J. Mol. Biol. (1992) 226, 263-269

Modulation

of Thrombin-Hirudin Interaction by Specific Ion Effects

Raimondo De Cristofarol,

John W. Fenton II2 and Enrico Di Ceralj’

‘Department of Biochemistry and Molecular Biophysics Washington University School of Medicine Box 8231, St Louis, MO 63110, U.S.A. ‘New York State Department of Health Wadsworth Center for Laboratory and Research Box 509, Albany, NY 12201, and Departments of Physiology and Biochemistry Albany Medical College of Union University Albany, NY 12208, U.S.A. (Received

19 November

1991; accepted 13 February

1992)

Kinetic studies of the inhibition of thrombin amidase activity by recombinant hirudin have been conducted as a function of salt concentration in the range 605 to 1 M, using NaCl. KU, NaBr and KBr. At the same ionic strength, the value of K, for thrombin-hirudin interaction is found to be different with different salts. The slope d In K,/d In a+, where a, is the mean ion activity, is constant in the range 0.05 to 05 M, is sensitive to the particular salt present in solution and is equal to 1.07+_009 (NaCl), 0*92+@10 (KCI), 1.37+0.10 (NaBr) and 056+0.10 (KBr). These results indicate that specific ion effects are involved in the modulation of thrombin-hirudin interaction in the form of ion release, as recently found in the case of thrombin interaction with its natural substrate fibrinogen. The linkage hierarchy for ion release found in the case of thrombin-fibrinogen interaction also applies in the case of thrombin-hirudin interaction, with the number of released ions decreasing in the order NaBr > NaCl > KC1 > KBr. It is proposed that the process of bridge-binding to the fibrinogen recognition site and the catalytic pocket of the enzyme, as seen in the case of fibrinogen and hirudin, is linked to ion release and controlled by modulation of the association rate constant. Keywords: thrombin;

hirudin;

salt effects; macromolecular

1. Introduction The anion-bonding-exosite involved in the recognition of fibrinogen by thrombin (Fenton et al., 1988) have been characterized in terms of crystallographic analysis as an insertion loop packed with a number of positively charged and hydrophobic residues (Rydel et al., 1990, 1991). Fibrinogen bridge-binds to this loop, the fibrinogen recognition site (FRSS), and the adjacent catalytic pocket, thereby triggering the subsequent physical steps leading to clot formation (Hantgan & Hermans, t Author to whom all correspondence should be addressed. $ Abbreviations used: FRS, fibrinogen recognition site; h.p.l.c., high pressure liquid chromatography; f.a.b.-m.s., fast atomic bombardment-mass spectrometry; S-2238, H-n-Phe-pipecolyl-Argp-nitroanilide; PEG, poly(ethylene glycol). 0022%2836/92/13026347

f03.00/0

interactions;

linkage

1979). Consistent with the structural findings, a recent investigation of thrombin-fibrinogen interaction under conditions of potential physiological interest has pointed out that specific ion effects play a key role in the modulation of this important reaction (De Cristofaro & Di Cera, 1992). When fibrinogen interacts with thrombin there is a net release of small ions to the solution, which provides a major thermodynamic driving force for binding. These ions can be released locally from the FRS and/or the complementary portion of fibrinogen interacting with it. Other regions of the enzyme and/or fibrinogen can also contribute to the linkage observed experimentally through ion-linked conformational transitions. In order to assess the nature of the structural perturbations involved in the linkage between macromolecular binding to the FRS and ion release, studies should be conducted with different molecules that bind to thrombin in a 263 @ 1992 Academic

Press Limited

R. De Cristofaro et al. fashion similar to fibrinogen. Hirudin, a small peptide isolated from the salivary glands of the medicinal leech Hirudo medicinalis European (Markwardt, 1970), is the most potent natural inhibitor of thrombin. Its inhibitory effect on thrombin-fibrinogen interaction is competitive and is achieved by bridge-binding to the FRS and the catalytic pocket (Rydel et al., 1990, 1991). In view of its peculiar binding properties, hirudin is an ideal probe for the energetics of the linkage between macromolecular binding to the FRS and ion release. In this study we explore the effect of salt concentration on the thrombin-hirudin interaction following the strategy recently used in the analysis of thrombin-fibrinogen interaction (De Cristofaro & Di Cera, 1992). The purpose of this study is to assess the thermodynamic aspects involved in the control of the thrombin-hirudin interaction and the way they compare with those recently observed in the case of the thrombin-fibrinogen interaction.

