Anal. Chem. 1992, 64, 181-186
181
Determination of Human Serum Alcohol Dehydrogenase Using Isozyme-Specific Fluorescent Substrates Jacek Wierzchowski,' Barton Holmquist, and Bert L. Vallee*
Center for Biochemical and Biophysical Sciences and Medicine, Harvard Medical School, 250 Longwood Avenue, Boston, Massachusetts 02115
Both class I and class I 1 alcohol dehydrogenase (ADH) activities are present In human serum. The contrlbutlon of each class can be measured using two class-specific, fluorogenic substrates, 4-methoxy-l-naphthaldehyde and 6-methoxy-2naphthaldehyde. The former is highly Selective for class I Isozymes, especially those contalnlng a or y subunits, whereas class I 1 ( 1 7 ) ADH preferenllally reduces the latter. Selective InhlbHlon of class I ADH by 4inethylpyrazde further Increases the speclfkky. Specfflclty, accuracy, and preclslon of the assay for serum measurements have been determined. The acllvity of class I ADH In normal human serum Is below the limit of detection of thls method, Le., 3 days for both class I and class I1 ADH. Storage at 4 "C provides some improvement in stability of class I but not of class 11, whereas storage of samples at -70 "C stabilizes both class I and class I1 activities for at least 1 month (Table 11). Precision. Repeatability of the serum ADH activity measurements with both substrates is given in Table III. Sera were chosen with activities between 6 and 145 nM/min for class I ADH and between 25 and 558 nM/min for class 11, values that encompass most activities encountered thus far in the sera of patients. One operator performed five measurements on each sample. The ADH activity of sera can vary
0.1 1 10 MOLAR RATIO ADDED
Figure 1. Mixed recovery experiment for class I and class I1 ADH. Mixtures of purified forms of both ADH classes were prepared at
different ratios and added to pooled inactive serum samples. Rates of reaction with IA and IIA as substrates were measured and activities calculated as described in the Expgrimental Section. Abscissa: composition of the prepared mlxtwe, expressed as molar ratb of class I/class I1 ADH added to serum. Ordinate: Composition calculated from the measured reaction rates. Concentrations of the reference samples, Containing pure class I or I1 ADH, were calculated on the basis of their activity alone in the presence of serum. Rates in this experiment were between 50 and 1700 nM/min for both substrates. The range for class I is 5-1000 nMlmin. 0, values determined when the class I1 ADH activity was heid constant at 200 nM/min. considerably, up to a few thousand times the limit of detection, and the routine 8-min recording time in such cases exceeds the recorder span. With high-activity samples the initial rate was therefore calculated from the linear tracing observed for 2 min of reaction giving sufficient precision (row 4 of Table
111). Accuracy. Since primary standards or reference sera with known amounts of ADH are not available, purified ADH isozymes were added to assays containing 50-fold-diluted, inactive serum to measure accuracy. The ADH isozymes employed were electrophoretically homogeneous with symmetrical HPLC elution profiles and had the highest constant specific activity. Their activity was established both by the standard A340assay using ethanol and by using substrates IA and IIA under conditions described above but in the absence of serum. Recovery experiments employed serum samples exhibiting little or no ADH activity (Table IV). Recoveries for both ADH classes were >75%. Efforts to increase the sensitivity using more concentrated sera showed that only 50% of the activity could be recovered when serum was diluted only 10-fold. Importantly, when serum is diluted 50-fold or more, the recovery of activity is constant, and thus, the activity observed is proportional to the amount of ADH added over the entire range of activities encountered. Mixed recovery experiments were carried out on purified class I1 ADH with a mixture of class I ADH isozymes to differentiate quantitatively between these activities. Such mixtures were added to 50-fold-diluted ADH-free serum samples. There was good agreement between the known
184
ANALYTICAL CHEMISTRY, VOL. 64, NO. 2, JANUARY 15, 1992
8 50
c1
n
100
a 25
50
75
ACTIVITY, %
50
-6
-4
-2
log [4-MeP] Figure 2. Inhibition of ADH activity in human serum by 4-methylpyrazole, compared to that for purified human liver ADH isozymes, as a function of inhibitor concentration: 0, serum activity, 0 , class 11, a a , A, p$,, V, y,y,. Panel A: activity measured with IA as substrate. Panel B: activity measured with I I A as substrate. The serum sample used in this experiment contained approximately 0.1 pM class I 1 and 0.03 p M class I ADH.
