The substrate specificity of human neutrophil colla- genase was examined using both monomeric and fi- brillar collagens. The neutrophil enzyme cleaved types.
T H EJOURNALOF BIOLOGICAL CHEMISTRY
Vol. 262, No. 21, Issue of July 25, pp. 10048-10052, 1987 Printed in U.S.A.
The Collagen Substrate Specificity of Human NeutrophilCollagenase* (Received for publication, February 5, 1987)
Karen A. Hasty$#, John J. Jeffreyll, Margaret S . Hibbsll, and Howard G . Welgus**$$ From the $Department of Anatomy and Neurobiology and the 11 Department of Medicine, University of Tennessee, Memphis, Tennessee 38163, the llDivision of Dermatology, Department of Medicine and the Department of Biological Chemistry, Washington University School of Medicine, St. Louis, Missouri 63110, and the **Divisionof Dermatology, Department of Medicine, Jewish Hospital a t Washington University Medical Center, St. Louis, Missouri 63110
The substratespecificity of human neutrophil colla- and one-quarter length fragments (1). Studies designed to genase was examined using both monomeric and fi- characterize the kinetic behavior of mammalian collagenase brillar collagens. The neutrophilenzyme cleaved types have been conducted utilizing the enzyme produced by human I, 11, and I11 collagens, but failed to attack types IV or skin fibroblasts as aprototype (2-5). Such investigations have V. Against monomeric collagen substrates at 25OC, the determined that human type I, 11, and I11 collagens are all neutrophil enzyme displayed values for the Michaelis susceptible to theaction of human fibroblast collagenase and constant (K,) of 0.6-1.8 x lo-‘ M, essentially indistin- display similar substrate-enzyme affinities ( K , = 0.7-2.1 X guishable from the substrate affinitiesthat character- 1O“j M); however, monomeric type I11 collageniscleaved ize human fibroblast collagenase. Catalytic rates,how- approximately 10-fold more rapidly than the other collagen ever, varied considerably; type I collagen was cleaved types (3). The triple helical structure of native collagen apwith a specificity (kceJKm)some 20-fold greater than pears to be critical for this type specificity since, following type 111. Type I1 collagen was degraded with interme- thermal denaturation,these interstitial collagens are degraded diate selectivity, ~ 2 5 % of the type I rate, but 450% at similar rates (6). This preferential selectivity for native that of type 111. This specificity contrasted markedly monomeric type I11 collagen has also been described with with that of human fibroblast collagenase, which collagenase obtained from human lung fibroblasts (7)and cleaved human type I11 collagen 15-fold faster than human alveolar macrophages (8, 9). By contrast, collagenase type I and>500-fold more rapidly than type 11. in crude neutrophil extract hasbeen reported to degrade type Interestingly, the 20-fold selectivity for typeI over I collagen much faster than type I11 collagen (7). The basic type I11 exhibited by neutrophil collagenase against kinetic parameters which characterize the action of neutrophil monomeric collagens was largely abolished following collagenase on monomeric collagens, however, have not been the reconstitution of these substrates into insoluble measured due to the difficulty in obtaining sufficient quanfibrils, falling to a value of just 1.5-fold. The distinc- tities of this enzyme in ahighly purified form for such kinetic tive and opposite preference by the human fibroblast determinations. Recently, we have purified human neutrophil collagenase enzyme for monomeric type I11 collagen over type I (15-fold) was similarly reduced to ~ 2 - f o l dupon sub- by immunoaffinity chromatography utilizing a monoclonal strate aggregation. The transitionfrom native soluble antibody (8, 10).This has enabled us to examine, inthe presentreport, the substrate specificity of the neutrophil collagen monomers into insoluble fibrils appeared to be handled by both the human neutrophil and fibroblast enzyme against humantype I, 11, and I11 collagens in solution, collagenases with similar facility on type I substrates. as well as the cleavage of guinea pig type I for comparison to By comparison, however, the neutrophil enzyme de- a heterologous mammalian species. Our results indicate that graded typeI11 collagen fibrils faster thanwould have the neutrophil and skin fibroblast enzymes exhibit similar been predicted from solution rates, while the fibroblast values for K,, but opposite and distinctive patterns of subenzyme