Assay of Proteolytic Enzymes by the Fluorescence Polarization Technique’ HIROSHI
MAEDA
Assay of proteolytic enzymes by the fluorescence polarization method is described in which fluorescein isothiocyanate-conjugated proteins are used as substrates. Hydrolysis of the natural substrates is accompanied by a decrease in the fluorescence polarization values reflecting time and also by an enzyme concentration-dependent decrease in the overall molecular weight of the substrates. Trypsin activity as low as 0.05 &ml at 20°C is detected at a fluorochrome concentration of IO to 20 pmol/ml. In this method. no separation procedure or further reaction for quantitation is necessary. Thus, using a computerized instrument, this method is extremely sensitive, simple. and rapid.
The fluorescence polarization method with its extremely high sensitivity appears to have versatile potential applications (2). The parameter utilized in this method is the fluorescence polarization value (P-value). which reflects the change in the molecular weight of a fluorochrome-labeled substance due to the altered Brownian motion. or the change in the viscosity of the solvent being measured. Such a change in the molecular weight or viscosity results in a change in the P-value. When the P-values of fluorescein isothiocyanate ( FIT0labeled2 proteins of different molecular weights were measured, the reciprocals of these P-values were found to be linearly correlated with the reciprocal values of their respective molecular weights in the range of 5000 to 160,000 for most of the globular and simple proteins (3). Empirically. the log molecular
weights and P-values were also found to exhibit a near-linear relationship. Based on this fact, quantitation of antigen or antibody became possible by the measurement of the increase in the molecular weight (and Pvalues) of the complex being formed (3-7). In this assay system, either one of the molecules. antigen or antibody, is labeled with FITC. This method is capable of detecting the interaction with extreme rapidity (about 2 min) and sensitivity ( IO-” mol/ml) (3.7). This paper reports on the application of the fluorescence polarization method to the assay of proteolytic enzymes using various protein substrates labeled with FITC. When a protein is hydrolyzed to smaller fragments, the apparent and overall molecular weight in the system diminishes. resulting in a decreased P-value and thus substantiating the proteolytic activity.
’ A preliminary account of this work was presented at the 50th General Meeting of the Japanese Biochemical Society, October 1977, Tokyo (I). ’ Abbreviations used: FITC. fluorescein isothiocyanate: TPCK, tosylphenylaline chloromethyl ketone; PBS. 0.01 M phosphate-buffered 0. I5 M NaCl solution, pH 7.0; Tris, tris(hydroxymethyl)aminomethane; Pvalue, fluorescence polarization value in arbitrary units; IgG. immunoglobulin G: F/P. Auorochrome per mole of protein.
0003.2697/79/010212-06$01.00/O CopyrIght All
right\
t of
1979
by
reproductmn
Academic m any
Press. form
Inc. rewrved.
MATERIALS trzstrlrmctlts.
polarimeter.”
AND METHODS
The fluorescence spectroModel MAC-2 Type HR-I
,’ The instrument is designed primarily for the analysis of fluorescein chromophore which has maximum excitation and emission spectra at about 490 and SZO nm, respectively.
7-l) ---
FLUORESCENCE
POLARIZATION
(Japan Immunoindustries Co., Ltd.. Takasaki), used for the measurements was equipped with three innovative features: (i) computerization which integrates 50 measurements and calculations of P-values before printout every 82 s; (ii) a rotating polarizer. which made the detection of fluorescence intensity of both parallel and perpendicular components in a single photomultiplier tube possible as a cosine function against the time of rotation (30 rps): and (iii) three-cavity filters (Detric Optics, Marborough, Mass.) with near-monochromatic performance, with a little extinction and no depolarization in the exciting or the emitting beam. A refrigerated water circulator Model RTE-8 (Neslab Inst. Inc., Portmouth, N.H.) was employed to control the temperature under operation constant within -+O.Ol”C of error. t5r;~r~rl:~.str/rd c~ilc~t~~icrrl.s. Proteolytic enzymes used were as follows: papain (E. Merck AG., Darmstadt. 60 U/mg); TPCKtreated trypsin and pepsin (both from Worthington Biochemical Corp.. Freehold. N. J.. 183.1 and 2530 Uimg, respectively): and Pronase (Kaken Chemical Co., Ltd.. Tokyo. 45 Uimg). Antipain and leupeptin were gifts from Dr. H. Umezawa and Dr. Y. Matsushima. Tokyo. Substrates used were obtained from the following sources: human fibrinogen from Midori Juji Company (Osaka): antihuman sheep IgG from Meloy Laiboratories (Springfield. Va.); and bovine pancreatic ribonuclease A and soybean trypsin inhibitor from Sigma Chemical Company (St. Louis, MO.). FITC was obtained from Dojin Chemical Company (Kumamoto). Other chemicals were from commercial sources. Ltrhclitlg the .srrhstrutc~s \i.itll FITS. The substrates were labeled in 0.1 M Na carbonate-bicarbonate buffer. pH 9.0, with about 20 to IO0 moles of FITC per amino group in proteins with stirring at room temperature for 11to 6 h. The labeled protein was separated from the reaction mixture with a column of Sephadex G-50 using 0.01 M
