- Email: [email protected]
Studies on liver alcohol dehydrogenase complexes
ARCHIVES
OF
BIOCHEMISTRY
Studies III. Multiple
AND
BIOPHYSICS
106, 243-251 (1964)
on Liver Alcohol
Inhibition
Dehydrogenase
Complexes
Kinetics in the Presence of Two Competitive T. YONETANI
From the Nobel Medical
Institute,
AND
H. THEORELL
Department of Biochemistry,
(Received
January
Inhibitors’
Stockholm,
Sweden
28, 1964)
A graphical method was devised to analyze the multiple inhibition of an enzyme by two competitive inhibitors. This method not only distinguishes whether two inhibitors interact with the same site or different sites of the enzyme, but also gives an interaction constant (a) between two inhibitors in the enzyme-inhibitor complex. This method was applied to the multiple inhibition of liver alcohol dehydrogenase by coenzyme-competitive inhibitor pairs among o-phenanthroline, adenosine diphosphate ribose, and adenosine-5’.monoand -diphosphates. The present kinetic analysis quantitatively reconfirmed our previous conclusion that o-phenanthroline and adenosine diphosphate ribose interact independently with different sites of the enzyme.
Competitive inhibitors are especially useful for elucidating the nature of the active center of enzymes, since they act on the center selectively and directly in competition with substrates and coenzymes. A good deal of information can be obtained by using rational combinations of these inhibitors in the study of enzyme mechanism. Recently Webb (1) reviewed the multiple inhibition in general terms and formulated kinetic equations dealing with enzyme inhibition by multiple inhibitors. Although actual experiments of multiple inhibition kinetics have been carried out only for some selected cases, they have already provided useful information which could not have been obtained with a single inhibitor. In 1952 Slater and Bonner (2) kinetically analyzed an unexpectedly strong inhibition of succinic dehydrogenase which was induced by a combination of fluoride and phosphate, both of which separately ,acted as weak succinate-competitive inhibitors. Recently Yagi and Osawa (3) were able to classify a series of coenzyme-competitive inhibitors of 1 Previous (5, 6).
papers
of this series are References 243
n-amino acid oxidase as either flavin-competitive or adenine-competitive by means of multiple inhibition kinetics. Both groups of workers (2, 3) employed different kinetic methods in analyzing their results. Webb (1) recommended a different method, i.e., Loewe’s isobols (4), for expressing multiple inhibition. It was felt that a more general approach to multiple inhibition would be useful in promoting future studies on enzyme inhibition. A simple graphical method is described in this paper which expresses the relation between two competitive inhibitors acting on an enzyme and permits a direct estimation of the kinetic constants involved. This method has been applied to the multiple inhibition of LADH2 by the coenzyme-competitive inhibitors, o-phenanthroline, ADPR, AMP, and ADP (5, 6). The relation of this 2 Abbreviations used: LADH, liver alcohol dehydrogenase; NAD+ and NADH, nicotiniumamide- and dihydronicotinamide-adenine dinucleotides, respectively; ADPR, adenosine diphosphate ribose; AMP and ADP, adenosine-5’-monoand -diphosphates, respectively; and K,” = Michaelis constant.
244
YONETANI
AND THEORELL
new graphical method to those of previous workers will be discussed. EXPERIMENTAL Crystalline LADH was prepared from horse liver according to a modification of Dalziel’s method (7). The purity of the enzyme preparation was found to be lOO’% on the basis of a specific extinction coefficient at 280 rnB of 0.42 ml per milligram X cm-l and a molecular weight of 34,000 per 2 coenzyme-binding sites (8). The enzyme concentration was expressed as N, the normality of coenzyme-binding capacities per liter, and was determined spectrophotometrically by titrating LADH with NAD+ in the presence of excess pyrazole (9). NAD+ (97% purity), NADH (99yo purity), ADPR, AMP, and ADP were purchased from Sigma Chemical Co. o-Phenanthroline was obtained from Eastman Kodak Co. The concentrations of these inhibitors were determined spectrophotometrically by the use of their extinction coefficients: 15.4, 15.4, 15.4, and 31.0 mM-r X cm-r for ADPR, AMP, and ADP at 259 mN, and o-phenanthroline at 264 rnti, respectively. The concentrations of coenzymes were expressed in terms of enzymically reducible NAD+ and oxidizible NADH by the use of a difference extinction coefficient (reduced minus oxidized) at 340 rnp of 6.22 mW1 X cm-l (10). Kinetic experiments were performed at 23.5”C in sodium phosphate buffer, pH 7.0, ionic strength 0.1. Initial rate measurements were made with a recording fluorometer (11) by observing the initial rate of fluorescence increase due to the reduction of NAD+ in the presence of LADH and ethanol. The ethanol concentration was 6.2 m&f, which is about the highest that does not cause any inhibition. Reactions were initiated by rapidly mixing IO-20 pl of 16 PN LADH into 4.0 ml of reaction mixtures (the final enzyme concentration = 4W30 ma). The initial reaction rates in the presence and absence of coenzyme-competitive inhibitors were plotted according to the method described in Rationale. RATIONALE
When two substrate-competitive inhibitors (II and 1,) and a substrate (S) react with an enzyme (E), the possible interactions among the components may be written in the following scheme provided 1112, I$, I&T, and IJ&fG are not formed. Obviously EI$, EIBS, and EIJzS are not formed, since I, and Iz are substrate-competitive.
h ES B (e-a-f-e--d)~(a)-E
+ Product
where e = [Etotar], s = IS], a = [ES], b = WI,],
c =
[Elzl,
d =
[ElJzl,
in =
[II],
i2 = [1& kl, k2, and ka = rate constants, K EIl> K,,, K,, ,Iz, and Ks, ,zj = dissociation constants of enzyme-inhibitor complexes, and (Y = an interaction constant between II and 1~ in the EIIIz complex. When II and I2 interact with the same site of E, they prevent each other from binding to E, .e., EIl12 is not formed and thus LY = 00. If 1, and 12 interact with different sites of E, m > Q! > 0. Initial rate equations for this inhibited system can be formulated as follows: It is assumed that i1 , iZ I and s >> e. The Michaelis constant of S(K,) is
The steady state treatment gives da - = kde dt
of the reaction
- a - b - c - d) (2) - (kz + ks)a = 0
:. (e - a - b - c - d) = uK,/s &z,
=
(3)
(e - a - b - c - d)il h .-. b=g-
aK,,A
(4)
RI1
a - b - c - d)i2 K Big = (e I: :.
c
=
UK,
s&zz
(5) i2
STUDIES
ON LADH
COMPLEXES.
