Human skin fibroblast collagenase. Assessment of activation energy and deuterium isotope effect with collagenous substrates.

The activation energy and solvent-deuterium kinetic isotope effect for human skin fibroblast collagenase were studied on Merent collagenous substrates. Values for activation energy (EA) were determined on the a1 and a2 polypeptide chains of denatured collagen (13,900 calories, 13,300 calories, respectively), collagen in solution (49,200 calories), and collagen in native fibrillar form (101,050 calories). The energy dependence of catalysis thus increased markedly with formation of the collagen triple helix and further with the assembly of collagen monomers into a fibrillar structure. The substitution of deuterium for hydrogen in solvent buffer (DzO:HzO = 0.9) slowed fibrillar collagenolysis 9-fold and degradation of solution monomers 2-fold. Enzyme-substrate binding (kl) and enzyme-substrate dissociation (k-3 were not affected by deuterium. These results indicate that proton transfer is involved in the slow step of collagen degradation. Evidence from this study and previous kinetic data strongly suggest that kz, or actual peptide bond cleavage, represents the ratelimiting step of collagen degradation by fibroblast collagenase. The activation energy and solvent-deuterium kinetic isotope effect for other collagenolytic enzymes were also investigated. In each case, the magnitude of the activation energy was directly proportional to that of the observed deuterium isotope effect, indicating an interrelationship between the accessibility of water molecules to the site of peptide bond hydrolysis in collagen and the energy requirements of enzymatic action. The data also suggest that the unusually large dependence of fibrillar collagen degradation upon both energy and hydrogen transfer are properties specific to mammalian collagenases, all of which catalyze only a single cleavage in native collagen, and may relate directly to the geographic location of the site of this cleavage in the collagen molecule.

The activation energy and solvent-deuterium kinetic isotope effect for human skin fibroblast collagenase were studied on M e r e n t collagenous substrates. Values for activation energy (EA) were determined on the a1 and a 2 polypeptide chains of denatured collagen (13,900 calories, 13,300 calories, respectively), collagen in solution (49,200 calories), and collagen in native fibrillar form (101,050 calories). The energy dependence of catalysis thus increased markedly with formation of the collagen triple helix and further with the assembly of collagen monomers into a fibrillar structure.
The substitution of deuterium for hydrogen in solvent buffer (DzO:HzO = 0.9) slowed fibrillar collagenolysis 9-fold and degradation of solution monomers 2-fold. Enzyme-substrate binding (kl) and enzyme-substrate dissociation ( k -3 were not affected by deuterium. These results indicate that proton transfer is involved in the slow step of collagen degradation. Evidence from this study and previous kinetic data strongly suggest that kz, or actual peptide bond cleavage, represents the ratelimiting step of collagen degradation by fibroblast collagenase.
The activation energy and solvent-deuterium kinetic isotope effect for other collagenolytic enzymes were also investigated. In each case, the magnitude of the activation energy was directly proportional to that of the observed deuterium isotope effect, indicating an interrelationship between the accessibility of water molecules to the site of peptide bond hydrolysis in collagen and the energy requirements of enzymatic action. The data also suggest that the unusually large dependence of fibrillar collagen degradation upon both energy and hydrogen transfer are properties specific to mammalian collagenases, all of which catalyze only a single cleavage in native collagen, and may relate directly to the geographic location of the site of this cleavage in the collagen molecule.