.,
isozyme concentrations and those calculated by means of the algorithm described in the Appendix (Figure 1). Limits of Detection. Reaction rates as low as 6.5 nM/min for IA and 25 nM/min for IIA were reproducible, with limits of detection, Le., when the fluorescence change after 8 min is twice that of the noise level, of 1and 5 nM/min, respectively. In serum, this corresponds to concentrations of 0.14 nM for class I and 0.17 nM for class I1 ADH, respectively. The sensitivity is limited primarily by background fluorescence of the serum. The four-assay procedure (see Appendix) provides only estimates of class I plpl and/or pz& isozyme activities. Based on their kcatvalues, the detection limit of these isozymes is -3 nM (Table I), i.e., considerably lower than that for the other isozymes, and hence, only in instances where the /3 forms greatly exceed the other class I or I1 isozymes would their contribution to the rates be significant. This has not been observed in any serum. Specificity and Possible Interferences. The potential of interference from spontaneous oxidation or that catalyzed by other oxidoreductases in serum (18) was examined by testing random samples and verifying that the products of the serum assays are indeed the alcohols IB and IIB and not the acid oxidation products. Assays were allowed to proceed for 1-2 h (to -40% completion), and the reaction mixture was then diluted 10-fold into 3 mL of carbonate buffer, pH 10.0, containing 0.5 mM NAD+. An ADH isozyme was then added (up to a concentration of 10 nM) and the decrease of fluorescence due to aldehyde production monitored. In all 10 samples the fluorescence a t 360-370 nm vanished after addition of ADH, demonstrating that the fluorescent product is indeed the expected alcohol. Only upon extended incubation (24 h) were secondary fluorescent products generated. Their spectra correspond to the 1,4- and 2,6-disubstituted methoxynaphthoic acids, presumably the result of low-level aldehyde dehydrogenase catalyzed oxidation. 4-Methylpyrazole and 4-pentylpyrazole inhibit the serum activity in a manner consistent with that of purified classes I and I1 ADH. The former is a potent inhibitor of class I but not class I1 ADH, while the latter inhibits both at micromolar concentrations (19). The inhibition of serum activity toward IA and IIA as a function of the 4-methylpyrazole concentration (Figure 2) compares well with that obtained for purified class I and class I1 ADH. Further, the activities of a large group of serum samples were examined for inhibition by 4-methylpyrazole and 4-
Flgure 3. Histograms of the degree of inhibition of fluorescent product formation observed in active human sera in the presence of 4methylpyrazole or Cpentylpyrazole: (A, top) Inhibition of the reaction of IA by 0.2 mM Cmethylpyrazole; (B, middle) inhibition of the reaction of I I A in the same conditions; (C, bottom) inhibition of the reaction of I I A by 0.2 mM by 4-pentylpyrazole. Only samples with activities of >10 nM/min toward IA and/or >50 nM/min toward IIA (10 times limit of sensitivity)are included. Total number of samples examined was 83 for A, 105 for B, and 44 for C.