measurements was typically 100 PM, and was dropped to 50 PM for measurements in the presence of NaCl at concentrations below 0175 M. The enzyme was pre-incubated with different hirudin concentrations to allow t’he thrombin-hirudin interaction to reach equilibrium. After pre-incubation, steady state measurements of synthetic substrate hydrolysis were started by addition of S-2238. A total of 7 curves were measured, using 7 substrate concentrations scaled by a factor of 2 in a range typically from 66 to about 40 PM. Only 1 curve was measured in the absence of hirudin and 6 other curves were collected in the presence of hirudin concentrations scaled by a factor of 1.5, starting with a concentration of 200~~. The increase in absorbance at 405 nm was linear over a time scale of 0.5 to 10 min, depending upon solution conditions. and was used to compute the steady state velocity as a function of hirudin and substrate concentration. (c) Data analysis

The relevant kinetic scheme for thrombin by hirudin is: E+S$

ES’: 1

2. Materials and Methods (a) Thrombin and hirudin preparation Human cr-thrombin was prepared as described elsewhere (Fenton et al., 1977), further purified and tested for activity as previously detailed (Di Cera et al., 1991). Thrombin solutions of 3 FM concentration were stored in 50 ~1 vials at -80°C until use. Recombinant CGP 39393 hirudin was a generous gift of Dr H. Grossenbacher (CibaGeigy Pharmaceuticals, Basel, Switzerland). The preparation was >95% pure, as shown by reverse-phase h.p.1.c. and f.a.b.-m.s. analysis and was used without further purification. Recombinant hirudin has the same sequence as the natural form but lacks the sulfate group on Tyr,,. The concentration of hirudin was calculated using a relative molecular mass of 695352. An extinction coefficient E zT5 = @42( +@02) ml/mg/cm was obtained from averaging the readings of 3 different hirudin solutions in water. Hirudin was stored in 250 ~1 glass vials at - 80°C until use. (b) Steady state measurements Steady state measurements of human a-thrombin amidase activity were made using the synthetic chromogenic peptide S-2238 purchased from KabiVitrum (Stockholm, Sweden). The concentration of S-2238 was measured at 342 nm using a Cary 3 dual-beam spectrophotometer and an extinction coefficient of 8270 M- ’ cm- ’ (Lottenberg & Jackson, 1983). Assays were performed by following the release of p-nitroaniline resulting from the hydrolysis of S-2238 as an increase in absorbance at 405 nm. The concentration of released p-nitroaniline was quantified by means of the extinction coefficient E,,,, = 97891+ 133461/(2.0+1), where Z is the ionic strength of the medium (De Cristofaro et al., 1992). Assays were performed using disposable polystyrene cuvettes under solution conditions of 50 maa-Tris.HCl, 61 To (v/v) PEG 8000 (pH 7.5) at 37”C, using NaCl, KCl, NaBr and KBr over a concentration range from 50 mM to 1 M. The preparation and titration of buffer solutions and the calculation of the mean ion activity, a*, were done as reported elsewhere (Di Cera et al., 1991; De Cristofaro & Di Cera, 1992). The concentration of human a-thrombin in steady state

E+I”r

inhibition

E+P.

“k:

EI. L where ‘k, is the acylation rate, while k: and km, are the rate constants for binding and dissociation for the substrate (S) and hirudin (H). E, is the free enzyme, P is the product and I the concentration. There are a number of equivalent ways to probe the kinetics of the reactions sketched above (Cha, 1975, 1976). Since we are interested in the equilibrium parameter K, = ‘km ,lHk:, that provides key information on the linkage between hirudin binding and ion release, we have chosen the simplest procedure that allows for resolution of the hirudin dissociation constant. In this procedure the enzyme is pre-incubated with hirudin to allow the reaction to reach equilibrium. Due to its extremely high affinity, the hirudin activity changes significantly upon binding to thrombin under conditions where these macromolecular components are present in comparable amounts. In this case the relevant equations need to be cast in terms of the total, rather than the free, amount of hirudin using the conservation relationship:

WI = h,--e,$&, I where eT is the total enzyme concentration, [H] and h, are the free and total concentrations of hirudin, and K, is the dissociation constant for thrombin-hirudin interaction. Solution of eqn (2) leads to:

WI =

hr-er-K,+Q

2

Pa)

~~-Q = db-eT

-K,)‘+4K,h,.