FOR
PROTEASE
ASSAY
223
phosphate-buffered 0.15 M NaCl solution (pH 7.0, PBS). Furthermore. dialysis against PBS at 4C for 2 days was carried out in order to ascertain the removal of free fluorescein chromophore. The labeled proteins had the following numbers of Buorochromes per mole of protein ( F/P) as determined spectroscopically at 490 and 280 nm. respectively: IgG, 5.5; soybean trypsin inhibitor, 1.O; fibrinogen. 1I .3: and ribonuclease, I .O. Reaction of FITC with proteins occurs primarily with the a-amino groups, subsequently with the e-amino groups, of the proteins depending upon their pX,, values and the pH of the reaction (8). The FITC-labeled proteins remained soluble after labeling. GI;JWlC wtrcti0u.v. Enzyme reactions were carried out in a cuvette placed in the instrument under the conditions described in the figure legends. The concentration of each substrate was 31 .Opg or more per milliliter with an absorbance of 0.005 to 0.015 at 490 nm, and the enzymes were used at various concentrations. P-Values were printed out directly on the recorder chart every 82 set during enzyme digestion. However. when the pH of the enzyme reaction is different from the pH of the measurements (pH 7 to 9). as in the case of pepsin digestion at pH 3, the pH of the reaction mixture or that of its aliquots used for the measurements was brought to an appropriate pH (about 8.6) in a buffer solution. This keeps both the quantum yield of fluorescence and the relaxation time of fluorescence-excitation constant, and the P-value thus obtained is reliable (1.7). RESULTS AND DISCUSSION P-Values obtained by the various enzyme digestions of FITC-labeled protein substrates were plotted, and the results are shown in Figs. I through 4. Figures I and 3 show that the polarization changes are small. In such cases, the accuracy of the assay suffers. Various enzyme concentra-
224
HIRCISHI
MAEDA
Pe# 100
Pm 100
I
4
I 6
I
I
12
16
I 20
1
900
1
IO
20
24
MIN FIG. I. Digestion of fibrinogen with trypsin. FITCLabe!ed human fibrinogen as the substrate, dissolved in 2 ml of 0.1 M Tris-HCI buffer containing 5 mM CaCI,, pH 7.8. in a cuvette, was placed in the instrument. The temperature of the cuvette was then allowed to equilibrate (20°C). after which varied amounts of trypsin were added and mixed. P-Values at each time period are plotted as percentage of f-value at zero time. The substrate concentration is constant for each while the enzyme assay (A 14,1 ,,,,,= 0.015 or 21 &ml). concentrations (microgram per milliliter) included 0.05 (C).0.1 (rl).0.2S(OL 0.75(A). I.O~m).2.0(r:---c). and 5.0 t@) and no trypsin. trypsin t5 pg/ml) plus soybean trypsin inhibitor (IO pg/ml). or leupeptin (5 pgiml) ((I!).
tions were tested in the time course studies which showed, however, a linear or almost linear decrease in P-value with time. This indicates the apparent decrease in the molecular weight of the labeled protein due to hydrolysis. At high enzyme/substrate ratios, the apparent rate of reaction was high. Enzyme inhibitors, leupeptin, soybean trypsin inhibitor, or antipain, inhibited the activity of trypsin (Fig. 1) or papain (Fig. 2) according to their known specificities. As shown in Fig. 1.. the trypsin activity in a quantity as small as 0.05 pgiml at 20°C. or 0.0125 pg/ml at 37°C (not shown), can be detected. Ribonuclease, which is relatively resistant against proteolytic digestion, was observed to be digested with Pronase (Fig. 4). A similar case with trypsin inhibitor is shown in Fig. 3. Although the above results indicate the dose (of enzyme) and time dependence of the measurement, it should be
4v
30
MIN FIG. 2. Digestion of IgG with papain. FITC-Labeled IgG was dissolved in 2 ml of 0.01 M phosphate-buffered 0. I5 M NaCl solution at pH 7.0 at 42 &ml. Cysteine was added to the buffer to give the concentration of 250 &ml. Digestions were carried out at 20°C. The enzyme was added to final concentrations tmicrogrampermilliliter)of2.5t~),5.0(0). lO(O).and IS(A), and no papain or papain (5 &ml) plus antipain (5 @ ml) (0).
