Baixe o app para aproveitar ainda mais
Prévia do material em texto
ANALYTICAL BIOCHEMISTRY l&,490-494 (1984) A Fluorescent Screening Assay for Collagenase Using Collagen Labeled with 2-Methoxy-2,4diphenyl-3(2H)-furanone’ ROBERTL. O'GRADY,ANDREWNETHERY,ANDNEILHUNTER Institute of Dental Research, United Dental Hospital, Chalmers Street, Surry Hills, New South Wales 2010. Australia Received December 27, 1983 This report describes the use of the compound 2-methoxy-2,4diphenyl-3(2H)-furanone to label collagen as a substrate for the detection of mammalian collagenase in a fluorescent assay which is suitable for screening huge numbers of samples. The compound 2-methoxy-2,4diphenyl- 3(2H)-furanone presents distinct advantages over other fluorophores, since both the unbound reagent and its hydrolysis products are nonfluorescent. The labeling procedure uses commercially available collagen, is fast and simple, and gives a 90% yield of labeled substrate. The fluorescent collagen substrate is stable and retains fluorescence over a wide range of pH. The assay detects, reproducibly, metal-dependent collagenase activity in microliter volumes of conditioned media from cultured neoplastic cells or in chromatographic fractions from such media. KEY WORDS: mammalian collagenase; collagen; fluorescence; MDPF. While a variety of assay systems have been developed for the detection of collagenases (1,2) few combine the simplicity, speed, and sensitivity desirable to screen large numbers of samples. In the past, we have used an as- say in which the substrate is polymeric col- lagen, labeled with fluorescein isothiocyanate (FITC)2 (3,4). The substrate in this form “is probably nearest to the state of collagen in the in situ fibre bundles of tissue” (5). However, the method is cumbersome and not well suited to the rapid examination of large numbers of samples, so we turned to the reconstituted col- lagen fibril assay of Baici et al. (6), in which the substrate is FITC-labeled calf skin collagen. While this method was useful for detecting enzyme activity in microliter volumes of tis- sue-culture samples, the preparation of labeled collagen was tedious, particularly due to the need to remove unbound FITC, which is itself ’ This work was supported, in part, by the N.S.W. State Cancer Council. * Abbreviations used: FITC, tluorescein isothiocyanate; MDPF, 2-methoxy-2,4diphenyL3(2H)-furanone; tlu- orescamine, 4-phenylspiro[furan-2(3H),I’-phthalanl-3,3’- dione. fluorescent, by gel filtration. Further, the yield of labeled collagen (20%) was poor. In the present report, we describe the use of 2-methoxy-2,4-diphenyl-3(2H)-fura- none (MDPF) (7) to label calf skin collagen as a fluorescent substrate for mammalian col- lagenase. MATERIALS AND METHODS Materials. The collagen used for labeling was from calf skin (Worthington Biochemical Corporation, Freehold, N. J.), supplied as a solution (7.5 mg/ml) in 75 mM sodium citrate, pH 3.7. The labeling compound, MDPF, was a gift from Dr. W. E. Scott and Dr. Peter F. Sorter of Hoffmann-LaRoche, Inc., Nutley, New Jersey. Bacterial collagenase, A grade (EC 3.4.4.19), was purchased from Calbiochem- Behring (Australia) Pty. Ltd., Sydney, Aus- tralia, and trypsin (EC 3.4.21.4) from Sigma Chemical Company, St. Louis, Missouri. Bio- Rad protein assay dye concentrate was ob- tained from Bio-Rad (Australia), Sydney, Australia. All other chemicals used were of analytical reagent grade. 0003-2697184 $3.00 Copyright 8 1984 by Academic Press, Inc. All rights of nproduction in any fomt reserved. 490 FLUOROMETRIC COLLAGENASE ASSAY 491 Labeling of collagen. Collagen solution (5 ml) was dialyzed against 50 mM borate buffer, pH 9.0, containing 100 mM NaCl, at 4OC overnight. The resultant gel was broken up by magnetic stirring in the above buffer (final volume 30 ml) and cooled in an ice bath, with continuous stirring, as 2 mg MDPF was added, dropwise, in 3 ml sodium sulfate-dried ace- tone, over a period of 60 min. Stirring was continued for a further 60 min. The mixture was then centrifuged at 5OOOg for 10 min at 4°C. The pellet was resuspended in 0.2% (v/ v) acetic acid, adjusted to pH 4 with NaOH, and dialyzed against 5 liters of this solution overnight at 4°C. The solution of labeled sub- strate, after dialysis, was adjusted to a final volume of 20 ml. This stock solution was stored at - 10°C in aliquots of 5 ml. The su- pematant from the 5000g centrifugation, con- taining fluorescent molecules which remained soluble at pH 9, was used in experiments to determine the stability of fluorescence of bound MDPF to changes in