Collagenases are characterized by the ability to initiate the specific degradation of the native triple helical collagen molecule. Studies designed to investigate the kinetics of collagenase action are essential in order to gain a better understanding of the mechanism by which this class of enzymes performs its unique function in nature. In a recent report characterizing the action of human skin fibroblast collagenase on fibrillar collagen (l), we have shown that collagenase binds very tightly to collagen fibrils and appears to remain bound to this substrate throughout ongoing collagen degradation at 37 "C, independent of subsequent dilution with buffer or the addition of exogenous collagen as a competitor. Thus, during the degradation of fibrillar collagen no equilibrium appears to exist between collagenase molecules bound to the fibrillar substrate and the external buffer of the reaction mixture. Studies of enzyme binding as a function of collagenase concentration indicated that only 10% of the total number of collagen molecules present in a fibrillar substrate gel are initially accessible to enzyme for binding prior to degradation of the substrate (1). The results were consistent with the available molecules occupying the surface of each fibril. The remaining 90% of collagen molecules, presumably located within the interior of the fibrils, apparently become accessible to the enzyme only after subsequent substrate catalysis.
In an accompanying paper (2), we have examined the collagen substrate specificity of human skin fibroblast collagenase as a function of both collagen type and species of substrate origin. Measurements of the basic kinetic parameters, K,,, and kcat, were performed on collagen types I-V from several animal species, using collagen in solution as substrate. Collagen types I, 11, and I11 of all species examined were successfully attacked by fibroblast collagenase and the measured enzyme-substrate affinity was similar in all cases, K,,, = 0.7-2.1 X 1O"j M. In contrast, large differences in catalytic rates were evident, ranging from 565 h" for the homologous human type I11 substrate to 1.0 h" for human type I1 collagen. Significant rate differences were also observed between collagens of the same type but of different species of origin. Human skin fibroblast collagenase was most specific for the homologous type I and I11 collagens.
In this communication, we have investigated the action of human skin fibroblast collagenase with respect to both temperature dependence of activity and solvent-deuterium isotope effect. These parameters have been studied utilizing fibrillar collagen, collagen in solution, and denatured collagen as substrates, in order to provide a more detailed understanding of the mechanism of collagenase action and specifically to help define the rate-limiting step of collagen degradation. For comparison, the collagenolytic proteases from the crustacean, Uca pugilator, and from the bacterium, Clostridium histolyticum, which are chemically unrelated to fibroblast collagenase, were also studied.

MATERIALS AND METHODS
Reagents-Acrylamide and bis-acrylamide were purchased from Eastman. Sodium dodecyl sulfate (99% pure) and deuterium oxide (99.8%) were obtained from Gallard-Schlesinger. Tris base, bovine pancreatic trypsin (type III), and soybean trypsin inhibitor were Grants Ah4 12129, HD 05291, and TO-AM 07284. The costs of * This work was supported by United States Public Health Service publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. procured from Sigma. All other chemicals used were reagent grade.
Culture Methods-Normal human skin fibroblasts (CRL 1187) were purchased from American Type Culture Collection. The cells were grown in the presence of 20% fetal calf serum and the medium was harvested as described by Bauer et al. (3).
Purification of Collagenases-Human skin fibroblast collagenase was purified to homogeneity from serum-containing medium by a combination of carboxymethylcellulose and Ultrogel AcA-44 chromatography, as described by Stricklin et al. (4,5). Pure collagenase from the hepatopancreas of the fiddler crab, U. pugilator, was prepared according to methods described by Eisen et al. (6). Bacterial collagenase (form IIk protease-free) from C. histolyticum was purchased from Advanced Biofactors.
Activation of Collagenase-The activation of fibroblast procollagenase was accomplished proteolytically by the addition of trypsin at 25 "C for 10 min. Further tryptic action was prevented by adding an 8-fold molar excess of SBTI.' Maximal collagenase activity was ensured by performing a trypsin titration on each batch of enzyme. Crustacean and bacterial collagenases did not require proteolytic activation.
