Indicator assay for amino acid decarboxylases

ANALYTICALBIOCHEMISTRY

181,59-65

(1989)

Indicator Assay for Amino Acid Decarboxylases’ Robert

M. Rosenberg,2

Departments

Received

of Chemistry

August

Richard

M. Herreid,

and Biochemistry,

University

George J. Piazza, of Wisconsin,

and Marion

Madison,

Wisconsin

H. O’Leary3 53706

151988

Indicator assays have been devised for glutamate decarboxylase and arginine decarboxylase as models for a general procedure for amino acid decarboxylases. Since the decarboxylation results in the absorption of a proton in the pH range of maximum enzymatic activity, the change in absorbance of an acid-base indicator can be used to follow the progress of the reaction. For glutamate decarboxylase, the substrate was used as the buffer, and l- l’-diethyl-2-2’-cyanine iodide was used as the indicator. For arginine decarboxylase, acetate was used as the buffer, and bromcresol green was used as the indicator. The change in absorption with extent of reaction is linear if the indicator pK is equal to the buffer pK. 0 1989 Academic Press, Inc.

ture bath to add the enzyme, it is not possible to study reactions at high enzyme concentrations; (ii) the solubility and supersaturation of the liberated CO2 lead to errors; and (iii) the efhux of CO, from the solution is slow. With the use of 14C-carboxyl-labeled substrate, the radioactivity of the CO, produced can be used to assay the enzyme (12), but this measurement is carried out at a single time and not continuously. In order to obtain a continuous record of reaction progress, Salvadori and Fasella (13) developed a pH-stat assay for bacterial glutamate decarboxylase based on continuous addition of acid to replace the proton taken up in the reaction -OOC-(CH&CH-COO-

+ H+ +

NH,+ -OOC-(CH,),-NH,+ Although the use of acid-base indicators to follow the progress of enzymatic reactions has a long history (l8), the method has not been used for measuring rates of enzymatic decarboxylations. We have adapted this technique to devise indicator assays for glutamate decarboxylase and arginine decarboxylase in order to have a continuous record of the time course of the reaction for active-enzyme sedimentation (9) and for other purposes and to provide a protocol for the assay of other decarboxylases. The assay of amino acid decarboxylases that are active at an acid pH traditionally has been carried out by measurement of the volume or pressure of the CO, liberated in the reaction as a function of time (lO,ll), but such measurements have several limitations: (i) Since manometric systems require about 100 s to return to thermal equilibrium after removal from the constant-tempera‘This research was supported by Grant NS-0’7657 from the National Institute of Neurological Diseases and Stroke and by a sabbatical leave from Lawrence University. ’ Present address: Department of Chemistry, Lawrence University, Appleton, WI 54912. 3 Present address: Department of Biochemistry, University of Nebraska, Lincoln, NE 68583. 0003-2697/89

$3.00

Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

+ COZ.

[l]

The glass electrode used as a detector responds more slowly than a photoelectric detector and is not suitable for rapid reactions. Zeman et al. (14) developed an automated&O, assay for glutamate, but this method uses a pCOZ electrode and has the same disadvantage of slow response as the glass electrode method. Spectrophotometric methods have sometimes been used for decarboxylases. Camier and Gonnard (15) developed an automatic spectrophotometric assay for Dopa decarboxylase, but their method depended on measuring the absorbance change in a solution of phenolphthalein when it absorbed the CO2 released by the enzymatic reaction. Cozzani (16) developed a spectrophotometric assay for glutamate decarboxylase based on a coupled enzyme system, a method not suitable for active enzyme centrifugation. Torchinskii et al. (17) have developed a spectrophotometric assay for glutamate decarboxylase based on the absorbance change at 220 nm, but many spectrophotometers and the monochromator and detector of the Model E ultracentrifuge have low sensitivity at this wavelength. We therefore undertook the development of a spectrophotometric assay based on the absorbance difference between the two forms of an acid-base indicator. 59

60 MATERIALS

ROSENBERG

AND

ET

AL.