cleaved such fibrils much slower than expected strate selectivity among the interstitial collagens in soluble from solution values. In exploring this phenomenon form. Thus, monomeric type I collagen is cleaved20-fold further, solvent deuterium isotope effects were meas- faster than type I11 by the neutrophil enzyme but type I11 is ured. The deuterium studies suggest that neutrophil attacked some 15-fold more rapidly than type I by the skin collagenase, acting on type I11 fibrils (kH,O/kD,O = fibroblast collagenase. Interestingly, this marked type Iuersus 5.0),is less sensitive to factors which govern the avail- I11 selectivity exhibited by each enzyme against such monoability of water at the relatively hydrophobic site of meric collagens is largely abolished when the substrates are peptide bond hydrolysis in the collagen molecule than present in the insoluble fibrillar form. In exploring this phenomenon more fully, solvent deuterium isotope effects were is fibroblast collagenase (kH,O/kD,O = 15.0). measured. These deuterium studies suggest that neutrophil collagenase is less sensitive to factors which govern the availability of water at the relatively hydrophobic site of peptide The degradation of native collagen is initiated by the action bond hydrolysis in the collagen molecule than is fibroblast of collagenases, a class of neutral metalloproteinases which collagenase. cleave the triple helical collagen molecule into three-quarter
* This work was supported in part by United States Public Health
Service Grants AI 22603 (K. A. H.), HD 05291 (J. J. J.), AM 01138 (M. S. H.), and AM 30805 (H. G . W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. To whom correspondence should be addressed Research Service (151), VA Medical Center, 1030 Jefferson St., Memphis, T N 38104. $4 Recipient of Research Career Development Award AM 01525.
MATERIALS AND METHODS
Reagents-Acrylamide and bisacrylamide were obtained from Eastman. Sodium dodecyl sulfate (99% pure) and deuterium oxide (99.8%) were purchased from Gallard-Schlesinger. Tris base, bovine pancreatic trypsin, and soybean trypsin inhibitor were obtained from Sigma. All other chemicals were reagent grade. Sources of Collagen-Human collagen preparations were kindly donated by Dr. Robert Burgeson, Shriner’s Hospital, Portland, OR (human placenta type I, human cartilage type 11, human placenta
10048
Collagen Specificity of Neutrophil Collagenase type 111). Guinea pig skin type I collagen was prepared in our laboratory as described by Gross (11). Purification of Collagenases-Human skin fibroblast procollagenase was purified from serum-containing medium by sequential chromatography on carboxymethylcellulose and Ultrogel AcA-44 matrices according to the methods of Stricklin et al. (12). The fibroblast zymogen was proteolytically activated by exposure to optimum concentrations of trypsin at room temperature for 10 min (12). Titrations were performed on each batch of enzyme to ensure maximal activation; generally, a 1:1weight ratio of trypsin to collagenase resulted in maximum activation. The tryptic activity was inhibited in all cases with a 20-fold molar excess of soybean trypsin inhibitor. Human neutrophil collagenase was prepared from culture supernatants of neutrophils that had been stimulated with phorbol myristate acetate to degranulate and release their collagenase (8). The enzyme was then purified by immunoaffinity chromatography using a monoclonal antibody specific for neutrophil collagenase (10). The collagenase was eluted from the immunoaffinity column with 3 M NaSCN containing 0.05 M Tris-HC1, 0.15 M NaCl, 0.005 M CaClZ, and 0.02% NaN3 a t pH 7.4, a step which converts the enzyme from the latent toactive form. were determined Assay Procedures-Values for K , and VmaX(4,) using collagen in solution a t 25 "C as substrate (3). For each collagen, 4, 6, 8, 10, 12, 16, and 20 pg of native substrate were diluted in 50 pl of buffer (0.05 M Tris, 0.01 M CaCl,,0.25 M NaCl, pH 7.5) and incubated with 50 pl of an appropriate concentration of active collagenase at 25 "C. The reaction mixtures were then stopped with 75 pl of sample buffer containing EDTA and dithiothreitol, boiled, and subjected to polyacrylamide gel electrophoresis according tothe method of King and Laemmli (13) in 9% slab gels. The gels were subsequently stained with 1% Coomassie Blue and densitometrically scanned using a Gilford Response spectrophotometer set at 600 nm equipped with a gel scanning linear transport