Fluorometric determination of bacterial protease activity using fluorescein isothiocyanate-labeled proteins as substrates

ANALYTICAL

BIOCHEMISTRY

133-137 (1990)

191,

Fluorometric Determination of Bacterial Protease Activity Using Fluorescein Isothiocyanate-Labeled Proteins as Substrates’ Karen

A. Homer

Hunterian

Dental

Received

February

and David

Research

Beighton

Unit, London

Hospital

Medical

College, Turner

Inc.

The sensitivity of gel electrophoresis as a technique for detecting protease activity has improved dramatically in recent years (l-3) and a recent study using a dot-blotting method to detect proteolytic enzymes (4) reports very low detection limits. Despite the ability to detect nanogram and subnanogram levels of proteases, such methods are either time consuming or utilize relatively large quantities of substrates. Synthetic substrates containing a chromophore, the hydrolysis of which may be followed spectrophotometrically, have also been improved in terms of their sensitivity (56) but usually require relatively large amounts of either substrate or enzyme preparation.

1 Supported 0003.2697/90 Copyright All rights

El 2AD, United

Kingdom

27, 1990

Intact fluorescein isothiocyanate-labeled proteins have relatively low background fluorescence at excitation and emission wavelengths of 495 and 525 nm, respectively. Degradation of these substrates leads to exposure of covalently linked fluorescein isothiocyanate molecules and to a concomitant increase in relative fluorescence at these wavelengths. The increase in relative fluorescence is proportional to the degree of protein degradation. This phenomenon provides the basis for a sensitive assay for bacterial protease activity. There is no requirement for the removal of undegraded substrate from the assay mixture prior to the measurement of fluorescence. Assays can be performed in 96-well microtiter trays, enabling a large number of samples and their respective controls to be processed simultaneously and repeated determinations of fluorescence values may be made on the same assay. o 1990 Academic Press,

Street, London

by MRC

Grant

GB8704557.

$3.00 0 1990 by Academic Press, of reproduction in any form

The development of synthetic protease substrates containing /I-naphthylamide (7) or 7-amino-4-methylcoumarin (8-10) as a leaving group have led to the development of protease assays in which activity may be detected fluorometrically, giving rise to increased sensitivity and the use of smaller quantities of materials. In the case of these substrates, although they may be tailored for the detection of a specific enzyme, they do not necessarily resemble native proteins encountered physiologically. Several methods of labeling native proteins with fluorescent leaving groups have been described, including those using fluorescein (l&12) or 2-methoxy-2,4-diphenyl-3(2H)-furanone (13). The sensitivity of such assays is excellent, being in the nanogram or subnanogram detection range, and relatively small amounts of materials are required. However, such techniques for measuring proteolysis necessitate the separation of the undegraded substrate from the fluorescent degradation products, which is time consuming and may introduce errors. The protease assay described here utilizes fluorescein isothiocyanate-labeled proteins as substrates and removes the necessity for the separation of substrate from degradation products. We have used Porphyromonas gingivalis in this study as it has been implicated as one of the causative organisms of periodontal disease in humans. Its proteolytic activity has been widely studied and it is proposed that this contributes to its virulence (14). MATERIALS

AND

METHODS

Assay of protease activity using fluorescein isothiocyanate-labeled proteins. Fluorescein isothiocyanatelabeled (FITC-labeled)2 bovine serum albumin (BSA), ’ Abbreviations used: FITC, fluorescein vine serum albumin; IgG, immunoglobulin

isothiocyanate; G; Tes

BSA, (buffer

boA), 133

Inc. reserved.