WI

After pre-incubation, the reaction is started by addition of substrate and the initial linear phase is used to measure the initial steady state velocity of substrate hydrolysis. In the presence of high enough concentrations of hirudin, the initial linear phase can typically be followed over 10 to 15 min, after which a transient evolution is observed into a faster linear phase. The 1st phase is slow and reflects the amidase activity of the free enzyme that is not inhibited by hirudin. The 2nd phase is fast since hirudin starts dissociating from the enzyme and more thrombin is available for catalysis, This behavior has been observed over

Thrombin-Hirudin the whole range of conditions studied and is to be expected whenever a relaxation between 2 steady states is linked to a slow process, as first pointed out by Frieden several years ago (Frieden, 1970). The slow process is the dissociation of hirudin from thrombin. The steady state velocity of substrate hydrolysis during the 1st phase is due to the activity of the free enzyme after pre-incubation (Cha, 1975, 1976; Williams & Morrison, 1979), which can be calculated from the relationship: Wl=eT&=eTh

I

T

(4)

-,“$+@

Since this is the maximum concentration of enzyme available for catalysis, the measured steady state velocity (w) is simply: kc,, PI

v=eTK,

2K, h,-e,+K,+Q’

Paradoxically, the effect of a tight binding inhibitor, such as hirudin. mimics that of a non-competitive inhibitor, since it only affects k,,, but not K, (Cha, 1975, 1976). The independent parameters k,,,, K, and K, can be resolved from analysis of experimental data obtained at different substrate and hirudin concentrations using eqn (5). All steady state determinations collected in a 7 x 7 matrix of hirudin and substrate concentrations were analyzed globally according to eqn (5) to resolve the independent parameters from a total of 49 experimental data points. All experimental points were weighted uniformly. Minimization was accomplished by non-linear least-squares using the Marquardt method and best-fit estimates were obtained by extensive search in the parameter space using different starting guesses. Convergence to a unique minimum was always obtained. Confidence intervals on the parameters were computed by F-testing at the cutoff of 1 standard deviation (68%). The total hirudin concentration, h,, was multiplied by a correcting factor, c, which was floated in the minimization procedure. This was done to detect any discrepancy between the nominal hirudin concentration and the actual concentration consistent with the experimental data. The correcting factor significantly improved the goodness of the fit in many cases and was found to be on average 1.33 f917 from a total of over 50 measurements. (d) Control experiments A number of control experiments were run to check the reproducibility and accuracy of the results. Particular attention was paid to ensure that equilibrium was truly reached in the pre-incubation of thrombin with hirudin. The time scale for equilibrium in a reaction such as (lb) is given by:

one in eqn (7) is given by (Di Cera, 1990):

aJ T-1

=

awl

where [El,, is the equilibrium eqn (4). Hence, 7-l = Hkf(hT-eT)+Hk-l

IE,=&’

value of [E] given by +2Hk:[E],,

= HkfQ.

(9)