METHODS

Materials Glutamate decarboxylase (EC 4.1.1.15) was isolated from Escherichia coli (ATCC 11246) by the method of Shukuya and Schwert (18) as modified by O’Leary (ll), with the following exceptions: nucleic acids were removed by digestion with ribonuclease and deoxyribonuclease in the initial extraction rather than by precipitation with protamine sulfate in a later step, and the fraction retained in the ammonium sulfate precipitation was that precipitated between 45 and 65% saturation with ammonium sulfate. Arginine decarboxylase (EC 4.1.1.19) was purified by the method of Boeker et al. (19). Pyridoxal-5’-phosphate, L-arginine hydrochloride, glutamic acid, and bromcresol green (3’3 “,5’,5”-tetrabromom-cresolsulfonephthalein) sodium salt were obtained from Sigma, dithiothreitol from Aldrich, and l,l’-diethyl-2,2’-cyanine iodide from Eastman Kodak. Enzyme-grade ammonium sulfate was from SchwarzMann. Other chemicals were of reagent grade, and all solutions were prepared in water purified by a Millipore Super Q system to a resistance greater than 18 MQ. Methods Enzyme activity was determined with a Gilson differential respirometer (11). Measurements for glutamate decarboxylase were made at 37°C in 0.1 M pyridine buffer, pH 4.9, containing 0.025 M glutamate and 0.07 M Cl-. Measurements. for arginine decarboxylase were made at 37°C in 0.05 M acetate buffer, pH 5.0, containing 0.025 M arginine. Three milliliters of buffer was placed in the main chamber of the Warburg flask, and 5-10 pg of enzyme in 5-20 ~1of buffer was placed in the sidearm. For glutamate decarboxylase, 20 ~1 of 1 X lo-’ M pyridoxal-5’-phosphate was also added to the sidearm. Protein concentration was determined by absorption at 280 nm, assuming that a 1 mg/ml solution of glutamate decarboxylase has an absorbance of 1.7 (20) and that a 1 mg/ml solution of arginine decarboxylase has an absorbance of 1.57 (21). Specific activities of the glutamate decarboxylase used varied between 2400 and 3200 ~1CO8 (100 s)-r (mg enzyme)-r. Crystalline glutamate decarboxylase has a specific activity of 3200 ~1 CO2 (100 s)-’ (mg enzyme)-l (20). The specific activity of arginine decarboxylase varied from 3700 to 13,000 ~1 CO2 (100 s)-’ Pure arginine decarboxylase has a (mg enzyme))l. specific activity of 17,400 ~1 COP (100 s))l (mg enzyme))l(19). The pK’s of buffers and indicators were determined by a pH-spectrophotometric titration of a solution containing both buffer and indicator. The pH of the solution was plotted against log[CJ(l - C,)] for both buffer and indicator, where C, is the degree of neutralization, with

5.0 Ip

3.0 -1.0

1

1

-0.5

0.0

0.5

1.0

log [cm-d] FIG. 1.

Plot of pH against log[C,/(l = 4.33 -t 0.05 for glutamic acid.

- C,)]

used to determine

a(buffer) = (moles NaOH added)/(total

pK

moles acid)

and C,(indicator) = (absorbance - absorbance in strong acid) (absorbance in strong base - absorbance in strong acid) The pK’s are the intercepts at log[C,/(l - C,)] = 0, as shown in Fig. 1 and Fig. 2. The spectrophotometric assays were carried out in the thermostated cell-holder of a Cary 118 spectrophotometer at 25°C at a wavelength of 520 nm for glutamate decarboxylase and at 610 nm for arginine decarboxylase. These wavelengths represent the maximum absorbance of the basic forms of the respective indicators. For glutamate decarboxylase, 3.0 ml of a solution that was 0.23 M in NaCl, 1.0 X lop5 M in indicator, and containing various concentrations of glutamic acid adjusted to pH 4.50 with standard NaOH solution, was placed in each of two spectrophotometer cells, one in the reference beam and one in the sample beam. After thermal equilibration, a 5-~1 sample of enzyme solution (l-2 mg/ml) was added to the cell in the sample beam. The change in absorbance was linear for a period of approximately 200 s. The corresponding assay solution for arginine decarboxylase contained 0.05 M acetate buffer at pH 5.0,0.025 M arginine, and 1.0 X low5 M indicator.

