Biochem. J. (1966) 100, 34

The Kinetics and Mechanism of Liver Alcohol Dehydrogenasewith Primary and Secondary Alcohols as Substrates

BY K. DALZIEL Am F. M. DICKINSONDepartment of Biochemi8try, University of Oxford

(Received 20 December 1965)

1. The activity of liver alcohol dehydrogenase with propan-2-ol and butan-2-olhas been confirmed. The activity with the corresponding ketones is small. Initial-rate parameters are reported for the oxidation of these secondary alcohols, andof propan-l-ol and 2-methylpropan-1-ol, and for the reduction of propionaldehydeand 2-methylpropionaldehyde. Substrate inhibition with primary alcohols is alsodescribed. 2. The requirements of the Theorell-Chance mechanism are satisfiedby the data for all the primary alcohols and aldehydes, but not by the data forthe secondary alcohols. A mechanism that provides for dissociation of eithercoenzyme or substrate from the reactive ternary complex is described, and shownto account for the initial-rate data for both primary and secondary alcohols, andfor isotope-exchange results for the former. With primary alcohols, the rapidrate of reaction of the ternary complex, and its small steady-state concentration,result in conformity of initial-rate data to the requirements of the Theorell-Chance mechanisms. With secondary alcohols, the ternary complex reacts more

slowly, its steady-state concentration is greater, and therefore dissociation ofcoenzyme from it is rate-limiting with non-saturating coenzyme concentrations.3. Substrate inhibition with large concentrations of primary alcohols is attributedto the formation of an abortive complex of enzyme, NADH and alcohol fromwhich NADH dissociates more slowly than from the enzyme-NADH complex.The initial-rate equation is derived for the complete mechanism, which includesa binary enzyme-alcohol complex and alternative pathways for formation of thereactive ternary complex. This mechanism would also provide, under suitableconditions, for substrate activation or substrate inhibition in a two-substratereaction, according to the relative rates of reaction through the two pathways.

Liver alcohol dehydrogenase (alcohol-NAD+oxidoreductase, EC 1.1.1.1), has an unusually widespecificity. It has been reported to catalyse theoxidation of a variety of alcohols, primary, secon-dary and cyclic, and the reduction of severalaldehydes and ketones (cf. Sund & Theorell, 1963).However, detailed kinetic studies with widelyvaried substrate and coenzyme concentrationshave been made only with ethanol and butan-l-oland the corresponding aldehydes (Dalziel, 1962b).Merrit & Tomkins (1959) reported that, with con-centrations of lmm-alcohol and 1001iM-NAD+ atpH:9-5, the rate of oxidation of butan-2-ol was36% of that of ethanol, whereas propan-2-ol wasinactive. Witter (1960) observed activity withpropan-2-ol at higher concentration. From suchmeasurements at single concentrations, it appearsthat primary alcohols and aldehydes are bettersubstrates than secondary alcohols and ketones,but it is not known, for example, whether this isdue to differences in the maximum rates, or the

Michaelis constants, or both. Van Eys (1961)showed that liver alcohol dehydrogenase alsocatalyses the isomerization of lactaldehyde toacetol, and of glyceraldehyde 3-phosphate todihydroxyacetone phosphate. These reactionsinvolve oxidation of a secondary alcohol group andreduction of an adjacent aldehyde group, andrequire only catalytic amounts of coenzyme.In the present work, we have confirmed the

activity of the enzyme with propan-2-ol and butan-2-ol and estimated initial-rate parameters for theiroxidation by detailed kinetic studies with wideranges of substrate and coenzyme concentrations.The oxidation of the primary alcohols propan-l-oland 2-methylpropan-1-ol and the reduction of thecorresponding aldehydes have been studied simi-larly. The main objectives of the work were tocompare the mechanisms of catalysis with primaryand secondary alcohols and to obtain data forsecondary alcohols that would aid interpretationof kinetic studies of the isomerase reactions.

34

MECHANISM OF LIVER ALCOHOL DEHYDROGENASEIt also seemed desirable to attempt a more

rigorous and detailed theoretical interpretation ofdata for this enzyme, including substrate inhibition,in terms of both abortive complex-formation andalternative pathways ofreaction, for which evidencehas been adduced from isotope-exchange experi-ments (Silverstein & Boyer, 1964).

MATERIALS AND METHODS

Materials. Reagent solutions and sodium phosphatebuffers were made up in glass-distilled water.

Crystalline alcohol dehydrogenase was prepared fromhorse liver and assayed as described by Dalziel (1961a).The substrates were obtained commercially and fractionallydistilled. NAD+ was purchased from C. F. Boehringer undSoehne G.m.b.H., Mannheim, Germany, and purified bychromatography (Dalziel, 1963b). NADH was prepared as

described by Dalziel (1962a).Initial-rate measurement8. The technique of initial-rate

measurements and the recording fluorimeter employedwere described by Dalziel (1962b). All measurements were

made at 23.50 and the reaction mixtures were I0-1 withrespect to sodium phosphate buffer, pH7.0. Kineticcoefficients in the initial-rate equation:

e-= 0o+ l+-02+vo [Si] [S2] [SI] [S2]

were obtained from primary and secondary plots of initial-rate data in the usual way, and are functions of rateconstants in the mechanism (Dalziel, 1957, 1962b). In thisequation, e is the concentration of enzyme active centres(twice the molar concentration) and Si and S2 are coenzymeand substrate respectively. The symbols #o etc. are usedfor kinetic coefficients for alcohol-NAD+ reactions, andO' etc. for those for aldehyde-NADH reactions. (Thisand the corresponding use of primed and unprimed velocityconstants is the reverse of the convention adopted inprevious papers, and is justified by typographical conveni-ence in this paper, which refers mainly to the alcohol-NAD+ reactions.)

All initial-rate measurements were made at least induplicate with a reproducibility of 5% in general, and atworst 15% with the smallest concentrations of bothsubstrate and coenzyme. Three separate complete experi-ments (cf. Figs. 1 and 2) were made with each of the twosecondary alcohols, and two with each of the other sub-strates, with different samples of coenzyme and substrate.o was reproducible in every case to within 5%, and theother parameters to within 20%.