When h, >>e, , eqn (9) reduces to eqn (6), as expected. In general, the time scale for equilibrium is set by the values of Hkf and Q. However, since Q > K, in all cases, then one necessarily has T- ’ > Hk- 1, which means that the time scale for equilibrium is at most as slow as the dissociation of hirudin from thrombin. We have already pointed out that progress curves of substrate hydrolysis measured after pre-incubation indicate a time scale for hirudin dissociation of at least 10 to 20 min, quite independent of salt concentrations. This finding is consistent with a saltindependent value of Hk- i of about 055 x 10e3 s-i, as reported previously (Stone & Hofsteenge, 1986). Hence, the maximum value for T is about 30 min. Control experiments were run at high salt concentrations (1 M), where T is expected to be slow, using different pre-incubation times. Analysis of data obtained after 45 min of preincubation yielded significantly different values of K, with respect to data obtained after 2 or 3 h of pie-incubation, but no difference was found among samples pre-incubated for 2 h or longer. These control experiments fully confirm the expectations drawn from considerations dealt with above. All steady state measurements reported in this study were performed after pre-incubation for 3 h, even when the predicted time scale for reaching equilibrium was of the order of a few minutes (e.g. at low salt concentrations where K, decreases). Additional control experiments were run to test thrombin amidase activity as a function of time in the absence of hirudin. The values of K, and k,, for the hydrolysis of S-2238 were identical, within experimental error, for thrombin samples pre-incubated at 37°C in a time scale ranging from 0 to 3 h, thereby indicating that no significant denaturation or autolysis occurs over the time scale of the steady state measurements in the presence of hirudin. coefficient of p-nitroaniline was The extinction measured in the presence of NaCl, KCI, NaBr and KBr over the concentration range 0 to 1 M, in order to assess any differential effect of salts on the spectral properties of the leaving group of the synthetic substrate. No significant difference was observed with different salts, which is consistent with the fact that the extinction coefficient is a function solely of the ionic strength.

3. Results

T-’ = Hk- I +Hkr[H], where T is the relaxation time. Eqn (6) only applies when binding of the ligand does not affect its activity in solution, even when the activity fluctuates around its deterministic value (Di Cera, 1991). When binding of the ligand significantly affects the amount free in solution, as in the case of thrombin-hirudin interaction, then the time evolution for the concentration of free enzyme in eqn (lb) is given by: J=d[E]/dt=Hk~IeT-{Hk~(hT-eT)+Hkk,}[E]-Hk~[E]Z, (7) and is a quadratic expression in [El. The relaxation for equilibrium in any l-component system such

265

Interaction

as

time the

A typical set of steady state velocity measurements as a function of hirudin concentration is shown in Figure 1. Even though hirudin is a competitive inhibitor, the effect observed affects “paradoxically” the maximum velocity and not K,. This is the result expected in the presence of a competitive inhibitor that binds tightly to the enzyme and does not dissociate significantly over the time scale of initial steady state velocity measurements. The data are accurate enough to allow for resolution of all parameters involved in equation (a), including the correcting factor for hirudin concentration. The

R. De Cristofaro

266

et al

150-c

112.5

i

m 75.c ‘;

37.5

-12.3;-,’

-6.5

-5.9

-4-6

-5.3

n’

”

-4.0

lag El CM)

Figure 1. Typical data set of steady state velocity measurements of S-2238 hydrolysis by human a-thrombin under solution conditions of 100 purr-thrombin, 50 mMTris’HCI, 0375~NaCI, @l% PEG at 37°C. in the presence of different concentrations of recombinant hirudin as indicated (from top to bottom): 0 PM, 2634 PM. 3951 PM, 5926 PM, 88%9 PM. The steady state velocity is expressed per unit enzyme concentration. Data were collected in a 7 x 7 matrix of substrate and hirudin concentrations. Only 5 curves are shown for the sake of clarity. Continuous lines were drawn using eyn (4) in the text with the best-fit parameter values: k,,, = 133.1 K, = 3.7( kO.2) /LM. (k2.2) s-1. K, = 2.9( f0.7) PM. c = 1.34( fGO4). The standard error of the fit is 2.9 8-I.

value of c is on average 1.3&W, which is comparable in absolute percent change to the value of 0.6 reported from analysis of progress curves (Stone & Hofsteenge, 1986). The effect of different salt,s on thrombin-hirudin interaction is shown in Figure 2 as the change in the logarithm of K, versus the logarithm of the mean ion activity. The values of K, are listed in Table 1. The dependence of K, and k,,, for the synthetic substrate S-2238 on a, and the different salts was found to be consistent with that observed in a previous study under identical solution conditions

Table 1 Effect of diferent salts on the K, of thrombin-hirudin interaction, under the experimental conditions reported in the text [I

(PM)