device. Collagen degraded (%) = 4/3 [TCA] + (4/3 [TCA] [a]). Fibrillar collagen assays were performed using reconstituted insoluble fibrils as substrate which were prepared by incubating monomeric guinea pig type I or humantype 111collagens overnight at 37 'C a t neutral pH (5). Guinea pig type I collagen was biosynthetically labeled with ["Clglycine to a specific activity of ~ 2 5 , 0 0 0cpm/mg (14).Human type 111 collagen was not isotopically labeled. The insoluble collagen gels (50 pl, ~1.5-2.0pg/pl) were then exposed to 50 pl of enzyme a t 37 "C for various periods of time, the maximum of which produced approximately 50% substrate lysis. The reaction mixtures were centrifuged at 10,000 X g to separate degraded from insoluble collagen. In the case of guinea pig type I collagen, the supernatants were counted in a liquid scintillation spectrometer. For human type 111, solubilized collagen in the supernatant and undegraded collagen in the precipitate fractions were each quantitated by hydroxyproline assay (15). Specific activity was calculated as micrograms of collagen degraded per mg of enzyme per min at 37 "C. Protein concentrations were measured spectrophotometrically at 224 and 233 nm using bovine serum albumin as a standard (16). The hydroxyproline content of the various collagens was determined by the method of Bergmann and Loxley (15). Solvent Deuterium Kinetic Isotope Effect-Reconstituted collagen fibrils were used to determine the effect upon enzyme catalytic rates of substituting deuterium oxide for water in reaction mixture buffer. Experiments were performed using methodology as described previously by Welgus et al. (17). Collagenases were dialyzed overnight against 0.05 M Tris-HC1 buffer, containing 0.01 M CaClZ,0.15 M NaCl, pH 7.5, in 99.8% deuterium oxide. Fibrillar collagen gels werewashed twice with DzO-containing buffer prior to incubation at 37 "C with collagenase. Catalytic rates were determined in buffer containing final reaction mixture concentrations of 90% DzOversus 100% H20.
10049
single cleavage in the a chains of type 1-111 collagens, resulting in the three-quarter-length TCA and one-quarter-length TCB fragmentscharacteristic of mammalian collagenase action (not shown). Neutrophil collagenase exhibited typical substrate saturation with increasing amounts of collagen, and linear Lineweaver-Burk plots were obtained (Fig. 1).Thus, under the conditions employed, the neutrophil collagenase displayed normal Michealis-Menten kinetics. Inspection of typical Lineweaver-Burk plots in Fig. 1 for types I and I11 collagens revealed a marked disparity in the neutrophil enzyme's specificity for these collagen types. In Table I, values for k,,,, K,, and the ratio kcat/Km, often utilized as a measure of substrate specificity, are compared for human neutrophil versus fibroblast collagenases against the various monomeric collagen types. The range of K,,,values exhibited by the neutrophil enzyme for all collagens examined . values for substrate was quite narrow, K , = 0.6-1.8 p ~ Such affinity werevery similar to those previously reported for human skin fibroblast collagenase (3) (Table I). Therefore, both the neutrophil and fibroblast proteinases demonstrated a relatively uniform affinity for susceptible collagen types. In contrast, measurements of catalytic rate (kc.,) and subI
:I O
A
8
+
'*'I
:ii
B
I/
3.0
I
RESULTS
The substrate specificity of human neutrophil collagenase against monomeric collagens was assessed by incubating increasing amounts of collagen substrate with a constant concentration of enzyme at 25 "C. The reaction process was terminated by the addition of EDTA, and the TCAand TCB products separated from intact CY chains by sodium dodecyl sulfate-polyacrylamide gel electrophoresis andquantitated densitometrically as described under "Materials and Methods." As expected, the neutrophil enzyme catalyzed only a
I / [Substrate. vM]
FIG. 1. Determination of K,,, and V,, (kat) on monomeric collagens. Lineweaver-Burk plots are shown for the degradation of monomeric guinea pig type I (A) and human type 111 ( B ) collagens by neutrophil collagenase at 25 "C. The data for type I cleavage have been replotted in B for purposes of comparison. Reaction velocity is expressed as thenumber of collagen moleculesdegraded per molecule of collagenase per h. Note that both Lineweaver-Burk plots are linear and also the marked difference in kcatexhibited by the human neutrophil enzyme for the cleavage of type I uersus 111 collagens ( B ) .
