DEVELOPMENT OF A MAMMALIAN COLLAGENASE ASSAY USING A SOLUBLE SUBSTRATE

LaPosta-Frazier, Nancy A. M.S., University of PJiode Island, 1980. Development of a Mammalian Collagenase Assay Using a Soluble Substrate. Major Professor: Dr. George C. Fuller. The classical assay for detection of collagenolytic activity has been the release of soluble peptide fragments from reconstituted radioactive gels following incubation with the enzyme. Because this assay system is dependent on diffusion of the enzyme through the gel, examination of the kinetic parameters of the enzyme has been hampered. In the present study, a soluble substrate for collagenase was developed by coupling purified Type I collagen to the fluorophor, 2-methoxy-2,4-diphenyl-3(2H) furanone (MDPF). MDPF-labeled collagen was incubated with varying amounts of collagenase for two hours and electrophoresed on polyacrylamide gels. The gels were scanned on a Gilford spectrophotometer and the resulting peaks quantitated on a Neumonic Electronic planimeter. It was found that the relationship between substrate disappearance and product disappearance was not stoichiometrical. After treatment of the fluorophor-labeled substrate with pepsin, the formatio~ of product and breakdown of substrate was stoichiometrical.

The classical assay for detection of collagenolytic activity has been the release of soluble peptide fragments from reconstituted radioactive gels following incubation with the enzyme. Because this assay system is dependent on diffusion of the enzyme through the gel, examination of the kinetic parameters of the enzyme has been hampered.
MDPF-labeled collagen was incubated with varying amounts of collagenase for two hours and electrophoresed on polyacrylamide gels. The gels were scanned on a Gilford spectrophotometer and the resulting peaks quantitated on a Neumonic Electronic planimeter. It was found that the relationship between substrate disappearance and product disappearance was not stoichiometrical. After treatment of the fluorophor-labeled substrate with pepsin, the formatio~ of product and breakdown of substrate was stoichiometrical.
ii To test the integrity of the triple-helix, MDPF-labeled Type 1 collagen was treated with trypsin. The results indicate that substrate treated with pepsin was more resistant to trypsin degradation than fluorescent collagen not pretreated with pepsin.
iii ACKNOWLEDGEMENTS I wish to thank Dr. George C. Fuller not only for his patience and guidance, but more importantly, for showing me how to think and reason scientifically.
Special thanks to Ronald Goldberg for the preparation of the Type ± collagen and to Dr. Paul S. Cohen for the preparation of the 14 c-tryptophan-labeled E. coli protein.
Lastly I wish to thank my husband, Michael, for putting up with grinders and cold pizzas so that this thesis could be written.
iv  (Grant and Prockop, 1972); however, recently several workers have proposed that increased collagen deposition may result from an interplay of increased synthesis and decreased degradation (Perez-Tamayo, 1978).
Defining the role of vertebrate collagenase in these collagen accumulation diseases has been hampered by lack of a sensitive, convenient assay which quantitates vertebrate collagenase. The classical assay for detection of vertebrate collagenase consists of applying a tissue sample or extract suspected of containing enzyme on a radiolabeled collagen gel and subsequently, counting solubilized peptides (Gross and Lapiere, 1962;Gross and Nagai, 1965). Problems with this assay procedure include lengthy preparation times, low specific activity of substrate and dependence on the rate of diffusion of the enzyme through the gel. Newer assay procedures have attempted to deal with these problems but a sensitive assay has yet to become the standard in the field.
It is the purpose of this present study to attempt to develo. p such an assay for vertebrate collagenase utilizing a purified substrate coupled to the fluorophor, 2-methoxy-2, 4-diphenyl-3-(2N)furanone (MDPF). Essentially, the procedure consists of incubating human skin collagenase with the fluorescent-labeled substrate followed by quantitation of the reaction products as isolated on polyacrylamide gels.

LITERATURE SURVEY
Collagen Heterogeneity Collagen, the major fibrous protein constituent of connective tissue, (Kivirikko and Risteli, 1976), exists as a right-handed triple helical molecule, con~isting of intertwined alpha chains. There are seven genetically distinct a-chains. Each a-chain is composed of approximately 1000 4 amino acid residues with an average molecular weight of 95000 daltons (Kivirikko and Risteli, 1976 is found in muscle, aorta and skin. It can be separated from Type I collagen by differential salt precipitation following pepsin digestion (Chung and Miller, 1974;Epstein, 1974) and differs from Type I in containing more 4-hydroxyproline and cysteine residues. The latter form disulfiae bonds within the helical portion of the molecule. Basement merobrane collagen, or Type IV, contains high amounts of hydroxylated amino acids and cysteine residues. Type IV collagen has a higher molecular weight than the other collagen types attributable to a 12.5% hexose content (Kefalides, 1971 (Minor, 1980). In its amino acid composition, Type V collagen is similar to Type IV, in that it contains a significant number of both 3-hydroxyproline residues and glycosylated hydroxylysines (Burgeson et al., 1976).