0,050 0.075 0125 0.175 0.250 0,375 0500 @750 1 .ooO

N&Cl

KC’1

W4fO.l os_fo.2 Q9kO.2 1.5 + 0.4 1.7kO.4 29fo.7 4G3+ 1.2 52k1.4 5,3_+ 1.4

2.1 +05 2.5 * 0% 3.7 + 1.0 7.0* I.7 8.1 k2.0 12.3+30 11.0 If: 2.7 12.5 f 3.0 20+3+31

N&r l.Of@% 1.2+@3 2.6 & 0.5 3.1+07 7.6* 1.8 11.1*2.s 131 *so 263 k 6.5 31.6k6.9

KBr 68k 1.7 8.0 * 2.0 92k2.3 14.1 f2.7 2POf4.9 20.0*51 17.5*45 22.5 + 56 22.9 * 49

Figure 2. Effect of (0) BaCl. (H) P;aBr, (0) Kc9 and (0) KBr on thrombin-hirudin interaction, under experimental conditions of 50 mM-Tris . HCl. 0.1 7); PEG (pH 7.5) and 37°C. Data are plotted as the logarithm of K, as a function of the logarithm of the mean ion activit,y. a,. Points in the 50 to 500 rn,Mrange of salt concentration w&e interpolated with a straight line, as shown. The slope of the curve for each salt is: 1.07 f @09 (NaCl), @92 + 0.10 (KCl), 1,37f(klO (NaBr). 056f010 (KBr).

(De Cristofaro $ Di Cera, 1992). The hirudin aflinit,v decreases with increasing salt concentration in a way that is strictly dependent on the particular ions present in solution. The dependence shows a striking similarity with that observed in the case of thrombin-fibrinogen int’eraction in t,he d0 t,o 500 mM range of salt concentration. At any given value of a+: the value of K, is lowest in the presence of NaCl. The logarithm of K, changes linearly witah the logarithm of the mean ion activity wit’h slopes of l.O7f0*09, 092f0.10, 1.37f0.10 and 0.56+0.10 for NaCl, KU, NaHr and KRr. respectively. The effects of ions on any equilibrium constant can arise from several different, sources (Record rt aZ., 1978). The relevant expression for the change of K, with a, is:

(l()) where A Y+ and A Y- denot’e the number of cations and anions exchanged upon hirudin binding to thrombin. AW is the differential hydration, o is a constant proportional to the molal concentration of the salt, y: and y& are the standard activit’y coeficients of E and ET in eqn (1 b), and a” is the hirudin activity. The different,ial hydra,tion t,erm usually contributes significantly only at salt concentrations ~0.5 M (Record et al., 1978). and can be dropped from equation (10) in the range of salt concentrations we are interested in (i.e. 50 to 500 mw). The

Thrombin-Hirudin

267

different salts at the same ionic strength unequivocally expresses the differential contribution of ion binding alone, since C(1) cancels out in eqn (10). A thermodynamic cycle can be constructed to depict the energetics of the linkage between macromolecular binding to the FRS and ion release, as shown in Figure 3. Each edge of the cycle in the direction A+B represents the change in Av observed when salt A is replaced by salt B, i.e. the difference

-0.30 -0.15

f0.30

Interaction

+0,8

AAv,,, T

I

KCI

I

I

J

$0-02 +O-36

I

Figure

3. Thermodynamic cycle for the linkage between ion binding and thrombin interaction with h&din and fibrinogen. Each edge of the cycle in the direction A+B represents the change in Av observed when salt A is replaced by salt B, i.e. the difference in eqn (11). Outer values refer to AAv A+B = Av,-Av, thrombin-hirudin interaction, while inner values were calculated from the slope d In K,/d In a, for thrombinfibrinogen interaction (De Cristofaro & Di Cera, 1992). Negative values indicate that ion release increases in the substitution A+B, while the opposite is observed with positive values. The cycle shows that differential cation effects (vertical edges) are anion dependent and vice versa (horizontal edges) for either hirudin or fibrinogen binding to the FRS. The diagonal arrow depicts the value of AAv when the salt substitution involves both the cationic and anionic components.

