Photochemistry und Phorubiology Vol. 51, No. I , pp. 37-43, 1990 Printed in Great Britain. All rights reserved
Copyright
0031-8655190 $03.00+0.00 P r e s plc
0 1990 Pergamon
PHOTOCOAGULATION OF HUMAN PLASMA: ACYL SERINE PROTEINASE PHOTOCHEMISTRY NED A . PORTER* and JOHN D. BRUHNKE Department of Chemistry, Duke University, Durham, NC 27706, USA (Received 16 June 1989;accepted 21 August 1989) Abstract-Human a-thrombin or bovine Factor Xa was acylated at the active site serine hydroxyl with a-methyl-2-hydroxy-4-diethylaminocinnamic acid. These modifed serine proteinase enzymes showed no plasma coagulation biological activity in the absence of light. Photolysis of the acyl serine proteinase enzymes in plasma for 1-35 s with monochromatic 366 nm light isolated from a high pressure mercury arc results in coagulation of the plasma. For example, photolysis of 3 NIH U of the acyl human athrombin for 5 s in human plasma results in a clot in 23 s. For comparison, 1 NIH U of unmodified human a-thrombin gave a clot in 21 s under the conditions of the assay but without photolysis. Appropriate controls showed that the coagulation is the result of the formation of active thrombin due to photodeacylation of the enzymes. The photoinduced clotting time measured is dependent on acyl thrombin concentration and photolysis time. Thus higher concentrations of acyl thrombin and longer photolysis times give a shorter clotting time. A kinetic scheme based upon Lineweaver-Burke analysis of the clotting process is developed.
be incapable of initiating coagulation while enzymes reactivated by photolysis could serve in this capacity. For blood coagulation, several zymogens are activated by serine proteinases formed earlier in the cascade and photoactivation of critical members of the cascade would lead to a clot. Thus, Factor Xa converts prothrombin to thrombin, while thrombin converts fibrinogen to fibrin and control of these and other proteinases by light would effectively control hemostasis. Several practical hurdles attend the application of this strategy for the coagulation of plasma. First, the chromophore of the modified enzyme must be such that it absorbs significant light intensity in plasma. Second, the photochemistry must be reasonably efficient so that effective concentrations of active enzyme can be generated during the photolysis period. Finally, the time course of the photoprocess must be reasonably short so that after absorption of a photon, the active enzyme is generated in a time period that is short relative to the coagulation assay. We report here the results of our studies with derivatives of a-methyl-2-hydroxy-4-diethylaminocinnamic acid, la. Serine proteinase enzymes of the coagulation cascade modified with this acid successfully contend with each of the practical hurdles outlined above. Thus, we demonstrate here the possibility of photocoagulation of plasma and we describe a kinetics scheme that accounts for the clotting times observed after photolysis of the acyl enzymes in plasma. 0
INTRODUCTION
The control of enzymatic activity by light has been the focus of much research activity over the past 25 yr (Martinek and Berezin, 1979; Erlanger, 1976). Several approaches have been taken to develop strategies for photo-control of enzyme catalytic activity. These approaches include: (1) control of p H and consequently enzyme activity by light; ( 2 ) control of the concentration of a coenzyme or critical cofactor by light; (3) control of the concentration of an enzyme competitive inhibitor by light; (4) control of the concentration of an irreversible enzyme inhibitor by light, and ( 5 ) modification of an enzyme with a photosensitive group that can be removed by light (Martinek et a / . , 1971; Berezin et a / . , 1980). We have recently reported a new approach for the control of activity of serine proteinase enzymes that falls into Category 5 (Turner et a/., 1987, 1988). This strategy relies on the conversion of the enzyme active site serine hydroxyl to an ohydroxy cinnamate substituted ester. The cinnamate ester is inactive since the critical active site serine hydroxyl is not available for catalytic activity but upon photolysis of the inactive acyl enzyme, enzyme activity returns. We have suggested that the photochemistry of enzyme activation in this case involves a photoisomerization-lactonization scheme that frees the active-site serine hydroxyl (see the “Discussion” section), and we have provided some evidence supporting this mechanism. We have also suggested that the application of this approach with the enzymes involved in the blood coagulation cascade offers the possibility of control of hemostasis with light. Thus, modified serine proteinases of the coagulation cascade would
la
*To whom correspondence should be addressed. tAbbreviafions: BSA. bovine serum albumin; RVV, Rus. sell’s viper venom; S-2222, N-Benzoyl-Ile-Glu-GlyArg-p-nitroanilide hydrochloride; S-2238, H-D-PhePip-Arg-p-nitroanilide dihydrochloride. PAP S1:l-D
R-H HCI
NH
1 b.
2
Scheme 1 31
38
NED A. PORTER and JOHN D. BRUHNKE MATERIALS AND METHODS
Reagents. Bovine serum albumin (BSA)t, Russell’s viper venom (RVV), and Sepharose 4B-CL were purchased from Sigma Chemical Co., St Louis, MO. Sephadex G-25 and G-100 were purchased from Pharmacia Fine Chemicals, Piscataway, NJ. Tris Ultra Pure was purchased from SchwarziMann Biotech, Cleveland, OH. Centricon 30 microconcentrators were purchased from Amicon Corp., Danvers, MA. The a-thrombin chromogenic substrate, H-D-Phe-Pip-Arg-p-nitroanilide dihydrochloride (S-2238), and the Factor Xa chromogenic substrate, N-Benzoyl-Ile-Glu-GIy-Arg-p-nitroanilide hydrochloride (and its methyl ester), S-2222, were purchased from Helena Laboratories, Beaumont, TX. Fresh frozen human plasma was obtained from the American Red Cross, Charlotte, NC. All other reagents were of the finest commercial grade available. Proteins. The human a-thrombin used was either purchased from Sigma Chemical Co., St Louis, MO or was the generous gift of Dr. Salvatore Pizzo, Duke University, Bovine and human Factor X were purchased from Sigma Chemical Co., St Louis, MO. Factor Xa was prepared by reacting sepharose-bound RVV-X enzyme with Factor X in the presence of Ca2+ at 37°C for 4 h in 50 mM Tris, 150 mM NaCI, pH 7.4. The solution was then centrifuged to remove the RVV-sepharose beads. The Factor X-activating fraction of Russell’s viper venom (RVV-X) was purified as described by Schiffman et al. (1969) and coupled to Sepharose 4B-CL according to the method of Porath et al. (1973). Protein concentrations. The activity of human a-thrombin was determined by chromogenic assay using S-2238 (KM=32.6 p M , k,.,=26.77 s-l). Typically, 100 FLe of 1 mM S-2238 (in H,O) were added to 20 pLe of enzyme solution in 1 me of 50 mM Tris, 250 mM NaCl pH 8.3 containing 1 mg me-’ BSA. The release of p-nitroanilide was monitored at 402 nm vs time. Using typical Michaelis-Menten kinetics, the AA4,>Jt is then proportional to the concentration of active a-thrombin (Hemker, 1983; Jackson and Coleman, 1981). The activity of Factor Xa was determined by chromogenic assay using S-2222 (KM=300 p M , k,,,=100 s-’) by an identical method used for thrombin. In vitro studies. A 5 M excess of 4’-amidinophenyl-4diethylamino-2-hydroxy-a-methylcinnamate, (N. A. Porter and J. D. Bruhnke, to be published) dissolved in Tris bufferimethanol, was added to thrombin or Factor Xa solution, and the activity was monitored as described above. When the activity was < 5%, the acyl enzyme was diluted with (1:l) 0.025 M Ca’+-Tris buffer pH 7.4 to the desired concentration used for coagulation studies (0.35 nM-0.175 p M ) . Coagulation studies. Clotting times were recorded on a COAG-A-MATE Dual channel instrument provided by Organon Teknika, Durham, NC. In all experiments, 100 pLe of human plasma were added to several sample cups in the instrument and allowed to reach 37°C. Then 400 pe of enzyme solution in (1:l) 0.25 M CaZ+-Tris buffer pH 7.4, at 39”C, were added and the timer started. The average clotting time and standard deviation for at least nine measurements per different concentration were determined. The enzyme concentrations ranged from 5 to 0.01 NIH U where 1 NIH U=0.028 p M in this procedure. A standard clotting time curve was made by plottin coa ulation time vs-l/[NIH U] for thrombin or I/& for Factor Xa. Photocoa&ation studies. Clotting times were recorded on a COAG-A-MATE Dual channel provided by Organon Teknika, Durham, NC. In all experiments, 100 pe of plasma were added to several sample cups in the instrument and allowed to reach 37°C. Then 400 p t of acyl enzyme solution in (1:l) 0.25 M CaZ+-Tris buffer pH 7.4, at 39”C, was added and allowed an 8 s mixing time before