The Triple Helix and Enzymatic Cleavage· of Collagen
One of the crucial steps in collagen biosynthesis is the association of a-chains into the triple helix. The formation and strength of the helix is dependent on both the disulfide bonds found in the N-terminal peptide extensions of the molecule (Kivirikko and Risteli, 1976) as well as to the hydroxylation of praline residues with subsequent synthesis of water bridges (Minor, 1980). The tightness of the triple helix is attributable to the presence of glycine as every third residue in the primary amino acid sequence •.
Because of its unique helical conformation, native collagen is resistant to nonspecific protease attack; however, it is susceptible to cleavage by two types of collagenases, each possessing a distinct specificity for the collagen molecule.
Bacterial collagenase (EC 3.4.24.3), commonly isolated from Clostridiurn histolyticum, cleaves the molecule at primary and secondary sites (Seifter and Harper, 1970). The most common point of attack occurs in the center of a four residue sequence consisting of Proline-Y-Glycine-Proline. The secondary sites of cleavage are numerous and appear to occur in four residue sequences where the second amino acid is usually hydroxyproline and the third residue is always glycine (Weiss, 1976).
Vertebrate and marnrnalian collagenases attack the collagen molecule at one specific locus between residues 772-773.
This cleavage site corresponds to the peptide bond between glycine and leucine in the etl(I) chain and to the amino acids glycine and iso-leucine in the a.2 chain (Gross et al., 1974). The two peptide products formed in the reaction comprise 75% and 25% of the original molecule and are called TCA and TCB' respectively (Gross and Nagai, 1965) .
Although fibrillar collagen possesses a denaturation temperature in excess of 55°C, collagen in solution denatures at 37°C. In contrast to these, TCA has a denaturation temperature (T ) of 32 C, while TCB has a T of 29°~. At m m physiological temperatures, then, the helical products would quickly denature and become susceptible to nonspecific protease attack.

·-
Early pioneers in the field (Sakai and Gross, 1967) hypothesized that collagenase was responsible for the further degradation of the cleavage products; however, later evidence indicates that this is not so. McCroskery et al. (1973) provided visual proof on polyacrylamide gels that collagenase cleaves native collagen into two characteristic products only. Furthermore, the enzyme is unable to cleave denatured collagen, that is, gelatins, under assay conditions identical to those in which the enzyme is active Instead, gelatin is cleaved by specific peptide hydrolases to dialyzable fragments (Weiss, 1976).
Physical Characteristics of Collagenase ~he molecular weight of collagenase varies between 31000 and 60000 daltons depending on the source of the enzyme.
Collagenase requires calcium as a cofactor and is inhibited by whole serum, a 2 -macroglobulin and metal chelators, such as EDTA and 1,10 phenanthroline (Birkedal-Hansen et al., 1976). Early reports indicated that human skin collagenase was inhibited by a 1 -anti-trypsin (Eisen et al., 1971); however, subsequent :data (Birkedal-Hansen et al., 1976) suggest that the antiprotease is inactive against the enzyme.
It is quite possible that the antiprotease used by Eisen was contaminated with a 40000 molecular weight protein which elutes after a 1 -antitrypsin on Sephadex G-200 (Woolley et al., 1975).
Collagenase has been isolated in a precursor f orrn but whether the latent enzyme exists as a zymogen or as an enzymeinhibitor complex has been the subject of debate. Early work (Eisen et al., 1971) indicated that a lag period of 24 to 48 hours occurred before the appearance of active collagenase in media of explants of human skin; however, irnrnunoreactive enzyme was detected within the first six hours. The authors hypothesized that the initial lack of enzyme activity was attributable to inhibition by serum anti-proteases. They did not consider the possibility of latent enzyme.
The question of whether collagenase existed as a proenzyme was further investigated in 1976. In that year, Birkedal-

Hansen et al. (1976) isolated a collagenase zymogen from
human fibroblasts which did not bind to a 2 -macroglobulin.
Upon activation with trypsin and loss of an 18000 molecular weight peptide, the enzyme reacted with the antiprotease. It should be noted that the human fibroblast enzyme was not unique in its susceptibility to trypsin activation: a collagenase isolated from human leukocytes was activated by both trypsin and rheumatoid synovial fluids. The proenzyme was released as a result of phagocytosis of aggregated Y-globulin by neutrophils (Oronsky et al., 1973).
In summary, the enzyme can be physiological ly inhibited in two ways: initially, it is synthesized as an inactive precursor, but upon activation, can still be inhibited by serum proteins such as a 2 macroglobulin.