RESULTS

Secondary alcohols. The results of initial-ratemeasurements with propan-2-ol in one experimentare shown in the primary plots of Figs. 1 and 2.Secondary plots of the intercepts and slopes are

shown in Figs. 3(a) and (b), the intercepts andslopes ofwhich are values for the kinetic coefficientsin eqn. (1). The results of one experiment withbutan-2-ol are similarly shown in Figs. 4-6. The

0)- 30

0)

0 005 0O10

1/[Propan-2-ol] (mM-1)

Fig. 1. Primary plots: variation of the reciprocal of thespecific initial rate at pH7-0 and 23.50 with the reciprocalof the propan-2-ol concentration for several constantNAD+ concentrations (1Mtm): 0, 517; *, 51-7; A, 15-8;A, 10-3; El, 4-3.

60

50

40.

0)

-90

a30

20

10

0 0-I 0-2

1/[NAD+] (,um-1)

Fig. 2. Primary plots: variation of the reciprocal of thespecific initial rate at pH7-0 and 23.50 with the reciprocalof the NAD+ concentration for several constant propan-2-ol concentrations (mm): 0, 400; *, 100; A, 40; A, 20;El, 10; *, 5.

Vol. 100 35

K. DALZIEL AND F. M. DICKINSON

40

30

1-:04)

~20

1/[Propan-2-ol] (mM-')10

0.1

1/[NAD+] (,xm-1)

005 0.101/[Propan-2-ol] (mM-1)

0.1 0-21/[NAD+] (um-1)

0.2

1500

1000

0

'4.44)

p4

0GQ

500 xxo_,

0 0-2 0 4 0-61/[Butan-2-ol] (mM'1)

Fig. 4. Primary plots: variation of the reciprocal of thespecific initial rate at pH7*0 and 23.50 with the reciprocalof the butan-2-ol concentration for several constantNAD+ concentrations (juM): 0, 457; *, 45.7; A, 22-8;A, 12.4.

40

0X15

0 3

Fig. 3. Secondary plots: variations of the intercepts (a)and slopes (b) of the Lineweaver-Burk plots in Figs. 1 and2 with the reciprocals of: the NAD+ concentration (0) andof the propan-2-ol concentration (0).

average values for the kinetic coefficients obtainedfrom three such experiments with each substrateare given in Table 1.The ranges of concentration of coenzyme and

substrates in these experiments were: NAD+,1-580uM; propan-2-ol, 5-400m; butan-2-ol, 18-375mm. The primary and secondary plots are

30c;

4)1-

X 20

10

0 0-02 0-04 0-06 0-081/[NAD+] (1uM-l)

Fig. 5. Primary plots: variation of the reciprocal of thespecific initial rate at pH7*0 and 23.50 with the reciprocalof the NAD+ concentration for several constant butan-2-olconcentrations (mM): 0, 375; 0, 75; A, 3-8; A, 1*88.

36

41)co

4)Qc;

fv

1966

L0

800

c)

0X 600x-49es

C 40000'4-

4)

v 2000

L0

MECHANISM OF LIVER ALCOHOL DEHYDROGENASE

,, 4-0C)Pk

0

" 2-0

0 0-2 0-4 0-61/[Butan-2-ol] (mw'1)

0 0-02 0-04 0-06

1/[NAD+] (,uM-1)80

0

X 60-

q4040

0

m 20xo

0 0-2 0-4 0-61/[Butan-2-ol] (mm-1)

0-8

1.

0-08

0-8

linear, to within the experimental error, over theseranges. The absence of substrate inhibition withsecondary alcohol concentrations up to 0-4M isnoteworthy.Prinary alcohols. The results of initial-rate

measurements with 0-1-2mm-2-methylpropan-1-olare shown in the primary plots of Figs. 7 and 8 andthe secondary plots of Fig. 9. Again the plots arelinear within the experimental error. With alcoholconcentrations greater than 2mM, substrate inhibi-tion occurs (see below).The results obtained with 0-025-2mm-propan-

1-ol (Figs. 10-12) show small but significantdeviations from linearity in the primary plots ofreciprocal initial rate against reciprocal NAD+

QC)

00

-0

cs

1.f

49

0 0-02 0-04 0X06 0-08

1/[NAD+] (/,tm-1)Fig. 6. Secondary plots: variations of the intercepts (a)and slopes (b) of the Lineweaver-Burk plots in Figs. 4 and5 with the reciprocals of the NAD+ concentration (-) andof the butan-2-ol concentration (0).

C)

C)Go

0 2-0 4-0 6-0 8-0 I0-01/[2-Methylpropan-l-ol] (mm-')

Fig. 7. Primary plots: variation of the reciprocal of thespecific initial rate at pH7-0 and 23-5° with the reciprocalof the 2-methylpropan-1-ol concentration for severalconstant NAD+ concentrations (juM): O, 530; *, 53; A,

26-5; A, 12-9; 5, 6-6; *, 3-3; x, 1-3.

Table 1. Kinetic coefficients for the oxidation of primary and secondary alcohols and aldehydesby NAD+ with liver alcohol dehydrogenase at 23.5° and pH 7

The kinetic coefficients are those in the initial-rate equation:

qo1 02 012

where Sl is NAD+ and S2 the substrate. The data for ethanol and butanol are those of Dalziel (1962b), the datafor the aldehydes those of Dalziel & Dickinson (1965). Other kinetic data were obtained as described in the text.01/o is the Michaelis constant for NAD+, and 02/00 that for the substrate.

SubstratesEthanolPropan-l-olButan-l-ol2-Methylpropan-l-olPropan-2-olButan-2-olAcetaldehydeButyraldehyde

#o(sec.)0-370-310-350-343-91-70-080-025

1 02(tm sec.) (,umsfeC.)