Assay Procedures: Determinution of E A on Fibrillar Collagen-Guinea pig skin type I collagen, prepared in this laboratory as previously described by Gross (7), was the source of all collagenous substrates employed in this study. The activation energy of each collagenase on fibrillar collagen was determined by measuring the rate of degradation of fibrillar substrate as a function of temperature. The temperature of the water bath incubator was controlled by a constant temperature circulator (Fisher Model 73) and monitored with a Fisher mercury thermometer with 1/10 "C subdivisions. In each assay, 50 pl of a 0.4% solution of native, reconstituted ["C] glycine-labeled collagen of specific activity 25,000 cpm/mg was allowed to gel at 37 "C overnight to permit completion of the aggregation process to occur. Following incubation of such collagen gels with enzyme, the reaction mixtures were centrifuged at 10,000 X g and the supernatant fraction was counted in a liquid scintillation spectrometer. At each temperature studied, samples were stopped at several time points to ensure linearity of reaction velocity with time and thus measurement of an initial velocity. Collagenase activity was then graphed as an Arrhenius plot of log reaction velocity versus l/temperature (K) and the activation energy Degradation of collagen was always t20% of the initial amount present, thereby enabling measurement of a true initial velocity (2). Incubation times varied from 3 min (36 "C) to 4 h (16 "C). The reaction process was stopped with EDTA, sodium dodecyl sulfate sample buffer added, and aliquots of each sample were then subjected to polyacrylamide gel electrophoresis using 8% gels cast in slab molds.
Following staining with 1% Coomassie blue, reaction velocity was quantitated by measuring the formation of single cleavage TCA products (% degradation = 4/3[TCA]/(4/3[TCA] + [a])) using a spectrophotometer equipped with a gel scanning linear transport device. The results were then graphed as an Arrhenius plot and EA was calculated from the formula given above.
Determination of EA on the crl and a2 Polypeptide Chains of Gelatin-Separation of the a1 and a2 polypeptide chains of denatured guinea pig skin collagen was accomplished by carboxymethylcellulose chromatography, as origindy described by Piez et aZ. (8). Pure human skin fibroblast collagenase initiates multiple cleavages in both gelatin chains at a low rate, including a cleavage indistinguishable by polyacrylamide gel electrophoresis from the 3 / 4 4 cleavage produced in native collagen (details of this gelatinolytic activity will be published in a future communication). In the case of the a2 chain, this 3 / 4 4 cleavage could be isolated as a single catalytic event, prior to any subsequent cleavages, by stopping the reaction process after onIy 15-20% of the substrate present had been degraded. The activation energy of fibroblast collagenase on the a2 chain of gelatin was there-' The abbreviation used is: SBTI, soybean trypsin inhibitor. fore determined by quantitating reaction velocity in terms of substrate conversion to %-length products, as described above for collagen in solution. Degradation was measured in 2 "C increments over the temperature range of 35-41 "C. At each temperature, 0.03 pg of collagenase was incubated with 20 pg of gelatin a2 chain in a total volume of 100 pl (buffer = 0.05 M Tris, 0.01 M CaClz, 0.15 M NaCl, pH 7.5). SBTI (I mg/ml) was used in each reaction mixture to maintain enzyme stability at all temperatures employed. Samples were incubated at each temperature for 5, 10, and 15 min to establish linearity of substrate degradation with time.
In the case of the a1 chain, a single preferred catalytic event could not be isolated before subsequent cleavages occurred. Therefore, the disappearance of a1 reactants, instead of the appearance of %-length products, was measured over the incubation times stated above; this disappearance of the reactants was linear with time until approximately 20-25% of the substrate was degraded. For studies on the a1 chain of gelatin, 0.12 pg of fibroblast collagenase was employed in each reaction mixture.
Solvent-Deuterium Kinetic Isotope Effect-For all experiments involving solvent-deuterium, procollagenase was dialyzed overnight against 0.05 M Tris, 0.01 M CaC12, 0.15 M NaC1, pH 7.5, containing 99.8% deuterium oxide. Trypsin and SBTI used to proteolytically activate the enzyme were dissolved in buffer which contained 99.8% D20. Fibrillar collagen, gelled from a 0.4% solution, was washed twice with DIO-containing buffer in order to replace internal water with deuterium. In this manner, for each experiment involving solventdeuterium, an exact ratio of deuterium:hydrogen could be attained in the total solvent buffer of the reaction mixture.
