was shown to decrease the rate of FPB release from normal fibrinogen by a factor of 5, normal fragment DI, although inhibiting clot formation of normal fibrinogen, did not influence the acceleration of FPB release. On the other hand, the presence of fragment D t did not enhance FPB release from fibrinogen London I, suggesting that interaction of D-domains in functional isolation with desAA-fibrin E-domains is not sufficient to enhance FPB release. Although clot formation was inhibited by the concentrations of fragment D 1 used, the formation of small desAA-fibrin oligomers was hardly affected. Thus, small fibrin polymers, but not desAA-fibrin monomers, act as optimal substrates for the release of FPB by thrombin.

Introduction

The conversion of fibrinogen to a fibrin clot by the action of thrombin involves two different reac- tions: enzymatic cleavage of fibrinopeptides A (FPA) and B (FPB) by thrombin, and polymeriza- tion of fibrin monomers. Fibrinopeptides A and B

Abbreviations: FPA, fibrinopeptides A; FPB, fibrinopeptides B.

Correspondence: G. Miiller-Berghaus, Clinical Research Unit for Blood Coagulation and Thrombosis, MPG, Gaffkystr. 11, D-6300 Giessen, F.R.G.

are released from the fibrinogen molecule simulta- neously, but with different rates [1-3]. Aa-chains are cleaved with the fastest rate in the initial phase of the thrombin-fibrinogen reaction, resulting in the formation of desAA-fibrin, from which most of the FPB are released [4]. Simultaneously to the enzymatic reaction, generated desAA-fibrin starts to polymerize. The formation of polymers results in an enhanced rate of FPB release compared to monomeric desAA-fibrin, as shown by the de- creased rate of FPB release in the presence of urea [5] or the synthetic tetrapeptide Gly-Pro-Arg-Pro [6]. Gly-Pro-Arg-Pro inhibits fibrin polymeriza-

0304-4165/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

170

tion by blocking polymerization sites in the distal domains (D-domains) of fibrinogen and fibrin molecules [7]. In Contrast to the results obtained with Gly-Pro-Arg-Pro, abnormal fibrinogens which reveal a molecular defect in the D-domain leading to a defect in fibrin polymerization have not been shown to exhibit a decreased rate of FPB release [8-10]. The high fibrinogen and thrombin concentrations used in those studies might explain the failure to detect abnormalities in FPB release. Not only blocking of distal polymerization sites, but also masking of central polymerization sites by the fibrinolytic fragment D 1 can keep fibrin in solution [11]. Furthermore, fragment D 1 is a use- ful tool in studies investigating the interaction of fibrin or fibrinogen D-domains with fibrin E-do- mains [12].

The aim of the present study was to further characterize the mechanism leading to the en- hanced rate of FPB release during fibrin polymeri- zation. For this purpose, the release of FPB from dysfunctional fibrinogen Ashford [13] and fibrinogen London I [14] was studied. Both of these fibrinogen variants can be considered to have primary polymerization abnormalities. That is, prolonged incubation with thrombin cleaves all of their fibrinopeptides, yet polymerization abnormalities remain. Fibrinogen London I may have its structural abnormality located in its D- domain [10]. Furthermore, the effect of fragment D 1 on FPB release, in both abnormal and normal fibrinogens, was investigated in order to help de- fine the mechanism of enhanced FPB release from desAA-fibrin.

Material and Methods

Materials All reagents used were analytical grade,

NaH2PO 4, Na2HPO 4 and acetonitrile were HPLC grade and obtained from Baker (Gross-Gerau, F.R.G.). Lysine-Sepharose and type I gelatine from porcine skin were from Sigma (Munich, F.R.G.), organo-mercuro-agarose (Affi-Gel 501) from Bio- Rad (Munich, F.R.G.) and Sepharose CL-4B from Pharmacia (Freiburg, F.R.G.). Fibrinopeptides (A and B) were purchased from Serva (Heidelberg, F.R.G.) and from Bachem (Bubendorf, Switzer- land). Gly-Pro-Arg-Pro was from Serva.