134

HOMER

AND

casein and immunoglobulin G (IgG) were purchased from Sigma Chemical Company (Poole, Dorset, UK). FITC-labeled transferrin was prepared by using a modification of the method described by Rinderknecht (15). Briefly, 5 mg of transferrin (Sigma) was dissolved in 1.5 ml of 50 mM sodium carbonate-bicarbonate buffer, pH 9.5, followed by the addition of 22 mg of FITC-cellite (Sigma). The solution was allowed to stand at room temperature for 3 min, after which period the unreacted FITC-cellite was removed by centrifugation. The supernatant was applied to a 0.5 X 25 cm column of G25 Sephadex (Sigma) which had been previously equilibrated with 50 mM N-[tris(hydroxymethyl)methyl]aminoethanesulfonic acid buffer (Sigma), pH 7.5 (buffer A). The FITC-labeled transferrin was separated from the free FITC by elution with buffer A and collected in a small volume (usually 1.0-3.0 ml). FITC-labeled proteins were diluted to a concentration of up to 200 pg/ml with buffer A and stored at -20°C in aliquots of approximately 5.0 ml. Substrates stored in this way were stable for periods of greater than 6 months. The reaction mixture used to determine protease activity contained 50 ~1 of bacterial suspension, 50 ~1 of FITC-labeled protein solution, and 100 ~1 of buffer A. Assays were prepared in duplicate in a clear 96-well microtiter tray with control assays containing no bacterial suspension to take into account the background fluorescence of the FITC-labeled protein molecules. Trays were incubated at 37°C for the required period, after which time fluorescence values were determined on a fixed scale of 10 at an excitation wavelength of 495 nm and an emission wavelength of 525 nm by using a Perkin-Elmer LS-3B fluorescence spectrometer fitted with a microtiter plate-reading attachment. Plotted data represent the mean of duplicate values. To determine the sensitivity and linearity of the protease assay, assays were set up as described above but either trypsin, chymotrypsin, elastase, or proteinase K (all purchased from Sigma) were incorporated at final concentrations of 5 rig/ml to 10 pg/ml (1 ng to 2 pg of enzyme in the assay). These assays were set up in both clear and opaque microtiter trays and were incubated at 37°C for up to 24 h. The detection limit of each assay was determined by calculating the mean fluorescence plus two standard deviations of control assays which incorporated no enzyme. In order to study the effect of a number of compounds on bacterial protease activity, dithiothreitol (DTT), ethylenediaminetetraacetic acid (EDTA), phenylmethylsulfonyl fluoride (PMSF), p-tosyl-L-lysine chloromethyl ketone (TLCK), and soybean trypsin inhibitor

N-[tris(hydroxymethyl)methyl]aminoethanesulfonic dithiothreitoh PMSF, phenylmethylsulfonyl syl-L-lysine chloromethyl ketone.

fluoride;

acid; TLCK,

DTT, p-to-

BEIGHTON

(Sigma) were prepared to a final concentration of 0.4 or 4.0 mM in Tes buffer. Aliquots (50 ~1) of each of these solutions were added to the protease assay in place of 50 ~1 of buffer A to give a final concentration of 0.1 or 1.0 mM. Appropriate controls were included to take into account any background fluorescence of these molecules and any quenching of the FITC fluorescence that might occur. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of native and degraded FITC-BSA. SDS-polyacrylamide gel electrophoresis was carried out by using a method based on that previously described (16). Briefly, to visualize the degradation of FITC-BSA, assay mixtures were extracted with sample buffer by placing them in a boiling water bath for 5 min, removing debris by centrifugation, and running 50 ~1 of extract on 9% polyacrylamide gels. Gels were viewed prior to staining on a uv transilluminator and green fluorescent bands of intact and degrade FITC-BSA were noted. Growth of P. gingivalis strain W83. P. gingivalis strain WS3 was grown at 37°C for 3 days on Fastidious Anaerobe Agar (Lab M, Bury, Lanes, UK) in an anaerobic chamber (Don Whitley, Shipley, W. York, UK). Cell suspensions were prepared by aseptically removing colonies of P. gingivalis strain W83 from agar plates into buffer A. Following protein estimation of small aliquots of the bacterial suspension using the method described below, buffer A was added to the original suspension to give a bacterial protein concentration of up to 200 pg/ml. Estimation of protein content of bacterial suspensions. Aliquots (1.0 ml) of bacterial suspension were centrifuged in order to pellet cells and the supernatant was discarded. Samples were hydrolyzed by boiling in 0.5 M sodium hydroxide, as described by Herbert et al. (17). Following removal of cell debris by centrifugation and adjustment of the pH of the solution to neutrality, protein concentrations were estimated by using the Coomassie blue dye-binding assay (Pierce, Rockford, IL) with bovine serum albumin (Sigma) as the standard.