INDICATOR

ASSAY

FOR

AMINO

ACID

DECARBOXYLASES

61

ited by its solubility, and the concentration of buffer should be as low as possible consistent with the condition cited above that the pH not change enough to cause a change in kinetic constants.4 Figure 3 shows a simulated plot of absorbance of indicator against degree of neutralization of buffer, with the latter linearly related to extent of reaction. A linear plot is obtained only if pKi, = pKB , as indicated by Eq. [ 21. RESULTS

log [a/CI-CJJ FIG. 2. Plot of pH against log[C,/(l = 4.29 + 0.03 for l,l’-diethyl-2,2’-cyanine

- C,)] used to determine iodide.

pK

De Voe (6), Gibbons and Edsall (7), and Khalifah (8) have shown that, in general, the rate of change in absorbance of the indicator is not equal to the rate of proton release in the enzymatic reaction, and the relationship between the two quantities is a complicated function of the ionization states of the buffer and the indicator as shown by dA -= dt

Ci,[c(In)

- e(HIn)]r

X

KB + [H+]’ WW'KB)~

CB

Kn

121

+ [H'l'

where C represents the total concentration of either buffer or indicator, t is the extinction coefficient of either indicator acid or base, A is the absorbance of the solution, r is the rate of the reaction, and K is the acid ionization constant of indicator or buffer. However, if the pK of the indicator is equal to the pK of the buffer, Khalifah (8) has shown that the rate of change in absorbance of the indicator and the rate of proton release in the enzyme solution are linearly related and that the change in absorbance is linear with time as long as there is no change in pH sufficient to cause a change in kinetic constants and as long as the substrate concentration remains large relative to K,,, . The resulting equation is dA/dT

= Ci, [c(In)

- t(HIn)]

r/C,.

131

It is clear that maximum sensitivity is obtained at maximum concentration of indicator and minimum concentration of buffer. The concentration of indicator is lim-

For glutamate decarboxylase, in order to obtain maximum sensitivity, we chose the substrate as buffer. The pK of the y-carboxyl of glutamic acid is in the pH range of maximum activity of the enzyme. After a considerable search, in which most indicators with pK’s in the appropriate range were found to be anionic and inhibitors of the enzyme, we chose as the indicator l,l’-diethyl-2,2’cyanine iodide (22), which we found to have a pK of 4.29 + 0.03 in 0.23 M NaCl; we found that glutamic acid has a pK of 4.33 f 0.05 in the same medium at 25”C.5 The protonated form of the indicator acid is essentially colorless, whereas the unprotonated form is orange, with an absorbance maximum at 520 nm. The value of t(In) at this wavelength is 5.48 X lo4 M-’ cm-‘. According to Eq. [3], the rate of change in absorbance should be a linear function of the rate of reaction, and the slope should be calculable from that equation. Figure 4 shows the results of a spectrophotometric titration in which aliquots of 0.100 M NaOH were added (before final dilution to volume) to a series of solutions in which the final concentration of total glutamate was 0.0101 M, the final concentration of NaCl was 0.23 M, and the final concentration of indicator was 0.989 X 10e5 M. This titration mimics the consumption of protons that occurs during an enzymatic decarboxylation. Linearity of this curve implies linearity of the enzyme assay. The experimental slope was 0.511 -t 0.034, and the value calculated from Eq. [2] was 0.542. For arginine decarboxylase, we used acetic acid as buffer, with an apparent pK of 4.73 by pH titration. The indicator, bromcresol green, had a pK of 4.75 by spectrophotometric titration. At 610 nm, E(In) is 3.44 X lo4 M-l cm-’ and t(HIn) is zero. Figure 5 shows the results of a spectrophotometric titration of bromcresol green in acetate buffer with NaOH, with the 4 Since the use of this assay requires that the pH of the solution must change in order for changes in the extinction coefficient of the indicator to be observed, the rate of reaction decreases when the pH changes beyond the region in which the activity of the enzyme is relatively independent of pH. The measured initial rate is that characteristic of the initial pH. 5 Glutamic acid has a larger pK value than the indicator at zero ionic strength, but the pK of glutamic acid decreases with increasing ionic strength and that of the indicator changes in the opposite direction, as predicted by the Debye-Huckel limiting law; the two values become equal at an ionic strength of 0.23.