Human fibrinogen (grade L) from KabiVitrum (Munich, F.R.G.) was further purified by affinity chromatography. After dissolving the lyophilized protein in double-distilled water, lysine-Sepharose suspended in Tris-buffered saline (100 mM NaC1, 50 mM Tris, pH 7.4) was added in order to remove contaminating plasminogen. After re- moval of the gel by centrifugation, gelatine-Seph- arose, which was prepared according to Miekka et al. [15], was added to the supernatant to remove contaminating fibronectin. A final incubation step with organo-mercuro-agarose was sufficient to re- move Factor XIII. After dialysis against a buffer containing 50 mM NaC1/50 mM Tris (pH 7.4), the protein solution was stored at -75 o C. Clotta- bility of the preparations used was at least 95% and SDS-polyacrylarnide gel electrophoresis did not reveal any degradation of Aa-chains upon incubation with streptokinase (KabiVitrum) in a final concentration of 100 U/ml and no "y-7-di- mer formation upon incubation with 1 U/ml thrombin in the presence of 10 mM CaC12 for 4 h at 37°C.

Fibrinogen London I and fibrinogen Ashford were prepared as described [14]. The lyophilized protein was dissolved in double-distilled water and dialysed against Tris-buffered saline (pH 7.4).

Human a-thrombin was obtained by incuba- tion of purified prothrombin with Echis carinatus venom as previously described [16]. Specific activi- ties of the thrombin preparations amounted to 2200-2600 NIH U/mg. Stock solutions of 100 NIH U/ml were stored at - 7 5 ° C in a buffer containing 0.1% (w/v) human serum albumin/100 mM NaCI/50 mM Tris (pH 7.4).

Fragment D 1 was prepared by incubation of 1 g of fibrinogen (KabiVitrum, Grade L) with 25 U of plasmin (KabiVitrum) in the presence of 10 mM CaC12 for 16 h. Lysine-Sepharose was used to remove the enzyme followed by an ion-exchange separation [17] of fragments E and D on DEAE- Sephacel (Pharmacia). A final gel-filtration step on Sephacryl S-200 (Pharmacia) was used to re- move contaminating fragment D-dimer and low molecular weight fragments from the fragment D 1 pool.

Methods SDS-polyacrylamide gel electrophoresis of

reduced fibrinogen samples was performed in gels containing 11% acrylamide utilizing the buffer sys- tem of Laemmli [18].

High performance liquid chromatography (HPLC). HPLC equipment from Beckman Instru- ments (Berkeley, CA, U.S.A.) consisting of a gradient controller (Model 420) connected to two pumps (model 112) and a UV detector (model 165) was used. Injections were performed by an automatic sample injector (Wisp 710B) from Waters (Milford, MA, U.S.A.). Prepacked Hibar (25 × 0.4 cm) columns filled with LiChrosorb RP 18 were purchased from Merck (Darmstadt, F.R.G.). Fibrinopeptides were separated with a linear acetonitrile gradient according to Kehl et al. [19] with the following modifications: instead of ammonium acetate buffer, 25 or 50 mM phos- phate buffers, pH 6.0 were used for solvents A or B, respectively. The gradient was run from 6.5 to 17% acetonitrile in 21 min with a flow rate of 1 m l / m in at room temperature. Peak absorbance was monitored at 210 nm, 0.02 absorption units full-scale, and peak areas were calculated by an SP 4100 integrator from Spectra Physics (Darmstadt, F.R.G.) Peak areas were converted into peptide concentrations by calibration curves, defined using commercially available FPA and FPB prepara- tions of known concentrations. An external stan- dard peptide preparation was analysed every tenth run in order to correct peak areas obtained at different days.