1-1 661-2 191-1 41-7 4082 4900041 90001-4 80000-6 400

012(mM2 sec.)0-00720-00120-00040-00326-20-950-260-015

Oi/Oo3-03.93-15-0

21241824

02/10(PM)1806111

118126005300

10000016000

012/02(PM)10968100801261053338

Vol. 100 37

K. DALZIEL AND F. M. DICKINSON

2.5

2 0

C)

01).-

-;Z

'.5

I*0

0.5

0 0-05 0-10 0 15

1/[NAD+] (,um-1)Fig. 8. Primary plots: variation of the reciprocal of theapecific initial rate at pH7*0 and 23.50 with the reciprocalof the NAD+ concentration for several constant 2-methyl-propan-l-ol concentrations (mM): *, 2; 0, 1; A, 05;A, 0.2; *, 0-1.

concentration (Fig. 11), and in the secondary plotof intercepts against reciprocal NAD+ concentra-tion (Fig. 12a). The deviations are towards greateractivity with the largest concentrations of propanoland NAD+, and make the estimates of kineticcoefficients more uncertain than usual. The small-ness of the deviations, and the fact that substrateinhibition occurs with still greater propanolconcentrations, make it difficult to draw any firmconclusions about the apparent activation.Average values for the kinetic coefficients from

the experiments with each of these primaryalcohols are given in Table 1, together with thosefor ethanol and butan-1-ol reported by Dalziel(1962b, 1963a).

Sub8trate inhibition with primary alcohol. Withall four primary alcohols, the rate decreases withincrease of alcohol concentration above about2mm. The plots of reciprocal rate against substrateconcentration (Fig. 13) show that inhibition ispartial, the rate approaching a constant value withlarge alcohol concentration in each case. Estimatesof these limiting rates, and ofthe alcohol concentra-tions at which the greatest rates are obtained witha saturating NAD+ concentration of 580M, are

given in Table 2.

1'5

0 1 0Go

05

L

0,Go

+ 1000

z

0mp4

.-0 500 I

L

0

1966

1/[2-Methylpropan-l-ol] (mm-1)

0 2 0-4 0 6

1/[NAD+] ((uM-1)0 8

5 10

1/[2-Methylpropan-l-ol] (mM-1)

0-25 0.501/[NAD+] (1hM-1)

0)

:L

"-1-0

Ca

es

0)f

104

CoxcoC>

0 75

Fig. 9. Secondary plots: variations of the intercepts (a)and slopes (b) of the Lineweaver-Burk plots in Figs. 7 and8 with the reciprocals of the NAD+ concentration (e) andof the 2-methylpropan-1-ol concentration (0).

Aldehydes. Secondary plots ofintercepts obtainedin the normal way from initial-rate measurementswith 2-methylpropionaldehyde and propion-aldehyde are shown in Figs. 14 and 15, and thekinetic coefficients obtained from them are com-

pared in Table 3 with those previously reportedfor acetaldehyde and butyraldehyde. 01k was not

estimated because at pH 7 it is small and can onlybe estimated approximately from very detailedmeasurements with small concentrations of bothsubstrate and coenzyme (Dalziel, 1962b).

38

MECHANISM OF LIVER ALCOHOL DEHYDROGENASE

3 0

-0@ci

0

1-0

0 5 101/[Propan-l-ol] (mM-l)

l 0

0-8

1 0-6

c4zd400

0-4

0-2Fig. 10. Primary plots: variation of the reciprocal of thespecific initial rate at pH7-0 and 23.50 with the reciprocalof the propan-l-ol concentration for several constantNAD+ concentrations (,uM): 0, 580; *, 58; A, 18.3; A, 9-6;

l, 4-5;E, 2-42.

3 0

-

0-brk

2-0

I 0

0 0.1 0-2 0 3 0 4

1/[NAD+] (,um-1)

Fig. 11. Primary plots: variation of the reciprocal of thespecific initial rate at pH 7 0 and 23 5° with the reciprocalof the NAD+ concentration for several constant propan-1-ol concentrations (mM): 0, 2; *, 0-4; A, 0'2; A, 041;Ol, 0 05; *, 0-025.

0

1/[Propan-l-ol] (mM-1)

0 0t1 0-2 0o3 0-4

1/[NAD+] (pM-')

c; 15

54)-.4

0

00

04 5

Go

0 2-0 4 0 6-0 8-0

1/[Propan-l-ol] (mM-1)

0 0 I 0 2 0 3 0-4

1/[NAD+] (/AM-')

10-0

0

OQ

500 1

+

250 4.20

P.4

0

0 5

Fig. 12. Secondary plots: variations of the intercepts (a)and slopes (b) of the Lineweaver-Burk plots in Figs. 10and 11 with the reciprocals of the NAD+ concentration(0) and of the propan-l-ol concentration (0).

Ketones. With 40,um-NADH, 10mi-acetone and0'15juM-enzyme, no oxidation of NADH could bedetected in the fluorimeter. The rate must there-fore be less than 1.5 uM-NADH oxidized/min., andthe turnover number less than 10/min.With butan-2-one, NADH oxidation was ob-

served but the progress curves were biphasic, a

rapid oxidation being followed by a slower lineardecrease of fluorescence. Repeated fractionaldistillation of the ketone showed the rapid initialphase to be due to an impurity, presumably an

aldehyde, and the amount of NADH oxidized inthe rapid phase indicated that the commercialketone preparations contained about 0.1% of thisimpurity. By repeated fractional distillationketone was obtained that showed no significantrapid phase, and gave a constant rate of NADHoxidation, with 0 75,zM-enzyme and 40,M-NADH,of 5,uM/min. with 50mM-ketone and 10,uM/min.with 930mM-ketone. This very slow reaction could

Vol. 100 39

K. DALZIEL AND F. M. DICKINSON

5

C)

Co

0

0 5

0 100 200[Alcohol] (mM)

500 1000

[Ethanol] (mM)1500

Fig. 13. Variation of the reciprocal of the specific initialrate of primary alcohol oxidation at pH7-0 and 23 5° andat constant NAD+ concentration (580Ozm) with the primaryalcohol concentrations: 0, 2-methylpropan-1-ol; 0,

ethanol; A, propan-l-ol; A, butan-l-ol.

Table 2. Maximum speciJic rates, limiting rates withinfinitely large substrate concentrations, and optimumsubstrate concentrations for primary alcohols at23.50 and pH17

Maximum specific rates, V = 1/Io, are from Table 1.Optimum substrate concentrations were estimated fromthe minima of plots of initial rate against reciprocal alcoholconcentrations, with 0O58mM-NAD+. Limiting rates withinfinitely large substrate concentrations, Vi, are fromFig. 13.