62

ROSENBERG

ET

AL.

0.60 0.50 0.40 z e5

0.30

:: 9

0.20 0.10

0.6

0.6

a (Buffer) FIG. 3. acid. pKs successive

Simulated plot of absorbance of indicator base relative = 4.31, Ci, = 1.0 X 10e5 M, c(In) - c(HIn) = 5.5 X 10e4 increments of 0.1. For Curve 1, pK1, = 5.91. For Curves

M-’

to indicator acid as a function of the degree of neutralization of the buffer cm -I. For Curve 7, pK,, = pKs. For Curves 6-2, pK is greater than pKs by 8-12, pKi, is less than pKs by increments of 0.1. For Curve 13, pK,, = 2.81.

linear behavior expected when the pK's of buffer and indicator are equal. The experimental slope was 0.00692 + 0.00013, and the slope calculated from Eq. [2] was 0.00662. The rate of the enzymatic reaction was calculated from the slope of the initial linear portion of the

plot of absorbance against time by means of Eq. [31, with the rate expressed as moles COZ per liter per unit time (100 s in this case). If the value of r calculated from Eq. [3] is divided by the enzyme concentration in the reaction mixture in milligrams per liter, the result is the

ml NaOH FIG. 4. M

NaOH

Plot of absorbance of 1.0 X lo-’ added, used as a test of Eq. [3].

M

l,l’-diethyl-2,2’-cyanine

iodide,

0.01

M

glutamate

buffer

solution,

as a function

of volume

of 0.100

INDICATOR

ASSAY

FOR

0

AMINO

ACID

63

DECARBOXYLASES

60

40

20

ml NaOH FIG. 5. Plot of absorbance of 1.036 added, used as a test of Eq. [3].

X 10e5 M bromcresol

green,

0.05 M acetate

usual u of steady-state enzyme kinetics, expressed in this case as moles CO2 (100 s)-l (mg enzyme)-‘. A plot of the rate of absorbance change against enzyme concentration for glutamate decarboxylase is shown in Fig. 6, demonstrating that the rate of change is a linear function of enzyme concentration. A double-reciprocal plot (23) of the data at 37°C and 0.23 M NaCl and an enzyme concentration of 3.02 mg/ liter is shown in Fig. 7. The resulting kinetic parameters

buffer

solution,

3.0

-

of 0.0948

V/Km = 3.4 X lo4 pm01 CO2 (100 s)-l (mg enzyme)-l

For the same enzyme in 0.1 M pyridine-HCl O’Leary et al. (24) obtained values at 37°C of

V = 145 pm01 CO2 (100 s)-’ (mg enzyme)-’ mM.

2.0 -

1.0 -

0.0

’

’

0.0

I 2.0

I 4.0

(E& /$g FIG.

6.

Plot

of dA/dt

against

concentration

of glutamate

M NaOH

M-’

Km = 2.0 mM.

Km = 0.95

-

of volume

V = 68 pm01 CO2 (100 s)-l (mg enzyme)-l

are

4.0

as a function

I

I 6.0

I

ml-9

decarboxylase,

C, = 1.0 X 1O-3 M, C,, = 8.8 X 10m6 M.

buffer,

ROSENBERG

l/[SJ, FIG. 7. mate liter.

/ 103Jmol-1

Double-reciprocal plot of steady-state kinetic decarboxylase at 37°C and an enzyme concentration

data for glutaof 3.02 mg/

In 0.1 M pyridine-HCI containing 0.2 M Cl-, Fonda (25) found Km approximately equal to 1 mM. Her value for V is not comparable since it is not corrected for the specific activity of her enzyme preparation. In a medium that is 0.1 M in NaCl and in which the substrate is the only buffer, Salvadori and Fassella (13), using a pH-stat method, obtained V = 160 hmol COZ (100 s)-’ (mg enzyme)-’ K,,, = 0.64 IIIM, with the value for V estimated from the supplier’s specific activity for the dried bacterial cells used in the experiments. Similar measurements for arginine decarboxylase at pH 4.73 with 1.04 X 10e5M bromcresol green gave a linear reciprocal plot and values of V = 223 pmol COZ (100 s)-l (mg enzyme))’ K,,, = 2.45 mM. The corresponding values of Blethen et al. (21) for measurements in succinate buffer, pH 5.2, are V = 833 pmol COP (100 s)-l (mg enzyme)-l K, = 0.65 mM.