Protein concentrations were determined spectrophotometrically by the method of Groves et al. (9). Albumin was used to establish a standard curve.

RESULTS
The EA of pure human skin fibroblast collagenase was determined on three forms of collagenous substrates: fibrillar collagen, collagen in solution, and denatured collagen (gelatin). EA was first measured on I4C-labeled fibrillar collagen over the temperature range of 35-39 "C. At all temperatures, collagen degradation was linear with time, the enzyme was thermostable, and essentially no substrate denaturation was demonstrable as assessed by trypsin blanks. Degradation of fibrillar collagen was measured in 1 "C increments over the temperature range of 35-39 "C. In each reaction mixture, 100 pl of a collagenase solution of concentration = 12 pg/ml was incubated with 200 pg of I4C-1abeled fibrillar collagen. At each temperature, samples were stopped at 10-15-min intervals to ensure linearity of collagen degradation with time. Log reaction velocity was arbitrarily set equal to 1.0 for the lowest temperature and values at the other temperatures were normalized accordingly. EA was calculated from the slope of the Arrhenius plot using the formula given under "Materials and Methods." B, EA on collagen in solution. Degradation of collagen in solution form was measured in 4 "C increments over the temperature range of 16-36 "C, as detailed under "Materials and Methods." Lengths of incubation at each temperature were as follows: 16 "c, 4 h; 20 "C, 90 m i n , 24 "C, 35 m i n , 28 "C, 35 mi n, 32 "C, 7 min; 36 "C, 3 min. Log reaction velocity was arbitrarily set equal to 1.0 for the lowest temperature and values at the other temperatures were normalized accordingly. EA was calculated using the formula given under "Materials and Methods." the resultant Arrhenius plot, whose slope indicates an EA of 101,050 calories. This activation energy is extremely high when compared with most enzyme-catalyzed reactions and, as a result, fibrillar collagen degradation by fibroblast collagenase would be expected to increase approximately 200-fold for each 10 "C increase in temperature.
The Arrhenius energy characterizing the action of fibroblast collagenase on collagen in solution was measured over the temperature range of 16-36 "C. Short incubation periods (<5 min) were utilized at the higher temperatures in order to avoid spontaneous gelation with conversion to the fibrillar form of substrate. Quantitation of TCA product formation was accomplished by running the final reaction mixtures on sodium dodecyl sulfate-polyacrylamide slab gels, staining with 1% Coomassie blue, and then scanning with a spectrophotometer equipped with a linear transport device (2). At each temperature studied, collagen degradation was linear with time over the entire assay period, thus ensuring measurement of an initial velocity. As shown in Fig. lB, the resultant Arrhenius plot indicated an EA of 49,200 calories. While this activation energy is considerably less than that which characterized the degradation of fibrillar collagen, cleavage of collagen monomers is nevertheless an unusually energy-dependent process. Collagenase activity on this form of substrate increased 16-fold per 10 "C.
As previously reported (4), human skin fibroblast collagenase purified to homogeneity exhibits a low level of proteolytic activity against gelatin. The activation energy of fibroblast collagenase on the a1 and a2 polypeptide chains of denatured type I collagen was examined from 35-41 "C, a temperature range high enough to prevent refolding of gelatin chains into a nonrandom secondary structure (10,ll) but yet low enough to maintain enzyme stability. Arrhenius energies of 13,900 and 13,300 calories were obtained for the a1 and the a2 chains, respectively (Fig. 2, A and B ) . The 2-fold change in reaction velocity per 10 "C indicated by these Arrhenius energies is far more typical of most enzyme-catalyzed reactions than the extraordinary temperature dependence which characterized the cleavage of native triple helical collagen both in fibrillar form and as solution monomers.