15 -
-3
-2
-1
0
1
log (AIMII)
Figure 4. Histogram of the ratio of class I to class I1 ADH activities, observed in active serum samples, plotted on a logarithmic scale. A total number of 73 samples, selected on the basis of their activity toward NDMA (7).were collected over a period of 7 weeks and analyzed by the four-assay procedure as described in the Experimental Section.
pentylpyrazole, both 0.2 mM. 4-Methylpyrazole inhibited more than 90% of the IA activity in the majority of samples (Figure 3A), but the activity toward IIA remained at 70% or greater (Figure 3B). In contrast, 4-pentylpyrazole inhibited activity of IA by more than 99% in all cases (not shown) and that of IIA by more than 90% in virtually all samples examined (Figure 3C). ADH Activity in Normal Human Serum. In duplicate measurements of serum samples obtained from healthy volunteers (n = 171, no activity toward IA was found. Thus, leaving aside activity due to the pp forms, the normal activity of class I ADH is 0.5 mM renders the assay essentially insensitive to class I11 ADH. Selective measurements of both class I and I1 ADH can be performed over a broad concentration range with both accuracy (Figure 1, Table 111) and precision (Table IV). The procedure can also provide information about the activity of the class I pp isozymes (either in the p1 or pZforms), which kinetically differ distinctly from other class I isozymes (Table I) by using the four-assay procedure (see Appendix). While isozymes reduce substrate IIA at a rate of 15-fold lower than class I1 ADH, unlike class 11, they are readily inhibited by 4-MeP. These isozymes do not make a significant contribution to the total activity of any of the sera examined. The four-assayprocedure is most rigorous in that it correcta for the overlap in specificity of the substrates, the relative sensitivity of the two classes to 4-MeP inhibition, and the presence of the pp isozyme, by means of a set of simultaneous equations which are easily solved. For nearly all samples, the less rigorous, but much faster and easier two-rate assay using IA in the absence, and IIA in the presence, of 4-MeP is quite satisfactory. To examine the validity of this simplified assay, the data obtained by the four-assay procedure were recalculated using the equations for the two-assay method. This simplification indeed results in valid data (Figure 5). When class I1 activity is more than 50-fold higher than that of class I, resulting in discrepancies, the two additional rate measurements of the four-assay procedure are then required. To a large degree the enhanced specificity and sensitivity of the method arises from the monitoring of the highly ADH specific formation of the product alcohols IB and IIB, rather than the nonspecific consumption or production of NADH, the conventional basis of ADH activity. There is neither a substrate nor a coenzyme background rate that requires correction. The method is from 1-2 orders of magnitude more sensitive than that using NDMA or assays based on NADH
185
absorption. It is optimized for use with a conventional fluorometer with standard 4-mL cuvettes and rectangular beam observation, the simplest arrangement presently suitable for routine measurements in an analytical laboratory. Thus,for the first time, an assay is available which measures class I1 ADH activity directly with high sensitivity and accuracy, in the presence of signifcant amounts of class I. Total ADH activity in serum has been measured previously using a variety of substrates and methods but none of them have been specific or selective for isozymes. Purified human isozymes in fact were not available for calibration or control as in this instance, and efforts to achieve indirect differentiation of class I1 from class I activity in serum and tissues based on differential 4-MeP inactivation have been inferential (10, 22). Thus, the present method cannot be compared with any existent one. Serum concentrations of 0.3vl), indicating very low levels of 4-MePsensitive (class I) ADH isozyme(s) relative to class I1 ADH. these cases, due to inaccuracy of the constant a become significant, and more accurate results are obtained when eq l a is replaced by linear combination of eqs l a and 2a: VI - VZ
= (1 - b)AI
+ ~cAII
(54
The equations were resolved using the spreadsheet program (Appleworks,Apple Computer Inc.), which also served for data storage. The resultant values of PP-ADH activity, A,,, must be treated cautiously, especially in those cases where its calculated value is much less than the measured v g rate (i.e., with IIA as substrate and no inhibitor in the assay). In such situations, the activity due to class I1 and/or class I will obscure the PP reaction and magnify experimental errors. Registry No. ADH, 9031-72-5;IA, 15971-29-6;IIA, 3453-33-6; 4-MeP, 7554-65-6.