1 -=vi
K ETz.I,
cil d
=
-
.'*
d=sK
aK,iliz 812K EIz,Il
12 = KEI$EI~ $11
03)
Then
K II1 .I2 z aKBr2, when KEW~ = C&W
(9)
From Eqs. 6,7, and 9,
d= Introducing
aK, il iz saKml Km
(10)
Eqs. 4, 5, and 10 into Eq. 3,
=
i&a
=
1 -I-
kse
(Km/s)
(12) 1,
EI2
EIl
El2
Rearrangements of Eq. 13 give 1 -= vi
++g&&-) m +sv~~~Il(l+&)i~
or
Rate equations similar to Eqs. 13, 14, and 15 were originally worked out by Slater and Bonner (2). Recently Webb (1) presented a more general form of Eq. 13 which included both competitive and noncompetitive inhibitors. Equation 14 indicates that, if l/vi is plotted against ir at fixed iz , a straight line will be obtained with a slope of
{ 1 + (&‘~EI~)]&&XJL~, and an intercept at l/vi axis of
Slope with I2 = 1 + iz ~ Slope without 12 C&I2
. . l+~+-$+ox~l$ El]
1 vwt
l/Vm + j 1 + (&/KET2)}WsVm . Thus the l/ui intercept is a linear function of iz independent of the LYvalue. When (x = 00, the slope will be constant (= Km/sV,JCBII) and thus (l/vi vs. iJ-plots at different iz will become parallel straight lines (Plot C of Fig. 1). When = > LY> 0, both slope and intercept of the (l/ui vs. &)-plots will increase as linear functions of 6 (Plot E of Fig. 1). crK,,, value will be calculated from slopes in the presence and absence of I2 according to Eq. 16:
The rate of the reaction (vi) is vi
245
(7)
Equations 6 and 7 give K&k,
III
(14)
(16)
The intersection of straight lines in the presence and absence of I2 will occur at an abscissa value of - aKEI1, because, according to Eq. 15, {l + (i~/cxK~l~)) X &. = (1 -t (il/~KEII)} X 0 = 0, at the intersection and thus il = -cuK,,, . Similar plots of (l/v; US. iz) in the presence and absence of iI will give Plots D or F of Fig. 1 depending upon the LY vaIue. When c~3> a! > 0, aKEI, and aKEr2 can be calcuIated from Plot F. Thus a combination of Plots E and F will give doublechecked values of &CEIl and CYK,,, . KEIl and PLI~ can be determined independently by either Dixon plot (12) (Plots A and B of Fig. 1) of Lineweaver-Burk plot (13) with a single inhibitor. The (Y value can be readily calculated from KBrl, K,, , c&x, , and aKxI, .
246
YONETANI
AND
When the interactions of II and Iz with independent of each other, a: = 1. When a positive attraction occurs between I1 and Iz in the EIlIs complex, 1 > (Y > 0. When I, and Tz interact repulsively in the EIIIz complex, 00 > LY> 1. In the present experiments with LADH, vo/vi instead of l/vi was plotted against il or iz, where v. = the initial rate in the absence of inhibitors: v. will be constant at a constant s. This modification was designed to permit a comparison of the present results to be made with those of Yagi-Ozawa plots (3) and to be used in the construction of Loewe-Webb isobograms (1, 4). We found, in agreement with earlier experiments (5, 14), the following values of some kinetic constants under these experimental conditions: sodium phosphate buffer, pH 7.0, ionic strength 0.1; 6.2 mM ethanol; and 23.5”:
E are strictly
V,/e = the maximal turnover
number per center = 2.95 X see-’ = 2.32 X see-’ at 50 p&l NAD+ vole and
K, of NAD+ = 14 /.LM RESULTS
ADPR (5, 15-17) and o-phenanthroline (6, 18-22) are known to inhibit LADH in competition with the coenzyme. ADP and AMP, structurally simpler analogs of ADPR, are less inhibitory than ADPR, but their effects are still strictly competitive with the coenzyme. Initial rates were measured at pH 7.0 at two fixed concentrations of NAD+ (10 and 50 PM) with 6.2 mM ethanol in the presence and absence of varied concentrations of each of these inhibitors and were plotted according to Dixon (12) to determine Kal of each inhibitor (cf. Plots A and B of Fig. 1). K,, values are given in Table I. The mutual independence of the LADHinteractions with ADPR and o-phenanthroline was previously demonstrated by equilibrium-competition experiments as well as by the crystallization of a mosaic complex of phenanthroline-enzyme-ADPR (5, 6). Judging from their structural analogy
THEORELL
to ADPR, it is most likely that ADP and AMP interact with LADH independently of the phenanthroline-binding site or the Zn site of the enzyme, and that ADP, AMP, and ADPR compete with each other for the same site or the ADPR-binding site of the enzyme. In order to prove this kinetically, multiple inhibitions of LADH by each pair of these inhibitors were graphically analyzed according to the present procedure (cf. Plots C, D, E, and F of Fig. 1). Plots with the ADPR-ADP and o-phenanthroline-ADP pairs are shown in Figs. 2 and 3, respectively. Types C or D plots were obtained with the ADPR-ADP pair, indicating o( = 00, while Types E or F plots were obtained with the phenanthrolineADPR pair, indicating 00 > (Y > 0. Determined (Y values for each pair among ADPR, ADP, AMP, and o-phenanthroline are given in Table II. Isobols are curves showing equi-eff ective combinations of active substances plotted on graphs whose coordinates are the concentrations of the substances (1). Loewe (4) used such curves to demonstrate the nature of the interactions of drug pairs on tissues. Webb (1) recommended the use of such curves to demonstrate the multiple inhibition of an enzyme. Isobograms of the ADPR-ADP and o-phenanthroline-ADP pairs were constructed from the data of Figs. 2 and 3, and are shown in Fig. 4. When the inhibitors prevented each other’s binding (a = x ), the isobols were linear (cf. the ADPR-ADP pair of Fig. 4). If the inhibitors interacted with different sites on the enzyme (a, > (Y > 0), the isobols became upward-concave (cf. the phenanthroline-ADP pair of Fig. 4). The degree of the concavity depended upon the (Y value: with decreasing a! values, the curves became more concave upwards. DISCUSSION
MULTIPLE
INHIBITION
KINETICS
OF LADH
LADH is not an ideal enzyme to demonstrate multiple inhibition kinetics, since it requires a substrate and a coenzyme (ethanol and NAD+ or acetaldehyde and NADH), and thus a rate equation more complex than
STUDIES
Dixon
ON LADH
III
-KEIz
i2
D
s=constont
Plot
-KEI,
Yonetani-Theorell
it
Plot
C a=a
247
COMPLEXES.