Kinetics of fibrinopeptide release. Kinetic experi- ments were performed at 37 °C in Tris-buffered saline, pH 7.4. A fibrinogen concentration of 0.36 #M and a thrombin concentration of 0.19 nM were chosen. No differences in the reaction con- stants were observed when the fibrinogen con- centration was decreased to 0.18 #M, suggesting that pseudo-first-order reaction conditions were fulfilled under the chosen experimental condi- tions. In a typical experiment, solutions were pre- incubated at 37°C for 15 min, the reaction was started by mixing thrombin and fibrinogen solu- tions and the reaction mixture was immediately transferred into several prewarmed polystyrol tubes. The reaction was stopped by heat precipita- tion followed by a centrifugation step to remove the precipitated protein. The supernatant contain- ing the fibrinopeptides was kept at - 2 0 °C until

171

analysis by HPLC. Infinite-time points for each experiment were obtained by incubating one tube of each series for at least 24 h. The concentration of fibrinopeptides A and B amounted to 0.714 ___ 0.05/~M and 0.736 + 0.07 #M (mean ___ S.D. for 31 infinite-time point determinations), respectively, showing a good correlation with the fibrinopep- tide concentration expected from the fibrinogen concentration (0.36 /tM) taken in the standard reaction mixture.

Calculations. Calculations were performed on a personal computer using standard routines for least-squares linear regression analysis and for non-parametric statistics (Wilcoxon test) [20].

R e s u l t s

The release of fibrinopeptides A (phosphoryl- ated FPA, FPA, desAla-FPA) and of FPB from fibrinogen by thrombin is shown in Fig. 1A. Reac- tion constants were determined from semi-loga- rithmic plots [21] of the residual substrate con- centration (ln[S]t/[S]0) versus time (Fig. 1B). The peptide concentrations determined at the infinite- time points were taken as So; S t was obtained by subtracting the peptide concentration at a given time point from S O . Thus, the kinetic analysis follows the formula:

St/S o = [Fibrmogen]t/[Fibrinogen]0

= ([FibrinopeptideLo

- [ Fibrinopeptide] t) / [ Fibrinopeptide] oo

The reaction constant for fibrinopeptide A release (kA) was determined from the slope of the initial linear part of the graph. Two linear parts were observed in the FPB release graph. The initial one was assumed to represent the release of FPB from fibrinogen and non-polymerized desAA-fibrin (kBi), the late one the release from polymerized desAA-fibrin (kB~). Control experiments with nor- mal and abnormal fibrinogen using a lower thrombin concentration ([T]) did not reveal any differences in the specificity constants (first order reaction constant/[T]), indicating that the forma- tion of desAA-fibrin polymers is not the rate-de- termining step for the release of FPB under stan-

0 . 8 -

0 0

0 . 6 -

o 0 . 4 -

= ~. 0 . 2 -

O-

A

[3

2'0 ,'o do Time (rain)

kBi

z ~ ' ~ x zx a B

-%

E - 1 -

172

- 2 -

I 0 2'o ~ 6'o

Time (min)

Fig. 1. Release of FPA (D) and FPB (zx) from normal fibrino- gen (0.36 #M) by the action of thrombin (0.19 nM). In (A), the concentration of fibrinopeptides released is plotted versus time. In (B), a semilogarithmic replot of the data given in (A) is depicted. The reaction constants for the release of FPA (k A), FPB in the initial phase (kBi) and the late phase (kBt) of the thrombin-fibrinogen reaction correspond to the slopes of the linear parts of the graphs and were calculated by least-squares

linear regression analysis.

dard experimental conditions (data not shown). In a first set of experiments, the specificity

constant for fibrinopeptides A release (kg/ [T]) was determined to be 6 . 7 _ 1.1-10 6 ( M - ] . s - 1 )

(mean _+ S.D., n = 7), 11-fold higher than the con- stant for the release of FPB in the initial phase of the thrombin-fibrinogen reaction. However, dur- ing progressive incubation of fibrinogen with thrombin, a 4-fold increase in the rate of FPB

TABLE I

SPECIFICITY CONSTANTS FOR FIBRINOPEPTIDE RE- LEASE F R O M N O R M A L A N D DYSFUNC TIONAL FIBRINOGEN U N D E R VARIOUS POLYMERIZATION CONDITIONS

Specificity constants (first-order rate constants / thrombin con- centration) for the release of FPA in the initial phase (SA) and for the release of FPB in the late phase of the thrombin-fibrinogen reaction (SB[) from fibrinogen London I (London I) and from normal fibrinogen (Normal) were de- termined at fixed concentrations of fibrinogen (0.35 /xM) and thrombin (0.19 nM). The values given for normal fibrinogen in the absence of inhibitors were determined in parallel with experiments performed with abnormal fibrinogens or in the presence of inhibitors of fibrin polymerization. The concentra- tions of added Gly-Pro-Arg-Pro (GPRP) and fragment D l (D1) are given in brackets, n indicates the number of indepen- dent experiments; means and standard deviations are given.