SubstrateEthanolPropan-l-olButan-l-ol2-Methylpropan-l-ol

V VI S]opt.(sec.-l) (sec.-l) (mM)2-7 1-23-2 1-02-9 0-652-9 2-1

8-02*01.03-0

be due to another impurity in the ketone, anddetailed kinetic studies would be unprofitable.Evidently butan-2-one is a very poor substrate.

DISCUSSION

These experiments confirm that propan-2-ol andbutan-2-ol are substrates for liver alcohol dehydro-genase. The data of Table 1 show that they arepoorer substrates than the isomeric primaryalcohols, and not only because they have largerMichaelis constants (#2/00). All four kinetic

0C)

40m

C)C)3.4

"

0 0.05 0 10 0.151/[2-Methylpropionaldehyde] (,Mm-1)

0 0-25 0 50 0 751/[NADH] (,Am-1)

Fig. 14. Secondary plots showing the variation of inter-cepts ofprimary Lineweaver-Burk plots with the reciprocalof the NADH concentration (0) and of the 2-methyl-propionaldehyde concentration (o).

0

c)

C)

0

I.4cz

0 0.1 0-2 0 3 0-41/[Propionaldehyde] (,tM-1)

0 0 25 0.50 0.75 1*001/[NADH] (,um-1)

Fig. 15. Secondary plots showing the variation of interceptsof primary Lineweaver-Burk plots with the reciprocals ofthe NADH concentration (e) and of the propionaldehydeconcentration (o).

coefficients in eqn. (1), and the Michaelis constantsfor the coenzyme (01/o), are substantially greaterwith secondary alcohols than with primary alcohols.For the four primary alcohols go and 01 are

40 1966

MECHANISM OF LIVER ALCOHOL DEHYDROGENASETable 3. Kinetic coefficients for the reduction ofaldehydes by NADH with liver alcohol dehydro-genase at 23.50 and pH7

The kinetic coefficients are those in the initial-rateequation:

e , 01 02 012o= # [Si'] +[S2 + [S'][S2]

where S' is NADH and S is aldehyde. The data foracetaldehyde and butyraldehyde are from Dalziel (1962b),and the other data were obtained as described in the text.

ES1

k+; +,,k-I S2

Si \ --

+ k_3\E ES1S2+ k-482 k+4

k'-2 Sik+2 +

ES2

ESi+ s"-I

S2 k'+,\k '

V+ 3

1

ESiS2 Ek' ~~~+Sj+4 SI

+Ie+ ES I

ES'2

SubstrateAcetaldehydePropionaldehydeButyraldehyde2-Methylpropion-aldehyde

(sec.)0-00750-00950-00750-009

Scheme 1.0-100-140-100-12

3-30-430-170-28

reasonably constant. The branched-chain 2-methyl.propan-1-ol is a poorer substrate than propan-l-ol,but among the three straight-chain alcohols #b2,and therefore the Michaelis constant for the sub-strate, decreases with increase of chain length.Similarly, butan-2-ol is a better substrate thanpropan-2-ol, but in contrast with the primaryalcohols these secondary alcohols give differentvalues for Oo and 01 as well as for 02-

Previous conclusions that the behaviour ofprimary alcohols and aldehydes could be describedby the Theorell-Chance mechanism were partlybased on the identity of 00, 01 4', and O' fordifferent alcohols and aldehydes, and on relationsbetween the eight kinetic coefficients for forwardand reverse reactions (Dalziel, 1962b, 1963b). Themost striking feature of the present work is thatsecondary alcohols do not satisfy several of thesecriteria, and that the enzyme therefore appears toexhibit different mechanisms with different sub-strates. Before attempting to interpret the data forsecondary alcohols, it is pertinent to summarizeprevious evidence of the mechanism for ethanol andbutan-1-ol oxidation and the reverse reactions, andto consider the extent to which the new data forpropan-l-ol and 2-methylpropan-1-ol and thecorresponding aldehydes, and the experiments on

substrate inhibition, support the previous work.Two-substrate or coenzyme-substrate reactions

may be considered as group-transfer reactions, forwhich there are two general types of mechanism.In the 'ternary-complex' or 'single-displacement'mechanism (Haldane, 1930; Koshland, 1955) bothsubstrates are simultaneously bound to the enzyme,and group transfer takes place directly within theternary complex. In the 'substituted-enzyme'or 'double-displacement' mechanism (Doudoroff,Barker & HaEsid, 1947; Koshland, 1955) the group

is transferred from the donor substrate to theenzyme, and then, after dissociation of the firstproduct, from enzyme to acceptor substrate.These two types of mechanism may be distin-guished by isotope-exchange studies, by studies ofthe reaction between stoicheiometric amounts ofenzyme and one substrate, and by initial-ratekinetics with rate-limiting concentrations of bothsubstrates. The latter distinction depends on thefact that, because ternary complexes are notformed, the initial-rate equation for the substituted-enzyme mechanism lacks the last term of eqn. (1),i.e. 012 = 0 (Dalziel, 1957). This mechanism istherefore not satisfactory for liver alcohol dehydro-genase with either primary (Dalziel & Theorell,1957) or secondary alcohols or for other nicotin-amide nucleotide-linked dehydrogenases (cf. Massey& Veeger, 1963).

Steady-state treatment of the ternary-complexmechanism with alternative pathways (Scheme 1)gives an initial-rate equation of the second degreein each substrate concentration (Dalziel, 1958).Non-linear reciprocal-rate plots (Lineweaver &Burk, 1934) of the substrate activation or inhibi-tion type, or even of more complex forms, aretherefore inherent in this general mechanism, anddo not necessarily indicate the presence of twoor more different, or identical but interacting,substrate-binding sites (Dalziel, 1957,1958). Thesepoints are worth re-emphasizing in view of thecurrent interest in such behaviour by enzymesconcerned in metabolic control.One of several special cases of this mechanism

(Dalziel, 1958) that does give a linear reciprocalinitial-rate equation of the form of eqn. (1) is therapid-equilibrium random-order mechanism, inwhich it is assumed that the interconversions ofthe ternary complexes (k, k') are rate-limiting, andbinary and ternary complexes are in equilibrium(Haldane, 1930). The kinetic coefficients arefunctions of k and the dissociation constants of thecomplexes (K1 =k1l/k+l etc.) as shown in Table 4.