The activation energy of pure crab collagenase and of bacterial collagenase on fibrillar collagen is shown in Fig. 3. An Arrhenius plot of fibrillar collagen cleavage by the crustacean enzyme, over the temperature range of 35-39 "C, indicated an EA of 24,300 calories (Fig. 3A). The activation energy of clostridial collagenase, assayed under the same conditions, was 19,400 calories (Fig. 3B). These values indicate that fibrillar collagen degradation by the crustacean and bac- terial enzymes is far less energy dependent than is the case for human skin fibroblast collagenase.
In order to provide further insight into the reaction mechanism of human skin fibroblast collagenase, the solvent-deuterium kinetic isotope effect for this enzyme was examined. Fig. 4 illustrates the rate of fibrillar collagen degradation as a function of an increasing DzO:HzO ratio in solvent buffer. When 90% of all water molecules was present as the heavier deuterium isotope, cleavage of collagen fibrils was slowed markedly, to only 11% of the rate observed in 100% HzO. This 9-fold reduction in enzyme activity was considerably greater than expected, since a 2-3-fold inhibition by solvent deuterium is generally observed for most enzyme-catalyzed reactions when the slow or rate-determining step of the reaction process involves proton transfer in the hydrolysis of a peptide bond (12-15). Further experiments in which both enzyme concen- in the presence of deuterium oxide (ratio of D20H20 in total solvent buffer = 0.90) was measured in 1 "C increments over the temperature range of 36-39 "C. At each temperature, 100 pl of a collagenase solution of concentration = 120 pg/ml were incubated with 200 pg of I4C-labeled fibrillar collagen. Log reaction velocity was arbitrarily set equal to 1.0 for the lowest temperature and values at the other temperatures were normalized accordingly. Ea was calculated from the Arrhenius plot using the formula shown under "Materials and Methods."

TABLE I
Effect of D20 on collagenase binding 100 pl of a collagenase solution of concentration = 20 pg/ml was bound to 200 pg of I4C-labeled fibrillar collagen in the presence of either 90% DzO or 100% Hz0 in total solvent buffer. Enzyme was bound at 25 "C for 3, 6, and 20 mi n, the latter period of time known to be sufficient to ensure completion of the binding process (1). Following binding, the collagen gels were washed and resuspended in new buffer such that the final reaction mixtures contained either 90% DzO or 100% HzO. The samples were then incubated at 37 "C X 1 hour, and centrifuged at 10, OOO X g, and the supernatant fractions were counted in a Liquid scintillation spectrometer.

cpm-blank cpm-blank
Bindat25OCx3mininH20 652 70 Bind at 25 "C X 3 min in D20 643 52 Bind at 25 "C X 6 min in Hz0 891 85 Bindat25OCx6mininD20 928 68 Bind at 25 "C x 20 min in H20 1262 108 Bind at 25 "C x 20 min in D20 1180 120 tration (10-100 pg/ml) and quantity of fibrillar substrate (100-300 pg) were varied in the presence of a constant deuterium: hydrogen ratio in solvent buffer resulted in the same kinetic isotope effect (not shown). Only the concentration of deuterium relative to hydrogen appeared to modulate the extent of enzyme inhibition. Although the absolute rate of collagenolysis was slowed by deuterium, the temperature dependence of the reaction process, as measured by EA, was largely unaffected by the heavier isotope (Fig. 5).
