3 A fluorescent screening assay for collagenase using collagen labeled with 2 methoxy 2,4 diphenyl 3(2H) furanone

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.