the experiment. For a photocoagulation experiment, after the 8 s mixing time, the plasma-acyl enzyme solution was photolyzed from above with 366 nm light from a high pressure Hg arc lamp via a quartz light pipe and monochromator. The duration of photolysis time was controlled by a shutter positioned between monochromator and light pipe.The average photo-induced clotting time and standard deviation for at least nine measurements per different concentration were determined. The acyl enzyme concentrations ranged from 0.28 nM to 0.14 p M . The light source was a 500 W Hg high pressure arc. A Bausch and Lomb monochromator with 5 mm slits was used and the exit radiation was focused on a 1/4 X 18” quartz light pipe. No attempts were made to measure Zo in these experiments, but detailed quantum yield studies will be published at a later date. RESULTS
Inhibition of serine proteinases by l b Time courses of inhibition were determined for various concentrations of inhibitor l b and the enzymes human a-thrombin and bovine Factor Xa. All experients were conducted at room temperature in 50 m M Tris, 150 mM NaCl, p H 7.4. Percent activity of a-thrombin and Factor Xa were determined by chromogenic assay using the chromogens S-2238 (thrombin) and S-2222 (Factor Xa). Activity vs time curves for thrombin and Factor Xa are presented in Figs. 1 and 2. To follow reactivation of inhibited enzyme in the absence of light, the enzyme was incubated with excess inhibitor for several hours and the inhibited enzyme was then purified at room temperature on BSA-treated Sephadex (3-25. The gel filtration procedure was performed on a 20 X 0.5 cm column. Purified ‘acyl enzymes’ showed an absorption in the UV between 340 and 440 nm with h,,,-360. The E for both thrombin and Factor Xa acyl derivatives was -2 x lo4 at 360-370 nm. The return of enzyme activity was monitored by chromogenic assay and the first order activation subsequently yielded k3,
0
20
40
60
80
1W
120
time(min)
Figure 1. Activity of human thrombin vs time during inhibition by l b in pH 7.4 Tris buffer. Concentration of thrombin was 1.4 x lo-” M . A 4 : l O inhibit0r:enzyme; +-18:10 inhibitor:enzyme; 0-36: 10 inhibit0r:enzyme.
Plasma photocoagulation
39
Table 3. Clotting times for plasma coagulation initiated by Factor Xa
0
60
30
120
90
time (min)
Figure 2. Activity of bovine Factor Xa vs time during inhibition by l b in pH 7.4 Tris buffer. Concentration of Factor Xa was 9.5 x lo-' M . A-2:1 inhibitor:enzyme; +5:1 inhibitor:enzyme; U 7 : l inhibitor:enzyme. Table 1. Rate constants for inhibition of thrombin and Factor Xa with l b , and first-order rate constants for enzyme deacylation Enzyme Thrombin Factor Xa
k,/K, ( M - '
SKI)
1200 72
k , (s-')
X
10"
1.4 2.4
Table 2. Clotting times for plasma coagulation initiated by human thrombin Thrombin ( M x lox)*
NIH units
Clotting time (s)
0.10 0.14 0.2 2.15 2.57 2.80 5.6 8.4 11.2
0.35 0.50 0.71 0.76 0.91 1.0 2.0 3.0 4.0
39.3 t 0.8 28.4 2 0.6 20.5 ? 0.4 23.3 ? 0.6 20.0 ? 1.1 17.6 ? 0.7 10.9 t 0.5 9.9 2 0.5 9.8 ? 0.6
*Final concentration of enzyme in coagulation assay. the rate constant for hydrolysis of the acyl enzyme. The measured k3 values and a simple numerical integration routine enabled us to curve-fit the time course plots (Figs. 1 and 2) and the second order rate constant of inhibition (Kam et al., 1988) was thus determined ( k 2 / K L )The . uncatalyzed pseudofirst order hydrolysis of the inhibitor was measured independently, (k=1.96 x lo-' s-I), and the numerical integration was corrected for this background hydrolysis. The rate constants determined for inhibition and enzyme deacylation are presented in Table 1 and the best-fit time course plots that use these rates in the numerical integration are presented as solid lines in Figs. 1 and 2.
Factor Xa ( M x lo")*
NIH units
5.6 2.8 1.4 1.4t 1.12 0.70 0.70t 0.28 0.21 0.21t 0.14 0.14t 0.07 0.07t 0.028
2.0 1.0 0.5 0.5 0.4 0.25 0.25 0.10 0.075 0.075 0.05 0.05 0.025 0.025 0.10
Clotting time
(s)
20.5 2 0.3 20.6 5 0.2 21.1 5 0.3 22.1 5 0.3 21.3 2 0.3 21.9 5 1.2 23.9 ? 0.6 25.3 2 1.6 26.8 5 0.4 28.1 t 0.6 28.2 2 0.3 30.9 2 0.6 30.8 2 0.6 35.7 2 0.4 40.5 5 0.8
*Final concentration of enzyme in coagulation assay. tHuman Xa where indicated, otherwise bovine Xa.
Thrombin clotting times and Factor Xa clotting times Standard clotting times were obtained for thrombin and Factor Xa initiated clotting. No appreciable difference in clotting behavior was observed between human and bovine Factor X a (although the human enzyme appeared to be marginally less effective) and the bovine enzyme was used in subsequent investigations of photocoagulation. The amount of thrombin and Factor Xa added to the plasma for a clotting time measurement was determined by standard chromogenic assay.
Photocoagulation of human plasma with acyl thrombin and acyl Factor Xa The serine proteinase enzyme was treated with a 50-fold excess of the ester l b and the resulting inactive enzyme was purified by gel filtration. Use of the acylated serine proteinase enzymes in the standard clotting assays and in the absence of light gave no coagulation during the time course of an assay (120 s). Five NIH U of an acyl thrombin that had been purified by gel filtration but had 9 % residual activity as determined by chromogenic assay led to a clotting time of 32 s. This is in accord with a residual activity of approx. 0.45 NIH U of residual activity ( 5 U x 0.09). Photolysis with 366 nm light of plasma to which inactive acyl enzyme had been added led to a subsequent formation of a clot. Photolysis of the plasma under the same conditions but without added acyl enzyme led to no clotting. A n alternate and experimentally simpler approach was to react the serine proteinase with a 5-fold excess of the inhibitor l b and then to follow the decline in enzyme activity over time until inhibition was essentially complete (enzyme activity
NED A. PORTER and JOHN D. BRUHNKE
40
< 2%). This enzyme solution was used directly without gel filtration purification. In this approach, the enzyme solution added to plasma had excess inhibitor present as well as the by-product of the inhibition process, p-amidinophenol. In all of our studies of photocoagulation, we were unable to detect any difference in the behavior of the acyl enzyme prepared by this method and the acyl enzyme prepared by gel filtration. For this reason most of our work was carried out with the acyl enzyme prepared by the simpler method. In Tables 4 and 5 are presented the clotting times determined for acyl enzyme photo-initiated plasma coagulation. Clotting times obtained after photolysis of acyl enzymes in plasma were dependent on the amount of acyl enzyme present and on the duration of photolysis of the plasma.