To define which step(@ in the degradation of fibrillar col-lagen was so markedly affected by deuterium, enzyme-substrate binding, enzyme-substrate dissociation, and enzymatic hydrolysis were independently examined for evidence of proton transfer. Enzyme-substrate binding was studied by incubating active collagenase with fibrillar collagen at 25 OC in solvent buffer containing either 9 0 % D20 or 100% H20. Incubations were performed for increasing lengths of time up to 20 min, a period known to ensure completion of the binding process (1). At this temperature, due to the high activation energy of fibroblast collagenase, binding occurs without measurable substrate degradation (1). Following enzyme binding, the respective substrate gels were washed, resuspended in new buffer, and then incubated at 37 "C. As shown in Table I, enzyme-substrate binding at 25 "C was not affected by the presence of either deuterium or hydrogen; collagenase activity, however, was critically dependent upon the isotope composition of the suspending buffer during collagen degradation at 37 "C. Quantitation of binding in D20 and Hz0 by calculation of the rate constant, kl, resulted in an identical value. 1.2 X lo3 M" s-', for both solvents. Therefore, enzyme-substrate binding exhibited no discernible kinetic isotope effect. Substrate cleavage subsequent to binding, however, was slowed markedly in the presence of the heavier deuterium isotope. Evidence for proton transfer in dissociation of the enzymesubstrate complex was next investigated. Fibroblast collagenase was bound to fibrillar collagen in buffer containing either DzO or H20, as described above. Following completion of binding, the gels were washed and resuspended in identical buffer at 25 "C for 1 h. Aliquots of each reaction mixture were then incubated with new collagen gels (solvent buffer = H20) at 37 "C to assay for the presence of dissociated enzyme. No significant collagenase activity was observed, consistent with little, if any, dissociation of enzyme from fibrillar substrate, regardless of whether bound enzyme was suspended in D20 or H20 (not shown). These results indicated that dissociation of the enzyme-substrate complex, like enzyme-substrate binding, did not exhibit any discernible kinetic isotope effect. Thus, only enzymatic hydrolysis showed evidence of proton transfer, since this step alone was slowed by the substitution of deuterium for hydrogen in solvent buffer. Interestingly, fibrillar collagen degradation by crab collagenase and clostridial collagenase was less profoundly affected by solvent deuterium than was the case for fibroblast collagenase. As illustrated in Table 11, fibrillar collagenolysis by the crustacean and bacterid enzymes was slowed 2.6-fold and 1.7-fold, respectively, by the presence of 90% deuterium in solvent buffer.
To assess the deuterium isotope effect for the cleavage of collagen in solution by human skin fibroblast collagenase, values for the K,,, and V, . were determined (2) in solvent buffer containing 90% DzO uersus 100% H20. The resultant data, plotted according to the method of Lineweaver and Burk, are shown in Fig. 6. The substitution of deuterium for hydrogen in solvent buffer resulted in a 50% reduction in Vmr, with no accompanying change in K,. Thus, the degradation

I1
Comparison of the activation energy and deuterium isotope effect on fibrillar collagen degradation for several collagenases Values for EA of human fibroblast, crab, and bacterial Collagenase on fibrillar collagen were obtained as described in Figs. 1A and 3, A  and 3. kH/kD represents the ratio of fibrillar collagen degradation in solvent buffer containing 100% Hz0 versus 90% D20. of collagen monomers by fibroblast collagenase was slowed by the heavier deuterium isotope, but to a lesser extent than characterized the cleavage of fibrillar collagen (2-fold uersus 9-fold).

DISCUSSION
This study examines the activation energy of human skin fibroblast collagenase on different forms of collagenous substrates. The values of EA on the a1 and a2 polypeptide chains of denatured collagen, collagen in solution, and collagen in native fibrillar form were 13,900, 13,300, 49,200, and 101,050 calories, respectively. Arrhenius plots were linear over the temperature ranges investigated, indicating no significant change in the structure of the reactants or in the rate-limiting step of enzyme catalysis in each case.
Recent work has demonstrated that the structure of collagen in solution consists of a mixture of single molecules and higher molecular weight aggregates (16, 17). We have not attempted to measure such aggregation in this study, and it is possible that the extent of microaggregate formation could affect the exact value of EA on this form of substrate. In our determination of EA, however, samples were incubated for different lengths of time and at different temperatures, conditions under which the level of collagen aggregation would be expected to change. Yet, linear Arrhenius plots always resulted, even when measured in 1 "C increments (not shown), again suggesting that such microaggregates may be indistinguishable to collagenase from collagen monomers (2).