10050
Collagen Specificity of Neutrophil Collagenase TABLEI
TABLEI1
Substrate specificity of human neutrophil versus humanfibroblast collagenase on monomeric collagens The K , and V,, (kcat) of humanneutrophilcollagenase were determined at 25 “C as described under “Materials and Methods.”
by Values for human skin fibroblast collagenase, reported previously Welgus et al. (3), are shown for comparison. The ratio, kcat/K,, is employed as an index of substrate specificity. Human neutrophil Human skin fibroblast collagenase
collagenase Collagen
Guinea pig Human I Human I1 Human 111
Km
kcat
pM
0.6 0.7 1.1 1.8
kcdKm
Km
h”
h” p M ’
pM
5.81 6.40 2.35 0.85
9.68 9.14 2.14 0.47
0.9 0.8
2.1 1.4
Specific activity of human neutrophil versus human fibroblast collagenase on fibrillar collagens &constituted collagen fibrils (100 pg; 50 pl) were incubated with
collagenase at 37 “C. Following centrifugation, supernatant and precipitate fractions wereanalyzed as detailed under “Materials and Methods.” Specific activity is expressed as micrograms of collagen degraded per mg of enzyme/min at 37 “C. activity Specific Collagen source Enzyme dmglmin
kcat h”
h” p M ’
22.5 53.4 1.0 565.0
25.0 66.7 0.5 403.6
LlKm
Neutrophil Fibroblast
Guinea pig I Human I11 Guinea pig I
Human I11
130 90 540 970
TABLEI11 Deuterium isotope effect on fibrillar collagen degradation Fibrillar collagendegradation by humanneutrophilcollagenase
strate specificity (kCat/Km) revealed marked differences between the two human collagenases. Neutrophil collagenase was measured inthe presence of solvent buffer containing100% H20 90% D20 (kH,O/kD,O) as detailedunder“Materials and cleaved human type I collagen 8-fold faster than the homol- versus Methods.” Values for human skin fibroblast collagenase, previously ogous type I11 substrate (kcat= 6.40 h-l uersus 0.85 h-’), and reported by Welgus et al. (5, 17),are shown for comparison. the magnitude of this specificity approached 20-fold when kH201kDn0 values of k,,,/Km were compared. As shown in TableI, and as Collagen Neutrophil Fibroblast Neutrophil/fibroblast reported previously (3), human fibroblast collagenase exhibited an opposite selectivity of roughly equivalent magnitude 7.2 9.0 0.80 Guinea pig I Human 111 5.0 15.0 0.33 for the monomeric human typeI11 substrate over type I. Thus, the neutrophil and fibroblast enzymes manifest strong, but opposite, selectivities in the cleavage of homologous type I tude of this difference was only 1.8-fold as compared to the uersus I11 collagens. While types I and I11 collagens varied 15-fold disparity observed with the same collagens in a soluble greatly in their susceptibility to degradation by the neutrophil monomeric form. Thus, the marked specificity exhibited by enzyme, the species of origin of the collagen substrate did not neutrophil collagenase for monomeric type I collagen and by appear toaffect proteolytic rates, at least with respect to typehuman skin fibroblast collagenase for type I11 largely disapI collagen. The kinetic parametersof guinea pig type I cleav- pear when the substrates areorganized into insoluble fibrillar age were virtually indistinguishable from those of human type arrays. I. An additional observation in this study was that, utilizing Differences in the catalytic properties of the neutrophil and the predominant collagen in the human organism, type I, as skin fibroblast enzymes were further apparent when monoa yardstick, the neutrophil collagenase appeared to possess meric type I1 collagen degradation was studied. Neutrophil only ~ 2 5 %of the catalytic activityof the fibroblast enzyme. collagenase attacked the type I1 substrate at a rate interme- This value was essentially constant whether degradationwas diate between its cleavage of type I and I11 collagens (Table measured againsteither monomeric (Table I) or