APPARENT MOL.WT. x 16 -13
P 400
300
200
5
I 10
I
20
1
30
I
40
I
50
1.7
MIN FIG. 3. Digestion of soybean trypsin inhibitor. FITCLabeled soybean trypsin inhibitor (120 &ml) was digested with varying amounts of pepsin in 0.1 M acetic acid, of which the pH was adjusted to 3.0, in a test tube at 37°C. Then an aliquot was taken out for the measurement of P-value at each time interval and placed in a cuvette containing 2 ml of 0. I M Tris-HCI buffer. pH 8.6. The concentrations of pepsin tmicrogram per milliliter) are as follows: 0.2 (0). 0.5 tfI). 1.0 (0). 5.0(O).and 10(A). Theenzymeisubstrateratiosare in the range of 1112 to 11600. P-Value, expressed in arbitrary units, of soybean trypsin inhibitor is exceptionally high and the value is outside of a linear relationship with molecular weights. When this protein is incubated in 0.1% of sodium dodecyl sulfate at 37°C. the value obtained is. however. that which is expected (7).
FLUORESCENCE
0
20
40 MIN
POLARIZATION
60
FIG. 4. Digestion of ribonuclease with Pronase. FITC-Labeled ribonuclease (A,!,,,,,,,, = 0.01, 0.18 ~moliml) was digested in 0.1 M Tris-HCI buffer, pH 7.8. 20°C at various concentrations of Pronase as follows (microgram per milliliter): IO(O). 30(C). and 9U (0). P-Values on the left ordinate are expressed in arbitrary units and the molecular weights corresponding to Pvalues are thown on the right ordinate.
noted that the present study was based on an assumption that the shape of the molecules (globular) and the spectroscopic property of the fluorochrome remain constant during the measurement. Therefore, if the substrates used are unique, the result may be delineated. For instance, if the fluorochrome is located in the vicinity of aromatic residues, the latter may affect the fluorescence rela.xation time of the fluorescein and thus interfere with the data. In the case of trypsinogen. in which four aspartic acid residues follow the N-terminal valine. the local environment of the labeled fluorescein chromophore (on valine I) could be very acidic. Thus, this may affect the fluorescence relaxation t,ime for a shorter period of time than that of the normal ionized state. Such an unusual peptide may also be likely to bind to trypsin by ionic interaction (very acidic peptide vs basic protein). even when cleaved at lysine 6. Since the protein concentration (substrate + enzyme) was so low that the change in viscosity was negligibly small, the P-values were not affected in this regard. In addition to the depolarization described above. an
FOR
PROTEASE
ASSAY
225
increase in the intensity of fluorescence was observed as proteolysis progressed. This could mean that fluorescence lifetime may be increasing. Its determination. therefore, will certainly aid the analysis and interpretation of the data. It was difficult to determine the precise total fluorescence intensity using the instrument described herein. The present method is so sensitive that the residual proteolytic activity contaminating the crude substrate preparation was detectable as a small change in the f-value. This was observed in the case of a commercial preparation of casein or low-grade plasminogen. Their residual proteolytic activity was inhibited by either kallikrein inhibitor or trypsin inhibitor (H. Maeda. unpublished data). Therefore, appropriate precautions must be taken when such labeled substrates are used in this method. Under a controlled condition. namely. when the pH of the reaction of protein and FITC is slightly above the pK,, of the N-terminal amino group, one primarily obtains N-terminal-labeled protein with FITC (8). With this substrate. proteolysis near the C-terminal end will effect little change in the P-value if such an enzyme as carboxypeptidase is used.4 The present method offers two advantages over other methods, simplicity and high sensitivity, the essential requirements for a routine assay. The present procedure does not require any process such as separation, precipitation, centrifugation. further reaction, etc. The data are usually obtained directly from the reaction mixture in the cuvette. The high sensitivity represented by 4 When only N-terminal amino groups are labeled with FITC. most zymogens. like plasminogen. and prohormones, like angiotensinogen, kininogen etc., will be the best candidates for this method. because they are known to be cleaved off near the amino-terminal ends rather than the carboxyl-terminal ends. They will yield small labeled peptide fragments and thus result in a pronounced decrease in P-values because the labeled parts are quite small in molecular weight compared with those of the original. The unlabeled larger section left is undetectable by fluorescence polarization.