Vol. 100 41

0if 0;(,umsec.) (jumsee.)

K. DALZIEL AND F. M. DICKINSONTable 4. Kinetic coefficients in the initial-rate equation for mechanisms for two-substrate reactions

The initial rate equation is:e 'ki #2 #12

v 0= 0° [S]+[ 2]+[S1][S1][S2]

The mechanisms are special cases of the general altemative-pathway mechanism shown in Scheme 1, to whichthe rate constants refer, as described in the text. A = 1+ k'/k ' 3.

Mechanism

1. Equilibrium random-order

2. General compulsory-order

3. Theorell-Chance

#0 1 021 K4 K3k k k

#12K1K3k

1 1 A 1 Ak-3+k k-1 (Ak-3+k)Zk Z3k' 3k k+j kk+3 k+lkk+3

1 Ak.3+k k-1 (Ak-3+k)k+j kk+3 k+1kk+3

If the lower alternative pathways in Scheme 1are omitted steady-state treatment yields eqn. (1)with kinetic coefficients as shown in Table 4. Afurther special case of this compulsory-ordermechanism arises when k'1 <<k, k't3 and kk' 3/k'(Dalziel, 1957). Ternary complexes are then notkinetically significant, and the maximum specificrate is 1/O=k'Q1. This is equivalent to themechanism first proposed by Theorell & Chance(1951) for liver alcohol dehydrogenase. Althoughthese workers did not actually show ternary com-plexes in their reaction scheme, the probableexistence of such intermediates in rapid steps wasrecognized (cf. Theorell, Nygaard & Bonnichsen,1955). The stereospecificity ofnicotinamide nucleo-tide-linked dehydrogenases (Vennesland, 1955)makes the formation of such complexes virtuallycertain, and the fact that the aldehyde-mutasecycle catalysed by liver alcohol dehydrogenaseexhibits a maximum rate is direct evidence fortheir presence (Dalziel & Dickinson, 1965). Thislimiting case of a compulsory-order mechanism, inwhich ternary complexes are formed but are notkinetically significant, is referred to here, as inprevious papers, as the Theorell-Chance mechan-ism. Some authors, however, restrict the termto a peroxidase-type mechanism in which ternarycomplexes are not formed (Bloomnfield, Peller &Alberty, 1962; Cleland, 1963; Wong & Hanes, 1964;Sanwal, Stachow & Cook, 1965).Primary alcohols and aldehydes. The main

evidence in support of the Theorell-Chance mechan-ism for the ethanol-acetaldehyde and butanol-butyraldehyde systems is as follows.

(1) Initial-rate data over wide ranges of substrateand coenzyme concentrations and pH satisfy eqn.(1) (Theorell & McKinley-McKee, 1961a; Dalziel,1962b, 1963a).

(2) The velocity constant kAl for combination

of enzyme with NADH measured directly by rapid-reaction technique (Theorell & Chance, 1951) agreesreasonably well with 1/0' (cf. Table 4). Con-firmation by the more sensitive rapid-reactiontechnique now available is desirable.

(3) Provided that pure coenzymes are used, themaximum specific rate in each direction (I/Io, I/O.)is equal to the rate of dissociation of the productcoenzyme calculated from initial-rate parametersfor the reverse reaction on the assumption ofa compulsory-order mechanism (k-1 = 012/0102,k'k012/0i/2). If the ternary complexes wererate-limiting, the maximum specific rate would besmaller than the latter function (Table 4). Previousevidence that the opposite inequality held (Dalziel& Theorell, 1957; Dalziel, 1962b), which was inter-preted as evidence that isomeric enzyme-coenzymecomplexes were formed (Mahler, Baker & Shiner,1962; Bloomfield et al. 1962; Wratten & Cleland,1963), was explained by the effects of competitiveinhibitors in commercial coenzyme preparations(Dalziel, 1961).

(4) The two alcohols gave the same values for#o and for #1, but different values for 02 and 012.The two aldehydes showed similar behaviour inthe reverse reaction (Dalziel, 1962b). The identityof the 01 values is expected for the general com-pulsory-order mechanism, that of the maximumrates only for the Theorell-Chance mechanism.Neither would be expected for a random-order oralternative-pathway mechanism.

(5) Values for the dissociation constantsof enzyme-coenzyme complexes calculatedfrom initial-rate data (K1 = 012/#2 = Ol/+0;

1=#12/2 = #1/I0) were the same for bothpairs of substrates (Dalziel, 1962b, 1963b) andagreed reasonably well with those obtained bydirect equilibrium measurements (Theorell &Winer, 1959; Theorell & McKinley-McKee, 1961b).

42 1966

1

kl 1

MECHANISM OF LIVER ALCOHOL DEHYDROGENASEA direct value for K1 by a more accurate methodis desirable.The results now reported for propan-l-ol and

2-methylpropan-1-ol and for the correspondingaldehydes (Tables 1 and 3) provide further supportfor the Theorell-Chance mechanism by substan-tiating these observations. In particular, theconstancy of qo and 01 is in striking contrast withthe differing values obtained for these parameterswith the secondary alcohols.

Sub8trate inhibition with primawry alcohol8 andalehyde8. Inhibition with ethanol concentra-tions greater than 8mM has been attributed tothe formation of the abortive ternary complexE NADH ROH (Theorell & McKinley-McKee,1961c), and direct fluorimetric evidence for theformation of this complex, and analogous com-plexes with other dehydrogenases, has beenreported (Winer & Schwert, 1958; Theorell &McKinley-McKee, 1961b). This interpretation issupported by the fact that the inhibition is mostmarked with saturating NAD+ concentrations(Dalziel, 1962b). Theorell & McKinley-McKee(1961c) considered that NADH dissociates from theabortive complex, at a lower rate than fromE .NADH, to give a binary enzyme-alcoholcomplex. The present results with four primaryalcohols substantiate this conclusion by showingthat the alcohols behave as partial inhibitors, i.e.that E.NADH.ROH is not a 'dead-end' complex;the limiting rate approached with very largealcohol concentrations can be identified with therate of dissociation of NADH from this complex,and decreases with increase of chain length of thealcohol (Table 2). The quantitative treatment ofthis inhibition mechanism is considered below.