series of reactions that produce a fibrin clot within a damaged blood vessel. Serine proteinase enzymes function by hydrolyzing peptide bonds and the mechanism of this hydrolysis involves acylation of the active site serine hydroxyl of the enzyme by the target amide acyl group (Scheme 2). Normal turnover of the enzyme involves hydrolysis of the acyl enzyme in the second step, k3 in Scheme 2a. Amidino esters act as inhibitors (Fujioka et al., 1982; Tanizawa et a l . , 1987; Markwardt et al., 1980) of the coagulation serine proteinase enzymes and
HzNR'
DISCUSSION
Thrombin and Factor Xa belong to an important subclass of serine proteinase enzymes that are vitamin K dependent. These enzymes, along with other serine proteinase clotting factors, mediate the
Scheme 2
Table 4. Clotting times for plasma coagulation initiated by photolysis of acyl thrombin NIH units of acyl thrombin*
Photolysis time (s)
Clotting time (s)
Calculated clotting time (s)+
5.0 5.0 5.0 3.0 3.0 3.0 3.0 2.0 1.o
13.5t 5.0 3.0 16.7t 5.0 4.0 3.0 23.lt 34.6t
13.5 ? 0.24 16.6 ? 0.48 18.7 ? 0.06 16.7 t 0.16 21.6 2 1.50 22.1 ? 1.80 25.8 ? 1.20 23.1 ? 0.22 34.6 ? 1.10
13.6 15.8 19.6 16.6 21.2 23.6 27.8 19.8 27.8
*1 NIH U = 0.028 p,M in this procedure. tcontinuous photolysis. $Calculated by numerical integration (see 'Discussion' section).
Table 5. Clotting times for plasma coagulation initiated by photolysis of acyl Factor Xa NIH units of acyl Factor Xa* 1.0 1 .o 1.0 1.0 0.25 0.25 0.25 0.25 0.1 0.075 0.075 0.075
0.05
Photolysis time (s)
Clotting time (s)
Calculated clotting time (s)+
21.91 3.0 2.0 1.0 25.0t 3.0 2.0 1.o 26.5t 27.9t 3.0 1.0 28.91.
21.9 ? 0.70 22.4 2 0.15 22.9 ? 0.21 26.2 t 0.20 25.0 ? 0.55 26.6 ? 0.44 28.2 ? 0.44 32.3 ? 0.95 26.5 t 0.52 27.9 0.52 31.7 ? 1.10 38.7 ? 1.10 28.9 t 1.20
21.3 22.5 23.2 24.8 24.1 26.6 28.0 31.3 27.1 28.3 33.2 41.9 30.4
*
'1 NIH U = 0.028 p M in this procedure. tcontinuous photolysis. $Calculated by numerical integration (see 'Discussion' section).
Plasma photocoagulation
Scheme 3 we have developed a strategy for photoactivation of these enzymes that relies on the use of amidinophenol esters of cinnamic acid to form hydrolytically stable acyl enzymes (Scheme 2b). Thus, we have previously reported that thrombin and Factor Xa esters of a-methyl-o-hydroxycinnamic acid can be prepared and that these acyl enzymes show no activity in standard chromogenic assays if kept in the dark (Turner e t a / . , 1987, 1988). Photolysis through Pyrex of these acyl enzymes with broad-band light from a high pressure mercury lamp returns enzyme activity. The proposed mechanism of this enzyme photoactivation is presented in Scheme 3 and involves photoisomerization of the cinnamate ester followed by lactonization of the cis cinnamate. This lactonization leads to a methyl coumarin and the active enzyme. The p-amidinophenyl ester of a-methyl-o-hydroxycinnamic acid was thus shown to inhibit thrombin and Factor Xa and the inhibition was reversed by photolysis with Pyrex filtered light from a high pressure mercury lamp. There were, however, several problems with this inhibition-photoactivation strategy utilizing the parent ester, Scheme 3 where R=H. The enzyme inhibition was temporary and enzyme activity returned in the course of a few hours in the dark, i.e. enzyme turnover was relatively fast. Furthermore, the photoactivation of these enzymes was slow (10-1.5 min) and required light intensities and wavelengths such that appreciable enzyme degradation occurred during photoactivation. Because of these problems, we have prepared several derivatives of a-methyl-o-hydroxycinnamic acid and have studied, in detail, their properties. Our best inhibitor, l b , where R=Et2N in Scheme 2b, forms an acyl enzyme that is hydrolytically more stable than the parent compound where R=H. While l b is a better inhibitor of thrombin than of Factor Xa, the thrombin acyl enzyme deacylates slower than the Factor Xa acyl enzyme. Furthermore, esters of this substituted cinnamate have an absorption spectrum with a ,,A at 360 nm, a region of the spectrum that is readily accessible and where protein has no significant absorption. In order to study the photocoagulation of plasma with the acyl enzymes, we first carried out standard coagulation assays of plasma with the active enzyme. Such standard coagulation assays have been studied in great detail and a kinetic analysis of coagulations initiated by one of the active enzymes in the cascade has been published (Hemker and Muller, 1968; Hemker et al., 1965). In this analysis, it is assumed that a clot is detected when
41
the concentration of a product in the cascade reaches a critical level. Thus, the clotting time, reaction velocity, and the critical product concentration are related as in Eq. 1 where t,=the clotting time,v=the reaction velocity and [P],=the critical concentration of product P formed in the enzymatic cascade.
r,
x v = [PIc
(1)
Using a modified Lineweaver-Burke analysis of the coagulation cascade, Hemker et af. (1965) have suggested that the clotting time and the concentration of active units of thrombin in the cascade, [Ha] in our presentation, can be expressed by Eq. 2 1, =
l/v =
+ B/[IIa] A/[PIc+ B/{[IIa] x [PI,.} A
(2)
(3)
with A and B being constants. Insertion of Eq. 1 into Eq. 2 gives an expression for l / v as a function of units of added enzyme, thrombin in our studies, and the critical concentration [PIc.A plot of clotting time vs l/[NIH U] for thrombin is presented in Fig. 3. The best fit parameters for the data presented for thrombin is t'. = 6.1 + 11.6 x l/[IIa] with a correlation coefficient of 0.99. The approach suggested by Hemker is thus shown to be valid in our system and this analysis is extended, vide infra, in our studies of photocoagulation. Factor Xa initiates a two component cascade in which thrombin is generated from prothrombin in the first step and fibrin is formed in the second step. For a two component cascade, Hemker and Muller (1968) suggest that the clotting time should be related to the active enzyme by an inverse squareroot relationship. Thus, as shown in Fig. 4,we find a best fit for the data of t, = 18.2 (2.17 x l/ 45
-I
0 0
1
a
1 3
Figure 3. Clotting times in seconds vs l/[thrombin] (top scale) and l/[NIH U thrombin) (bottom scale). Best-fit correlation is clotting time = 6.1 2 11.6 X l/[thrombin] where [thrombin] is the number of NIH units of thrombin (rz > 0.99).
NED A. PORTER and JOHN D. BRUHNKE
42
5o
= 2.303 +Illel (B. Faust, personal communication; Zepp, 1978).
1
d[IIa]ldt = -d[Acyl IIa]/dt = k[Acyl IIa]
"1 10
0
4 0
.
I
2
.
,
.
4
I
6
.
I
8
.