fibrillar I). In contrast, typeI1 was degraded by the fibroblastenzyme (Table 11) forms of type I collagen. at only 2% therate of type I and 0.2% therate of the Previous evidence obtained for human fibroblast collagenhomologous type I11 collagen. Thus, the substrate specificity ase indicates that peptide bond hydrolysis of the fibrillar exhibited by neutrophil collagenase against human type I, 11, collagen molecule is the rate-limiting stepof degradation (2and I11 collagens was totally different from that observed for 6, 17). Sincethe collagenolytic activity of theneutrophil the human fibroblastenzyme. enzyme against typeI11 collagen fibrils was greater thanwould Although study of the proteolytic cleavage of monomeric have been expected from solution values (Tables I, 11, and collagens is amenable t o analysis by Michaelis-Menten kinet- IV), thesolventdeuteriumkinetic isotopeeffectfor this ics, it mustbe remembered that during the in uiuo degradation enzyme was examined. The cleavage of fibrillarcollagens, of collagen such monomeric forms do not exist in the extra- types I and 111, was measured following the substitution of cellular space where the substrate isorganized into insoluble deuterium forhydrogen insolvent buffer of thereaction fibrillar arrays. To examine the activity of neutrophil colla- mixtures. As shown in Table 111, the neutrophil and fibroblast genase on collagen fihrils, monomeric type I and I11 collagens enzymes both exhibited similar values for kH20/kD20 on type were allowed to polymerize overnight at 37 ”C prior to incu- I collagen fibrils (7.2 uersus 9.0, respectively). However, bation with enzyme. Purified type I1 collagen does not form against the typeI11 fibrillar substrate, the neutrophilenzyme reconstituted fibrils and therefore could not be evaluated.As was far less affected by deuterium oxide than its fibroblast shown in Table 11, the extraordinary type I uersus I11 speci- counterpart (kH20/kD20= 5.0 uersus 15.0, respectively). ficity exhibited by the neutrophil collagenase for such monomeric collagens largely disappeared with the fibrillar subDISCUSSION strates. Although the neutrophil enzyme still degraded type I The data presented in this paper provide further evidence collagen fibrils at a higher specific activity than type111, this ratio was only 1.5-fold, as compared to 20-fold against the that the collagenases derived from human neutrophils and solution monomers. An analogous diminution in the magni- from human fibroblasts are distinct proteases. Previous retude of specificity on fibrillarcollagen substrates has also ports have documented a number of differences between the been observed with human fibroblast collagenase (5) (Table two molecules, including disparate molecular weights, isoelec11). While type I11 collagen fibrils were still more susceptible tric points, andimmunologic determinants (8, 10, 12, 18, 19). to cleavage by the fibroblast enzyme than type I, the magni- Additionally, neutrophil collagenase is stored within the spe-
Collagen Specificity of Neutrophil Collagenase
10051
cific granule compartment of the cell (20), whereas thefibro- blast enzyme the ratio is less than 2, and for the neutrophil blast secretes the vast majority of its synthesizedenzyme (21). collagenase, the ratio is greater than100, indicating that the Thedatainthisreportdocumentyetfurther differences neutrophil collagenase is in some way significantly less conby the aggregation of type I11 monomers between these two “interstitial” collagenases, revealing them strained in its activity enzyme. to have fundamentally different catalytic properties with re- into fibrils than is the fibroblast It should be noted that our previous studies, using the spect to nativecollagen substrates. Oneconsistentdifference between the two enzymes, fibroblast collagenase, revealed