Mechanism of Action of the Flavoenzyme Lactate Oxidase*

SUMMARY Lactate oxidase from Mycobacferium smegmatis forms a detectable intermediate during reaction with each of the substrates, L-lactate, /3-phenyl L-lactate, L-ac-hydroxy-/3-methylvalerate and r-ar-hydroxyisovalerate. This intermediate, while present greater amounts during anaerobic reactions, is also seen in the course of aerobic catalysis. The intermediate formed from each of reaction to turnover. A spectrally identical intermediate can be made by mixing reduced enzyme with an of the corresponding keto acid. The intermediate produced by mixing reduced enzyme and pyruvate, forms the normal catalytic Free

Sutton suggested that the enzyme forms an intermediate with pyruvate which decarboxylates in the presence of oxygen but dissociates in its absence. Our mechanism agrees with this hypothesis.
Takemori et al. (4) observed what appeared to be two intermediates on reacting enzyme with lactate anaerobically.
We have looked at the reaction of enzyme with various substrates in the stopped flow apparatus and have seen just one intermediate. This difference is due to the fact that we used 0.01 M imidazole-HCl buffer; while Takemori et al. worked in the presence of high concentrations of phosphate, a competitive inhibitor of the enzyme. Evidence is presented that the intermediate participates in catalysis and that it is a complex made up of reduced enzyme and ketoacid.
The mechanism is like that of D-ark0 acid oxidase, with product dissociation being the rate-determining step (5). MATERIALS AND METHODS L-cY-Hydroxyisovaleric acid was purchased from Calbiochem, sodium [2-14C]pyruvate from Nuclear-Chicago.
All other reagents were obtained from Sigma Chemical Company.
Enzyme ,was purified essentially according to the method of Sullivan (6) with the modification that the bacteria were broken in a Manton-Gaulin homogenizer at 6000 pounds per square inch and then treated with 20 mg of crude DNase per liter to reduce viscosity.
All experiments were done at 25" in 0.01 M imidazole-HCl, .pH 7.0, unless otherwise indicated. Separation and Identification of Acetic, Pyruvic, and Lactic Acids-The reaction mixture (2.6 ml) was applied to the top of a 12ml Dowex l-X8 (Cl-) column in the cold room and eluted with an acid gradient, 250 ml of 10 rnM HCI versus 250 ml of water.
Acetic acid came off first at 140 to 144 ml, always in just one tube, followed by lactic acid, and lastly by pyruvic acid (400 to 500 ml). In control runs, made to establish the size column and volume of eluant needed to get good separation of 10V5 moles of each reagent, acetic acid was identified by gas chromatography.
A glass column (4 feet X )C inch, inner dimeter) packed with Chromosorb 102 (mesh size 100 to 120, Johns-Manville) was used. This packing has the advantage that acids can be applied in water solution.
F & M 402 gas chromatograph was operated at 200". Pyruvate and lactate were assayed with lactic dehydrogenase (7).
In the enzyme reaction with 12-14C]pyruvate, acids were identified by comparing the elution pattern of the radioactive peaks to that of controls. Contaminants accounted for about 1 to 2 y0 of the total pyruvate counts and came off in two small broad peaks between acetic and pyruvic acids and clearly separated from both of them. The scintillation fluid for aqueous samples (10 ml of fluid plus 1 ml of sample per vial) consisted of 100 mg of 1,4-bis [2-(5.phenyloxazolyl)]benzene (POPOP) in 1 liter of toluene, plus 4 g of 2,5-diphenyloxazole (PPO) and 500 ml of Triton X-100. Turnover Measurement-The rate of the lactate oxidase-catalyzed reaction of oxygen with L-lactate, or one of its analogs, was measured by following oxygen consumption in the oxygen electrode (Yellow Springs Instruments, model 53) and by stopped flow measurements of absorbance changes in the enzyme itself. The oxygen electrode data were obtained in a simple straightforward manner from initial velocities of oxygen consumption at lo%, 21%, 50'%, and 100% oxygen saturation at various lactate concentrations.
The reaction generally dropped off after the first minute, having by then consumed about 20% of the dissolved oxygen. Enzyme concentration is expressed in terms of FMN absorbance at 450 nm with a determined molar extinction coefficient of 10,600 M-l cm-l.
Turnover is micromoles of O2 consumed per min per pmole of enzyme Aavin.
In the stopped flow apparatus turnover was calculated from the absorbance changes of the enzyme with time at 460 nm or at 540 nm according to the principles of Chance (8) and as detailed by Gibson et al. (9). The calculation is possible because in a system where the concentration of oxygen is much lower than that of substrate, the flavin is a built-in indicator of oxygen presence or absence. In its presence flavin is predominantly oxidized but once all the oxygen has been used up the flavin is reduced by excess substrate.
The difference in extinction between oxidized and reduced flavin enzyme allows one to measure the time elapsed till all the oxygen is consumed.
A Gibson-Milnes (10) stopped flow apparatus was used in these studies. The electronics of the Gibson-Milnes stopped flow instrument were replaced by a solid state circuit which included a logarithmic amplifier that produced an output directly proportional to absorbance rather than to transmittance.' This greatly simplifies calculations and operation of the instrument.
Peroxide Assay-In the presence of horseradish peroxidase, hydrogen peroxide reacts with o-dianisidine to give a product with a molar extinction coefficient at 460 nm of 1.16 X lo4 ~-l cm+ (11). A stock solution of 10 mM o-dianisidine in 3 mM EDTA is stable at 0". Just before use 0.2 ml of 20% Triton X-100 is added to 0.8 ml of o-dianisidine-EDTA stock solution. Since the dye decomposes rapidly in Triton, the latter solution is discarded at the end of the day. The Triton is used to keep in solution the rather insoluble chromogen formed by the peroxidatic reaction with o-dianisidine.
The assay mixture contains 0.12 ml of the orthodianisidine-EDTA-Triton solution, 0.10 ml of 0.3 M phosphate buffer, pH 7.0, 50 ~1 of horseradish peroxidase, 1 mg per ml (added last), water, and sample to total 3.0 ml.
Spectrophotometers-Static spectra were taken in a Cary 17 recording spectrophotometer at 25". Assays involving changes in absorbance at one wave length were done with a Beckman-Gilford spectrophotometer.