Assays for the Detection of Collagenase Activity
The classical assay for the detection of collagenase activity has been the release of soluble peptide fragments from reconstituted radioactive gels following incubation with the enzyme (Gross and Lapiere, 1962) . In these earliest experiments, explants of amphibian tail fins were utilized as the source of enzyme. In subsequent assays (Gross and Nag~i, 1965;Nagai et al., 1966;Sakai and Gross, 1967) enzyme purified by ammonium sulfate precipitation and Sephadex gel filtration was used.
In addition to the solubilization of the radioactive gels, earliest researchers capitalized on two physiochemical properties of collagen: its viscosity and high negative optical rotation. Upon incubation with collagenase (Gross and Nagai, 1965), collagen in solution undergoes a drop in viscosity without any change in optical rotation.
The authors interpreted this to mean that collagenase cleaved the collagen molecule without changing the triple hel i cal nature of the substrate or products. Since this experiment was conducted at 20°C, well below the denaturation temperature of collagen, TCA or TCB their early interpretation has proved correct. Visualization of the products of the reaction was made on polyacrylamide gels and by segment long spacing (SLS) electron microscope (Gross and Nagai, 1965).
Although the gel lysis method was a milestone in collagenase research, there are drawbacks to its utilization on a wide scale. First, the labeled procedure for collagen as outlined by Gross, calls for the injection of the radio- ac ive ~so _ ope, C-g ycine, into guinea pigs an su sequent extraction and purification of the collagen. This results in a substrate with low specific activity. Secondly, the _actual preparation of the gels is time-consuming: collagen in solution must be incubated for twelve hours at 37° after which, the gels must be disrupted with a steel needle and reincubated an additional hour (Nagai et al., 1966). In addition, the assay suffers from high background and a sensitivity which varies depending on pH, ionic strength and ionic species present in test samples (Terato et al., 1976).
Lastly, results of the assay are highly dependent on rate of diffusion of the enzyme through the gel preventing kinetic analysis of the enzyme. As a result, many investigators have sought to develop new assay systems which would prove easier and more sensitive to use.
One of the first modifications of Gross's assay system involved the formation of collagen gels after incubation with the enzyme (Sakamoto et al., 1972 , 1978), production of free radicals in evacuated quartz tubes and subse. quent reaction with 3 H 2 s (Labrosse et al., 1976) and acetylation reactions of acid soluble collagen and (l-14 c) acetic anhydride (Gisslow and McBridge, 1975). In all cases, the labeled collagen produced was higher in specific activity; however, other problems still persisted.
In th_ e collagenase reactions of l5oth Labrosse· et· al. (19 76) and Gisslow  In testing his substrate for susceptibility to nonspecific protease attack, 24°C was selected as the incubation temperature, instead of the 37°C used in those experiments designed to prove the increased sensitivity of this system.
Other current assays for collagenase include a film microassay procedure (Levenson, 1976) and dioxane extraction of in vivo labeled reaction products (Terato et al., 1976).
Again, the latter method suffers from low specific activity of the collagen substrate. Likewise, data was not presented concerning the substrate's resistance to nonspecific protease attack.
Recently, Goldberg and Fuller (1978) described a method for quantitating denatured collagen in solution.
In their system, purified collagen a 1 -chains were coupled to the fluorophor, 2-methoxy-2,4-diphenyl-3-(2H)-furanone (MDPF) and quantitated after separation by electrophoresis in polyacrylarnide gels. One of the distinct advantages to MDPF is that the molecule itself is not fluorescent, but becomes so only after coupling to primary amines. Any MDPF that remains unreacted is subsequently hydrolyzed (Weigle et al., 1973). This current study proposes to quantitate the separated products of the collagenase reaction by using MDPFlabeled collagen as the substrate. EXPERIMENT.AL

Materials
The MDPF and human skin collagenase used in these experiments were gifts from Hoffman-LaRoche Laboratories, Nutley, New Jersey. All chemicals used in the Neville gel system were of electrophoretic grade quality (Eastman). All other chemicals used were analytical reagent grade.

Selection of Protease Inhibitors
The purity of the human skin collagenase was tested by Type I collagen is biochemically separated from Type III collagen by differential salt precipitation. Lyophilized collagen was dissolved in 0.05M Tris containing l.OM NaCl.
The sodium chloride concentration was then adjusted to l.7M and kept at 4° for 24 hours to precipitate Type III collagen.
The mixture was then centrifuged for 2 hours at 35000 x g and the NaCl concentration of the resulting supernate raised to 2.5M to precipitate the Type I collagen. After a 24 hour settling period, the collageri solution was centrifuged at 35000 x g for 2 hours and the resulting pellet suspended in 1% acetic acid. After exhaustive dialysis against 1% acetic acid, the solution was centrifuged for 1 hour at 55,000 To denature the collagen further, the tubes then were heated for thirty minutes at 56°C. The standard curve for fluorescent collagen was constructed by using the above assay procedure in the absence of enzyme.