that the aggregation of collawhether using soluble or fibrillar substrates, is that the spe- gen molecules into fibrils always resulted in a sharp decrease cific activity of the neutrophil collagenase is approximately in the activity of the enzyme from the values predicted by one-quarter thatof the fibroblastcollagenase. It is notpossible rates in solution (2, 3, 17). These values would predict that at this time to assign the reason for this difference; it is the ratio of specific activity to kc,, illustrated in Table IV conceivable that a lower catalytic efficiency is a n intrinsic should exceed 500, instead of the observed 23. In the case of property of the granulocyteenzyme. However,the purification the fibroblast enzyme, the decrease in activity observed with procedure for the granulocyte collagenase is harsh, since the type I11 fibrils was considerably greater than that observed elution of the enzyme from the immunoaffinity column re- with type I fibrils as substrate (5). In order to attempt to quires the use of a chaotropic agent (3 M NaSCN), resulting explain these large decreases in catalytic activity, we have of peptide bond in a loss of 50 percent of the enzymatic activity in the eluant postulated that theaccess of water to the site (10). It is possible that the enzyme molecules inactivated by hydrolysis is severely restrictedin fibrillar collagen. This of this elution method are responsible for the apparently low explanation was suggested by the fact that the substitution specific activity of the resultant preparation. An alternative deuterium oxide for water in reaction mixtures resulted in purification scheme for the granulocytecollagenase, one not extremely large decreases in the rateof collagenolysis. These involving the use of potentially denaturing reagents,would be decreases, occasionally as high as 40-fold, were too large to be explained by classical kinetic deuterium isotope effects (22), of considerable help in resolving this issue. Although there may be discrepancies in thespecific activity and were invariably correlated with the organization of the of these proteinases, it is clear that their relative catalytic substrates into higher order structures. Thus, for example, rates against themonomeric forms of collagen types I, 11, and when denatured collagen, or gelatin, chains were utilized as a I11 show distinctive differences (Table I). Against such sub- substrate for the enzyme, no deuterium isotope effect was strates the neutrophil collagenase exhibits a selectivity to- observed on the catalytic rate. When the same chains were wards type I versus type I11 in excess of 7-fold, while the organized in native triple helical configuration, a deuterium fibroblast enzyme prefers type I11 over type I by a factor of isotope effect was now measurable, ranging in value from 1.5 more than 10. Differences of a n even greater magnitude are to 2.0. Aggregation of collagen molecules into fibrils then observed for the cleavage of type I1 collagen. Human type I1 resulted in very high values for kH20/kD20,ranging from 10 collagen is attacked by the neutrophil enzyme at approxi- to 40 (4,5, 17). In an effort to better understand the ofnature this disparity mately 40% the rate of type I and 275% that of type 111. On the other hand, the fibroblast collagenase degrades type I1 between the action of the neutrophil and fibroblastcollagenonly 2% as rapidly as type I and at 0.2% the rate of type 111. ases on typeI11 collagen, the effect of substituting deuterium Clearly, then, the neutrophil and fibroblast collagenases dis- oxide for water on the rates of collagenolysis was examined. fibroblast play very different selectivities for the interstitial collagen The data presentedin Table I11 reveal that both the and the neutrophil enzymes exhibit a similardegree of