pH. Aliquots of this solution were prepared, with pH values ranging from 1 to 12, and their fluorescence was measured. Characterization of MDPF-collagen. The concentration of collagen in the stock MDPF- collagen solution was determined by the method of Duhamel et al. (8), using acid-sol- uble rat tail tendon collagen (9) as standard. Collagen for assay was dissolved or diluted in 6 M urea, 20 mM Na2HP04, 5% (v/v) 2-mer- captoethanol, pH 7.2. The concentration of collagen-bound MD- PF was determined from the 385~nm absor- bance of the stock solution using 6 z 6500 M-’ Cm-’ (7). Fluorescence spectra were recorded using an Aminco-Bowman SPF-500 spectrofluo- rometer operating in single-beam mode. The slit width for varied wavelengths was 1 nm. CoIlagenase screening assay. Aliquots (50 ~1) of labeled collagen (approximately 85 pg collagen) were placed into plastic centrifuge tubes containing 1 ml of assay buffer, 50 mM Tris, 10 mM CaC&, 100 mM NaCl, pH 7.5. These were incubated for 6 h at 35°C to en- courage the reconstitution of collagen fib&. At the end of this period, the tubes were di- vided into two series. To one series, duplicate samples to be tested for collagenase activity were added with EDTA buffer (50 mM Tris, 100 mM NaCl, 200 mrvr EDTA, pH 7.5) to give a final EDTA concentration of 20 mM. The other series contained identical samples plus assay buffer instead of EDTA buffer. The final volume varied between 1.2 and 2.0 ml, but was constant throughout each assay. In- cubations generally were performed for 16 h at 35°C but in some experiments the tem- perature was reduced to 25 ‘C. The assay tubes were then centrifuged at 5000g for 10 min at lO”C, and the fluorescence in 1 .O-ml aliquots of supematants was measured immediately, using excitation and emission wavelengths of 385 and 480 nm, respectively. The divalent metal ion-dependent collagenase activity was expressed as the difference in fluorescence in the presence and absence of EDTA. For assay, bacterial collagenase and trypsin were dissolved in assay buffer. To confirm that samples showing activity in the screening assay contained a specific mammalian collagenase, they were incubated with unlabeled collagen at 25 “C and the prod- ucts of incubation were separated by electro- phoresis on sodium dodecyl sulfate-poly- acrylamide gels, as described previously ( 10). RESULTS In each labeling procedure, 37.5 mg calf skin collagen was used. The concentration of MDPF-collagen in the final stock solution was 1.7 mg/ml, so that 34 mg was recovered, in- dicating a yield of 90%, sufficient for about 400 samples. The MDPF-collagen was stable for at least 3 months when stored at -10°C. From measurements of MDPF absorbance in the final preparation, it was estimated that an average of six MDPF molecules were bound per tropocollagen molecule (assumed M, 300,000). Figure 1 shows the fluorescence spectra of MDPF-collagen in 0.2% acetic acid, pH 4.0. 492 G’GRADY, NETHERY, AND HUNTER ,A, : \ ti ‘: 5 , : :: , & , \ 9 : 2 ? ‘\ ‘\ ‘. ) ‘.._ -.- t 260 350 450 550 650 WAVELENGTH (nm)FIG. 1. Excitation (solid line) and emission (broken line) spectra for an 85-fig ml-’ solution of MDPF-coUagen in 0.2% acetic acid, pH 4.0. Both the overall spectral pattern and the ex- citation and emission maxima (385 and 480 nm, respectively) were identical, at pH 7.5, for hydrolyzed MDPF-collagen in the super- natant after incubation with 4 pg of bacterial collagenase under assay conditions. The max- ima are similar to those observed for other MDPF-protein conjugates (7,ll). The molecules which remained in the su- pematant at the end of the labeling procedure exhibited fluorescence over a wide range of pH. The fluorescent response was relatively constant between pH 1.0 and pH 10.0 but at pH 12.0 it dropped by about 50%. The assay was used to detect collagenase activity produced by neoplastic cell lines. En- zyme activity was detected readily in 10 ~1 of culture medium from a mammary carcinoma cell line which had been shown previously to secrete a specific mammalian collagenase (10). The assay has been used also to detect enzyme activity in chromatographic fractions during purification of the tumor collagenase. Figure 2 illustrates the use of the assay to screen for collagenase activity in dilute protein-contain- ing fractions from gel-filtration chromatog- raphy. While the assay was intended primarily as a fast, qualitative screening procedure for mammalian collagenase, it was very sensitive to bacterial collagenase. It appeared that about 10 gg of bacterial collagenase hydrolyzed the -0.04 -0.03 E c z N -0.02 : 5 z 0 ii -0.01 < I \ \ I : ’ --\ : 20 40 60 60 100 120 FRACTION NUMBER FIG. 2. Elution profile from gel filtration of a partially purified tumor collagenase preparation on Ultrogel AcA 34 (85 X 1.6 cm) eluted at 6 ml h-’ with 10 mM Tris, 5 mM CaClr, 100 mM NaCl, 0.02% (w/v) sodium azide, pH 7.5, at 5°C. (0) Absorbance at 280 nm; (0) fluorescence released in collagenase screening assay f variance. Fractions were 1.5 ml each, and 0.2-ml samples of selected fractions were assayed as described. The collagenase sample had been prepared for gel filtration by ammonium sulfate precipitation and ion-exchange chromatography. The peaks of 280-nm absorbance at fractions 46, 65, and 100, which are closest to the peaks of collagenam activity, correspond to M, for globular proteins of approximately 300,000, 160,000, and 40,000, respectively. FLUOROMETRIC COLLAGENASE ASSAY 493 85 pg of substrate, as increases in the amount of enzyme (up to 35 pg) did not increase, sig- nificantly, the release of fluorescence due to divalent metal ion-dependent enzyme activity. Approximately 10% of this maximum fluo- rescence was released by 0.3 pg of bacterial collagenase under the same conditions. The background fluorescence obtained by incubating substrate with buffer alone varied between different batches of substrate but was 5-10% of the total fluorescence that could be released by bacterial collagenase. This indi- cated that at least 90% of the substrate was in the gel form, with 5-10% remaining in so- lution. Trypsin (20 rg/ml), when incubated with substrate at 35°C released further flu- orescence, giving values that were consistently double the controls. These figures are similar to those reported for radioactively labeled rat tail collagen (12). However, the release of flu- orescence by trypsin was unaffected by EDTA, and even up to 100 pg of trypsin released negligible fluorescence due to divalent metal ion-dependent enzyme activity. When the in- cubation temperature was lowered to 25’C, trypsin released no fluorescence above the controls. Both mammalian and bacterial col- lagenases were active at 25 “C, but more slowly than at 35°C. The tumor collagenase, which we are using in ongoing studies, has proved positive in the MDPF assay at both temperatures, and it pro- duces the characteristic 3/4 and l/4 fragments of collagen, as visualized in polyacrylamide gels. This last step is necessary to confirm the pres- ence of a specific mammalian collagenase, which cannot be claimed purely on the results of the screening assay. DISCUSSION A major aim of our experiments is to purify specific mammalian collagenases known to be produced in vim by a number of malignant tumors (10,13). Collagenase activity from some of these tumors was demonstrated pre- viously by digestion of polymeric collagen, while 3/4/1/4 fragments of collagen were dem- onstrated by electrophoresis on polyacryl- amide gels (10,13). Because these methods were unsuitable for screening large numbers of samples, a fast, reproducible assay using fluorescent, rather than radiolabeled, collagen was developed. Both FITC (4,6) and 4-phenylspiro[furan- 2(3H), I’-phthalanl-3,3’dione (fluorescamine) ( 14) have been used in collagenase assays in the past, but MDPF was thought to be more suitable than either of these other two com- pounds for our particular requirements. A distinct disadvantage in the use of FITC- labeled substrates for enzyme assays is the need for rigorous purification to remove all un- reacted fluorescent reagent. Failure to achieve this results in very high background readings. Preparation of the polymeric collagen sub- strate (3,4) involves washing for at least 2 weeks to remove unreacted FITC, while the method of Baici et al. (6) requires a chro- matographic step to remove unreacted FITC and, in our hands, took a week to prepare. The preparation of MDPF-collagen is com- pleted in, at most, 40 h. In 1972, Weigele et al. (15) reported that 2-oxysubstituted 3(2H)-furanones produce highly fluorescent compounds upon reaction with primary amines. One of these, fluores- camine, has been used to label collagen for the assay of collagenase activity (14). A major disadvantage of this fluorophore is the loss of fluorescent activity of the bound form after a few hours, so that these workers labeled each aliquot of the mixture, individually, at the end of the incubation period. This procedure is time consuming, so that it is not ideally suited to assays of large numbers of samples and there may be a risk of nonuniformity of labeling. Fluorescamine is also