Fibrinogen Additives n S A (10 -6) SBI (10 6) variant ( M - l - s -1) ( M - ] - s -1)

Normal 5 5.1___0.5 2.7+0.4 London I 5 5.3+0.6 1.2_+0.1 London I D] (5.8/~M) 4 5.5_+0.6 1.3_+0.1 Normal GPRP 3 5.2 _+ 0.7 0.5 _+ 0.1

(360 ttM)

Normal D 1 (5.8/~M) 5 5.9-+0.9 2.9_+0.7 Normal 5 5.6 _+ 0.9 2.6 _+ 0.4

Normal D 1 (11.5/~M) 3 5.5 _+ 0.2 2.2 _+ 0.4 Normal 3 5.2_+0.7 2.2_+0.2

release was observed, expressed as an increase in the specificity constant from kBi/[T] = (0.6 + 0.2) • 10 6 ( M -1 • S -1) to kBI/[T ] = (2.2 ___ 0.5) • 1 0 6

(M - l - s-]) . The addition of a 1000-fold molar excess of Gly-Pro-Arg-Pro over fibrinogen effec- tively prevented fibrin polymerization, as judged by monitoring the turbidity of the reaction mix- ture at 350 nm. The presence of the tetrapeptide prevented the enhanced rate of FPB release, as demonstrated by a 4-5-fold reduction in KBI/[T ] (Table I). The rate of FPB release in the presence of the tetrapeptide did not exceed the rate of FPB release in the initial phase of the thrombin- fibrinogen reaction, determined in the previous experiments in the absence of inhibitors of fibrin polymerization.

Upon reaction with thrombin, fibrinogen London I exhibited no difference in FPA release compared with normal fibrinogen, indicating nor-

o

_~c

FPB: normal Fbg + GPRP

Fbg London I

Fibr inopept ides A

FPB: normal Fbg

2'o 4'0 e'o Time (min)

Fig. 2. Semilogarithmic plot of the release of FPA ( - , It, O) and FPB (zx, t2, o ) from normal fibrinogen ( - , zx), fibrinogen London I (0, ©) and normal fibrinogen in the presence of a 1000-fold molar excess of Gly-Pro-Arg-Pro (11, 12). Experimen-

tal conditions were identical to those given in Fig. 1.

mal enzyme-substrata interaction. However, the rate of FPB release from fibrinogen London I was significantly (P < 0.05) decreased compared with normal fibrinogen (Fig. 2, Table I). Identical re- sults were obtained with fibrinogen Ashford (ka l / [T ] = (1.4 _ 0.2) • 106 M -1 • s -1, n = 3), which also exhibits a primary polymerization abnormality and a normal amount of total releasa- ble fibrinopeptides A and B [13].

In a turbidity assay at 350 nm, the fragment D1 preparations used inhibited clot formation during

L O -1 -

c

FPB: Fbg London I + D I FPB: Fbg London I

FPB: normal Fbg

FPE: normal Fbg+ 01

'o 2'0 3'0 ,'o ~'o ~'o Time (rain)

Fig. 3. Semilogarithmic plots of the release of FPB from normal fibrinogen (ll, 12) and fibrinogen London I (o, o ) in the absence ([3, o ) and presence of a 16-fold (11, O) molar excess of fragment D] over fibrinogen. Values of one typical experiment are given. Experimental conditions were identical to those

described in Fig. 1.

173

the time period of the experiment when a 16-fold molar excess over fibrinogen was applied, and for at least 24 h when a 32-fold excess was used. Although clot formation was inhibited, the ad- dition of fragment D 1 in a 16- and 32-fold molar excess over normal fibrinogen affected neither the rate of FPA release in the initial phase of the thrombin-fibrinogen reaction nor the accelerated rate of FPB release (Table I, Fig. 3). Furthermore, addition of fragment D 1 in a 16-fold molar excess over fibrinogen London I did not alter the rate of FPB release (Table I, Fig. 3).