activity types in solution. Although the examination of monomeric collagen degrada- loss on type I fibrillar collagen when DzO is substituted for mixtures. The ratios of activities in tion allows for the measurement of important kinetic param- water in the reaction eters, it must be emphasized that this form of the substrate water versus thosein DzO (kH,O/kD,O) are 9.0 and 7.2, respectively. In contrast, when type I11 collagen fibrils are is essentiallynonexistentintheextracellular spacewhere collagen is presentas insolublefibrils. It is, therefore, of used as substrate, degradation by the neutrophil enzyme is considerable physiologic interest that following the reconsti- slowed only 5-fold, compared to a 15-fold reduction in the rate of the fibroblast enzyme. Thus, the neutrophil collagentution of solublecollagen monomersinto insolublefibrils, much of the selectivity for collagen genetic type exhibited by ase appears tobe significantly better able to handle,in some both the neutrophil and the fibroblastenzymes is lost. Thus, way, the problem of water access to therelatively hydrophobic although the neutrophilcollagenase cleaves type I monomers site of peptide bond hydrolysis within the collagen helix of a t approximately 20 times the rateof type I11 (kcat/& = 9.68 type 111, but not type I, fibrils. The result, presumably, is a more rapid cleavage of type I11 fibrils than would be expected versus 0.47), this ratio is reduced to 1.5 when the fibrillar in form of these two collagens is employed as substrate. Simi- from solutionrates.Theserelationshipsareillustrated fibroblast larly, the preference of the fibroblast enzyme for soluble type Table IV. Clearly then,theneutrophilandthe I11 over type I exceeds 15-fold, whereas thispreference is less collagenases deal withthe transition of type I11 collagen from than two with the corresponding fibrillar substrates (Table monomers tofibrils in a fundamentally different manner. The exact basisfor the catalytic discrepancies between the 11). Further analysis of the data reveals that these disparities two enzymes, whereby they interact with substrates in sucha are essentially all attributable to the characteristics of type very different way, is notknown. Since much of the difference I11 degradationandnotthat of type I (Table IV). This between the two molecules is related to type I11 collagen, it conclusion is derived from the fact that the ratioof fibrillar will be of considerable interest to further examine this aspect to monomeric degradation rates for type I collagen is approx- of comparative enzymology. We have observed previously,for example, that type I11 collagens from various animal species imately the same for both the neutrophil and fibroblast enappear to differ in the stabilityof their helix in the region of zymes(specific activitylk,,, = 22-24), indicating that both handlethetransition frommonomer to fibril essentially the collagenase cleavage site (5). The human type I11 collagen which the neutrophilenzyme equally. In the caseof type I11 collagen, however, these ratios employed in this study, and that are markedly different for the two collagenases. For thefibro- would encounter in vivo, falls into the category of “loose”
Collagen Specificity of Neutrophil Collagenase
10052
TABLEIV Effect upon kinetic parameters of substrate transition from collogen monomers to fibrils The ratio of each collagenase’s catalytic activity on guinea pig type I collagen fibrils at 37 “C (specific activity) uersus guinea pig type I solution monomers at 25 “C (kat) is derived from the data in Tables I and 11. Similarly, the ratios are shown of catalytic activity on human type 111 collagen fibrils at 37 “C (specific activity) uersus human type I11 solution monomers at 25 “C (kcat).Note the similarity in the ratio of catalytic rates (22.4,24.0)and for both enzymes against type I collagen. Such values exhibit marked disparity deuterium isotope effects (7.2, 9.0) with regards to type 111 collagen. Type I Enzyme