L4 I v;
- a KEI,
in=i,’
I
if ii
Vi
i2=0
i, =O 34
i2
-ah2
if
.1 '2 aKEl = 2 slopetyri$) slope(l2=0)
,
i2 aK4=
'I slope(il=i{I slope(il=O)
_ ,
FIG. 1. Graphical methods to determine competitive-inhibitor constants (Kell and KEr2) and an interaction constant (a) between two competitive inhibitors (II and 12) in an enzyme-inhibitor complex (EZ,Z,). Plots A and B, Dixon procedure with a single inhibitor (12). Plots C, D, E and F, the present procedure with two inhibitors.
that discussed in this paper is required for the complete analysis of its multiple inhibition. However, LADH has a number of coenzyme-competitive inhibitors which are fully reversible and whose actions are well defined. Furthermore, when the substrate concentration is kept at a nonrate-limiting value and the coenzyme concentration is kept at sufficiently low values, the initial rate of LADH reactions can be considered to be controlled by the coenzyme concentration only. Under such conditions, LADH may be treated as a single-substrate enzyme, and Eqs. 13, 14, and 15 can be applied to LADH reactions. The validity of this approximation in the study of the coenzymecompetitive inhibition of LADH was proved by the fact that coenzyme-competitive inhibitor-constants of o-phenanthroline and
TABLE
I
DISSOCIATION CONSTANTS COENZYME~OMPETITIVE
LADH Inhibitor
o-phenanthroline ADPR ADP AMP
AT pH
(KEl) OF SOME INHIBITORS OF
7.0
~KEI
bM)
9 26” 390 140
a This value is somewhat higher than those reported previously (= 16 - 20 PM) (5), but agrees with our more recent KxI value of the LADH-ADPR complex, which was determined spectrophotometrically on the basis of absorn5 tion changes due to their interaction (23).
ADPR determined kinetically at nonratelimiting concentrations of substrates (at
248
YONETANI
AND
THEORELL
i2 (mM I 2.5 ::8
I 0
I 95
I 190
I 285
il
I 390
I 475
I 0
I 0.5
(FM)
I 1.0
I 1.5
i2
I 2.0
I 2.5
(mM1
FIG. 2. Multiple inhibition of LADH by ADPR and ADP plotted according to the present procedure. The plots are of Types C and D (of Fig. 1). Thus a = m , indicating that two inhibitors compete for the same site of the enzyme. Yagi-Ozawa plot (3) is indicated with dotted lines (See Discussion).
i, =[o-Phenanthroline] i2=
iI1 ,‘@
F
i2 (mM'
E
i, =[o-Phenanthroline]
[ADP]
ip+
[ADP]
s = 50pM NAD+
s = 50pM NAD+ 15
--z
IO J!?
IO !!?
“i
“i
! P
43.5
I/
34.8
Ei-
I
I
0
&, FIG.
to the cating lated cated
8.7
I
I
I
I
124 261 348 43.5 ii (PM)
L
&lp
i2 (mM)
3. Multiple inhibition of LADH by o-phenanthroline and ADP plotted according present procedure. The plots are of Types E and F (of Fig. 1). Thus 00 > 01 > 0, indithat two inhibitors interact with different sites of the enzyme. The o! value is calcuto be 0.5 (a positive attraction between two inhibitors). Yagi-Ozawa plot (3) is indiwith dotted lines (See Discussion).
6.2 mM ethanol or 1.1 mM acetaldehyde) agreed well with dissociation constants of these inhibitors which were determined directly (5, 6). Thus, it is justified to apply Eqs. 14 and 15 to the multiple inhibition
of LADH by two coenzyme-competitive inhibitors, if initial rates are measured at a nonrate-limiting concentration of ethanol (6.2 mM) and a sufficiently low concentration of NAD+ (50 PM). In fact, the experi-
249
STUDIES ON LADH COMPLEXES. III
mentalpoints shownin Figs. 2 and 3 obey 2 and 3 with dottedlines,Yagi-Ozawaplots Eqs. 14 and 15 well, though a fairly wide range of initial rates were determined. These results support the above-mentioned approximation. PLOTTING
NETHODS OF MULTIPLE INHIBITION
Previously Yagi and Ozawa (3) employed the vO/vi as. (il + &) plot to identify the interaction type of a given pair of competitive inhibitors. When vo/vi is plotted against (6 + &), a straight line and a second-order curve are obtained with CY= cc1and co > (Y > 0, respectively. As indicated in Figs. TABLE 01 VALUES FOR THE COENZYME-COMPETITIVE Inhibitor
II
LADH
INHIBITION INHIBITOR PAIRS
pair
Type of plota
o-phenanthroline-ADPR o-phenanthroline-ADP o-phenanthroline-AMP ADPR-ADP ADPR-AMP ADP-AMP
E and EandF E and C and C and C and
o cf. Fig. 1 for types
F F D D D
BY
(I Value
1.0 0.5 0.3 m 00 m
of plots,
are special examples of our present plots. Recently Webb (1) suggested a possible application of Loewe’s isobols to the analysis of multiple inhibition kinetics. As shown in Figs. 24, both Yagi-Ozawa (3) and LoeweWebb (1, 4) plots are simple and clear methods to demonstrate the interaction type of a given pair of inhibitors, but they fail to give the (Y value directly. The (Y value has to be calculated indirectly from rate equations such as Eqs. 13, 14, and 15 by substituting Uo/‘Ui,K,, V,, s, K,, , and values. K ET2 with known experimental Slater and Bonner (2) and Yagi and Ozawa (3) calculated their cy values in this manner. If the present plotting method is cornbined with Dixon’s (la), K,,, , K,,, , OK,,, , and aKe,, are graphically determined. Thus the evaluation of the a! value is straightforward and may be more reliable. A summary of these graphical methods is given in Table III. Previously the mutual independence of the LADH-interactions with ADPR and o-phenanthroline was demonstrated directly by an equilibrium experiment as well as by the crystallization of the complex, phenan-
l-----J i, = [O-Phenonthroline] i2 [ADP] s = 50 /LM NAD+
’
’
q
l
o
0
95
190
4
285
380
( /LLMADPR 1
=from Plot E =from Plot F
475
4
CpMo-Phenanthroline)
FIG. 4. Isobograms of multiple inhibition of LADH by ADPR-ADP and o-phenanthroline-ADP pairs. Loewe’s isobograms (4) are constructed from the data of Figs. 2 and 3. Isobols of relative activity (vi/v,) = 0.2, 0.3, 0.4, 0.5, and 0.6 are shown. The ADPR-ADP pair shows linear isobols, indicating that two inhibitors compete for the same site of the enzyme (01 = m), while the o-phenanthroline-ADP pair exhibits upward-concave curves (m > 01 > 0), indicating that two inhibitors interact with different sites of the enzyme.