Discussion

Thrombin cleaves the N-terminal Aa- and Bfl- chains in the central domain of the fibrinogen molecule with different rates, resulting in a typical sequential pattern of FPA and FPB release at low thrombin concentrations [5]. The release of FPA in the initial phase of the thrombin-fibrinogen reaction was shown to be about l l - fo ld faster than the release of FPB. This difference is some- what higher than the 4-5-fold difference [2,3] determined under second-order reaction condi- tions, indicating that under these conditions des- AA-fibrin polymers might have formed in the initial phase of the reaction. Kinetic evidence was presented by Higgins et al. [4] that FPB is released from fibrinogen with a much lower rate than from monomeric desAA-fibrin. Therefore, the rates of FPB release determined in our study for the initial phase of the thrombin-fibrinogen reaction most likely represent the release from desAA-fibrin monomers.

The acceleration of FPB release depends on undisturbed fibrin polymerization, as shown by the slow rate of release in the presence of urea [5] or Gly-Pro-Arg-Pro [6], which prevents fibrin polymerization by blocking C-terminal polymeriz- ation sites [7]. Polymerization studies with purified D-domains of fibrinogen London I have suggested that the primary polymerization defect of this variant is located in its D-domain [10]. Therefore, the polymerization abnormality of fibrinogen London I may be compared in some respect to the inhibition of polymerization of normal fibrinogen by Gly-Pro-Arg-Pro. Under the experimental con- ditions chosen in our study, a significantly de-

174

creased rate of FPB release from fibrinogen London I could be demonstrated in the late phase of its reaction with thrombin compared with the rate determined for normal fibrinogen, and similar results were obtained with fibrinogen Ashford. However, the rates of FPB release from fibrinogen London I and from fibrinogen Ashford were higher compared to experiments performed in the pres- ence of Gly-Pro-Arg-Pro, indicating that poly- merization in these fibrinogen preparations is im- paired but not totally absent. These findings can be reconciled with the presence of a heterozygous defect, as reported in most of the fibrinogen variants described. Earlier studies with fibrinogen London I [10] as well as with other abnormal fibrinogens exhibit a C-terminal polymerization defect [8,9] did not reveal any differences in FPB release compared to normal fibrinogen. Because those studies were performed at higher thrombin and fibrinogen concentrations, the decreased rate of FPB release might not have become apparent. Indeed, studies by Southan et al. [22] on the release of FPB from a fibrinogen variant with an N-terminal polymerization defect have shown that a decreased rate of FPB release was highly depen- dent upon thrombin concentration.

Unlike the decreased rate of FPB release under conditions where C-terminal polymerization sites are occupied (in the presence of Gly-Pro-Arg-Pro), blocking of N-terminal polymerization domains with fragment D 1 did not alter the rate of FPB release from normal fibrinogen. Inhibition of polymerization by Gly-Pro-Arg-Pro is achieved by binding of the tetrapeptide to distal domains of fibrinogen prior to the addition of thrombin, thereby initially keeping desAA-fibrin in an essen- tially monomeric state [6]. In contrast, binding of fragment D1 to N-terminal polymerization do- mains cannot occur prior to their exposure by removal of FPA. Therefore, D-domains of neigh- boring desAA-fibrin molecules compete with the inhibitory action of fragment D1, resulting in the formation of small fibrin oligomers during normal fibrinogen-fibrin conversion [11]. As demon- strated by the present findings, these fibrin oligomers formed in the presence of fragment D1 then serve as optimal substrates for the enzymatic attack of thrombin on the Bfl-chain, resulting in accelerated FPB release. In addition, fragment D~

did not affect FPB release from fibrinogen London I. Since qualitative differences in binding proper- ties of fragment D 1 to fibrin E-domains have not yet been described [12], the failure of fragment D 1 to enhance FPB release from fibrinogen London I may not be caused by its ineffective binding to the central domain. Rather, it suggests that the D - E interaction is not sufficient in itself to facilitate the enzymatic attack of thrombin on the Bfl-chain.