1
10
1 v'[ Factor X a l
Figure 4. Clotting times in seconds w l / q F a c t o r Xa]. Best-fit correlation is clotting time = 18.2 f (2.17 x I / where [Xa] is the number of NIH units of Factor Xa (Tz > 0.99).
m)
m)
where [Xa] i-s the number of N I H U of Factor Xa (9> 0.99). Photocoagulation in plasma is efficient using the acyl modified thrombin or Factor Xa (Tables 4 and 5). Photolysis of 5 U of the acyl thrombin for 5 s, for example, results in a clot in 16.6 s, a clotting time shorter than that obtained from 1 U of the fully active enzyme in the standard assay. Photolysis of 1 U of acyl Factor Xa for 1 s results in a clotting time of 26 s. Furthermore, the light delivery system used in these studies was not maximized and we have every reason to believe that significant improvements in light delivery could lead to shorter required photolysis times. Analysis of the clotting times obtained during photolysis by a Lineweaver-Burke approach analogous to that of Hemker et al. (1965) is instructive. If we assume that the concentration of active species P at any time t is given by [PI = t x v (see Eq. 1) and that l l v is given by Eq. 3 for thrombin, we obtain Eq. 4. When [P]/[PIc = 1, a clot forms and t = t,. The additional feature
(4) to this analysis compared to Hemker is that in the photocoagulation experiment, [IIa] is initially zero and increases during the period of photolysis. This must be taken into account in our analysis of photocoagulation. At the low concentrations of acyl enzyme used in this study, the increase in concentration of thrombin or Factor Xa is first order dependent on the concentration of the acyl enzyme [Acyl IIa] or [Acyl Xa] and also depends on the quantum yield for the photoprocess 4 by which active enzyme is generated, the intensity of incident irradiation, I,,, the molar extinction coefficient of the acyl enzyme, e , and the pathlength of the sample. This relationship is given in Eq. 5, where k
(5)
By the use of Eqs. 4 and 5, we have developed a simple numerical integration computer analysis of thrombin photocoagulation. The numerical integration calculates the increase of [IIa] and decrease of [Acyl Ha] over a small time increment, At, after guessing a value of k . The increase in [P]/[PIcis calculated over At using Eqs. 4 and 5 and this process is then repeated over another small time increment. When [P]/[PIc= 1, the integration is stopped and the time is reported as the clotting time. No increase in [IIa] is calculated if the light is off and the clotting time can thus be calculated for continuous photolysis throughout the course of the assay or for short-term photolysis and a subsequent dark period. With this numerical integration approach and guessing a value for k in the photoactivation scheme, we were able to get a reasonably good fit to the data presented in Table 4 for thrombin photoactivation. The calculated clotting times using k = 0.07 s-I are presented in Table 4 for thrombin photoactivation and the results are remarkably good given the number of assumptions that are made in the analysis. A similar analysis is used in the photocoagulation initiated by photolysis of the acyl Factor Xa enzyme. In this approach, a numerical integration program has been written analogous to the one described above for thrombin photoactivation, but with Eq. 4 replaced by Eq. 6. With this
approach and guessing a value for k of 0.12 s-' in the photoactivation scheme, we were able to get a reasonably good fit to the data provided in Table 5 for Factor Xa photoactivation. The results are also excellent for this analysis given the assumptions, and our success in fitting the photolysis data using a l/[IIa] correlation for the one-step thrombin process and l/mcorrelation for the two-step cascade lends credibility to the analysis. Comparison of the best fit k values for thrombin (0.07) and Factor Xa (0.12) also indicates no dramatic difference for the two analogous photochemical processes. The differences noted could be the result of more efficient photoprocess for Factor Xa and/or a higher molar absorptivity for the Factor Xa acyl enzyme, since k = 2.303 +Il,El. CONCLUSIONS
Serine proteinase enzymes of the coagulation cascade modified at the active site serine hydroxyl can be activated efficiently with light. Furthermore,
, '
Plasma photocoagulation
coagulation of plasma can result from photolysis of an acyl enzyme in plasma during the time-course of the clotting experient. Thus, photolysis of 1 U of a modified Factor Xa enzyme for as short a time as 1 s with monochromatic 366 nm light results in clot formation in 26 s in a plasma where 1 U of thrombin gives a clotting time of 18 s. The photochemical time course is therefore short enough and the efficiency of photoactivation is high enough so that this strategy can lead to a significant biological response to a modest photostimulus. Acknowledgemenrs-We gratefully acknowledge support from NIH (HL-17921) and gifts of serine proteinase enzymes from Dr. S . Pizzo. We also thank Dr. B . Faust for stimulating discussions. REFERENCES
Berezin, I . V., N. F. Kazanskaya, R. B. Aisina and E. V. Lukasheva (1980) Detection of light signals by microencapsulated cis-cinnamoyl chymotrypsin. Enzyme Microb. Technol. 2, 15G154. Erlanger, B. F. (1976) Photoregulation of biologically active macromolecules. Annu. Rev. Biochem. 45, 267-283. Fujioka, T., K. Tanizawa and Y. Kanoaka (1982) Inverse substrates-XV. Spectrometric properties of fluorescence-labeled acyl trypsins. Chem. Pharmacol. Bull. 30, 230-236. Hemker, H . C. (1983) Handbook of Synthetic Substrates for the Coagulation and Fibrinolytic System, pp. 27-44. Martinus-Nijhoff, Boston, MA. Hernker, H . C., P. W. Hernker and E.A. Loeliger (1965) Kinetic aspects of the interaction of blood clotting enzymes-I. Derivation of basic formulas. Thromb. Diath. Haemorrh. 13, 155-175. Hemker, H. C. and A. D . Muller (1968) Kinetic aspects
43
of the interaction of blood clotting enzymes-V. The reaction mechanism of the extrinsic clotting system as revealed by the kinetics of one-stage estimations of coagulation enzymes. Thromb. Diath. 19, 368-382. Jackson, C. M. and P. L. Coleman (1981) Assay of coagulation proteases using peptide chromogenic and gluorogenic substrates. Methods Enzymol. 80, 341-361. Kam, C.-M., K. Fujikawa and J . C. Powers (1988) Mechanism-based isocoumarin inhibitors for trypsin and blood coagulation serine proteases: new anticoagulants. Biochemistry 21, 2547-2557. Markwardt, F., G . Wagner, J. Sturzebecher and P. Walsmann (1980) N-a-arylsulfonyl-w-(4-amidinophenyl)aminoalkylcarboxylic acid amides-novel selective inhibitors of thrombin. Thromb. Res. 17, 425-430. Martinek, K. and I. V. Berezin (1979) Artificial lightsensitive enzymatic systems as chemical amplifiers of weak light signals. Photochem. Photobiol. 29, 637-649. Martinek, K., S. D. Varfolomeyev and I. V. Berezin (1971) Interaction of a-chymotrypain with N-cinnamoylimidazole, substrate sensitive to light. Eur. J . Biochcm. 19, 242-249. Porath, J., K. Aspberg, H. Drevin and R. Axen (1973) Preparation of cyanogen bromide-activated agarose gels. J . Chromatogr. 86, 53-56. Schiffman, S., I . Theodor and S . I. Rapaport (1969) Separation from Russell’s viper venom of one fraction reacting with Factor X and another reacting with Factor V. Biochemistry 8, 1397-1405. Tanizawa, K . , Y. Kanaoka and W. B. Lawson (1987) Inverse substrates for trypsin and trypsin-like enzymes. Acc. Chem. Res. 20, 337-342. Turner, A. D . , S. V. Pizzo, G . W. Rozakis and N. A . Porter (1987) Photochemical activation of acylated a thrombin. J . Am . Chem. Soc. 109, 1274-127s. Turner, A. D . , S . V. Pizzo, G . Rozakis and N. A . Porter (1988) Photoreactivation of irreversibly inhibited serine proteinases. J . Am. Chem. Soc. 110, 244-250. Zepp, R. G . (1978) Quantum yields for reaction of pollutants in dilute aqueous solution. Environ. Sci. Technol. 12, 327-329.