250
YONETANI
AND THEORELL
TABLE III A COMPARISON OF PLOTTING METHODS FOR ENZYME INHIBITION
I
Plotting
-
BY Two COMPETITIVE-INHIBITORS
Results Evaluation of the LI value
T
_VO vs. Vi
(it
+
iz)
il vs. i2 at con-
IStraight line
Straight lines
vs. il at fixed it
Vi
or VO - vs. in at fixed il
Yagi and Ozawa (3)
Second-order curve Indirect calculation from rate-equations
stant 2 VO -VO
Investigators (References)
Parallel straight lines
Loewe (4) Webb (1)
Outward-concave curves
Straight lines with different slopes
Vi
Direct calculation from &I, , &I~ , ~K.EI, and (YKEI*
Yonetani and Theorell (this paw
-
throline-enzyme-ADPR (5, 6). This was possible because both inhibitors have relatively small KEI’s (5, 6, 23) and the interaction of the LADH-bound Zn with o-phenanthroline was directly monitored by specific absorption changes appearing as a result of their combination (5, 6, 24). Such fortunate cases as these are not always available for competitive inhibition studies. Then multiple inhibition kinetics becomes the only way to analyze the interaction of an enzyme with multiple inhibitors. INTERACTION
CONSTANT,
LY
The determined cy value of unity for the phenanthroline-ADPR pair (cf. Table II) quantitatively confirms our previous conclusion that the interactions of LADH with these inhibitors are independent of one another (5, 6). When 00 > o( > 0 and CY# 1, positive (a < 1) or negative (0~ > 1) interactions exist between II and I2 in the EIJz complex. A number of causes such as ion-dipole, interdipole, hydrophobic, and interionic! hydrophilic interactions as well as simple steric hindrance and protein-conformational changes may be involved. CYValues of 0.0034 for the succinic dehydrogenase inhibition by the fluoride-
phosphate pair and 0.2 for the n-amino acid oxidase inhibition by the riboflavin5’-sulfate-adenosine-5’-sulfate pair were reported by Slater and Bonner (2) and Yagi and Ozawa (3), respectively. In the present experiments, OLvalues of 0.5 and 0.3 were obtained for the phenanthroline-ADP and phenanthroline-AMP pairs, respectively. At present it remains unexplained why such positive interactions occur between these pairs of inhibitors in the EIIIf complexes, but not between phenanthroline and ADPR. The lack of interaction in the last case may be an occasional result of cancelling positive and negative interactions. The determination of a: values for pairs phenanthroline-derivatives and between ADPR-derivatives may give useful information concerning the mechanism of these interactions as well as the distance between the Zn site and the ADPR-binding site of LADH (5, 6). It should be pointed out that the CYvalue is closely related to the “stabilization” (= (Y) factors (= l/a!) or “destabilization” operating between the coenzyme and inhibitor in ternary complexes of the enzyme (14). For example, in the LADH-NADHisobutyramide complex the following relation was found:
ONLADHCOMPLEXES. III STUDIES K,, K,.
N.4DH/Kb-isobutyrarnid~, NADH = isobutyramide/KE-NADH, isobutyramide
= 60 zz
the stabilization factor = l/a!. Thus CY= 0.016. ACKNOWLEDGMENTS The authors are indebted for financial support given by Statens Medicinska Forskningr%d, Knut och Alice Wallenbergs Stiftelse, Svenska Maltdrycksforskningsinstitutet, and the Rockefeller Foundation. The skillful technical assistance of Miss Gisela Schreck and the constructive suggestions of Dr. George Waller are gratefully appreciated. REFERENCES 1. WEBB, J. L., “Enzyme and Metabolic Inhibitors.” Academic Press, New York, 1963. 2. SLATER, E. C., AND BONNER, W. D., JR., Biochem. J. 62, 185 (1952). 3. YAGI, K., AND OZAWA, T., Biochim. Biophys. Acta 42, 381 (1960). 4. LOEWE, S., Pharmacol. Rev. 9, 237 (1957). 5. YONETANI, T., Acta Chem. Stand. 17, Suppl.
1, 96 (1963). 6. YONETANI, T., Biochem. 2. 338, 300 (1963). 7. DALZIEL, K., Acta Chem. &and. 12,459 (1958).
251
8. EHRENBERG, A., AND DALZIEL, K., Acta Chem. &and. 12, 465 (1958). 9. THEORELL, H., AND YONETANI, T., Biochem. Z. 338, 537 (1963). 10. HORECKER, B. L., AND KORNBERG, A., J. Biol. Chew 176, 385 (1948). 11. THEORELL, H., PYTYGAARD, A. P., AND 9, BONNICHSEN, R., Acta Chem. &and.
1148 (1955). 12. DIXON, M., Biochem. J. 66, 170 (1953). 13. LINEWEAVER, H., AND BURB, D., J. Ajn. Chem. Sot. 66, 658 (1934). 14. THEORELL, H., AND MCKINLEY MCKEE, J. S., Acta Chem. Stand. 16, 1797 (1961). 15. DALZIEL, K., Nature 191, 1098 (1961). 16. DALZIEL, K., Biochem. J. 84, 240 (1962). 17. DALZIEL, K., J. Biol. Chem. 237, 1538 (1963). 18. WALLENFELS, K., AND SUND, H., Biochem. 2. 329, 48 (1957). 19. VALLEE, B. L., WILLIAMS, R. J. P., AND HOCH, F. L., J. Biol. Chem. 234,262l (1959). 20. PLANE, R. A., AND THEORELL, H., Acta Chem. Stand. 16, 1866 (1961). 21. MAHLER, H. R., BAKER, R. H., JR., AND SINER, V. J., JR., Biochemistry 1, 47 (1962). 22. DALZIEL, K., Nature 197, 462 (1963). 23. THEORELL, H., AND YONETANI, T., Arch. Riothem. Biophys. 106, 252 (1964). 24. VALLEE, B. L., COOMBS, T. L., AND WILLIAMS, R. J. P., J. Am. Chem. Sot. 80, 397 (1958).