RESULTS Effect of bacterial protein concentration on FITC-BSA degradation. Degradation of FITC-labeled BSA, as measured by relative fluorescence, was found to be dependent on the quantity of bacterial protein added to the assay. Cleavage of FITC-BSA by P. gingivalis strain W83 was linear with bacterial protein in the assay up to 20 pg/ml after 1 h of incubation at 37°C (Fig. 1). At protein concentrations above this value the relationship ceased to be linear. The nonTime course of FITC-BSA degradation. disruptive nature of the assay enables readings of relative fluorescence to be recorded on the same assay,

FLUOROMETRIC

DETERMINATION

OF

BACTERIAL

PROTEASE

135

ACTIVITY

1

oy I 0

10 Bacterial

1 1 20 30 protein concentration

FIG. 1. Hydrolysis of FITC-BSA a final concentration of 25 pg/ml incubated in clear microtiter trays

I SO (ug;ml)

3

50

by P. gingiualis. Assays contained FITC-BSA in buffer A and were for 30 min at 37°C.

without further treatment, over a time course. Figure 2 shows that degradation of FITC-BSA is dependent on both the concentration of bacterial protein added to the assay and the period of incubation at 37°C. In order to ensure that the increase in relative fluorescence using this assay system is related to the degree of degradation when suspensions of P. gingivalis are incubated with FITC-BSA, assays containing 50 pg/ml of bacterial pro-

140/ 20 -

4

-

0

20

40

D

A

I

I

I

60

m

100

Time (min) FIG. The ml (A), and

2. Time-dependent hydrolysis of FITC-BSA by P. gingivalis. assays contained FITC-BSA at a final concentration of 12.5 pgl in buffer A. P. gingivalis was added at final concentrations of 2.5 5.0 (B), 10.0 (C), 30.0 (D), and 50.0 (E) pg/ml bacterial protein the assays were incubated in clear microtiter trays at 37°C.

20

40 60 Time (min)

50

100

FIG. 3. Time-dependent hydrolysis of FITC-casein, FITC-IgG, and FITC-transferrin by P. gingiualis. The assays contained FITCIgG (A), FITC-casein (B), or FITC-transferrin (C) at a final concentration of 25 pglml and P. gingivalis at 50 pg/ml in buffer A. Assays were incubated in clear microtiter trays at 37°C.

tein were incubated for a period of up to 90 min and subjected to SDS-polyacrylamide gel electrophoresis. As the incubation period increased, so the molecular weight of the degradation products decreased (data not shown), indicating hydrolysis of the FITC-BSA molecule. After 90 min no undegraded FITC-BSA was visible and the fluorescent degradation products were observed to run with the dye front. These results indicate that the increase in relative fluorescence is paralleled by an increase in the breakdown of the FITC-BSA molecule. Degradation of other FITC-labeled proteins. With FITC-labeled casein, IgG, or transferrin as substrates, the increase in relative fluorescence when incubated with P. gingivalis is related to bacterial protein concentration and incubation period, as for FITC-BSA. There were, however, marked differences in the rate of increase in relative fluorescence for the four substrates (Fig. 3). The increase in fluorescence was found to be the slowest with FITC-IgG as a substrate over an incubation period of 90 min. The rate of increase in relative fluorescence was greater for FITC-casein than for FITC-IgG, but still not as high as that observed for FITC-transferrin. Results indicate that of the four substrates, FITC-BSA was the substrate which gave the greatest increase in relative fluorescence when supplied as a protease substrate for P. gingivalis. Sensitivity and linearity of assays. The relationship between enzyme concentration and FITC-BSA degradation (measured as relative fluorescence) was essentially linear over a wide range of enzyme concentrations following l-h incubation at 37°C dependent upon the par-

136

HOMER

AND

BEIGHTON TABLE

4a

1

Detection Limits of Commercially

Available Enzymes“

A Incubation period Substrate /

Protease

FITC-BSA

Proteinase Trypsin Chymotrypsin Elastase Proteinase Trypsin Chymotrypsin Elastase Proteinase Trypsin Chymotrypsin Elastase

: D

,/

FITC-transferrin

; 0

FITC-casein 4

2 Enzyme

6 concentration

= ‘Gujlmii

12

10

800 T=

c; c; ” : z : ;j 3 c

600

I

0 C

/

-

Q

I

K

K

24 h

50 50 100 200 500 500 500 500 10 50 100 100

tl tl
Note. Assays were incubated at 37°C in clear microtiter trays and change in fluorescence is reported after l- and 24-h incubation. ‘As nanograms of enzyme detected when included in assay with total volume of 200 ~1.