Copyright
0031-8655190 $03.00+0.00 P r e s plc
0 1990 Pergamon
PHOTOCOAGULATION OF HUMAN PLASMA: ACYL SERINE PROTEINASE PHOTOCHEMISTRY NED A . PORTER* and JOHN D. BRUHNKE Department of Chemistry, Duke University, Durham, NC 27706, USA (Received 16 June 1989;accepted 21 August 1989) Abstract-Human a-thrombin or bovine Factor Xa was acylated at the active site serine hydroxyl with a-methyl-2-hydroxy-4-diethylaminocinnamic acid. These modifed serine proteinase enzymes showed no plasma coagulation biological activity in the absence of light. Photolysis of the acyl serine proteinase enzymes in plasma for 1-35 s with monochromatic 366 nm light isolated from a high pressure mercury arc results in coagulation of the plasma. For example, photolysis of 3 NIH U of the acyl human athrombin for 5 s in human plasma results in a clot in 23 s. For comparison, 1 NIH U of unmodified human a-thrombin gave a clot in 21 s under the conditions of the assay but without photolysis. Appropriate controls showed that the coagulation is the result of the formation of active thrombin due to photodeacylation of the enzymes. The photoinduced clotting time measured is dependent on acyl thrombin concentration and photolysis time. Thus higher concentrations of acyl thrombin and longer photolysis times give a shorter clotting time. A kinetic scheme based upon Lineweaver-Burke analysis of the clotting process is developed.
be incapable of initiating coagulation while enzymes reactivated by photolysis could serve in this capacity. For blood coagulation, several zymogens are activated by serine proteinases formed earlier in the cascade and photoactivation of critical members of the cascade would lead to a clot. Thus, Factor Xa converts prothrombin to thrombin, while thrombin converts fibrinogen to fibrin and control of these and other proteinases by light would effectively control hemostasis. Several practical hurdles attend the application of this strategy for the coagulation of plasma. First, the chromophore of the modified enzyme must be such that it absorbs significant light intensity in plasma. Second, the photochemistry must be reasonably efficient so that effective concentrations of active enzyme can be generated during the photolysis period. Finally, the time course of the photoprocess must be reasonably short so that after absorption of a photon, the active enzyme is generated in a time period that is short relative to the coagulation assay. We report here the results of our studies with derivatives of a-methyl-2-hydroxy-4-diethylaminocinnamic acid, la. Serine proteinase enzymes of the coagulation cascade modified with this acid successfully contend with each of the practical hurdles outlined above. Thus, we demonstrate here the possibility of photocoagulation of plasma and we describe a kinetics scheme that accounts for the clotting times observed after photolysis of the acyl enzymes in plasma. 0
INTRODUCTION
The control of enzymatic activity by light has been the focus of much research activity over the past 25 yr (Martinek and Berezin, 1979; Erlanger, 1976). Several approaches have been taken to develop strategies for photo-control of enzyme catalytic activity. These approaches include: (1) control of p H and consequently enzyme activity by light; ( 2 ) control of the concentration of a coenzyme or critical cofactor by light; (3) control of the concentration of an enzyme competitive inhibitor by light; (4) control of the concentration of an irreversible enzyme inhibitor by light, and ( 5 ) modification of an enzyme with a photosensitive group that can be removed by light (Martinek et a / . , 1971; Berezin et a / . , 1980). We have recently reported a new approach for the control of activity of serine proteinase enzymes that falls into Category 5 (Turner et a/., 1987, 1988). This strategy relies on the conversion of the enzyme active site serine hydroxyl to an ohydroxy cinnamate substituted ester. The cinnamate ester is inactive since the critical active site serine hydroxyl is not available for catalytic activity but upon photolysis of the inactive acyl enzyme, enzyme activity returns. We have suggested that the photochemistry of enzyme activation in this case involves a photoisomerization-lactonization scheme that frees the active-site serine hydroxyl (see the “Discussion” section), and we have provided some evidence supporting this mechanism. We have also suggested that the application of this approach with the enzymes involved in the blood coagulation cascade offers the possibility of control of hemostasis with light. Thus, modified serine proteinases of the coagulation cascade would
la
*To whom correspondence should be addressed. tAbbreviafions: BSA. bovine serum albumin; RVV, Rus. sell’s viper venom; S-2222, N-Benzoyl-Ile-Glu-GlyArg-p-nitroanilide hydrochloride; S-2238, H-D-PhePip-Arg-p-nitroanilide dihydrochloride. PAP S1:l-D
R-H HCI
NH
1 b.
2
Scheme 1 31
38
NED A. PORTER and JOHN D. BRUHNKE MATERIALS AND METHODS
Reagents. Bovine serum albumin (BSA)t, Russell’s viper venom (RVV), and Sepharose 4B-CL were purchased from Sigma Chemical Co., St Louis, MO. Sephadex G-25 and G-100 were purchased from Pharmacia Fine Chemicals, Piscataway, NJ. Tris Ultra Pure was purchased from SchwarziMann Biotech, Cleveland, OH. Centricon 30 microconcentrators were purchased from Amicon Corp., Danvers, MA. The a-thrombin chromogenic substrate, H-D-Phe-Pip-Arg-p-nitroanilide dihydrochloride (S-2238), and the Factor Xa chromogenic substrate, N-Benzoyl-Ile-Glu-GIy-Arg-p-nitroanilide hydrochloride (and its methyl ester), S-2222, were purchased from Helena Laboratories, Beaumont, TX. Fresh frozen human plasma was obtained from the American Red Cross, Charlotte, NC. All other reagents were of the finest commercial grade available. Proteins. The human a-thrombin used was either purchased from Sigma Chemical Co., St Louis, MO or was the generous gift of Dr. Salvatore Pizzo, Duke University, Bovine and human Factor X were purchased from Sigma Chemical Co., St Louis, MO. Factor Xa was prepared by reacting sepharose-bound RVV-X enzyme with Factor X in the presence of Ca2+ at 37°C for 4 h in 50 mM Tris, 150 mM NaCI, pH 7.4. The solution was then centrifuged to remove the RVV-sepharose beads. The Factor X-activating fraction of Russell’s viper venom (RVV-X) was purified as described by Schiffman et al. (1969) and coupled to Sepharose 4B-CL according to the method of Porath et al. (1973). Protein concentrations. The activity of human a-thrombin was determined by chromogenic assay using S-2238 (KM=32.6 p M , k,.,=26.77 s-l). Typically, 100 FLe of 1 mM S-2238 (in H,O) were added to 20 pLe of enzyme solution in 1 me of 50 mM Tris, 250 mM NaCl pH 8.3 containing 1 mg me-’ BSA. The release of p-nitroanilide was monitored at 402 nm vs time. Using typical Michaelis-Menten kinetics, the AA4,>Jt is then proportional to the concentration of active a-thrombin (Hemker, 1983; Jackson and Coleman, 1981). The activity of Factor Xa was determined by chromogenic assay using S-2222 (KM=300 p M , k,,,=100 s-’) by an identical method used for thrombin. In vitro studies. A 5 M excess of 4’-amidinophenyl-4diethylamino-2-hydroxy-a-methylcinnamate, (N. A. Porter and J. D. Bruhnke, to be published) dissolved in Tris bufferimethanol, was added to thrombin or Factor Xa solution, and the activity was monitored as described above. When the activity was < 5%, the acyl enzyme was diluted with (1:l) 0.025 M Ca’+-Tris buffer pH 7.4 to the desired concentration used for coagulation studies (0.35 nM-0.175 p M ) . Coagulation studies. Clotting times were recorded on a COAG-A-MATE Dual channel instrument provided by Organon Teknika, Durham, NC. In all experiments, 100 pLe of human plasma were added to several sample cups in the instrument and allowed to reach 37°C. Then 400 pe of enzyme solution in (1:l) 0.25 M CaZ+-Tris buffer pH 7.4, at 39”C, were added and the timer started. The average clotting time and standard deviation for at least nine measurements per different concentration were determined. The enzyme concentrations ranged from 5 to 0.01 NIH U where 1 NIH U=0.028 p M in this procedure. A standard clotting time curve was made by plottin coa ulation time vs-l/[NIH U] for thrombin or I/& for Factor Xa. Photocoa&ation studies. Clotting times were recorded on a COAG-A-MATE Dual channel provided by Organon Teknika, Durham, NC. In all experiments, 100 pe of plasma were added to several sample cups in the instrument and allowed to reach 37°C. Then 400 p t of acyl enzyme solution in (1:l) 0.25 M CaZ+-Tris buffer pH 7.4, at 39”C, was added and allowed an 8 s mixing time before