OF
BIOCHEMISTRY
Studies III. Multiple
AND
BIOPHYSICS
106, 243-251 (1964)
on Liver Alcohol
Inhibition
Dehydrogenase
Complexes
Kinetics in the Presence of Two Competitive T. YONETANI
From the Nobel Medical
Institute,
AND
H. THEORELL
Department of Biochemistry,
(Received
January
Inhibitors’
Stockholm,
Sweden
28, 1964)
A graphical method was devised to analyze the multiple inhibition of an enzyme by two competitive inhibitors. This method not only distinguishes whether two inhibitors interact with the same site or different sites of the enzyme, but also gives an interaction constant (a) between two inhibitors in the enzyme-inhibitor complex. This method was applied to the multiple inhibition of liver alcohol dehydrogenase by coenzyme-competitive inhibitor pairs among o-phenanthroline, adenosine diphosphate ribose, and adenosine-5’.monoand -diphosphates. The present kinetic analysis quantitatively reconfirmed our previous conclusion that o-phenanthroline and adenosine diphosphate ribose interact independently with different sites of the enzyme.
Competitive inhibitors are especially useful for elucidating the nature of the active center of enzymes, since they act on the center selectively and directly in competition with substrates and coenzymes. A good deal of information can be obtained by using rational combinations of these inhibitors in the study of enzyme mechanism. Recently Webb (1) reviewed the multiple inhibition in general terms and formulated kinetic equations dealing with enzyme inhibition by multiple inhibitors. Although actual experiments of multiple inhibition kinetics have been carried out only for some selected cases, they have already provided useful information which could not have been obtained with a single inhibitor. In 1952 Slater and Bonner (2) kinetically analyzed an unexpectedly strong inhibition of succinic dehydrogenase which was induced by a combination of fluoride and phosphate, both of which separately ,acted as weak succinate-competitive inhibitors. Recently Yagi and Osawa (3) were able to classify a series of coenzyme-competitive inhibitors of 1 Previous (5, 6).
papers
of this series are References 243
n-amino acid oxidase as either flavin-competitive or adenine-competitive by means of multiple inhibition kinetics. Both groups of workers (2, 3) employed different kinetic methods in analyzing their results. Webb (1) recommended a different method, i.e., Loewe’s isobols (4), for expressing multiple inhibition. It was felt that a more general approach to multiple inhibition would be useful in promoting future studies on enzyme inhibition. A simple graphical method is described in this paper which expresses the relation between two competitive inhibitors acting on an enzyme and permits a direct estimation of the kinetic constants involved. This method has been applied to the multiple inhibition of LADH2 by the coenzyme-competitive inhibitors, o-phenanthroline, ADPR, AMP, and ADP (5, 6). The relation of this 2 Abbreviations used: LADH, liver alcohol dehydrogenase; NAD+ and NADH, nicotiniumamide- and dihydronicotinamide-adenine dinucleotides, respectively; ADPR, adenosine diphosphate ribose; AMP and ADP, adenosine-5’-monoand -diphosphates, respectively; and K,” = Michaelis constant.
244
YONETANI
AND THEORELL
new graphical method to those of previous workers will be discussed. EXPERIMENTAL Crystalline LADH was prepared from horse liver according to a modification of Dalziel’s method (7). The purity of the enzyme preparation was found to be lOO’% on the basis of a specific extinction coefficient at 280 rnB of 0.42 ml per milligram X cm-l and a molecular weight of 34,000 per 2 coenzyme-binding sites (8). The enzyme concentration was expressed as N, the normality of coenzyme-binding capacities per liter, and was determined spectrophotometrically by titrating LADH with NAD+ in the presence of excess pyrazole (9). NAD+ (97% purity), NADH (99yo purity), ADPR, AMP, and ADP were purchased from Sigma Chemical Co. o-Phenanthroline was obtained from Eastman Kodak Co. The concentrations of these inhibitors were determined spectrophotometrically by the use of their extinction coefficients: 15.4, 15.4, 15.4, and 31.0 mM-r X cm-r for ADPR, AMP, and ADP at 259 mN, and o-phenanthroline at 264 rnti, respectively. The concentrations of coenzymes were expressed in terms of enzymically reducible NAD+ and oxidizible NADH by the use of a difference extinction coefficient (reduced minus oxidized) at 340 rnp of 6.22 mW1 X cm-l (10). Kinetic experiments were performed at 23.5”C in sodium phosphate buffer, pH 7.0, ionic strength 0.1. Initial rate measurements were made with a recording fluorometer (11) by observing the initial rate of fluorescence increase due to the reduction of NAD+ in the presence of LADH and ethanol. The ethanol concentration was 6.2 m&f, which is about the highest that does not cause any inhibition. Reactions were initiated by rapidly mixing IO-20 pl of 16 PN LADH into 4.0 ml of reaction mixtures (the final enzyme concentration = 4W30 ma). The initial reaction rates in the presence and absence of coenzyme-competitive inhibitors were plotted according to the method described in Rationale. RATIONALE
When two substrate-competitive inhibitors (II and 1,) and a substrate (S) react with an enzyme (E), the possible interactions among the components may be written in the following scheme provided 1112, I$, I&T, and IJ&fG are not formed. Obviously EI$, EIBS, and EIJzS are not formed, since I, and Iz are substrate-competitive.
h ES B (e-a-f-e--d)~(a)-E
+ Product
where e = [Etotar], s = IS], a = [ES], b = WI,],
c =
[Elzl,
d =
[ElJzl,
in =
[II],
i2 = [1& kl, k2, and ka = rate constants, K EIl> K,,, K,, ,Iz, and Ks, ,zj = dissociation constants of enzyme-inhibitor complexes, and (Y = an interaction constant between II and 1~ in the EIIIz complex. When II and I2 interact with the same site of E, they prevent each other from binding to E, .e., EIl12 is not formed and thus LY = 00. If 1, and 12 interact with different sites of E, m > Q! > 0. Initial rate equations for this inhibited system can be formulated as follows: It is assumed that i1 , iZ I and s >> e. The Michaelis constant of S(K,) is
The steady state treatment gives da - = kde dt
of the reaction
- a - b - c - d) (2) - (kz + ks)a = 0
:. (e - a - b - c - d) = uK,/s &z,
=
(3)
(e - a - b - c - d)il h .-. b=g-
aK,,A
(4)
RI1
a - b - c - d)i2 K Big = (e I: :.
c
=
UK,
s&zz
(5) i2
STUDIES
ON LADH
COMPLEXES.