400-

markedly inhibited protease activity, while PMSF and trypsin soybean inhibitor had little effect. When EDTA was added to the assay at a concentration of 0.1 or 1.0 mM, slight inhibition of protease activity was observed.

200-

DISCUSSION

.a, II

K

lh

0

1

I

2

4 Enzyme

I 6 concentration

I 8 (ug/ml)

FIG. 4. Relationship between concentration of able enzymes (A, proteinase K; B, chymotrypsin; trypsin) and degradation of FITC-BSA at a final fig/ml after l-h incubation at 37’C; (a) assays trays; (b) assays performed in opaque trays.

I

I

10

12

commercially availC, elastase; and D, concentration of 25 performed in clear

titular enzyme (Fig. 4a). Similar linear relationships were found for the degradation of FITC-casein and FITC-transferrin while none of the four enzymes degraded the FITC-IgG significantly. The detection limits for the degradation of FITCBSA, FITC-transferrin, and FITC-casein by each enzyme after 1 and 24 h in clear microtiter trays are shown in Table 1. The sensitivity of the assays can be increased by approximately 20-fold by performing assays in opaque microtiter trays (Fig. 4b). Influence of effecters on FITC-BSA degradation by P. gingiualis. The effects of potential inhibitors and stimulators of P. gingiualis protease activity are shown in Table 2. Dithiothreitol added at a concentration of 0.1 or 1.0 mM markedly stimulated protease activity. TLCK

We have described an assay for the detection of protease activity using FITC-labeled proteins as substrates. Proteolytic activity may be detected by using concentrations of bacterial protein as low as 2.5 pg/ml (corresponding to a total of 0.5 pg of bacterial protein in a 200-~1 assay), or of less than 1 ng of a commercially available protease, proteinase K, depending on the choice of substrate. FITC-BSA resulted in the greatest increase in relative fluorescence and this would be the

TABLE Stimulatory

and Inhibitory Hydrolysis

2 Compounds

by

of FITC-BSA

P. gingiualis Concentration

DTT EDTA PMSF TLCK Soybean

(mM)

Compound

0.1

1.0

trypsin

189.2 87.1 87.5 10.5 101.7

243.8 71.2 83.9 5.3 98.3

inhibitor

Note. Compounds were added to the standard protease assay at the concentrations shown. Results are recorded as the percentage of the reaction obtained in the absence of any of the compounds.

FLUOROMETRIC

DETERMINATION

substrate of choice for routine screening of protease activity. The sensitivity of the assay is such that that proteolytic activity has recently been detected, with this method, in oral streptococci (18), bacteria with levels of proteolytic activity considerably lower than P. gingivalis. The sensitivity of the assay may be increased in a number of ways, including increasing fluorometer sensitivity, using opaque microtiter trays to increase fluorescence detection in the assay wells, and increasing the assay incubation period. The assay is nondisruptive, removing the requirement for the time-consuming physical separation of macromolecules from breakdown products which is often an integral part of other protease assays (12,13). This enables the increase in relative fluorescence to be followed over a time course and hence the rate of degradation of proteins to be calculated. An additional advantage of the assay is the rapid and easy labeling of substrates (5 mg of protein labeled with FITC in under 0.5 h) or their commercial availability (12,15). This system could, in theory, be utilized to label any number of native proteins with FITC for use as substrates (12). With the choice of an appropriate substrate (e.g., FITCBSA for P. gingivalis) proteolysis may be detected in under 0.5 h for a highly proteolytic organism, which is far more rapid than the electrophoretic detection systems commonly employed (l-3). The ease with which protease activity can be detected by using the assay method described here provides a rapid method for screening compounds which may act as protease inhibitors. Data presented here indicated that the protease activity of P. gingivalis was stimulated by the presence of DTT, suggesting that the activity is thiol-dependent. Activity was markedly inhibited by TLCK, a protease inhibitor which acts by binding to histidine residues at the active site, while PMSF and trypsin soybean inhibitor, both serine protease inhibitors, had little effect. Recently (19), a cysteine proteinase has been isolated from P. gingivalis with properties which could account, at least in part, for the observed inhibition of protease activity in our study. Protease inhibitors have attracted increasing interest in recent years and are important because of their action against enzymes which are potential virulence factors for periodontopathogens. Compounds which exhibit protease inhibition in vitro may then be studied for the ability to reduce the efficacy of such pathogens, and hence the level of disease, in vivo. The main disadvantage of protease assays which employ a native substrate labeled with a fluorescent leaving group is the necessity for controls which take into account quenching of fluorescence by compounds added to the assay. This is often tedious and has been considered to render the assay unsuitable for the routine measurement of large numbers