two terms reflecting the change in the activity coefficients of the enzyme intermediates, as well as hirudin, can be expressed as C(I) since they only depend on Debye-Hiickel screening parameters and hence ionic strength, independent of the particular salt being used. The only term that is salt-specific and reflects binding equilibria of either cations and anions is Av. that gives the net number of ions thrombin-hirudin interaction. exchanged upon Since the effects observed experimentally in principle reflect the contribution of specific binding and ionic strength, a great deal of experimental work must be conducted with different salts before a meaningful conclusion can be drawn on the nature of the ion effects observed. In a previous study (Stone et al., 1989) it has been suggested that thrombin-hirudin interaction is controlled solely by ionic strength in a way that is consistent with the Debye-Hiickel theory. However, this conclusion has been drawn from studies conducted in the presence of NaCl only. On the other hand, when thrombinhirudin interaction is studied as a function of different salts, then one arrives at a quite different conclusion. The data shown in Figure 2 demonstrate clearly that specific ion effects are involved in the modulation of thrombin-hirudin interaction, since different effects are observed with different salts at the same ionic strength. Furthermore, the difference between values of d In K,/d In a, obtained with

= Avs-Av,.

(11)

For example, the substitution of NaCl with NaBr gives a value of AAv = -630, which indicates that upon thrombin-hirudin interaction, the net release of ions is larger in the presence of NaBr than NaCl. The horizontal edges of the cycle give the differential anion effect, while the vertical edges give the differential cation effect. Either effect depends on the particular counterion present, which implies that significant interactions must, exist between cation and anion binding sites, as found in the case of thrombin-fibrinogen interaction (De Cristofaro & Di Cera, 1992).

4. Discussion We have recently documented the importance of specific ion effects in the modulation of thrombin activity (De Cristofaro & Di Cera, 1990; Di Cera et aE., 1991; De Cristofaro et al., 1992) and thrombinfibrinogen interaction (De Cristofaro & Di Cera. 1992). The results reported here provide additional support to the general conclusion that specific interactions of small ions with the enzyme are critical driving forces for binding and catalytic events. They also shed light on the structural components involved in the effects observed experimentally. A linkage hierarchy seems to emerge from analysis of thrombin-fibrinogen and thrombin-hirudin interaction. Ion release decreases in the order NaBr >NaCl>KCl >KBr. The existence of such a hierarchy suggests that the enzyme is probably the major source of ions released upon binding to the FRS. If a significant contribution comes from either fibrinogen or hirudin, then it probably originates from a structural component complementary to the FRS that is simlar for both macromolecules. This picture seems to be supported by a number of considerations. In studies of thrombin amidase activity as a function of salt concentration, it has been shown that the structural perturbation of the FRS in y-thrombin significantly reduces the linkage between ion and substrate binding (De Cristofaro et al., 1992) compared with the native enzyme (Di Cera et aE., 1991). We have proposed that the structural basis for the reduced linkage observed in y-thrombin may be identified with a perturbation of a putative anion binding site formed by the triangle of basic residues Arg67, Arg75 and Arg77A in the FRS (Rydel et al., 1990, 1991), due to the proteolytic cleavage by trypsin at Arg73. An additional and important observation is that the value of k,,,