the experiment. For a photocoagulation experiment, after the 8 s mixing time, the plasma-acyl enzyme solution was photolyzed from above with 366 nm light from a high pressure Hg arc lamp via a quartz light pipe and monochromator. The duration of photolysis time was controlled by a shutter positioned between monochromator and light pipe.The average photo-induced clotting time and standard deviation for at least nine measurements per different concentration were determined. The acyl enzyme concentrations ranged from 0.28 nM to 0.14 p M . The light source was a 500 W Hg high pressure arc. A Bausch and Lomb monochromator with 5 mm slits was used and the exit radiation was focused on a 1/4 X 18” quartz light pipe. No attempts were made to measure Zo in these experiments, but detailed quantum yield studies will be published at a later date. RESULTS
Inhibition of serine proteinases by l b Time courses of inhibition were determined for various concentrations of inhibitor l b and the enzymes human a-thrombin and bovine Factor Xa. All experients were conducted at room temperature in 50 m M Tris, 150 mM NaCl, p H 7.4. Percent activity of a-thrombin and Factor Xa were determined by chromogenic assay using the chromogens S-2238 (thrombin) and S-2222 (Factor Xa). Activity vs time curves for thrombin and Factor Xa are presented in Figs. 1 and 2. To follow reactivation of inhibited enzyme in the absence of light, the enzyme was incubated with excess inhibitor for several hours and the inhibited enzyme was then purified at room temperature on BSA-treated Sephadex (3-25. The gel filtration procedure was performed on a 20 X 0.5 cm column. Purified ‘acyl enzymes’ showed an absorption in the UV between 340 and 440 nm with h,,,-360. The E for both thrombin and Factor Xa acyl derivatives was -2 x lo4 at 360-370 nm. The return of enzyme activity was monitored by chromogenic assay and the first order activation subsequently yielded k3,
0
20
40
60
80
1W
120
time(min)
Figure 1. Activity of human thrombin vs time during inhibition by l b in pH 7.4 Tris buffer. Concentration of thrombin was 1.4 x lo-” M . A 4 : l O inhibit0r:enzyme; +-18:10 inhibitor:enzyme; 0-36: 10 inhibit0r:enzyme.
Plasma photocoagulation
39
Table 3. Clotting times for plasma coagulation initiated by Factor Xa
0
60
30
120
90
time (min)
Figure 2. Activity of bovine Factor Xa vs time during inhibition by l b in pH 7.4 Tris buffer. Concentration of Factor Xa was 9.5 x lo-' M . A-2:1 inhibitor:enzyme; +5:1 inhibitor:enzyme; U 7 : l inhibitor:enzyme. Table 1. Rate constants for inhibition of thrombin and Factor Xa with l b , and first-order rate constants for enzyme deacylation Enzyme Thrombin Factor Xa
k,/K, ( M - '
SKI)
1200 72
k , (s-')
X
10"
1.4 2.4
Table 2. Clotting times for plasma coagulation initiated by human thrombin Thrombin ( M x lox)*
NIH units
Clotting time (s)
0.10 0.14 0.2 2.15 2.57 2.80 5.6 8.4 11.2
0.35 0.50 0.71 0.76 0.91 1.0 2.0 3.0 4.0
39.3 t 0.8 28.4 2 0.6 20.5 ? 0.4 23.3 ? 0.6 20.0 ? 1.1 17.6 ? 0.7 10.9 t 0.5 9.9 2 0.5 9.8 ? 0.6
*Final concentration of enzyme in coagulation assay. the rate constant for hydrolysis of the acyl enzyme. The measured k3 values and a simple numerical integration routine enabled us to curve-fit the time course plots (Figs. 1 and 2) and the second order rate constant of inhibition (Kam et al., 1988) was thus determined ( k 2 / K L )The . uncatalyzed pseudofirst order hydrolysis of the inhibitor was measured independently, (k=1.96 x lo-' s-I), and the numerical integration was corrected for this background hydrolysis. The rate constants determined for inhibition and enzyme deacylation are presented in Table 1 and the best-fit time course plots that use these rates in the numerical integration are presented as solid lines in Figs. 1 and 2.
Factor Xa ( M x lo")*
NIH units
5.6 2.8 1.4 1.4t 1.12 0.70 0.70t 0.28 0.21 0.21t 0.14 0.14t 0.07 0.07t 0.028
2.0 1.0 0.5 0.5 0.4 0.25 0.25 0.10 0.075 0.075 0.05 0.05 0.025 0.025 0.10
Clotting time
(s)
20.5 2 0.3 20.6 5 0.2 21.1 5 0.3 22.1 5 0.3 21.3 2 0.3 21.9 5 1.2 23.9 ? 0.6 25.3 2 1.6 26.8 5 0.4 28.1 t 0.6 28.2 2 0.3 30.9 2 0.6 30.8 2 0.6 35.7 2 0.4 40.5 5 0.8
*Final concentration of enzyme in coagulation assay. tHuman Xa where indicated, otherwise bovine Xa.
Thrombin clotting times and Factor Xa clotting times Standard clotting times were obtained for thrombin and Factor Xa initiated clotting. No appreciable difference in clotting behavior was observed between human and bovine Factor X a (although the human enzyme appeared to be marginally less effective) and the bovine enzyme was used in subsequent investigations of photocoagulation. The amount of thrombin and Factor Xa added to the plasma for a clotting time measurement was determined by standard chromogenic assay.
Photocoagulation of human plasma with acyl thrombin and acyl Factor Xa The serine proteinase enzyme was treated with a 50-fold excess of the ester l b and the resulting inactive enzyme was purified by gel filtration. Use of the acylated serine proteinase enzymes in the standard clotting assays and in the absence of light gave no coagulation during the time course of an assay (120 s). Five NIH U of an acyl thrombin that had been purified by gel filtration but had 9 % residual activity as determined by chromogenic assay led to a clotting time of 32 s. This is in accord with a residual activity of approx. 0.45 NIH U of residual activity ( 5 U x 0.09). Photolysis with 366 nm light of plasma to which inactive acyl enzyme had been added led to a subsequent formation of a clot. Photolysis of the plasma under the same conditions but without added acyl enzyme led to no clotting. A n alternate and experimentally simpler approach was to react the serine proteinase with a 5-fold excess of the inhibitor l b and then to follow the decline in enzyme activity over time until inhibition was essentially complete (enzyme activity
NED A. PORTER and JOHN D. BRUHNKE
40
< 2%). This enzyme solution was used directly without gel filtration purification. In this approach, the enzyme solution added to plasma had excess inhibitor present as well as the by-product of the inhibition process, p-amidinophenol. In all of our studies of photocoagulation, we were unable to detect any difference in the behavior of the acyl enzyme prepared by this method and the acyl enzyme prepared by gel filtration. For this reason most of our work was carried out with the acyl enzyme prepared by the simpler method. In Tables 4 and 5 are presented the clotting times determined for acyl enzyme photo-initiated plasma coagulation. Clotting times obtained after photolysis of acyl enzymes in plasma were dependent on the amount of acyl enzyme present and on the duration of photolysis of the plasma.