The increase in the activation energy of fibroblast collagenase which accompanies an increasing level of organization in the collagen substrate itself is of particular interest. Nearly all reported values for the Arrhenius energy of enzyme-catalyzed reactions are 10-20,OOO calories (14), indicating that reaction velocity changes 1.7-3.0-fold per 10 "C. While such an activation energy characterized fibroblast collagenase degradation of gelatin CY chains, cleavage of native triple helical collagen molecules in solution was a far more energy dependent process, E A 2 : 50,000 calories, while further ordering of collagen into fibrils raised the Arrhenius energy to 100, OOO calories. This very high activation energy for fibrillar collagen, the physiologic form of substrate, may be of considerable significance to the animal organism. For each 2 "C increase in body temperature, collagenase activity would be expected to increase approximately 3-fold. A 5 "C change in temperature, which not uncommonly accompanies febrile illnesses and local inflammation (18), would result in a 14-fold increase in the rate of fibrillar collagen degradation, while a change of 10 "C, albeit most unlikely, would increase collagenase activity 200fold.
A similar substrate-dependent process was evident when deuterium was substituted for hydrogen in solvent buffer. Whereas cleavage of collagen monomers by fibroblast collagenase decreased 2-fold in the presence of 90% deuterium oxide, when collagen was in the fibrillar form, the rate of degradation was slowed 9-fold. This kinetic isotope effect for fibrillar collagen cleavage is considerably greater than the 2-%fold reduction in activity reported for most enzyme-catalyzed reactions where peptide bond hydrolysis is rate-limiting (12)(13)(14)(15). Although the nature of the rate-determining step of collagenolysis has not been previously defined for any collagenase, it seems clear that proton transfer must be implicated at the slow step of fibrillar collagen degradation by fibroblast collagenase. Since enzyme-substrate binding ( k l ) and enzymesubstrate dissociation (k+ were both unaffected by solvent deuterium, the results are most consistent with actual enzymatic hydrolysis, represented by the rate constant, k2, as the rate-determining step of the reaction process. While a smaller, more typical deuterium isotope effect was observed for the cleavage of collagen molecules in solution by fibroblast collagenase, this 2-fold reduction in Vmax without any accompanying change in K, is also consistent with product formation as the rate-limiting step. In a previous study (l), we reported the equilibrium con- suggests that K, s Kd, indicating that the ratio k2/kl must be negligible, and further implicating peptide bond cleavage (k2) as the rate-limiting step of collagen degradation. Additionally, as reported in the accompanying paper (2), large differences in the V,,, (kcat) of fibroblast collagenase are found for collagen substrates of nearly identical affinity (K,). Thus, although the collagen molecule is large, extremely asymmetric, and contains only a single catalytic site for human fibroblast collagenase, there is no evidence of any difficulty in obtaining effective collisions between enzyme and substrate. In view of the activation energy and deuterium isotope effect which characterize the action of fibroblast collagenase on collagen fibrils and solution monomers, it appears that water is involved at the rate-limiting step of collagen degradation, and, with increased ordering of the collagen substrate, additional energy is required to properly position and/or utilize this water. Since the native collagen molecule exists as a triple helix with a hydrophobic core, the interposition of water molecules inside of this helical structure may become increasingly difficult as the substrate assumes higher orders of organization. Thus, in the case of the collagen fibril, the water of hydrolysis must traverse a liquid-solid phase boundary in addition to gaining access into the interior of the triple helix. However, the nature of the collagen substrate alone does not determine the activation energy and solvent-deuterium kinetic isotope effect for all collagenolytic enzymes. The degradation of fibrillar collagen by both clostridial and crab collagenases was characterized by a lower Arrhenius energy (19,400 and 24,300 calories, respectively) and smaller kinetic isotope effect (1.7-and 2.6-fold inhibition by 90% DzO, respectively), compared to the values for human skin fibroblast collagenase on this same substrate (101,050 calories and 9fold, respectively). Interestingly, however, in each case the magnitude of the Arrhenius energy was directly proportional to that of the deuterium isotope effect (Table 11), lending further support to the possible interrelationship between the accessibility of water molecules to the site of peptide bond hydrolysis in collagen and the energy requirements of enzymatic action.