OF

BACTERIAL

PROTEASE

137

ACTIVITY

of samples (20). However, since this assay may be set up in small volumes in a 96-well microtiter tray, assays and their respective controls may easily be included in the same tray. A method for the detection of protease activity has been described which employs FITC-labeled native proteins as substrates. Any number of proteins may be utilized as substrates and there is a potential use for the assay in detecting protease activity in a number of bacteria or even purified enzymes. The assay is simple and rapid and has a low blank value and measurements are nondisruptive. The small assay volumes used means that the amounts of substrate utilized are low and the assay is therefore economical. ACKNOWLEDGMENT We thank Dr. H. Shah of the Oral Microbiology London Hospital Medical College for supplying P. gingivalis strain W83.

Department of the the stock culture of

REFERENCES 1. Brown,

T. L., Yet,

M.-G.,

and

Wold,

F. (1982)

Anal.

Biochem.

122,164-172. 2. Kelleher, P. J., and Juliano, R. L. (1984) Anal. Biochem. 136, 470-475. 3. Ohlsson, B. G., Westrom, B. R., and Karlsson, B. W. (1986) Anal. Biochem. 162,239-244. 4. Schlage, W. K. (1988) Hoppe-Seyler’s 2. Physiol. Chem. 369,357363. 5. Harper, J. W., Ramirez, G., and Powers, J. C. (1981) Anal. Biothem. 118,382-387. 6. Kuwada, M., and Katayama, K. (1984) Anal. Biochem. 139,438443. 7. De Lumen, B. O., and Tappel, A. L. (1972) Anal. Biochem. 48, 378-385. 8. Zimmerman, M., Yurewicz, E., and Patel, G. Anal. Biochem. 70, 258-262. 9. Zimmerman, M., Ashe, B., Yurewicz, E. C., and Patel, G. (1977) Anal. Biochem. 78,47-51. 10. Kuromizu, K., Shimokawa, Y., Abe, O., and Izumiya, N. (1985) Anal. Biochem. 151,534-539. 11. De Lumen, B. O., and Tappel, A. L. (1970) Anal. Biochem. 36, 22-29. 12. Twining,

S. S. (1984)

13. Wiesner, R., and 14. Van Winkelhoff, J. (1988) J. Clin. 15. Rinderknecht, H.

16. Laemmli, 17. Herbert,

Anal.

Biochem.

143,30-34.

Troll, W. (1982) Anal. Biochem. 121, 290-294. A. J., Van Steenbergen, T. J. M., and de Graaf, Periodontol. 15, 145-155. (1962) Nature (London) 193, 167-168.

U. K. (1970)

Nature

(London)

227,680-685.

D., Phipps, P. J., and Strange, R. E. (1971) in Methods in Microbiology (Navis, R. J., and Ribbons, D. W., Eds.), Vol. 5B pp. 210-344, Academic Press, London.

18. Homer, biol. L&t.

K. A., Whiley, 67, 257-260.

R., and Beighton,

D. (1990)

FEMS

Micro-

19. Shah, H. N., Gharbia, S. E., Kowlessur, D., Wilkie, E., and Brocklehurst, K. (1990) Biochem. Sot. Trans. 18, 578-579. 20. Evans, C. H., and Ridella, J. D. (1984) Anal. Biochem. 142,411~

420.