series of reactions that produce a fibrin clot within a damaged blood vessel. Serine proteinase enzymes function by hydrolyzing peptide bonds and the mechanism of this hydrolysis involves acylation of the active site serine hydroxyl of the enzyme by the target amide acyl group (Scheme 2). Normal turnover of the enzyme involves hydrolysis of the acyl enzyme in the second step, k3 in Scheme 2a. Amidino esters act as inhibitors (Fujioka et al., 1982; Tanizawa et a l . , 1987; Markwardt et al., 1980) of the coagulation serine proteinase enzymes and
HzNR'
DISCUSSION
Thrombin and Factor Xa belong to an important subclass of serine proteinase enzymes that are vitamin K dependent. These enzymes, along with other serine proteinase clotting factors, mediate the
Scheme 2
Table 4. Clotting times for plasma coagulation initiated by photolysis of acyl thrombin NIH units of acyl thrombin*
Photolysis time (s)
Clotting time (s)
Calculated clotting time (s)+
5.0 5.0 5.0 3.0 3.0 3.0 3.0 2.0 1.o
13.5t 5.0 3.0 16.7t 5.0 4.0 3.0 23.lt 34.6t
13.5 ? 0.24 16.6 ? 0.48 18.7 ? 0.06 16.7 t 0.16 21.6 2 1.50 22.1 ? 1.80 25.8 ? 1.20 23.1 ? 0.22 34.6 ? 1.10
13.6 15.8 19.6 16.6 21.2 23.6 27.8 19.8 27.8
*1 NIH U = 0.028 p,M in this procedure. tcontinuous photolysis. $Calculated by numerical integration (see 'Discussion' section).
Table 5. Clotting times for plasma coagulation initiated by photolysis of acyl Factor Xa NIH units of acyl Factor Xa* 1.0 1 .o 1.0 1.0 0.25 0.25 0.25 0.25 0.1 0.075 0.075 0.075
0.05
Photolysis time (s)
Clotting time (s)
Calculated clotting time (s)+
21.91 3.0 2.0 1.0 25.0t 3.0 2.0 1.o 26.5t 27.9t 3.0 1.0 28.91.
21.9 ? 0.70 22.4 2 0.15 22.9 ? 0.21 26.2 t 0.20 25.0 ? 0.55 26.6 ? 0.44 28.2 ? 0.44 32.3 ? 0.95 26.5 t 0.52 27.9 0.52 31.7 ? 1.10 38.7 ? 1.10 28.9 t 1.20
21.3 22.5 23.2 24.8 24.1 26.6 28.0 31.3 27.1 28.3 33.2 41.9 30.4
*
'1 NIH U = 0.028 p M in this procedure. tcontinuous photolysis. $Calculated by numerical integration (see 'Discussion' section).
Plasma photocoagulation
Scheme 3 we have developed a strategy for photoactivation of these enzymes that relies on the use of amidinophenol esters of cinnamic acid to form hydrolytically stable acyl enzymes (Scheme 2b). Thus, we have previously reported that thrombin and Factor Xa esters of a-methyl-o-hydroxycinnamic acid can be prepared and that these acyl enzymes show no activity in standard chromogenic assays if kept in the dark (Turner e t a / . , 1987, 1988). Photolysis through Pyrex of these acyl enzymes with broad-band light from a high pressure mercury lamp returns enzyme activity. The proposed mechanism of this enzyme photoactivation is presented in Scheme 3 and involves photoisomerization of the cinnamate ester followed by lactonization of the cis cinnamate. This lactonization leads to a methyl coumarin and the active enzyme. The p-amidinophenyl ester of a-methyl-o-hydroxycinnamic acid was thus shown to inhibit thrombin and Factor Xa and the inhibition was reversed by photolysis with Pyrex filtered light from a high pressure mercury lamp. There were, however, several problems with this inhibition-photoactivation strategy utilizing the parent ester, Scheme 3 where R=H. The enzyme inhibition was temporary and enzyme activity returned in the course of a few hours in the dark, i.e. enzyme turnover was relatively fast. Furthermore, the photoactivation of these enzymes was slow (10-1.5 min) and required light intensities and wavelengths such that appreciable enzyme degradation occurred during photoactivation. Because of these problems, we have prepared several derivatives of a-methyl-o-hydroxycinnamic acid and have studied, in detail, their properties. Our best inhibitor, l b , where R=Et2N in Scheme 2b, forms an acyl enzyme that is hydrolytically more stable than the parent compound where R=H. While l b is a better inhibitor of thrombin than of Factor Xa, the thrombin acyl enzyme deacylates slower than the Factor Xa acyl enzyme. Furthermore, esters of this substituted cinnamate have an absorption spectrum with a ,,A at 360 nm, a region of the spectrum that is readily accessible and where protein has no significant absorption. In order to study the photocoagulation of plasma with the acyl enzymes, we first carried out standard coagulation assays of plasma with the active enzyme. Such standard coagulation assays have been studied in great detail and a kinetic analysis of coagulations initiated by one of the active enzymes in the cascade has been published (Hemker and Muller, 1968; Hemker et al., 1965). In this analysis, it is assumed that a clot is detected when
41
the concentration of a product in the cascade reaches a critical level. Thus, the clotting time, reaction velocity, and the critical product concentration are related as in Eq. 1 where t,=the clotting time,v=the reaction velocity and [P],=the critical concentration of product P formed in the enzymatic cascade.
r,
x v = [PIc
(1)
Using a modified Lineweaver-Burke analysis of the coagulation cascade, Hemker et af. (1965) have suggested that the clotting time and the concentration of active units of thrombin in the cascade, [Ha] in our presentation, can be expressed by Eq. 2 1, =
l/v =
+ B/[IIa] A/[PIc+ B/{[IIa] x [PI,.} A
(2)
(3)
with A and B being constants. Insertion of Eq. 1 into Eq. 2 gives an expression for l / v as a function of units of added enzyme, thrombin in our studies, and the critical concentration [PIc.A plot of clotting time vs l/[NIH U] for thrombin is presented in Fig. 3. The best fit parameters for the data presented for thrombin is t'. = 6.1 + 11.6 x l/[IIa] with a correlation coefficient of 0.99. The approach suggested by Hemker is thus shown to be valid in our system and this analysis is extended, vide infra, in our studies of photocoagulation. Factor Xa initiates a two component cascade in which thrombin is generated from prothrombin in the first step and fibrin is formed in the second step. For a two component cascade, Hemker and Muller (1968) suggest that the clotting time should be related to the active enzyme by an inverse squareroot relationship. Thus, as shown in Fig. 4,we find a best fit for the data of t, = 18.2 (2.17 x l/ 45
-I
0 0
1
a
1 3
Figure 3. Clotting times in seconds vs l/[thrombin] (top scale) and l/[NIH U thrombin) (bottom scale). Best-fit correlation is clotting time = 6.1 2 11.6 X l/[thrombin] where [thrombin] is the number of NIH units of thrombin (rz > 0.99).
NED A. PORTER and JOHN D. BRUHNKE
42
5o
= 2.303 +Illel (B. Faust, personal communication; Zepp, 1978).
1
d[IIa]ldt = -d[Acyl IIa]/dt = k[Acyl IIa]
"1 10
0
4 0
.
I
2
.
,
.
4
I
6
.
I
8
.