While the activation energy and kinetic isotope effect for bacterial and crustacean collagenases on fibrillar collagen were much smaller than for human fibroblast collagenase on this same substrate, it should be noted that these three collagenases are chemically and functionally distinct from one another. Only the fibroblast enzyme is a member of the group of specific vertebrate collagenases, which are all metalloenzymes that catalyze only a single cleavage in the native collagen molecule (19). The site of this cleavage, which has been determined for tadpole and a mammalian tumor collagenase, is a Gly-Ile bond (775-776) in the a1 chain (20,21) and a Gly-Leu bond in the a2 polypeptide chain and results in typical %and %-length products, TCA and TCB (22). By contrast, the clostridial collagenase, although also a metalloenzyme, catalyzes multiple cleavages in the native collagen molecule and, in addition, manifests a very different bond specificity. The bacterial collagenase cleaves the Y-Gly bond in sequences such as Gly-Pro-Y-Gly-Pro-Z or Gly-Pro-Y-Gly-Z-Hyp, resulting in the formation of numerous NHz-terminal glycine residues in the collagen chains (23). Of interest in comparing the bonds broken by human skin and clostridial collagenases is the geographical position of the bond cleaved in each case. For the human skin fibroblast enzyme, the nucleophilic attack by a water molecule during hydrolysis is at the carboxyl carbon of glycine, which is positioned most closely to the hydrophobic center of the triple helix. For the bacterial enzyme, on the other hand, this addition of water occurs at a carbonyl group located on the outside of a turn of the collagen helix (24). It is possible, then, that this relatively more accessible location of the bond for hydrolysis is related to the lower activation energy and deuterium isotope effect for the bacterial collagenase. The higher values of these parameters for the human enzyme may be the effect of the hydrophobic location of the glycine carbonyl in the triple helix. Crab collagenase is a chymotrypsin-like serine protease which catalyzes multiple cleavages in the native collagen molecule (6,25). The collagen bond specificity of this enzyme is presently not known.
There is evidence suggesting that an unusually high activation energy characterizes not only human skin fibroblast collagenase but also other vertebrate collagenases which catalyze only a single %-% cleavage in native collagen. Hayashi et al. (26) have reported the activation energy of tadpole collagenase on types I, 11, and I11 collagen monomers in solution (EA = 41,000, 39,000, and 63,000 calories, respectively). These values are similar to the activation energy of fibroblast collagenase on type I collagen monomers (EA = 49,200 calories). In addition, Harris and McCroskery (18) have reported a 4-fold increase in the rate of cleavage of cartilage collagen fibrils by crude human rheumatoid synovial cell collagenase at 36 "C compared to 33 "C (from their data, EA calculated on this basis would be 85,000 calories). Finally, our preliminary experiments indicate that rat uterus collagenase has an EA = 75,000 calories on fibrillar collagen. Following an exhaustive search of the literature, we have been unable to find any other class of enzymes which are characterized by such a high activation energy on their physiologic substrates. Therefore, an unusually high activation energy on native collagen may be a property which is common to specific vertebrate collagenases as a group, and, as discussed above, is likely to relate to the difficulty in bringing water molecules to the site of the %-Vi length cleavage initiated by these enzymes in the native collagen molecule.
The enzymatic cleavage of native collagen by human skin fibroblast collagenase represents an enzymatic process, which, in certain respects is different from most enzyme-catalyzed reactions described to date. The unique physical structure of the collagen molecule, so necessary for its functional role in the animal organism, has probably presented nature with equally unique problems regarding its degradation. Studies designed to further investigate the extreme dependence of this process upon both energy and hydrogen transfer will be vital toward ascertaining a more complete understanding of the kinetics of collagen degradation by human skin fibroblast collagenase.