(1) Vallee, 8. L.; Bazzone, T. J. Curr. Top. B i d . Med. Res. 1983, 8 ,
219-244. (2) Mftrdh, G.; Luehr, C. A.; Vallee, B. L. f r o c . Natl. Acad. Sci. U . S . A . 1985, 82, 4979-4982. (3) Consalvi, V.; Milrdh, G.; Vallee, B. L. Biochem. Siophys. Res. Commun. 1988, 739, 1009-1016. (4) Mrdh. G.; Dingley. A. L.; Auld, D. S.; Vallee, B. L. Roc. Natl. Acad. Sci. U . S . A . 1988, 8 3 , 8908-8912. (5) McEvlly, A.; Holmquist B.: Vallee, B. L. Biochemistry 1988, 2 7 , 4284-4288. (6) Koivusalo, M.; Baumann, M.; Uotila, L. FEBS Lett. 1989, 257, 105-109. (7) Skursky, L.: Kovar, J.; Stachova, M. Anal. Biochem. 1979, 99, 65-71. (8) Agarwal, D. P.; Stapelfeldt, H.; Meier-Tackmann, D.; G o m e , H. W. Z . Anal. Chem. 1982, 371, 314. (9) Kato, S.; Ishil, H.; Kano, S.; Hagihara, S.; Todoroki, T.; Nagata, S.; Takahashl, H.: Nagasaka, M.; Sato, J.; Tsuchlya, M. Clin. Chem. 1984, 30, 1817-1820. (IO) Khayrollah, A,; ACTamer, Y. Y.; Taka, M,; Skursky, L. Ann, C/jn. Bjochem. 1982, 79, 35-42. (11) Wlerzchowskl, J.; Dafeldecker, W. P.; Holmquist, B.; Vallee, B. L. Anal. Biochem. 1989, 778, 57-62. (12) Tolf, B.-R.; Plechaczek, J.; Dahlblom, R.; Theorell, H.; Akeson, A,; Lundqulst. G. Acta Chem. Scand. B 1979, 33, 483-487. (13) McEvlly. A.; Holmquist, B.; Vallee, B. L. Siochromatography 1990, 5, 13-17. (14) Wagner, F. W.; Burger, A. R.; Vallee, B. L. Biochemistry 1983, 2 2 , 1857- 1863. (15) Fr-lich, P. Application of Luminescence Spectroscopy to Quantitative Analyses in Clinical and Biological Samples. I n Modern Fhorescence Spectroscopy; Wehry, E. L., Ed.; Plenum Press: New York. 1976; Vol. 2, pp 49-90. (16) Deetz, J. S.;Luehr, C. A.; Vallee. B. L. Biochemistry 1984, 23, 6822-6828. (17) Wolfbeis, 0.S.; Leiner, M. Anal. Chim. Acta 1983, 167, 203-215. (18) Wermuth, 8. In Enzymology of Carbonyl Metabolism; Flynn, T. G., Weiner, H., Eds.; A. R. L i s : New York, 1985; Vol. 2, pp 209-230. (19) DfilOW, c. c.; Holmquist, 8.: Morelock, M. M.; V a h , B. L. Biochemlstry 1984, 2 3 , 6363-6368. (20) Jbrnvall, H.; Hempel, J.; Vallee, B. L. Enzyme 1987, 37, 5-18. (21) Strydom, D. J.; Vallee, B. L. Anal. Biochem. 1982, 723, 422-429. (22) Skursky, L.; Khayrollah, A. Drug AlcohoiDepend. 1980, 6 . 187-190. (23) Dalziel, K. Acta Chem. Scand. 1957, 8 , 1706. (24) Bosron, W. F.; Li, T.-K. Enzyme 1987, 37, 19-28. (25) Mlrdh, G.;Vallee, B. L. Biochemistry 1888, 25. 7279-7282.
RECEIVED for review July 2, 1991. Accepted October 7,1991. This work has been supported by a grant from the Samuel Bronfman Foundation, Inc., to the Endowment for Research in Human Biology, with funds provided by Joseph H. Seagram & Sons, Inc.