1
10
1 v'[ Factor X a l
Figure 4. Clotting times in seconds w l / q F a c t o r Xa]. Best-fit correlation is clotting time = 18.2 f (2.17 x I / where [Xa] is the number of NIH units of Factor Xa (Tz > 0.99).
m)
m)
where [Xa] i-s the number of N I H U of Factor Xa (9> 0.99). Photocoagulation in plasma is efficient using the acyl modified thrombin or Factor Xa (Tables 4 and 5). Photolysis of 5 U of the acyl thrombin for 5 s, for example, results in a clot in 16.6 s, a clotting time shorter than that obtained from 1 U of the fully active enzyme in the standard assay. Photolysis of 1 U of acyl Factor Xa for 1 s results in a clotting time of 26 s. Furthermore, the light delivery system used in these studies was not maximized and we have every reason to believe that significant improvements in light delivery could lead to shorter required photolysis times. Analysis of the clotting times obtained during photolysis by a Lineweaver-Burke approach analogous to that of Hemker et al. (1965) is instructive. If we assume that the concentration of active species P at any time t is given by [PI = t x v (see Eq. 1) and that l l v is given by Eq. 3 for thrombin, we obtain Eq. 4. When [P]/[PIc = 1, a clot forms and t = t,. The additional feature
(4) to this analysis compared to Hemker is that in the photocoagulation experiment, [IIa] is initially zero and increases during the period of photolysis. This must be taken into account in our analysis of photocoagulation. At the low concentrations of acyl enzyme used in this study, the increase in concentration of thrombin or Factor Xa is first order dependent on the concentration of the acyl enzyme [Acyl IIa] or [Acyl Xa] and also depends on the quantum yield for the photoprocess 4 by which active enzyme is generated, the intensity of incident irradiation, I,,, the molar extinction coefficient of the acyl enzyme, e , and the pathlength of the sample. This relationship is given in Eq. 5, where k
(5)
By the use of Eqs. 4 and 5, we have developed a simple numerical integration computer analysis of thrombin photocoagulation. The numerical integration calculates the increase of [IIa] and decrease of [Acyl Ha] over a small time increment, At, after guessing a value of k . The increase in [P]/[PIcis calculated over At using Eqs. 4 and 5 and this process is then repeated over another small time increment. When [P]/[PIc= 1, the integration is stopped and the time is reported as the clotting time. No increase in [IIa] is calculated if the light is off and the clotting time can thus be calculated for continuous photolysis throughout the course of the assay or for short-term photolysis and a subsequent dark period. With this numerical integration approach and guessing a value for k in the photoactivation scheme, we were able to get a reasonably good fit to the data presented in Table 4 for thrombin photoactivation. The calculated clotting times using k = 0.07 s-I are presented in Table 4 for thrombin photoactivation and the results are remarkably good given the number of assumptions that are made in the analysis. A similar analysis is used in the photocoagulation initiated by photolysis of the acyl Factor Xa enzyme. In this approach, a numerical integration program has been written analogous to the one described above for thrombin photoactivation, but with Eq. 4 replaced by Eq. 6. With this
approach and guessing a value for k of 0.12 s-' in the photoactivation scheme, we were able to get a reasonably good fit to the data provided in Table 5 for Factor Xa photoactivation. The results are also excellent for this analysis given the assumptions, and our success in fitting the photolysis data using a l/[IIa] correlation for the one-step thrombin process and l/mcorrelation for the two-step cascade lends credibility to the analysis. Comparison of the best fit k values for thrombin (0.07) and Factor Xa (0.12) also indicates no dramatic difference for the two analogous photochemical processes. The differences noted could be the result of more efficient photoprocess for Factor Xa and/or a higher molar absorptivity for the Factor Xa acyl enzyme, since k = 2.303 +Il,El. CONCLUSIONS
Serine proteinase enzymes of the coagulation cascade modified at the active site serine hydroxyl can be activated efficiently with light. Furthermore,
, '
Plasma photocoagulation
coagulation of plasma can result from photolysis of an acyl enzyme in plasma during the time-course of the clotting experient. Thus, photolysis of 1 U of a modified Factor Xa enzyme for as short a time as 1 s with monochromatic 366 nm light results in clot formation in 26 s in a plasma where 1 U of thrombin gives a clotting time of 18 s. The photochemical time course is therefore short enough and the efficiency of photoactivation is high enough so that this strategy can lead to a significant biological response to a modest photostimulus. Acknowledgemenrs-We gratefully acknowledge support from NIH (HL-17921) and gifts of serine proteinase enzymes from Dr. S . Pizzo. We also thank Dr. B . Faust for stimulating discussions. REFERENCES
Berezin, I . V., N. F. Kazanskaya, R. B. Aisina and E. V. Lukasheva (1980) Detection of light signals by microencapsulated cis-cinnamoyl chymotrypsin. Enzyme Microb. Technol. 2, 15G154. Erlanger, B. F. (1976) Photoregulation of biologically active macromolecules. Annu. Rev. Biochem. 45, 267-283. Fujioka, T., K. Tanizawa and Y. Kanoaka (1982) Inverse substrates-XV. Spectrometric properties of fluorescence-labeled acyl trypsins. Chem. Pharmacol. Bull. 30, 230-236. Hemker, H . C. (1983) Handbook of Synthetic Substrates for the Coagulation and Fibrinolytic System, pp. 27-44. Martinus-Nijhoff, Boston, MA. Hernker, H . C., P. W. Hernker and E.A. Loeliger (1965) Kinetic aspects of the interaction of blood clotting enzymes-I. Derivation of basic formulas. Thromb. Diath. Haemorrh. 13, 155-175. Hemker, H. C. and A. D . Muller (1968) Kinetic aspects
43
of the interaction of blood clotting enzymes-V. The reaction mechanism of the extrinsic clotting system as revealed by the kinetics of one-stage estimations of coagulation enzymes. Thromb. Diath. 19, 368-382. Jackson, C. M. and P. L. Coleman (1981) Assay of coagulation proteases using peptide chromogenic and gluorogenic substrates. Methods Enzymol. 80, 341-361. Kam, C.-M., K. Fujikawa and J . C. Powers (1988) Mechanism-based isocoumarin inhibitors for trypsin and blood coagulation serine proteases: new anticoagulants. Biochemistry 21, 2547-2557. Markwardt, F., G . Wagner, J. Sturzebecher and P. Walsmann (1980) N-a-arylsulfonyl-w-(4-amidinophenyl)aminoalkylcarboxylic acid amides-novel selective inhibitors of thrombin. Thromb. Res. 17, 425-430. Martinek, K. and I. V. Berezin (1979) Artificial lightsensitive enzymatic systems as chemical amplifiers of weak light signals. Photochem. Photobiol. 29, 637-649. Martinek, K., S. D. Varfolomeyev and I. V. Berezin (1971) Interaction of a-chymotrypain with N-cinnamoylimidazole, substrate sensitive to light. Eur. J . Biochcm. 19, 242-249. Porath, J., K. Aspberg, H. Drevin and R. Axen (1973) Preparation of cyanogen bromide-activated agarose gels. J . Chromatogr. 86, 53-56. Schiffman, S., I . Theodor and S . I. Rapaport (1969) Separation from Russell’s viper venom of one fraction reacting with Factor X and another reacting with Factor V. Biochemistry 8, 1397-1405. Tanizawa, K . , Y. Kanaoka and W. B. Lawson (1987) Inverse substrates for trypsin and trypsin-like enzymes. Acc. Chem. Res. 20, 337-342. Turner, A. D . , S. V. Pizzo, G . W. Rozakis and N. A . Porter (1987) Photochemical activation of acylated a thrombin. J . Am . Chem. Soc. 109, 1274-127s. Turner, A. D . , S . V. Pizzo, G . Rozakis and N. A . Porter (1988) Photoreactivation of irreversibly inhibited serine proteinases. J . Am. Chem. Soc. 110, 244-250. Zepp, R. G . (1978) Quantum yields for reaction of pollutants in dilute aqueous solution. Environ. Sci. Technol. 12, 327-329.