326
HIROSHI
trypsin (Fig. 1) could be comparable to the most sensitive methods using synthetic chromogenic substrates where trypsin is used at 0.01 to 0.1 pgiml (9) or to those utilizing radiolabeled substrates. which allow the determination of trypsin at 0.065 Fug/ml (IO). Another advantage to be emphasized is that with the present method. natural proteins (or enzymes) can be utilized as the substrate in minute quantities. Intricate and elaborate substrate requirements of unique or unidentified proteases, such as intracellular proteases. renin (I I ). or proteases in inflammatory processes (12). can be readily analyzed using natural protein substrates rather than synthetic artificial substrates. APPENDIX Molecular
Weight
and P-Value
The relationship between P-values of various FITC-labeled proteins and their molecular volumes can be expressed by Perrin’s equation ( 13) as l/P = l/P,, + (l/P,,
-
113) x
[(RT)IV]
x r/n,
[l]
where the protein under consideration is spherical, P is the observed P-value, P,, is the constant obtained by extrapolating P to ~/q = 0 for each fluorochrome (about 0.45 for fluorescein), R is the gas constant, T is the absolute temperature, T is the lifetime of fluorescence [for fluorescein. about 5 x lo-!’ s (14)]. 77is the viscosity of the solvent, and V is the molecular volume. When lip and l/V are replaced by Y and X. respectively, Eq. [l] is transformed into Y=
Kx
+ lip,,.
PI
where K is a constant which combines all of the above constants. Since, the molecular weight (M) and the molecular volume ( V) have the following relationship: v = u x M.
MAEDA
2
4
I
6
a(I/MIX
lo”
FIG. 5. Perrin plot of FITC-labeled proteins. A linear relationship between I/P and l/.%1 is shown. (0) Primarily globular proteins; (cl) glycoproteins. Abbreviations and molecular weights are as follows: IgG. immunoglobulin G (I.6 Y lo”): TF. transfertin I9 x IO’); ALB, human serum albumin (6.8 x IO’. free of noncovalently bound fluorescein): Fc, fragment c of IgG (5 x IO”); CTG. chymotrypsinogen (7.6 x IO”): TG. trypsinogen (3.5 x IO’): RN. bovine pancreatic ribonuclease (I.3 Y IO’): NCS. neocarzinostatin (I.1 x IO’); AGP. (Y,acid glycoprotein (4 * IO’); OVM. chicken ovomucoid (2.9 x IO’). AGP and OVM contain about 35 and 2.5’; carbohydrates. Measurements are made at 20°C. CTG. TG, RN. and NCS are measured in 0.75 M boratebuffered (pH 8.8) 0.35 M NaCl and others were in 0.01 M phosphate-buffered tpH 7.5) 0. I5 hr NaCI. All proteins were obtained from commercial sources except neocarzinostatin. cu,-acid glycoprotein. and ovomucoid which were from our own preparation. gifts of Dr. K. Schmid. Boston University, and Dr. R. E. Feeney. University of California, Davis. California, respectively.
Thus. l/P and l/M exhibit a linear relationship when K remains constant. and u is the partial specific volume of the protein (0.70-0.75 ml/g). The results obtained based on Eq. [2] are shown in Fig. 5 which indicates P 3~M. Therefore, the decrease in molecular weight is reflected in the decrease in P-value as described in the results of the preceding sections. Nonglobular glycoprotein. ovomucoid. and n,-acid glycoprotein, exhibited larger P-values than expected. This may be explained by the hydration effect in these proteins (thus, larger molecular volume). Reduced and alkylated FITC-chymotrypsinogen showed a more rapid decrease in P-value when digested with trypsin than with native FITC-chymotrypsinogen indicating enhanced susceptibility due to loss
FLUORESCENCE
of tight folding by the disruption bridges.
POLARIZATION
of disulfide
REFERENCES I. 2. 3. 4.
Maeda. H. ( 1977) Srikri&~ 49,990. [In Japanese] Weber. G. (1953)Adv. Protein C‘he~z. 8, 415-459. Maeda, H., ( 1978) (‘/in. C‘/jrtn. 24, in press. Dandlik’x. W. B. (1977) in Immunochemistry of Proteins (Atassi. M. Z.. ed.). Vol. 1. Chap. 3. pp. 131-262. Plenum, New York. 5. Dandliklsr. W. B.. Shapiro. H. C.. Maduski. J. W.. Alonr,o. R.. Feigen. G. A., and Hamrick, J. R.. Jr. ( 1’964) I,,~~r~~tr?c~~./zc~n~i.rtr\. 1. 16% 191. 6. Harber. H., and Bennett. J. C. (1962) Proc,. Not. Ac,~rd S
MAEDA
Assay of proteolytic enzymes by the fluorescence polarization method is described in which fluorescein isothiocyanate-conjugated proteins are used as substrates. Hydrolysis of the natural substrates is accompanied by a decrease in the fluorescence polarization values reflecting time and also by an enzyme concentration-dependent decrease in the overall molecular weight of the substrates. Trypsin activity as low as 0.05 &ml at 20°C is detected at a fluorochrome concentration of IO to 20 pmol/ml. In this method. no separation procedure or further reaction for quantitation is necessary. Thus, using a computerized instrument, this method is extremely sensitive, simple. and rapid.