1 -=vi
K ETz.I,
cil d
=
-
.'*
d=sK
aK,iliz 812K EIz,Il
12 = KEI$EI~ $11
03)
Then
K II1 .I2 z aKBr2, when KEW~ = C&W
(9)
From Eqs. 6,7, and 9,
d= Introducing
aK, il iz saKml Km
(10)
Eqs. 4, 5, and 10 into Eq. 3,
=
i&a
=
1 -I-
kse
(Km/s)
(12) 1,
EI2
EIl
El2
Rearrangements of Eq. 13 give 1 -= vi
++g&&-) m +sv~~~Il(l+&)i~
or
Rate equations similar to Eqs. 13, 14, and 15 were originally worked out by Slater and Bonner (2). Recently Webb (1) presented a more general form of Eq. 13 which included both competitive and noncompetitive inhibitors. Equation 14 indicates that, if l/vi is plotted against ir at fixed iz , a straight line will be obtained with a slope of
{ 1 + (&‘~EI~)]&&XJL~, and an intercept at l/vi axis of
Slope with I2 = 1 + iz ~ Slope without 12 C&I2
. . l+~+-$+ox~l$ El]
1 vwt
l/Vm + j 1 + (&/KET2)}WsVm . Thus the l/ui intercept is a linear function of iz independent of the LYvalue. When (x = 00, the slope will be constant (= Km/sV,JCBII) and thus (l/vi vs. iJ-plots at different iz will become parallel straight lines (Plot C of Fig. 1). When = > LY> 0, both slope and intercept of the (l/ui vs. &)-plots will increase as linear functions of 6 (Plot E of Fig. 1). crK,,, value will be calculated from slopes in the presence and absence of I2 according to Eq. 16:
The rate of the reaction (vi) is vi
245
(7)
Equations 6 and 7 give K&k,
III
(14)
(16)
The intersection of straight lines in the presence and absence of I2 will occur at an abscissa value of - aKEI1, because, according to Eq. 15, {l + (i~/cxK~l~)) X &. = (1 -t (il/~KEII)} X 0 = 0, at the intersection and thus il = -cuK,,, . Similar plots of (l/v; US. iz) in the presence and absence of iI will give Plots D or F of Fig. 1 depending upon the LY vaIue. When c~3> a! > 0, aKEI, and aKEr2 can be calcuIated from Plot F. Thus a combination of Plots E and F will give doublechecked values of &CEIl and CYK,,, . KEIl and PLI~ can be determined independently by either Dixon plot (12) (Plots A and B of Fig. 1) of Lineweaver-Burk plot (13) with a single inhibitor. The (Y value can be readily calculated from KBrl, K,, , c&x, , and aKxI, .
246
YONETANI
AND
When the interactions of II and Iz with independent of each other, a: = 1. When a positive attraction occurs between I1 and Iz in the EIlIs complex, 1 > (Y > 0. When I, and Tz interact repulsively in the EIIIz complex, 00 > LY> 1. In the present experiments with LADH, vo/vi instead of l/vi was plotted against il or iz, where v. = the initial rate in the absence of inhibitors: v. will be constant at a constant s. This modification was designed to permit a comparison of the present results to be made with those of Yagi-Ozawa plots (3) and to be used in the construction of Loewe-Webb isobograms (1, 4). We found, in agreement with earlier experiments (5, 14), the following values of some kinetic constants under these experimental conditions: sodium phosphate buffer, pH 7.0, ionic strength 0.1; 6.2 mM ethanol; and 23.5”:
E are strictly
V,/e = the maximal turnover
number per center = 2.95 X see-’ = 2.32 X see-’ at 50 p&l NAD+ vole and
K, of NAD+ = 14 /.LM RESULTS
ADPR (5, 15-17) and o-phenanthroline (6, 18-22) are known to inhibit LADH in competition with the coenzyme. ADP and AMP, structurally simpler analogs of ADPR, are less inhibitory than ADPR, but their effects are still strictly competitive with the coenzyme. Initial rates were measured at pH 7.0 at two fixed concentrations of NAD+ (10 and 50 PM) with 6.2 mM ethanol in the presence and absence of varied concentrations of each of these inhibitors and were plotted according to Dixon (12) to determine Kal of each inhibitor (cf. Plots A and B of Fig. 1). K,, values are given in Table I. The mutual independence of the LADHinteractions with ADPR and o-phenanthroline was previously demonstrated by equilibrium-competition experiments as well as by the crystallization of a mosaic complex of phenanthroline-enzyme-ADPR (5, 6). Judging from their structural analogy
THEORELL
to ADPR, it is most likely that ADP and AMP interact with LADH independently of the phenanthroline-binding site or the Zn site of the enzyme, and that ADP, AMP, and ADPR compete with each other for the same site or the ADPR-binding site of the enzyme. In order to prove this kinetically, multiple inhibitions of LADH by each pair of these inhibitors were graphically analyzed according to the present procedure (cf. Plots C, D, E, and F of Fig. 1). Plots with the ADPR-ADP and o-phenanthroline-ADP pairs are shown in Figs. 2 and 3, respectively. Types C or D plots were obtained with the ADPR-ADP pair, indicating o( = 00, while Types E or F plots were obtained with the phenanthrolineADPR pair, indicating 00 > (Y > 0. Determined (Y values for each pair among ADPR, ADP, AMP, and o-phenanthroline are given in Table II. Isobols are curves showing equi-eff ective combinations of active substances plotted on graphs whose coordinates are the concentrations of the substances (1). Loewe (4) used such curves to demonstrate the nature of the interactions of drug pairs on tissues. Webb (1) recommended the use of such curves to demonstrate the multiple inhibition of an enzyme. Isobograms of the ADPR-ADP and o-phenanthroline-ADP pairs were constructed from the data of Figs. 2 and 3, and are shown in Fig. 4. When the inhibitors prevented each other’s binding (a = x ), the isobols were linear (cf. the ADPR-ADP pair of Fig. 4). If the inhibitors interacted with different sites on the enzyme (a, > (Y > 0), the isobols became upward-concave (cf. the phenanthroline-ADP pair of Fig. 4). The degree of the concavity depended upon the (Y value: with decreasing a! values, the curves became more concave upwards. DISCUSSION
MULTIPLE
INHIBITION
KINETICS
OF LADH
LADH is not an ideal enzyme to demonstrate multiple inhibition kinetics, since it requires a substrate and a coenzyme (ethanol and NAD+ or acetaldehyde and NADH), and thus a rate equation more complex than
STUDIES
Dixon
ON LADH
III
-KEIz
i2
D
s=constont
Plot
-KEI,
Yonetani-Theorell
it
Plot
C a=a
247
COMPLEXES.