ET

AL.

carboxylase enzymes that are active at acid pH. Since glutamate decarboxylase is the only decarboxylase for which the substrate is a buffer in the pH range of enzyme activity, arginine decarboxylase and other decarboxylases require an added buffer. To obtain linearity in the assay, as indicated above, an indicator must be found with a pK within 0.1 pH unit of that of the buffer. For arginine decarboxylase, acetate buffer and bromcresol green were found to be suitable. Linearity of the absorbance with extent of reaction is essential for active enzyme sedimentation, but those who use an indicator assay for measuring initial rates have additional flexibility in choosing buffer-indicator combinations since an empirical calibration can provide a relation between initial absorbance change and initial rate of proton uptake. As can be seen from Fig. 3, one can even obtain increased sensitivity when pKs does not equal pKI,. The equality of pKs and pK1, for the glutamate decarboxylase assay applies only at an ionic strength of 0.23, and variation of ionic strength is suggested as a method of achieving equality of pK values when it is necessary.6 Bromcresol green was tested for use with glutamate decarboxylase, but it was found to be an inhibitor of the enzyme (26), along with many other anionic indicators. Although we do not have evidence for other decarboxylases, our experience with glutamate decarboxylase and arginine decarboxylase suggests that an anionic indicator is less likely to inhibit an enzyme with a cationic substrate and vice versa. As indicated by Eq. [3], the sensitivity of this assay is proportional to the difference between the extinction coefficients of the acidic and basic forms of the indicator and to the concentration of the indicator, and is inversely proportional to the concentration of the buffer. The maximum concentration of indicator is limited by its solubility in aqueous solution, and as dilute a buffer as feasible should be used. The use of an acid-base indicator for which E(In) - e(HIn) = 5.94 X lo4 M-’ cm-’ provides a greater sensitivity than the direct observation at 220 nm of the difference between the absorbance of glutamic acid and the absorbance of y-aminobutyric acid, if the ratio of E(In) to f(B) is sufficiently great. The difference in extinction coefficients for the direct spectral difference method is 17.0 M-’ cm-’ (8). The use of the direct spectral difference also has the disadvantage of using a spectral region in which there is a considerable background absorption and in which most spectrophotometers have a low sensitivity. Although our measurements were carried out only at pH 4.5 for glutamate decarboxylase and at pH 4.73 for arginine decarboxylase, it should be possible to carry out

DISCUSSION

The method described here for glutamate decarboxylase and arginine decarboxylase is applicable to other de-

s Chloride is an activator of glutamate decarboxylase, plateau in the variation of activity with Cl- concentration near 0.2 M NaCl(26).

and there is a in the region

INDICATOR

ASSAY

FOR

AMINO

the assays in a range of fl pH unit about the pK of the indicator, with differences in sensitivity and linearity as indicated in Fig. 3. We obtained a linear double-reciprocal plot for the steady-state kinetics of glutamate decarboxylase using the indicator assay at substrate concentrations between 0.67 and 5.0 Km. At lower substrate concentrations, points deviated from the line characteristic of higher concentrations, perhaps because buffering was not adequate. This problem should not arise for those decarboxylases for which an independent buffer is added. A spectrophotometric assay is essential for active enzyme centrifugation since an optical method of detecting the sedimentation of the active species is necessary. In addition, spectrophotometric methods have the advantages of speed and convenience over the conventional manometric assays of decarboxylases, and the range of enzyme concentrations that can be used is not limited by the capacity of the respirometer. In principle, very high concentrations of enzyme can be studied by means of stopped-flow kinetics, whereas the respirometer readings are not reliable before 100 s after mixing. The indicator assay is, in general, less sensitive than the manometric assay or the coupled enzyme assay. The criteria for choosing a buffer-indicator-enzyme system are the following: (i) The indicator must not be an inhibitor zyme in the manometric assay procedure.