268

R. De Cristofaro et al.

for the hydrolysis of amide substrates by thrombin decreases significantly (30 to 50%) when Na+ salts are replaced by K+ salts (Orthner & Kosow, 1980; Di Cera et al., 1991; De Cristofaro & Di Cera, 1992). These findings imply that both cations and anions are involved in the control of thrombin activity. The carboxyl-terminal of hirudin contains a number of acidic residues and participates in numerous salt bridges with basic residues of the FRS (Rydel et al., 1990, 1991). It is very likely that a similar structural organization is present in the region of fibrinogen responsible for binding to the FRS (Naski et al., 1990; Schmitz et al., 1991). We have proposed that the complementary region on the fibrinogen molecule may provide a binding site for cations (De Cristofaro & Di Cera, 1992). Binding to the FRS would involve at least three classes of ion binding sites. Anions can be released from the FRS, while cations can be released from either thrombin and fibrinogen or hirudin. The interactions among cation and anion binding sites implied by the thermodynamic cycle in Figure 3 are readily understood in terms of this scheme. The anion binding site in the FRS may be allosterically coupled to the cation binding site controlling the value of k,,, for the enzyme. Also, when fibrinogen or hirudin bind to thrombin, the groups with opposite charges present in the FRS and its complementary region must be exposed with a resulting release of counterions. The release of anions from the FRS may favor the release of cations from the complementary region of fibrinogen or hirudin by simple electroThe quantitative differences static repulsion. observed with different salts between fibrinogen and hirudin can be accounted for by different affinities of the cation binding to the region complementary to the FRS, as well as differences in the affinities of cation and anion binding sites of thrombin in the thrombin-fibrinogen and thrombin-hirudin complexes. It is clear that the FRS plays a critical role in the linkage observed experimentally, since it is coupled to cation binding sites of either the enzyme and hirudin or fibrinogen. This explains why perturbation of the FRS in y-thrombin not only reduces the effect of salt concentration on binding and catalysis of synthetic amide substrates (De Cristofaro et al., 1992), but it also markedly reduces the affinity for hirudin and the sensitivity of K, to changes in NaCl concentration (Stone & Hofsteenge, 1991). The results reported here on thrombin-hirudin interaction can be used to further understand the energetics of components involved in thrombinfibrinogen interaction. The values for thrombinfibrinogen interaction reported in Figure 3 strictly reflect the change of K, for fibrinogen as a function of salt concentration. It should be pointed out that these values may actually reflect a change of the fibrinogen dissociation constant. The expression for K, in the case of fibrinogen can be written as FK, = FK,(l + Fk,/Fk- 1), where FK, is the fibrinogen dissociation constant, while Fkz and Fk-, are the rates for acylation and dissociation. The acylation

corresponds to k,,, for amide substrates. Two reasonable assumptions can be made as follows. One can assume that acylation has the same salt dependence for fibrinogen and synthetic amide substrates. and that the same situation holds for the dissociation rate in the case of fibrinogen and hirudin. The first assumption is supported by the similarity of the values of k,,, obtained for fibrinogen (Mihaly, 1988) and synthetic amide substrates (De Cristofaro & Di Cera, 1992). The second assumption is supported by the common mechanism of bridgebinding to the FRS and the catalytic pocket between fibrinogen and hirudin. Since the effect of salt concentration on k,,, for synthetic amide substrates saturates at 91 M, while the dissociation rate of hirudin is practically salt-independent, then one expects the ratio Fk2/Fkp i to be constant for salt concentrations above 91 M. Therefore, in the range 905 to 0.5 M, the change in K, for fibrinogen may, indeed, reflect a change in K,. This conclusion would apply regardless of whether fibrinogen is a sticky substrate or not, i.e. independent of the actual value of FkzfFk- 1. In this picture, bridgebinding to the FRS and the catalytic pocket would be linked to ion release through modulation of the association rate constant. A number of synthetic inhibitors have been developed that bridge-bind to thrombin in a fashion similar to fibrinogen and hirudin (Di Maio et al., 1990; Maraganore et al., 1990). Analysis of the action of these inhibitors as a function of salt concentration along the lines discussed here will be useful to test the molecular picture that seems to emerge for thrombinfibrinogen interaction. We are grateful to Drs Carl Frieden, Linda Kurz and Timothy Lohman for very helpful discussions. This work was supported in part by National Institutes of Health research grant HL 13160 (J.W.F.), National Institutes of Health research support grant RR05389 and a grant from the Lucille P. Markey Charitable Trust (E.D.C.).

References Cha, S. (1975). Tight-binding inhibitors I. B&hem. Pharmacol. 24, 2177-2185. Cha, S. (1976). Tight-binding inhibitors III. B&hem. Pharmacol. 25, 2695-2702. De Cristofaro, R. & Di Cera, E. (1990). Effect of protons on the amidase activity of human cc-thrombin. J. Mol. Biol. 216, 1077-1085. De Cristofaro, R. & Di Cera, E. (1992). Modulation of thrombin-fibrinogen interaction by specific ion effects. Biochemi8try, 31, 257-265. De Cristofaro, R., Fenton, J. W. & Di Cera, E. (1992). Linkage between proton binding and amidase activity in human y-thrombin. Biochemistry, in the press. Di Cera, E. (1990). Negative binding capacity and bistability in one-component systems. J. Chem. Phys. 92, 3241-3243. Di Cera, E. (1991). Stochastic linkage: effect of random fluctuations on a two-state process. J. Chem. Whys. 95, 5082-5086.