The fluorescence polarization method with its extremely high sensitivity appears to have versatile potential applications (2). The parameter utilized in this method is the fluorescence polarization value (P-value). which reflects the change in the molecular weight of a fluorochrome-labeled substance due to the altered Brownian motion. or the change in the viscosity of the solvent being measured. Such a change in the molecular weight or viscosity results in a change in the P-value. When the P-values of fluorescein isothiocyanate ( FIT0labeled2 proteins of different molecular weights were measured, the reciprocals of these P-values were found to be linearly correlated with the reciprocal values of their respective molecular weights in the range of 5000 to 160,000 for most of the globular and simple proteins (3). Empirically. the log molecular
weights and P-values were also found to exhibit a near-linear relationship. Based on this fact, quantitation of antigen or antibody became possible by the measurement of the increase in the molecular weight (and Pvalues) of the complex being formed (3-7). In this assay system, either one of the molecules. antigen or antibody, is labeled with FITC. This method is capable of detecting the interaction with extreme rapidity (about 2 min) and sensitivity ( IO-” mol/ml) (3.7). This paper reports on the application of the fluorescence polarization method to the assay of proteolytic enzymes using various protein substrates labeled with FITC. When a protein is hydrolyzed to smaller fragments, the apparent and overall molecular weight in the system diminishes. resulting in a decreased P-value and thus substantiating the proteolytic activity.
’ A preliminary account of this work was presented at the 50th General Meeting of the Japanese Biochemical Society, October 1977, Tokyo (I). ’ Abbreviations used: FITC. fluorescein isothiocyanate: TPCK, tosylphenylaline chloromethyl ketone; PBS. 0.01 M phosphate-buffered 0. I5 M NaCl solution, pH 7.0; Tris, tris(hydroxymethyl)aminomethane; Pvalue, fluorescence polarization value in arbitrary units; IgG. immunoglobulin G: F/P. Auorochrome per mole of protein.
0003.2697/79/010212-06$01.00/O CopyrIght All
right\
t of
1979
by
reproductmn
Academic m any
Press. form
Inc. rewrved.
MATERIALS trzstrlrmctlts.
polarimeter.”
AND METHODS
The fluorescence spectroModel MAC-2 Type HR-I
,’ The instrument is designed primarily for the analysis of fluorescein chromophore which has maximum excitation and emission spectra at about 490 and SZO nm, respectively.
7-l) ---
FLUORESCENCE
POLARIZATION
(Japan Immunoindustries Co., Ltd.. Takasaki), used for the measurements was equipped with three innovative features: (i) computerization which integrates 50 measurements and calculations of P-values before printout every 82 s; (ii) a rotating polarizer. which made the detection of fluorescence intensity of both parallel and perpendicular components in a single photomultiplier tube possible as a cosine function against the time of rotation (30 rps): and (iii) three-cavity filters (Detric Optics, Marborough, Mass.) with near-monochromatic performance, with a little extinction and no depolarization in the exciting or the emitting beam. A refrigerated water circulator Model RTE-8 (Neslab Inst. Inc., Portmouth, N.H.) was employed to control the temperature under operation constant within -+O.Ol”C of error. t5r;~r~rl:~.str/rd c~ilc~t~~icrrl.s. Proteolytic enzymes used were as follows: papain (E. Merck AG., Darmstadt. 60 U/mg); TPCKtreated trypsin and pepsin (both from Worthington Biochemical Corp.. Freehold. N. J.. 183.1 and 2530 Uimg, respectively): and Pronase (Kaken Chemical Co., Ltd.. Tokyo. 45 Uimg). Antipain and leupeptin were gifts from Dr. H. Umezawa and Dr. Y. Matsushima. Tokyo. Substrates used were obtained from the following sources: human fibrinogen from Midori Juji Company (Osaka): antihuman sheep IgG from Meloy Laiboratories (Springfield. Va.); and bovine pancreatic ribonuclease A and soybean trypsin inhibitor from Sigma Chemical Company (St. Louis, MO.). FITC was obtained from Dojin Chemical Company (Kumamoto). Other chemicals were from commercial sources. Ltrhclitlg the .srrhstrutc~s \i.itll FITS. The substrates were labeled in 0.1 M Na carbonate-bicarbonate buffer. pH 9.0, with about 20 to IO0 moles of FITC per amino group in proteins with stirring at room temperature for 11to 6 h. The labeled protein was separated from the reaction mixture with a column of Sephadex G-50 using 0.01 M