Substrate inhibition has also been observed inthe reverse reaction with butyraldehyde and2-methylpropionaldehyde concentrations greaterthan about 1mm, and is explained by the formationof the analogous complex E.NAD+ RCHO, whichis, however, the active ternary complex of thealdehyde-dehydrogenase reaction catalysed bythis enzyme (Dalziel & Dickinson, 1965).Secondary alcohol8. The large values of 02 for

secondary alcohols could be due to their relativelysmall affinity for E .NAD+, or slow intramolecularreaction in the ternary complex (k). The constancyof the ratio 012/02 for both primary and secon-dary alcohols and for aldehydes in the aldehyde-dehydrogenase reaction (Dalziel & Dickinson, 1965)is striking in view of the enormous variations of thevalues of each of these parameters for the threesets of oxidizable substrates (Table 1). A satis-factory mechanism for secondary alcohols mustclearly retain the relation 012/02 =K1. The lowvalues of this ratio for aldehydes compared withthose for alcohols is explained by the smaller

phosphate buffer concentrations used in thealdehyde-mutase studies (Dalziel & Dickinson,1965).The different values for #1 for the two secondary

alcohols are not consistent with a simple com-pulsory-order mechanism, and suggest that withthese substrates the dependence of rate on NAD+concentration is determined not by the rate ofcombination of NAD+ with enzyme but by dis-sociation ofNAD+ from the active ternary complexE.NAD.R1R2CHOH. Evidence for such dis-sociation of coenzyme, without prior dissociation ofsubstrate, was presented by Silverstein & Boyer(1964), and is discussedbelow. Supporting evidenceis the analogous dissociation of the abortivecomplex E NADH.RCH20H indicated by thepartial character of substrate inhibition, andkinetic evidence from studies of the mutase cyclethat the active ternary complex of the aldehyde-dehydrogenase reaction, E-NAD+.RCHO, candissociate in this way (Dalziel & Dickinson, 1965).With primary alcohols, this alternative pathwayof reactant dissociation would not be rate-limitingin initial-rate studies because the ternary-complexinterconversion and aldehyde-product dissociationare fast, and therefore with saturating alcoholconcentrations the ternary-complex concentrationin the steady state will be negligible. However,with secondary alcohols the greater values of 0osuggest that the intramolecular reaction of theternary complex is slower than with primaryalcohols, and is rate-limiting under maximum-rateconditions. The different values for qo obtainedwith the two secondary alcohols support this view,as in the aldehyde-dehydrogenase reaction (Dalziel& Dickinson, 1965). With saturating concentra-tions of secondary alcohols the steady-stateconcentration of the ternary complex would thenbe large, and dissociation of NAD+ from it couldbe the limiting factor that determines 01.Thus a consistent account of the kinetics of

oxidation of primary and secondary alcohols (andof aldehydes) can be obtained on the assumptionof alternative pathways of dissociation of reactantsfrom the ternary complex, and different rates ofoxidation within the ternary complex.

Evidence of alternative pathway8 and enzyme-alcohol complexe8. Silverstein & Boyer (1964)adduced evidence for the existence of alternativepathways from the observation that isotopeexchange between NAD+ and NADH persisted ata significant rate, in the presence of liver alcoholdehydrogenase, up to very large equilibrium con-centrations of ethanol and acetaldehyde. Ifalternative pathways of dissociation of ternarycomplexes were forbidden, as in a compulsory-order mechanism involving only the upper path-ways of Scheme 1, the enzyme would be entirely

Vol. 100 43

K. DALZIEL AND F. M. DICKINSONin the form of the ternary complexes under theseconditions, and NAD+=KNADH exchange wouldbe blocked. The abortive complex E.NADH.ROHand the active ternary complex of the aldehyde-dehydrogenase reaction, E.NAD RCHO, wouldalso be present, however, and it was recognizedthat these complexes, and not the active complexesof the alcohol-dehydrogenase reaction, might beresponsible for the exchange (Wong & Hanes,1964; Dalziel & Dickinson, 1965). However, theevidence presented above for dissociation of NAD+from the ternary complex E.NAD R1R2CHOHmakes it likely that this can also occur withprimary alcohols.

Silverstein & Boyer (1964) found the rate ofacetaldehyde -ethanol exchange, R', to be 160times that of the NAD+=INADH exchange, R,with excess of reactants at equilibrium, and con-cluded that ternary-complex interconversion couldnot be a slow step in the catalysis, in accordancewith conclusions from initial-rate studies. Fromthe relations given by these authors, this resultshows that, in Scheme 1, k>kI4 or k'>lc4, orboth, and that k 3>k1 or k' 3k1c4 or both.However, it sets no restriction on the relativevalues of k_3 and k or k' 3 and k', and therefore,though ternary-complex interconversion cannotbe rate-limiting in the lower pathway of reaction,it could be in the upper pathway.The data do, however, provide other evidence,

which seems to have been overlooked, that thedissociation of the product complex E.NADH(k'A1) is in fact rate-limiting in the overall ethanol-NAD+ reaction. The maximum initial rate of theNADH-acetaldehyde reaction, V, was 1-44 timesthe rate of the acetaldehyde =-ethanol exchange,R'. At the pH of these experiments, the maximuminitial rate of the reverse reaction, V', is only one-fiftieth of V (Dalziel, 1963a) and therefore one-thirtieth of R'. From the relations given bySilverstein & Boyer (1964) for R', this shows that

k' 1 determines V', in accordance with conclusionsfrom initial-rate studies.Wong & Hanes (1964) have pointed out that the

isotope-exchange results could be accounted for bya compulsory-order mechanism with dead-endbranches (Scheme 2). The exchange rates withexcess of reactants do not depend on the rates ofreaction of free enzyme with coenzyme or substrate.It follows that the relative values of R and R'give no information about the relative rates of netreaction through the upper and lower reactantpathways of Scheme 1. Thus, though the existenceof enzyme-alcohol complexes is indicated by theisotope-exchange experiments, by the partialcharacter of substrate inhibition and by the largevalues of 01 for secondary alcohols, there is noevidence that they are kinetically significant in thereaction. It remains to show that a mechanismthat provides for the formation of such complexeswill account for all the experimental results.Proposed mechanism for alcohol dehydrogenase.