Thrombin-Hirudin Di Cera, E., De Cristofaro, R., Albright, D. J. & Fenton, J. W. (1991). Linkage between proton binding and amidase activity in human a-thrombin: effect of ions and temperature. Biochemistry, 30, 7913-7924. Di Maio, J., Gibbs, B., Munn, D., Lefebvre, J., Ni, F. & Konishi, Y. (1990). Bifunctional thrombin inhibitors based on the sequence of hirudin45-65. J. Biol. Chem. 265, 21698821703. Fenton, J. W., II, Fasco, M. G., Stackrow, A. B., Aronson, D. L., Young, A. M. & Finlayson, J. S. (1977). Human thrombins. J. Biol. Chem. 252, 35873598. Fenton, J. W., II, Olson, T. A., Zabinski, M. P. & Wilner, G. D. (1988). Anion-binding exosite of human a-thrombin and fibrin(ogen) recognition. Biochemistry, 27, 7106-7112. Frieden, C. (1970). Kinetic aspects of regulation of metabolic processes. J. Biol. Chem. 245, 5788-5799. Hantgan, R. R. & Hermans, J. (1979). Assembly of fibrin. J. Biol. Chem. 254, 11272-11281. Lottenberg, R. & Jackson. C. M. (1983). Solution composition dependent variation in extinction coefficients for p-nitroaniline. Biochim. Biophys. Acta, 742, 558-564. Maraganore, J. M., Bourdon, P., Jablonski, J., Ramachandran, K. L. & Fenton, J. W. (1990). Design and characterization of hirulogs: a novel class of bivalent peptide inhibitors of thrombin. Biochemistry, 29, 7095-7 101. Markwardt. F. (1970). Hirudin as an inhibitor of thrombin. Methods Enzymol. 19, 924-932. Mihaly, E. (1988). Clotting of bovine fibrinogen. Biochemistry, 27, 967-976. Naski, M. C.. Fenton, J. W., Maraganore, J. M., Olson, 6. T. & Shafer, J. A. (1990). The COOH-terminal

Edited

Interaction

269

domain of hirudin. J. Biol. Chem. 265, 13484-13489. Orthner, C. L. & Kosow, D. P. (1980). Evidence that human a-thrombin is a monovalent cat’ion-activated enzyme. Arch. Biochem. Biophys. 202, 63-75. Record, M. T., Anderson, C. F. & Lohman, T. M. (1978). Thermodynamic analysis of ion effects on the binding and conformational equilibria of proteins and nucleic acids: the roles of ion association or release, screening, and ion effects on water activity. Quart. Rev. Biophys. 11, 103-178. Rydel, T. J., Ravichandran, K. G., Tulinsky, A., Bode, W., Huber, R., Roitsch, C. & Fenton, J. W. (1990). The structure of a complex of recombinant hirudin and human u-thrombin. Science, 249, 277-280. Rydel, T. J., Tulinsky, A., Bode, W. & Huber, R. (1991). Refined structure of the hirudin-throm bin complex. J. Mol. Biol. 221, 583-601. Schmitz, T., Rothe, M. & Dodt, J. (1991). Mechanism of the inhibition of cr-thrombin by hirudin-derived fragments hirudirre4’ and hirudin45-65. Eur. J. Biochem. 195, 251-256. Stone, S. R. & Hofsteenge, J. (1986). Kinetics of the inhibition of thrombin by hirudin. Biochemistry, 25, 4622-4628. Stone, S. R. & Hofsteenge, J. (1991). Basis for the reduced affinity of fir- and yr-thrombin for hirudin. Biochemistry, 30, 395063955. Stone, S. R., Dennis, S. & Hofsteenge, J. (1989). Quantitative evaluation of the contribution of ionic interactions to the formation of the thrombinhirudin complex. Biochemistry, 28, 6857-6863. Williams, J. W. & Morrison, J. F. (1979). The kinetics of reversible tight-binding inhibition. Methods Enzymol. 63, 437-467.

by F. E. Cohen