FOR
PROTEASE
ASSAY
223
phosphate-buffered 0.15 M NaCl solution (pH 7.0, PBS). Furthermore. dialysis against PBS at 4C for 2 days was carried out in order to ascertain the removal of free fluorescein chromophore. The labeled proteins had the following numbers of Buorochromes per mole of protein ( F/P) as determined spectroscopically at 490 and 280 nm. respectively: IgG, 5.5; soybean trypsin inhibitor, 1.O; fibrinogen. 1I .3: and ribonuclease, I .O. Reaction of FITC with proteins occurs primarily with the a-amino groups, subsequently with the e-amino groups, of the proteins depending upon their pX,, values and the pH of the reaction (8). The FITC-labeled proteins remained soluble after labeling. GI;JWlC wtrcti0u.v. Enzyme reactions were carried out in a cuvette placed in the instrument under the conditions described in the figure legends. The concentration of each substrate was 31 .Opg or more per milliliter with an absorbance of 0.005 to 0.015 at 490 nm, and the enzymes were used at various concentrations. P-Values were printed out directly on the recorder chart every 82 set during enzyme digestion. However. when the pH of the enzyme reaction is different from the pH of the measurements (pH 7 to 9). as in the case of pepsin digestion at pH 3, the pH of the reaction mixture or that of its aliquots used for the measurements was brought to an appropriate pH (about 8.6) in a buffer solution. This keeps both the quantum yield of fluorescence and the relaxation time of fluorescence-excitation constant, and the P-value thus obtained is reliable (1.7). RESULTS AND DISCUSSION P-Values obtained by the various enzyme digestions of FITC-labeled protein substrates were plotted, and the results are shown in Figs. I through 4. Figures I and 3 show that the polarization changes are small. In such cases, the accuracy of the assay suffers. Various enzyme concentra-
224
HIRCISHI
MAEDA
Pe# 100
Pm 100
I
4
I 6
I
I
12
16
I 20
1
900
1
IO
20
24
MIN FIG. I. Digestion of fibrinogen with trypsin. FITCLabe!ed human fibrinogen as the substrate, dissolved in 2 ml of 0.1 M Tris-HCI buffer containing 5 mM CaCI,, pH 7.8. in a cuvette, was placed in the instrument. The temperature of the cuvette was then allowed to equilibrate (20°C). after which varied amounts of trypsin were added and mixed. P-Values at each time period are plotted as percentage of f-value at zero time. The substrate concentration is constant for each while the enzyme assay (A 14,1 ,,,,,= 0.015 or 21 &ml). concentrations (microgram per milliliter) included 0.05 (C).0.1 (rl).0.2S(OL 0.75(A). I.O~m).2.0(r:---c). and 5.0 t@) and no trypsin. trypsin t5 pg/ml) plus soybean trypsin inhibitor (IO pg/ml). or leupeptin (5 pgiml) ((I!).
tions were tested in the time course studies which showed, however, a linear or almost linear decrease in P-value with time. This indicates the apparent decrease in the molecular weight of the labeled protein due to hydrolysis. At high enzyme/substrate ratios, the apparent rate of reaction was high. Enzyme inhibitors, leupeptin, soybean trypsin inhibitor, or antipain, inhibited the activity of trypsin (Fig. 1) or papain (Fig. 2) according to their known specificities. As shown in Fig. 1.. the trypsin activity in a quantity as small as 0.05 pgiml at 20°C. or 0.0125 pg/ml at 37°C (not shown), can be detected. Ribonuclease, which is relatively resistant against proteolytic digestion, was observed to be digested with Pronase (Fig. 4). A similar case with trypsin inhibitor is shown in Fig. 3. Although the above results indicate the dose (of enzyme) and time dependence of the measurement, it should be
4v
30
MIN FIG. 2. Digestion of IgG with papain. FITC-Labeled IgG was dissolved in 2 ml of 0.01 M phosphate-buffered 0. I5 M NaCl solution at pH 7.0 at 42 &ml. Cysteine was added to the buffer to give the concentration of 250 &ml. Digestions were carried out at 20°C. The enzyme was added to final concentrations tmicrogrampermilliliter)of2.5t~),5.0(0). lO(O).and IS(A), and no papain or papain (5 &ml) plus antipain (5 @ ml) (0).