L4 I v;
- a KEI,
in=i,’
I
if ii
Vi
i2=0
i, =O 34
i2
-ah2
if
.1 '2 aKEl = 2 slopetyri$) slope(l2=0)
,
i2 aK4=
'I slope(il=i{I slope(il=O)
_ ,
FIG. 1. Graphical methods to determine competitive-inhibitor constants (Kell and KEr2) and an interaction constant (a) between two competitive inhibitors (II and 12) in an enzyme-inhibitor complex (EZ,Z,). Plots A and B, Dixon procedure with a single inhibitor (12). Plots C, D, E and F, the present procedure with two inhibitors.
that discussed in this paper is required for the complete analysis of its multiple inhibition. However, LADH has a number of coenzyme-competitive inhibitors which are fully reversible and whose actions are well defined. Furthermore, when the substrate concentration is kept at a nonrate-limiting value and the coenzyme concentration is kept at sufficiently low values, the initial rate of LADH reactions can be considered to be controlled by the coenzyme concentration only. Under such conditions, LADH may be treated as a single-substrate enzyme, and Eqs. 13, 14, and 15 can be applied to LADH reactions. The validity of this approximation in the study of the coenzymecompetitive inhibition of LADH was proved by the fact that coenzyme-competitive inhibitor-constants of o-phenanthroline and
TABLE
I
DISSOCIATION CONSTANTS COENZYME~OMPETITIVE
LADH Inhibitor
o-phenanthroline ADPR ADP AMP
AT pH
(KEl) OF SOME INHIBITORS OF
7.0
~KEI
bM)
9 26” 390 140
a This value is somewhat higher than those reported previously (= 16 - 20 PM) (5), but agrees with our more recent KxI value of the LADH-ADPR complex, which was determined spectrophotometrically on the basis of absorn5 tion changes due to their interaction (23).
ADPR determined kinetically at nonratelimiting concentrations of substrates (at
248
YONETANI
AND
THEORELL
i2 (mM I 2.5 ::8
I 0
I 95
I 190
I 285
il
I 390
I 475
I 0
I 0.5
(FM)
I 1.0
I 1.5
i2
I 2.0
I 2.5
(mM1
FIG. 2. Multiple inhibition of LADH by ADPR and ADP plotted according to the present procedure. The plots are of Types C and D (of Fig. 1). Thus a = m , indicating that two inhibitors compete for the same site of the enzyme. Yagi-Ozawa plot (3) is indicated with dotted lines (See Discussion).
i, =[o-Phenanthroline] i2=
iI1 ,‘@
F
i2 (mM'
E
i, =[o-Phenanthroline]
[ADP]
ip+
[ADP]
s = 50pM NAD+
s = 50pM NAD+ 15
--z
IO J!?
IO !!?
“i
“i
! P
43.5
I/
34.8
Ei-
I
I
0
&, FIG.
to the cating lated cated
8.7
I
I
I
I
124 261 348 43.5 ii (PM)
L
&lp
i2 (mM)
3. Multiple inhibition of LADH by o-phenanthroline and ADP plotted according present procedure. The plots are of Types E and F (of Fig. 1). Thus 00 > 01 > 0, indithat two inhibitors interact with different sites of the enzyme. The o! value is calcuto be 0.5 (a positive attraction between two inhibitors). Yagi-Ozawa plot (3) is indiwith dotted lines (See Discussion).
6.2 mM ethanol or 1.1 mM acetaldehyde) agreed well with dissociation constants of these inhibitors which were determined directly (5, 6). Thus, it is justified to apply Eqs. 14 and 15 to the multiple inhibition
of LADH by two coenzyme-competitive inhibitors, if initial rates are measured at a nonrate-limiting concentration of ethanol (6.2 mM) and a sufficiently low concentration of NAD+ (50 PM). In fact, the experi-
249
STUDIES ON LADH COMPLEXES. III
mentalpoints shownin Figs. 2 and 3 obey 2 and 3 with dottedlines,Yagi-Ozawaplots Eqs. 14 and 15 well, though a fairly wide range of initial rates were determined. These results support the above-mentioned approximation. PLOTTING
NETHODS OF MULTIPLE INHIBITION
Previously Yagi and Ozawa (3) employed the vO/vi as. (il + &) plot to identify the interaction type of a given pair of competitive inhibitors. When vo/vi is plotted against (6 + &), a straight line and a second-order curve are obtained with CY= cc1and co > (Y > 0, respectively. As indicated in Figs. TABLE 01 VALUES FOR THE COENZYME-COMPETITIVE Inhibitor
II
LADH
INHIBITION INHIBITOR PAIRS
pair
Type of plota
o-phenanthroline-ADPR o-phenanthroline-ADP o-phenanthroline-AMP ADPR-ADP ADPR-AMP ADP-AMP
E and EandF E and C and C and C and
o cf. Fig. 1 for types
F F D D D
BY
(I Value
1.0 0.5 0.3 m 00 m
of plots,
are special examples of our present plots. Recently Webb (1) suggested a possible application of Loewe’s isobols to the analysis of multiple inhibition kinetics. As shown in Figs. 24, both Yagi-Ozawa (3) and LoeweWebb (1, 4) plots are simple and clear methods to demonstrate the interaction type of a given pair of inhibitors, but they fail to give the (Y value directly. The (Y value has to be calculated indirectly from rate equations such as Eqs. 13, 14, and 15 by substituting Uo/‘Ui,K,, V,, s, K,, , and values. K ET2 with known experimental Slater and Bonner (2) and Yagi and Ozawa (3) calculated their cy values in this manner. If the present plotting method is cornbined with Dixon’s (la), K,,, , K,,, , OK,,, , and aKe,, are graphically determined. Thus the evaluation of the a! value is straightforward and may be more reliable. A summary of these graphical methods is given in Table III. Previously the mutual independence of the LADH-interactions with ADPR and o-phenanthroline was demonstrated directly by an equilibrium experiment as well as by the crystallization of the complex, phenan-
l-----J i, = [O-Phenonthroline] i2 [ADP] s = 50 /LM NAD+
’
’
q
l
o
0
95
190
4
285
380
( /LLMADPR 1
=from Plot E =from Plot F
475
4
CpMo-Phenanthroline)
FIG. 4. Isobograms of multiple inhibition of LADH by ADPR-ADP and o-phenanthroline-ADP pairs. Loewe’s isobograms (4) are constructed from the data of Figs. 2 and 3. Isobols of relative activity (vi/v,) = 0.2, 0.3, 0.4, 0.5, and 0.6 are shown. The ADPR-ADP pair shows linear isobols, indicating that two inhibitors compete for the same site of the enzyme (01 = m), while the o-phenanthroline-ADP pair exhibits upward-concave curves (m > 01 > 0), indicating that two inhibitors interact with different sites of the enzyme.