unstable at low and high pH. Another of these substituted furanones, MDPF, reacts rapidly with primary amines to form highly fluorescent, stable fluorophores (7). MDPF is nonfluorescent, and excess re- agent is hydrolyzed to the hydroxyfuranone, which is also nonfluorescent. The rate of the fluorogenic reaction is strongly pH dependent 494 G’GRADY, NETHERY, AND HUNTER but, unlike FITC or fluorescamine, the re- sultant compounds retain fluorescence over a wide range of pH (7). This proved to be ap plicable also to MDPF-collagen, so that it would be a suitable fluorophore for use in assays which use soluble collagen as substrate. These assays are used for kinetic studies and unreacted substrate is removed by trichloro- acetic acid or dioxane at the end of the in- cubation. If samples to be tested in the MDPF assay contain significant nonspecific protease activ- ity, the incubation should be performed at 2YC. However, the nature of our experiments is such that a shorter incubation at 35°C is preferred, provided that the effect of trypsin is minimal and that positive samples are able to produce the characteristic 3/4 and ‘14 frag- ments of collagen on polyacrylamide gels. The concentration of collagen in the screening as- say is such that complete gelation may not occur with some preparations, leaving a large proportion in solution. This potential problem was not apparent in our experiments, as only 5-10% of the substrate remained in solution. If such a problem did arise, it could beover- come by decreasing the total volume of the assay mixture or by adding more substrate. In addition to the electrophoretic experi- ments mentioned above, some samples of cul- ture media and chromatographic fractions which showed activity in the MDPF assay were tested in several other collagenase assay sys- tems: the polymeric collagen assay (4) and the viscometric assay (l), followed by the dem- onstration of 3/4/9i fragments of collagen by electrophoresis on sodium dodecyl sulfate- polyacrylamide gels. In all cases examined (n = 6), activity in the MDPF assay correlated with activity in these established assay pro- cedures. Samples showing no activity in the MDPF assay were not active in the other sys- tems, so that divalent metal ion-dependent hydrolysis of MDPF-calf skin collagen fibrils appears to be a reproducible indicator of mammalian collagenase activity. MDPF has been used previously (7,11, 16,17) and we report that, when bound to collagen, it is suitable for assaying collagenase and that it presents distinct advantages over both FITC and fluorescamine. Also, it is likely to be a suitable compound for labeling the other types of collagen which are used as sub- strates for collagenases and may well find ap plication in other areas of biochemistry. I. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. REFERENCES Harris, E. D., and Vater, C. A. (1982) in Methods in Enzymology (Cunningham, L. W., and Freder- iksen, D. W., eds.), Vol. 82, pp. 428-439, Academic Press, New York. Harris, E. D., and Vater, C. A. ( 1980) in Collagenase in Normal and Pathological Connective Tissues (Woolley, D. E., and Evanson, J. M., eds.), pp. 37-48, Wiley, Chichester. Steven, F. S. (1975) Connect. Tissue Res. 4, l-6. Steven, F. S., and Lowther, D. A. (1975) Connect. Tissue Rex 4, 7-10. Weiss, J. B. (1976) in International Review of Con- nective Tissue Research (Hall, D. A., and Jackson, D. S., eds.), Vol. 7, pp. 101-157, Academic Press, New York. Baici, A., Cohen, G., Fehr, K., and Boni, A. (1980) Anal. Biochem. 108,230-232. Weigele, M., De Bemardo, S., Leimgruber, W., Clee- land, R., and Grunberg, E. (1973) B&hem. Bio- phys. Res. Commun. 54, 899-906. Duhamel, R. C., Meexan, E., and Brendel, K. (198 1) J. Biochem. Biophys. Meth. 5, 67-74. Baxin, S., and Delaunay, A. (1976) in The Meth- odology of Connective Tissue Research (Hall, D. A., ed.), pp. 14-15, Joynson-Bruvvers, Oxford. G’Grady, R. L., Upfold, L. I., and Stephens, R. W. (198 1) Int. J. Cancer 28, 509-5 15. Weisner, R., and Troll, W. (1982) Anal. Biochem. 121,290-294. Birkedal-Hansen, H., and Dano, K. (1981) Anal. Biochem. 115, 18-26. O’Grady, R. L., Harrop, P. J., and Cameron, D. A. (1982) Pathology 14, 135-l 38. Macartney, H. W., and Tschesche, H. (1980) FEBS Lett. 119, 327-332. Weigele, M., De Bemardo, S. L., Tongi, J. P., and Leimgruber, W. (1972) J. Amer. Chem. Sot. 94, 5927-5928. Barger, B. O., White, F. C., Pace, J. L., Kemper, D. L., and Ragland, W. L. ( 1976) Anal. B&hem. 70, 327-335. Chen-Kiang, S., Stein, S., and Udenfriend, S. (1979) Anal. B&hem. 95, 122-126.
Compartilhar