APPARENT MOL.WT. x 16 -13
P 400
300
200
5
I 10
I
20
1
30
I
40
I
50
1.7
MIN FIG. 3. Digestion of soybean trypsin inhibitor. FITCLabeled soybean trypsin inhibitor (120 &ml) was digested with varying amounts of pepsin in 0.1 M acetic acid, of which the pH was adjusted to 3.0, in a test tube at 37°C. Then an aliquot was taken out for the measurement of P-value at each time interval and placed in a cuvette containing 2 ml of 0. I M Tris-HCI buffer. pH 8.6. The concentrations of pepsin tmicrogram per milliliter) are as follows: 0.2 (0). 0.5 tfI). 1.0 (0). 5.0(O).and 10(A). Theenzymeisubstrateratiosare in the range of 1112 to 11600. P-Value, expressed in arbitrary units, of soybean trypsin inhibitor is exceptionally high and the value is outside of a linear relationship with molecular weights. When this protein is incubated in 0.1% of sodium dodecyl sulfate at 37°C. the value obtained is. however. that which is expected (7).
FLUORESCENCE
0
20
40 MIN
POLARIZATION
60
FIG. 4. Digestion of ribonuclease with Pronase. FITC-Labeled ribonuclease (A,!,,,,,,,, = 0.01, 0.18 ~moliml) was digested in 0.1 M Tris-HCI buffer, pH 7.8. 20°C at various concentrations of Pronase as follows (microgram per milliliter): IO(O). 30(C). and 9U (0). P-Values on the left ordinate are expressed in arbitrary units and the molecular weights corresponding to Pvalues are thown on the right ordinate.
noted that the present study was based on an assumption that the shape of the molecules (globular) and the spectroscopic property of the fluorochrome remain constant during the measurement. Therefore, if the substrates used are unique, the result may be delineated. For instance, if the fluorochrome is located in the vicinity of aromatic residues, the latter may affect the fluorescence rela.xation time of the fluorescein and thus interfere with the data. In the case of trypsinogen. in which four aspartic acid residues follow the N-terminal valine. the local environment of the labeled fluorescein chromophore (on valine I) could be very acidic. Thus, this may affect the fluorescence relaxation t,ime for a shorter period of time than that of the normal ionized state. Such an unusual peptide may also be likely to bind to trypsin by ionic interaction (very acidic peptide vs basic protein). even when cleaved at lysine 6. Since the protein concentration (substrate + enzyme) was so low that the change in viscosity was negligibly small, the P-values were not affected in this regard. In addition to the depolarization described above. an
FOR
PROTEASE
ASSAY
225
increase in the intensity of fluorescence was observed as proteolysis progressed. This could mean that fluorescence lifetime may be increasing. Its determination. therefore, will certainly aid the analysis and interpretation of the data. It was difficult to determine the precise total fluorescence intensity using the instrument described herein. The present method is so sensitive that the residual proteolytic activity contaminating the crude substrate preparation was detectable as a small change in the f-value. This was observed in the case of a commercial preparation of casein or low-grade plasminogen. Their residual proteolytic activity was inhibited by either kallikrein inhibitor or trypsin inhibitor (H. Maeda. unpublished data). Therefore, appropriate precautions must be taken when such labeled substrates are used in this method. Under a controlled condition. namely. when the pH of the reaction of protein and FITC is slightly above the pK,, of the N-terminal amino group, one primarily obtains N-terminal-labeled protein with FITC (8). With this substrate. proteolysis near the C-terminal end will effect little change in the P-value if such an enzyme as carboxypeptidase is used.4 The present method offers two advantages over other methods, simplicity and high sensitivity, the essential requirements for a routine assay. The present procedure does not require any process such as separation, precipitation, centrifugation. further reaction, etc. The data are usually obtained directly from the reaction mixture in the cuvette. The high sensitivity represented by 4 When only N-terminal amino groups are labeled with FITC. most zymogens. like plasminogen. and prohormones, like angiotensinogen, kininogen etc., will be the best candidates for this method. because they are known to be cleaved off near the amino-terminal ends rather than the carboxyl-terminal ends. They will yield small labeled peptide fragments and thus result in a pronounced decrease in P-values because the labeled parts are quite small in molecular weight compared with those of the original. The unlabeled larger section left is undetectable by fluorescence polarization.