250
YONETANI
AND THEORELL
TABLE III A COMPARISON OF PLOTTING METHODS FOR ENZYME INHIBITION
I
Plotting
-
BY Two COMPETITIVE-INHIBITORS
Results Evaluation of the LI value
T
_VO vs. Vi
(it
+
iz)
il vs. i2 at con-
IStraight line
Straight lines
vs. il at fixed it
Vi
or VO - vs. in at fixed il
Yagi and Ozawa (3)
Second-order curve Indirect calculation from rate-equations
stant 2 VO -VO
Investigators (References)
Parallel straight lines
Loewe (4) Webb (1)
Outward-concave curves
Straight lines with different slopes
Vi
Direct calculation from &I, , &I~ , ~K.EI, and (YKEI*
Yonetani and Theorell (this paw
-
throline-enzyme-ADPR (5, 6). This was possible because both inhibitors have relatively small KEI’s (5, 6, 23) and the interaction of the LADH-bound Zn with o-phenanthroline was directly monitored by specific absorption changes appearing as a result of their combination (5, 6, 24). Such fortunate cases as these are not always available for competitive inhibition studies. Then multiple inhibition kinetics becomes the only way to analyze the interaction of an enzyme with multiple inhibitors. INTERACTION
CONSTANT,
LY
The determined cy value of unity for the phenanthroline-ADPR pair (cf. Table II) quantitatively confirms our previous conclusion that the interactions of LADH with these inhibitors are independent of one another (5, 6). When 00 > o( > 0 and CY# 1, positive (a < 1) or negative (0~ > 1) interactions exist between II and I2 in the EIJz complex. A number of causes such as ion-dipole, interdipole, hydrophobic, and interionic! hydrophilic interactions as well as simple steric hindrance and protein-conformational changes may be involved. CYValues of 0.0034 for the succinic dehydrogenase inhibition by the fluoride-
phosphate pair and 0.2 for the n-amino acid oxidase inhibition by the riboflavin5’-sulfate-adenosine-5’-sulfate pair were reported by Slater and Bonner (2) and Yagi and Ozawa (3), respectively. In the present experiments, OLvalues of 0.5 and 0.3 were obtained for the phenanthroline-ADP and phenanthroline-AMP pairs, respectively. At present it remains unexplained why such positive interactions occur between these pairs of inhibitors in the EIIIf complexes, but not between phenanthroline and ADPR. The lack of interaction in the last case may be an occasional result of cancelling positive and negative interactions. The determination of a: values for pairs phenanthroline-derivatives and between ADPR-derivatives may give useful information concerning the mechanism of these interactions as well as the distance between the Zn site and the ADPR-binding site of LADH (5, 6). It should be pointed out that the CYvalue is closely related to the “stabilization” (= (Y) factors (= l/a!) or “destabilization” operating between the coenzyme and inhibitor in ternary complexes of the enzyme (14). For example, in the LADH-NADHisobutyramide complex the following relation was found:
ONLADHCOMPLEXES. III STUDIES K,, K,.
N.4DH/Kb-isobutyrarnid~, NADH = isobutyramide/KE-NADH, isobutyramide
= 60 zz
the stabilization factor = l/a!. Thus CY= 0.016. ACKNOWLEDGMENTS The authors are indebted for financial support given by Statens Medicinska Forskningr%d, Knut och Alice Wallenbergs Stiftelse, Svenska Maltdrycksforskningsinstitutet, and the Rockefeller Foundation. The skillful technical assistance of Miss Gisela Schreck and the constructive suggestions of Dr. George Waller are gratefully appreciated. REFERENCES 1. WEBB, J. L., “Enzyme and Metabolic Inhibitors.” Academic Press, New York, 1963. 2. SLATER, E. C., AND BONNER, W. D., JR., Biochem. J. 62, 185 (1952). 3. YAGI, K., AND OZAWA, T., Biochim. Biophys. Acta 42, 381 (1960). 4. LOEWE, S., Pharmacol. Rev. 9, 237 (1957). 5. YONETANI, T., Acta Chem. Stand. 17, Suppl.
1, 96 (1963). 6. YONETANI, T., Biochem. 2. 338, 300 (1963). 7. DALZIEL, K., Acta Chem. &and. 12,459 (1958).
251
8. EHRENBERG, A., AND DALZIEL, K., Acta Chem. &and. 12, 465 (1958). 9. THEORELL, H., AND YONETANI, T., Biochem. Z. 338, 537 (1963). 10. HORECKER, B. L., AND KORNBERG, A., J. Biol. Chew 176, 385 (1948). 11. THEORELL, H., PYTYGAARD, A. P., AND 9, BONNICHSEN, R., Acta Chem. &and.
1148 (1955). 12. DIXON, M., Biochem. J. 66, 170 (1953). 13. LINEWEAVER, H., AND BURB, D., J. Ajn. Chem. Sot. 66, 658 (1934). 14. THEORELL, H., AND MCKINLEY MCKEE, J. S., Acta Chem. Stand. 16, 1797 (1961). 15. DALZIEL, K., Nature 191, 1098 (1961). 16. DALZIEL, K., Biochem. J. 84, 240 (1962). 17. DALZIEL, K., J. Biol. Chem. 237, 1538 (1963). 18. WALLENFELS, K., AND SUND, H., Biochem. 2. 329, 48 (1957). 19. VALLEE, B. L., WILLIAMS, R. J. P., AND HOCH, F. L., J. Biol. Chem. 234,262l (1959). 20. PLANE, R. A., AND THEORELL, H., Acta Chem. Stand. 16, 1866 (1961). 21. MAHLER, H. R., BAKER, R. H., JR., AND SINER, V. J., JR., Biochemistry 1, 47 (1962). 22. DALZIEL, K., Nature 197, 462 (1963). 23. THEORELL, H., AND YONETANI, T., Arch. Riothem. Biophys. 106, 252 (1964). 24. VALLEE, B. L., COOMBS, T. L., AND WILLIAMS, R. J. P., J. Am. Chem. Sot. 80, 397 (1958).