In conclusion, as long as fibrin is kept in an essentially monomeric state, no acceleration of FPB release occurs. Furthermore, fragment D 1 interacting with E-domains of monomer ic desAA-fibrin is not capable of enhancing FPB release. However, the formation of small fibrin polymers results in the release of FPB at a high rate. Thus, from the results presented here, the interaction of native D-domains with E-domains is a minimum prerequisite for enhanced FPB re- lease. During alignment of desAA-fibrin mole- cules in a polymer, one D - E contact is usually tightened by a reciprocal contact between the same molecules. Alignment in the polymer may there- fore involve a long-range conformational change of the desAA-fibrin molecule, leading to enhanced FPB release.

References

1 Martinelli, R.A. and Scheraga, H.A. (1980) Biochemistry 19, 2343-2350.

2 Hanna, L.S., Scheraga, H.A., Francis, C.W. and Marder, V.J. (1984) Biochemistry 23, 4681-4687.

3 Liu, C.Y. and Sepe, O. (1986) in Fibrinogen and Its Deriva- tives (Miiller-Berghaus, G., Scheefers-Borchel, U., Selmayr, E. and Henschen, A., eds.), pp. 49-52, Elsevier/North-Hol- land, Amsterdam.

4 Higgins, D.L., Lewis, S.D. and Shafer, J.A. (1983) J. Biol. Chem. 258, 9276-9282.

5 Blomb~ick, B., Hessel, B., Hogg, D. and Therkildsen, L. (1978) Nature 275, 501-505.

6 Hurlet-Jensen, A., Cummins, H.Z., Nossel, H.L. and Liu, C.Y. (1982) Thromb. Res. 27, 419-427.

7 Laudano, A.P. and Doolittle, R.F. (1980) Biochemistry 19, 1013-1019.

8 Reber, P., Furlan, M., Rupp, C. Kehl, M., Henschen, A., Mannucci, P.M. and Beck, E.A. (1986) Blood 67, 1751-1756.

9 Reber, P., Furlan, M., Henschen, A., Kaudewitz, H., Barbui, T., Hilgard, P., Nenci, G.G., Berrettini, M. and Beck, E.A. (1986) Thromb. Haemost. 56, 401-406.

10 Lane, D.A., Ireland, H., Thompson, E., Southan, C. and Henschen, A. (1985) in Fibrinogen-Structural Variants and Interactions (Henschen, A., Hessel, B., McDonagh, J. and

Saldeen, T., eds.), pp. 197-206, De Gruyter, Berlin, New York.

11 Knoll, D., Hantgan, R., Williams, J., McDonagh, J. and Hermans, J. (1984) Biochemistry 23, 3708-3715.

12 Cierniewski, C.S., Kloczewiak, M. and Budzynski, A.Z. (1986) J. Biol. Chem. 261, 9116-9121.

13 Southan, C., Thompson, E. and Lane, D.A. (1987) Br. J. Haematol. 65, 469-473.

14 Lane, D.A., Cuddigan, B., VanRoss, M. and Kakkar, V.V. (1980) Br. J. Haematol. 44, 483-494.

15 Miekka, S.I., Ingham, K.C. and Menache, D. (1982) Thromb. Res. 27, 1-14.

16 Preissner, K.T., Delvos, U. and MiiUer-Berghaus, G. (1987) Biochemistry 26, 2521-2528.

175

17 Chela, J.P., Shurley, H.M. and Vickroy, M.F. (1974) Bio- chem. Biophys. Res. Commun. 61, 66-71.

18 Laemmli, U.K. (1970) Nature 227, 680-685. 19 Kehl, M., Lottspeich, F. and Henschen, A. (1981) Hoppe-

Seyler's Z. Physiol. Chem. 362, 1661-1664. 20 Sachs, L. (1984) Angewandte Statistik, pp. 230-235,

Spdnger-Verlag, Berlin. 21 Segel, I.H. (1975) Enzyme Kinetics, pp. 57-59, John Wiley

and Sons, New York. 22 Southan, C., Lane, D.A., Bode, W. and Henschen, A.

(1985) Eur. J. Biochem. 147, 593-600.