The preceding qualitative interpretation of initial-rate data for primary and secondary alcohols canbe given quantitative expression by the generalmechanism of Scheme 3, of which the mechanisms

Si+

E'~

SI+

ES+I

82 SI52~~~~~~~~~~~

ES1S2 I' ESiS2

Si SI+ +S2 ES2

Scheme 2.

S1

St+

E

ES1kk,-1

+1 S.,

Si "\ k+3+ k-3\ A- k'-3 - 1

ESIS2 ES'S', s I + ES' S8I ++ ±- k

Is4I/ +

2'-2

k+,2V + \& kES'S2

i-6

S' + ES2I+ 4L-+,s~~~~~~~~~~~Scheme 3.

E

ES.,

1

AA 1966

E,

MECHANISM OF LIVER ALCOHOL DEHYDROGENASEfor the two types of alcohol are special cases. Thisdiffers from Scheme 1 by inclusion of the abortivecomplex ES'S2 and omission of the alternativepathway of product dissociation through ES'.The latter omission is justified for primary alcoholsby evidence from isotope-exchange experimentsand initial-rate studies that k'1c>>k- and that themaximum specific rate (excluding substrate in-hibition) is k'.1 The upper pathway of productdissociation is also likely for secondary alcohols,especially as ketones are such poor substrates. Onsimilar grounds it may be assumed that (i) k-3> k-4.The initial-rate equation in the absence of products,derived by steady-state treatment of Scheme 3,is then:

se (A +1 B + Ak-4k+lVOT+Z3 kl+ kk+4

k-2 + (C + DIE) k+4[S0] + k+2[S2] 1k-2 + k+4[Sl] + Ek+2[S2] Jk+1[S1]

|Ak_3 k_2 + Ck+4[Sl]k k2 + k+4[S + Ek+2[S2] k+3[S2]

+ Ak_3+ k-2 + Ck+4[Sl] + k+2[S2]lk k-2 + k+4S1] + Ek+2[S2]

k+lk+3[Sl] [S2] (2)

with:

A k+k 3 k+jS2]+k_5+k-6k' ' (k-61k' )k+,[S]2+k-,+k-6

a k-r,k+k-6(k_6/kL'1) k+J[S2] + k_5 + k_6

D = (k_6/k'1)k+5[S21(k-61k-L) k+2[S2] + k_5 + k-6'

C+D= landE =-+4k+1

If k-6 < k' 1, B, C and D are substrate-inhibitionfactors arising from ES,S2 formation with largeprimary alcohol concentrations. With smallerconcentrations (less than 8mm) such that (ii)k+5[S2] k-5+ k-6 then B =C= 1 and D = 0. Theresulting equation [which can also be derived fromthat given by Dalziel (1958) for Scheme 1] is notof the linear reciprocal form of eqn. (1) becauseof the alternative pathways for formation ofES1S2.

If also (iii) Ek+2[S2]< k-2 + k+4[S1] and (iv)k+2[S2]< k-2+ k+4[Sl], a linear reciprocal equationof the form of eqn. (1) is obtained, namely:

vo (k k_-3 k_21

+ Ak-4S+ 1 1 + 02 + 012kk+4 k+1J [Sl] [S21 [Sl] [S21 (3)

where #2 and 012 are the same functions of velocityconstants as for a compulsory-order mechanism inTable 4. Condition (iii) together with condition (i)means that the rate of reaction through the bottompathway (vb) is negligible compared with thatthrough the top pathway (vt), since it can beshown by steady-state analysis that:

Vb (k-4/k_3) k-2+ (k+4/k+l) k+2[S2]Vt - k-2+ k+4[S1]

If k+4 k+j, condition (iv) is implicit in condition(iii). However, if k+4 < k+j, condition (iii) may besatisfied but not condition (iv), with sufficientlylarge substrate concentrations. In that case thecoefficients of the second and fourth terms ofeqn. (3) would include the inhibition factors1+ k+2[S2]/(k-2+ k+4[S2]), indicating competitiveinhibition of Si by S2. This is another possiblecause of substrate inhibition, arising from theinefficiency of the pathway through ES2 (Dalziel,1957). It may also be mentioned that a morerestrictive assumption than conditions (iii) and(iv), which also yields eqn. (3), is that k+2= k-2 = 0.This corresponds to a dead-end-complex mechan-ism in which ES2 cannot dissociate and is formedonly via ES1S2 (cf. Scheme 2).Eqn. (3) accounts for the data for primary and

secondary alcohols and that for aldehydes in thealdehyde-dehydrogenase reaction. The constancyof 012/02 for all these substrates is predicted.For primary alcohols, k> k-4 and 01= l/k+l=1-l,pMsec., and #o=I/k'L1=0-35sec. 1, as dis-cussed above. For secondary alcohols, the largervalues of #o indicate that l/k'-1 is small comparedwith A/k+ 1/k' 3. It also seems likely that k'13 willbe at least as great for ketone dissociation as foraldehyde dissociation and that Oo= A/k for secon-dary alcohols, i.e. ternary-complex interconversionis rate-limiting. Similarly, the large values for Oiindicate that 01= Ak_4/kk+4, l/k+, being negligible.The very different values for 0o and 01 obtained

previously for the aldehyde-dehydrogenase reactioncan be interpreted in the same way as for secondaryalcohols (Dalziel & Dickinson, 1965). Accordingto this interpretation, #1/#o = k_4/k+4, i.e. theMichaelis constant for the coenzyme is equal toK4, the dissociation constant for NAD+ from theternary complex. The constancy of this ratio fordifferent secondary alcohols and aldehydes isstriking in view of the variation of each of theseparameters (Table 1), and indicates that K4= 20,uMindependent of the nature of the substrate. This