15.0 7.2
Fibroblast Neutrophil
Fibrillar rate (specific activity) Solution rate (Lt)
1.72
24.0 22.4
helical type I11 collagens. Other species possess type 111 collagen with considerably tighter helices; it is possible that the integrity of this helical region can affect the ability of water to gain access to the site of hydrolysis in fibrillar substrates. If this is indeed the case, the fibroblast enzyme appears unable to exploit such substrate-mediated differences in water availability. It may be, however, that the granulocyte collagenase is considerably more sensitive to the presence of water and can utilize it much more effectively in a loose helix fibrillar type I11 collagen. This hypothesis can be tested by examining the behavior of the two enzymes on fibrils of type 111collagens with both loose and “tight” helices at their collagenase cleavage site. Such studies should include an assessment of type I11 fibril diameters compared to those of type I, since fibril diameter can clearly affect the number of collagen molecules available for enzyme binding. In this regard, it will also be necessary to examine the characteristics of binding of both the neutrophil and fibroblast collagenases to type 111collagen fibrils, as has previously been accomplished for the fibroblast enzyme and type I fibrils (2). This, and other aspects of the neutrophil collagenase’s mechanisms of catalysis will demand further investigation. Acknowkdgment-We gratefully acknowledge the excellent technical assistance of Catherine Fliszar. REFERENCES 1. Gross, J., andNagai, Y. (1965)Proc. Natl. Acad. Sci. U. S. A. 54, 1197-1204 2. Welgus, H. G., Jeffrey, J. J., Stricklin, G. P., Roswit, W. T., and Eisen, A. Z. (1980)J. Biol. Chem. 255,6806-6813
Type I11
kH20 kD20
rate Fibrillar (specific activity) Solution rate kt)
kH20 kD20
9.0 5.0
3. Welgus, H. G., Jeffrey, J. J., and Eisen, A. Z. (1981)J. Biol. Chem. 266,9511-9515 4. Jeffrey, J. J., Welgus, H. G., Burgeson, R. E., and Eisen, A. Z. (1983)J. Biol. Chem. 268,11123-11127 5. Welgus, H. G., Burgeson, R. E., Wootten, J. A. M., Minor, R. R., Fliszar, C., and Jeffrey, J. J. (1985)J. Biol. Chem. 260, 10521059 6. Welgus, H. G., Jeffrey, J. J., Stricklin, G. P., and Eisen, A. Z. (1982)J. Biol. Chem. 257, 11534-11539 7. Horowitz, A. L., Hance, A. J., and Crystal, R.G. (1977)Proc. Natl. Acad. Sci. U.S. A. 74,897-901 8. Hasty, K. A., Hibbs, M. S., Mainardi, C. L., and Kang, A. H. (1984)J. Exp. Med. 169, 1455-1463 9. Welgus, H. G., Campbell, E. J., Bar-Shavit, Z., Senior, R.M., and Teitelbaum, S. L. (1985)J. Clin. Invest. 76, 219-224 10. Hasty, K. A., Hibbs, M. S., Kang, A. H., and Mainardi, C.L. (1986)J. Biol. Chem. 261,5645-5650 11. Gross, J. (1958)J. Exp. Med. 108, 215-226 12. Stricklin, G. P., Bauer, E. A., Jeffrey, J. J., and Eisen, A. Z. (1977)Biochemistry 16, 1607-1615 13. King, J., and Laemmli, V.K. (1971)J. Mol. Biol. 62,465-477 14. Nagai, Y.,Lapiere, C. M., and Gross, J. (1966)Biochemistry 6, 3123-3130 15. Bergmann, I., and Loxley, R. (1963)Anal. Chem. 35, 1961 16. Groves, W. E., Davis, F. C., Jr., and Sells, B. (1968)Anal. Biochem. 22. 195-210 17. Welgus,H.G.; Jeffrey, J. J., and Eisen, A. Z. (1981)J. Biol. Chem. 256,9516-9521 18. Hastv. K. A.. Stricklin. G. P.. Hibbs. M. S.. Mainardi. C. L.. and King, A. H . (1986)Arthritk Rheum., in press 19. Stricklin, G. P., Eisen, A. Z., Bauer, E. A., and Jeffrey, J. J. (1978)Biochemistry 17,2331-2337 20. Murphy, G., Reynolds, J. J., Bretz, V., and Baggiolini, M.(1977) Bwchem. J. 162,135-197 21. Valle, K.-J., and Bauer, E. A. (1979)J. Biol. Chem. 254,1011510122 22. Walsh, C. (1979)Enzymatic Reaction Mechanism, Vol. 112 ff, Freeman Publications, New York