326
HIROSHI
trypsin (Fig. 1) could be comparable to the most sensitive methods using synthetic chromogenic substrates where trypsin is used at 0.01 to 0.1 pgiml (9) or to those utilizing radiolabeled substrates. which allow the determination of trypsin at 0.065 Fug/ml (IO). Another advantage to be emphasized is that with the present method. natural proteins (or enzymes) can be utilized as the substrate in minute quantities. Intricate and elaborate substrate requirements of unique or unidentified proteases, such as intracellular proteases. renin (I I ). or proteases in inflammatory processes (12). can be readily analyzed using natural protein substrates rather than synthetic artificial substrates. APPENDIX Molecular
Weight
and P-Value
The relationship between P-values of various FITC-labeled proteins and their molecular volumes can be expressed by Perrin’s equation ( 13) as l/P = l/P,, + (l/P,,
-
113) x
[(RT)IV]
x r/n,
[l]
where the protein under consideration is spherical, P is the observed P-value, P,, is the constant obtained by extrapolating P to ~/q = 0 for each fluorochrome (about 0.45 for fluorescein), R is the gas constant, T is the absolute temperature, T is the lifetime of fluorescence [for fluorescein. about 5 x lo-!’ s (14)]. 77is the viscosity of the solvent, and V is the molecular volume. When lip and l/V are replaced by Y and X. respectively, Eq. [l] is transformed into Y=
Kx
+ lip,,.
PI
where K is a constant which combines all of the above constants. Since, the molecular weight (M) and the molecular volume ( V) have the following relationship: v = u x M.
MAEDA
2
4
I
6
a(I/MIX
lo”
FIG. 5. Perrin plot of FITC-labeled proteins. A linear relationship between I/P and l/.%1 is shown. (0) Primarily globular proteins; (cl) glycoproteins. Abbreviations and molecular weights are as follows: IgG. immunoglobulin G (I.6 Y lo”): TF. transfertin I9 x IO’); ALB, human serum albumin (6.8 x IO’. free of noncovalently bound fluorescein): Fc, fragment c of IgG (5 x IO”); CTG. chymotrypsinogen (7.6 x IO”): TG. trypsinogen (3.5 x IO’): RN. bovine pancreatic ribonuclease (I.3 Y IO’): NCS. neocarzinostatin (I.1 x IO’); AGP. (Y,acid glycoprotein (4 * IO’); OVM. chicken ovomucoid (2.9 x IO’). AGP and OVM contain about 35 and 2.5’; carbohydrates. Measurements are made at 20°C. CTG. TG, RN. and NCS are measured in 0.75 M boratebuffered (pH 8.8) 0.35 M NaCl and others were in 0.01 M phosphate-buffered tpH 7.5) 0. I5 hr NaCI. All proteins were obtained from commercial sources except neocarzinostatin. cu,-acid glycoprotein. and ovomucoid which were from our own preparation. gifts of Dr. K. Schmid. Boston University, and Dr. R. E. Feeney. University of California, Davis. California, respectively.
Thus. l/P and l/M exhibit a linear relationship when K remains constant. and u is the partial specific volume of the protein (0.70-0.75 ml/g). The results obtained based on Eq. [2] are shown in Fig. 5 which indicates P 3~M. Therefore, the decrease in molecular weight is reflected in the decrease in P-value as described in the results of the preceding sections. Nonglobular glycoprotein. ovomucoid. and n,-acid glycoprotein, exhibited larger P-values than expected. This may be explained by the hydration effect in these proteins (thus, larger molecular volume). Reduced and alkylated FITC-chymotrypsinogen showed a more rapid decrease in P-value when digested with trypsin than with native FITC-chymotrypsinogen indicating enhanced susceptibility due to loss
FLUORESCENCE
of tight folding by the disruption bridges.
POLARIZATION
of disulfide
REFERENCES I. 2. 3. 4.
Maeda. H. ( 1977) Srikri&~ 49,990. [In Japanese] Weber. G. (1953)Adv. Protein C‘he~z. 8, 415-459. Maeda, H., ( 1978) (‘/in. C‘/jrtn. 24, in press. Dandlik’x. W. B. (1977) in Immunochemistry of Proteins (Atassi. M. Z.. ed.). Vol. 1. Chap. 3. pp. 131-262. Plenum, New York. 5. Dandliklsr. W. B.. Shapiro. H. C.. Maduski. J. W.. Alonr,o. R.. Feigen. G. A., and Hamrick, J. R.. Jr. ( 1’964) I,,~~r~~tr?c~~./zc~n~i.rtr\. 1. 16% 191. 6. Harber. H., and Bennett. J. C. (1962) Proc,. Not. Ac,~rd S