Vol. 100 45

46 K. DALZIEL AND F. M. DICKINSON 1966is smaller than K1= #12/k2, the dissociation con-stant of E -NAD+ (Table 1), i.e. these substratesincrease the affinity ofthe enzyme for the coenzyme.The similarity between the behaviour of secondaryalcohols and aldehydes may be correlated withtheir structural similarity if the aldehydes react inthe hydrated form. Thus the different values of#o and '1 are explained by different rates ofoxidation within the temary complex (k), whichincrease in the order propan-2-ol, butan-2-ol,acetaldehyde, butyraldehyde. The aldehydes givebigger maximum rates than the primary alcoholsbecause the product complex E .NADH is rapidlyreoxidized by the reverse alcohol-dehydrogenasereaction, whereas the dissociation velocity ofE NADH, k'1, limits the maximum rate forprimary alcohols. For the latter, the Michaelisconstant for the coenzyme, 0j/1 =k' j1k+1, is nota dissociation constant.

Substrate inhibition with large concentrations(greater than 8mM) of primary alcohols and ofNAD+ is explained by increase of the first termof eqn. (3) from qb = 1/k' 1 to c,b0=B/k'1, theother terms being negligible under these conditions.The limiting value of B/k' 1 with very large alcoholconcentrations is 1/k-6. The absence of substrateinhibition with secondary alcohol concentrationsup to 400mM will be partly due to the fact thatA/k and not B/k' 1 is the dominant term in 00. Inaddition, the affinity of E.NADH for secondaryalcohols could be smaller, or k-6 larger, than forprimary alcohols.

It has been pointed out above that another causeof substrate inhibition could operate if k+4 < k+j;it would affect the second and fourth terms ofeqn. (3) and therefore be most apparent with smallconcentrations of NAD+. If k+4> k+j, substrateactivation would be possible. The importance ofthe relative values of k+1 and k+4 in explaining suchdeviations from linear reciprocal plots has beenpointed out previously in connexion with a simpleralternative-pathway mechanism (Dalziel, 1957).Substrate activation would also occur if k-6 > k' 1,and would be most marked with large coenzymeconcentrations. This is found experimentally foryeast alcohol dehydrogenase with ethanol as sub-strate (Nygaard & Theorell, 1955), and also forliver alcohol dehydrogenase with cyclohexanol assubstrate (K. Dalziel & F. M. Dickinson, unpub-lished work). It is apparent that the alternative-pathway and abortive-complex mechanism ofScheme 3 for a two-substrate reaction will readilyaccount for substrate activation and inhibition,and possibly also for the more complex types ofbehaviour reported for citrate-condensing enzyme(Kosicki & Srere, 1961) and various regulatoryenzymes.

We thank the Medical Research Council for financialaid, and Miss J. Holland for technical assistance.

REFERENCES

Bloomfield, V., Peller, L. & Alberty, R. A. (1962). J. Amer.chem. Soc. 84, 4367.

Cleland, W. W. (1963). Biochim. biophys. Ada, 67, 104.Dalziel, K. (1957). Acta. chem. scand. 11, 1706.Dalziel, K. (1958). Trans. Faraday Soc. 54, 1247.Dalziel, K. (1961a). Biochem. J. 80, 440.Dalziel, K. (1962a). Biochem. J. 84, 240.Dalziel, K. (1962b). Biochem. J. 84, 244.Dalziel, K. (1963a). J. biol. Chem. 238, 1538.Dalziel, K. (1963b). J. biol. Chem. 238, 2850.Dalziel, K. & Dickinson, F. M. (1965). Nature, Lond.,

206,255.Dalziel, K. & Theorell, H. (1957). Biochem. J. 66, 34P.Doudoroff, M., Barker, H. A. & Hassid, W. Z. (1947).

J. biol. Chem. 168, 725.Haldane, J. B. S. (1930). Enzymes, p. 83. London: Long-mans, Green and Co. Ltd.

Koshland, D. E. (1955). Disc. Faraday Soc. no. 20, 142.Kosicki, G. W. & Srere, P. A. (1961). J. biol. Chem. 236,

2560.Lineweaver, H. & Burk, D. (1934). J. Amer. chem. Soc.

56, 658.Mahler, H. R., Baker, R. H. & Shiner, V. J. (1962). Bio-

chemistry, 1, 47.Massey, V. & Veeger, C. (1963). Annu. Rev. Biochem. 32,

579.Merrit, A. D. & Tomkins, G. M. (1959). J. biol. Chem.

234, 2778.Nygaard, A. P. & Theorell, H. (1955). Acta chem. scand.

9, 1300.Sanwal, B. D., Stachow, C. S. & Cook, R. A. (1965).

Biochemistry, 4, 410.Silverstein, E. & Boyer, P. D. (1964). J. biol. Chem. 239,

3908.Sund, H. & Theorell, H. (1963). In The Enzymes, vol. 7,

p. 25. Ed. by Boyer, P. D., Lardy, H. & Myrback, K.New York: Academic Press Inc.

Theorell, H. & Chance, B. (1951). Acta chem. scand. 5, 1127.Theorell, H. & McKinley-McKee, J. S. (1961a). Acta chem.

scand. 15, 1797.Theorell, H. & McKinley-McKee, J. S. (1961b). Acta chem.

scand. 15, 1811.Theorell, H. & McKinley-McKee, J. S. (1961c). Acta chem.

8cand. 15, 1834.Theorell, H., Nygaard, A. P. & Bonnichsen, R. (1955).Ada chem. scand. 9, 1148.

Theorell, H. & Winer, A. D. (1959). Arch. Biochem.Biophys. 83,291.

Van Eys, J. (1961). J. biol. Chem. 236, 1531.Vennesland, B. (1955). Disc. Faraday Soc. no. 20, 240.Winer, A. D. & Schwert, G. W. (1958). Biochim. biophys.Ada, 29, 424.

Witter, A. (1960). Acta chem. scand. 14, 1717.Wong, J. T. & Hanes, C. S. (1964). Nature, Lond., 203,492.Wratten, C. C. & Cleland, W. W. (1963). Biochemistry, 2,

935.