of the en-

(ii) It will be convenient if the indicator-buffer combination shows a linear relationship between absorbance and amount of base added, as in Fig. 4. This is essential for active enzyme centrifugation. (iii) The pK of buffer and indicator mined as in Figs. 1 and 2.

should be deter-

(iv) The extinction coefficients of the acidic and basic forms of the indicator (usually at the wavelength of maximum difference in extinction coefficient) must be determined, so that the results of a graph such as Fig. 4 can be compared with Eq. [3 ] and so that the rate of reaction can be compared with that obtained with a manometric assay under the same conditions.

ACID

65

DECARBOXYLASES

REFERENCES H., Croxatto, R., and Huidobro, F. (1939) Anal. Biol. 5, 1. Croxatto, 55-65. 0. H., Roberts, N. R., Wu, M.-L., Nixon, W. S., and Craw2. Lowry, ford, E. J. (1954) J. Biol. Chem. 207,19-37. 3. Wajzer, J. (1949) Compt. Rend. 229,1270-1272. 4. Crane, R. K., and Sols, A. owick, S. P., and Kaplan, demic Press, New York. 5. Darrow, R. A., and Colowick, ogy (Colowick, S. P., and 235, Academic Press, New

(1960) in Methods in Enzymology N. O., Eds.), Vol. I, pp. 277-286,

(ColAca-

S. P. (1962) in Methods in EnzymolKaplan, N. O., Eds.), Vol. V, pp. 226York.

6. DeVoe, H. (1960) Ph.D. dissertation, Harvard University. 7. Gibbons, B. H., and Edsall, J. T. (1963) J. Biol. Chem. 238,35023507. 8. Khalifah, R. G. (1971) J. Biol. Chem. 264,2561-2573. 9. O’Leary, M. H., and Rosenberg, R. M. (1976) Division of Biological Chemistry, Amer. Chem. Sot., San Francisco, August. 10. Najjar, V. A., and Fischer, J. (1954) J. Biol. Chem. 206,215-219. 11. O’Leary, M. (1969) Biochemistry 8,1117-1122. 12. Strausbach, P. H., Fischer, E. H., Cunningham, C., and Hager, L. F. (1967) Biochem. Biophys. Res. Commun. 28,525-530. 13. Salvadori, C., and Fasella, P. (1970) Ital. J. Biochem. 19,193-203. 14. Zeman, G. H., Sobocinski, P. Z., and Chaput, R. L. (1973) Anal. Biochem. 52,63-68. 15. Camier, M., and Gonnard, P. (1969) Bull. Sot. Chim. Biol. 50, 933-937. 16. Cozzani, I. (1970) Anal. Biochem. 33,125-131. 17. Torchinskii, Yu. M., Mekhanik, M. L., and Kogan, G. A. (1972) Dok. Akad. Nuuk SSR 205(3), 731-733, Consultant’s Bureau Translation, pp. 276-278. 18. Shukuya, R., and Schwert, G. (1960) J. Biol. Chem. 235, 16491652. 19. Boeker, E. A., Fischer, E. H., and Snell, E. E. (1969) J. Biol. Chem. 244,5239-5245. 20. Strausbach, P. H., and Fischer, E. H. (1970) Biochemistry 9,226233. 21. Blethen, S. L., Boeker, E. A., and Snell, E. E. (1968) J. Biol. Chem. 243.1671-1677. 22. Feldman, L. H., Herz, A. H., and Regan, T. H. (1968) J. Phys. Chem. 72,2008-2031. 23. Lineweaver, H., andBurk, D. (1934) J. Amer. Chem. Sot. 56,658666. 24. O’Leary, M. H., Richards, D. T., and Hendrickson, D. W. (1970) J. Am. Chem. Sot. 92,4435-4440. 25. Fonda, M. L. (1972) Biochemistry 11,1304-1309. 26. Rosenberg, R. M., Campbell, R. C., and Doherty, E. H. (1977) Unpublished observations.