Electrophoresis and Quantitation of Gels
The collagenase reaction samples were then electrophoresed according to the method of Neville (1971). The concept of the Neville gel system is based on a sodium dodecyl sulfate-protein complex electrophoresed in a discontinuous buffer.
The gels utilized in this procedure were composed of a 6.9% acrylamide: bisacrylamide running gel overlayed with 0.1 ml of a 3.2% acrylamide:bisacrylamide concentrating gel.
1.5 X 10 - The reaction of collagen and its degradative enzyme, collagenase, produces specific cleavage products when these are isolated on polyacrylamide gels as shown in Figure 2.
The last peak on the right corresponds to TCA, the three-quarter length reaction product of the collagenase reaction.
The remaining one-quarter fragment of the collagen molecule, designated TCB' could not be detected on the acry lamide gels used in this investigation.
To compensate for this, the amount of TCA detected was multiplied by a correction factor, 1.33, in order to describe the amount of collagen cleaved.
The results obtained when human skin collagenase was incubated with MDPF-labeled Type I qol.!agen are reported in Tables I through IV The relationship between substrate disappearance and product appearance was calculated and found not to be stoichiometrical.
Examination of the data indicated that more collagen was cleaved than could be accounted for in the appearance of product.
To test the stability of the substrate, MDPFlabeled collagen was incubated with trypsin. This resulted in a significant hydrolysis of the fluorescent substrate (Tables III and VI) . This could indicate the presence of denatured collagen molecules which unlike native collagen would be susceptible to nonspecific protease attack. To circumvent this problem, 25 U of pepsin was added to aliquots of labeled collagen and the mi x ture d i al y zed overnight against l.ON acetic acid. After recovery of the pepsin resistant collagen, the percent hydrolysis b y trypsin of the resulting substrate was reduced (Tables II I and  for both substrate disappearance and product appearance.
The correlation between uncleaved and cleaved collagen is also linear for both substrates as observed in Figures 4 and 5, indicating that the problem of nonspecific hydrolysis of the fluorescent substrate was somewhat resolved.

DISCUSSION
In a recent review, Perez-Tomayo (1978) states that a complete assay for collagenase activity would include not only measurement of a change in a physiochemical or biological property of collagen but also identification of the specific products of enzymatic cleavage. Traditionally, this would require two separate techniques. The physiochemical aspect could be measured as changes in viscosity, optical rotation, melting temperature or as the release of radioactively labeled collagen peptides from insoluble fibrils (Hu et al., 1978;Labrosse et al., 1978;Gisslow and McBride, 1975). Identification of the specific cleavage products could be achieved by electrophoresis on polyacrylamide gels or through formation and detection of SLS crystallites.
The quantitation and evaluation of the collagenase reaction by most of these procedures is time-consuming and tedious.
The technique developed in this study, to follow the collagenase reaction through the quantitation of cleavage products from a fluorescent labeled substrate, alleviates the need for a two-part assay system. Quantitation of the collagenase reaction can be made by direct scanning of polyacrylamide gels for fluorescence and subsequent area analysis of the resulting peaks. Visualization and identification of TCA and TCB on the gels can be made under ultraviolet light.
This system was found to compare favorably with the degree of sensitivity reported for detecting vertebrate collagenases using other assay procedures (Hu et al., 1978;Terato et al., 1976;Labrosse et al., 1976). Moreover, MDPFlabeled collagen has an additional advantage. Since an aliquot of the reaction mixture is used for electrophoresis, both cleaved and intact collagen can be measured. Thus, collagen recoveries can be monitored in each experiment.
Previously described systems can only detect the amount of collagen cleaved.
Quantitation of MDPF-labeled a 1 -chains is linear down to the 10 nanogram range with a precision of + 6% (Goldberg and Fuller, 1978). In the present assay system, the lowest  Type 1 collagen was incubated at 27 C for two hours with the indicated amounts of human skin collagenase. After electrophoresis on acrylamide gels, the collagen was quantitated as indicated in the methods.
For uncleaved collagen, r = -0.848, slope= -6.39 For cleaved collagen, r = 0.935, slope = 5.61. Type 1 collagen, pepsinized as described in the methods, was incubated for 2 hours at 27°C with the indicated amounts of enzyme and quantitated as described in the methods.
For uncleaved collagen, r = -0.934, slope= -6.32 For cleaved collagen, r = +0.98, slope= +6.89 The indicated amounts of human skin collagenase were added to Type 1 collagen and incubated for 2 hours at 27°C. Aliquots of the reaction products were run on 7% acryamide gels and quantitated.