RESULTS
Anaerobic Reduction X&dies-When lactate oxidase and Llactate are mixed under anaerobic conditions there is a rapid appearance of a long wave length-absorbing species, followed 1 D. P. Ballou and G. Ford, to be published. The dissociation constant for the enzyme-lactate complex can be obtained from the slope/intercept of this plot and is 5 x 1OV M.
The value of kz is obtained from the intercept and is 14,000 min-I.
As will be shown in a later section, the turnover number for the enzyme-catalyzed reaction under the same conditions is 6,250 min-I.
Thus it is clear that the intermediate is formed sufficiently rapidly for it to be involved in catalysis; however, its slow anaerobic decay (2.5 min-r) is far too slow for the product of that reaction to be catalytically viable.
The above results were obtained in the presence of 0.01 M imidazole buffer, pH 7.0. This buffer was chosen because of the sensitivity of lactate oxidase to inhibition by a wide variety of anions. The use of phosphate buffer in a previous rapid reaction study of this enzyme led to erroneous interpretation of the results (4,13). Phosphate is an inhibitor of the lactate oxidase reaction, competitive with lactate and with a KC of 1.0 to 1.6 x lo+ M. The formation of an enzyme-phosphate complex can be monitored readily by the rat,her extensive spectral changes which occur on the addition of phosphate (Fig. 2), characterized by an increase in absorbance in the region of 340 to 487 nm and a decrease in absorbance at wave lengths greater than 487 nm. The inset of Fig. 2 shows absorbance ohanges at 450 and 510 nm as a function of phosphate concentration, plotted by the method of Benesi and Hildebrand (14). The results give a value for the dissociation constant of the enzyme-phosphate complex of 1.25 x 1O-2 M, very similar to that of the Ki value obtained from inhibition studies.
The effect of added phosphate on the course of anaerobic reduction of the enzyme by lactate is shown in Fig. 3. In the absence of phosphate a simple biphasic reaction is observed. The changes at 540 nm have already been described above; these changes are reflected by a biphasic decay of absorbance at 460 nm, with the same rate constants. When 0.01 M phosphate is present, the results become much more complicated.
The absorbance changes are now triphasic in nature.
Approximately half of the rapid absorbance changes which occur in the absence of phosphate still occur, with approximately the same apparent rate constant.
At 540 nm this is followed by a slow further increase in absorbance, which, however, never reaches the same level as in the absence of phosphate, presumably because its rate of formation is comparable with the normal slow decay of the intermediate.
When the phosphate concentration is increased to 0.1 M, less than 10% of the reaction goes in the fast phase, as monitored either at 540 or 460 nm. On the assumption that the extent of rapid reduction is due to uncomplexed enzyme, the fraction of enzyme in the form complexed with phosphate can be determined. The inset in Fig. 3b shows treatment of the data in this way, yielding a value for Kd of 0.9 X 10v2 M. The rate of the slow formation of intermediate observed in the presence of phosphate is independent of phosphate concentration and is 7.2 min-1.
The simplest interpretation of these results is that lactate cannot reduce the enzyme flavin when phosphate is bound, and that the rate of the slow formation of the intermediate in the presence of phosphate is in fact merely the rate at which phosphate is released from the enzyme-phosphate complex (k,ff).
If the above hypothesis is correct, an interesting prediction may be made. The Kd for the enzyme-phosphate complex is given by the ratio koff :kon. The spectrophotometric data of Fig. 2 show that Kd = 1.25 X 10e2 M. If the observed slow formation of the long wave length intermediate in the presence of phosphate is equal to lc,,ff thenit would be predicted that k,, would also be quite small. This was tested directly by following the rate of absorbance increase at 450 nm and absorbance decrease at 500 nm on mixing enzyme with phosphate.
Identical kinetics was obtained at both wave lengths; Fig. 4 shows the results obtained at 450 nm. The absorbance changes were indeed slow. Both the extent and rate of change varied with phosphate concentration. The apparent first order rate constant was not directly proportional to phosphate concentration, but was when the contribution of korf was subtracted (Table I). This is the situ- with enzyme 1.94 X 1w5 M with respect to enzyme-bound FMN and with an initial 02 concentration of 6.5 X 1w4 M. All experiments were performed in 0.01 M imidazole, pH 7.0, at 25". ation expected for approach to equilibrium, where kobs = k,, + koff (15). From these results, the second order rate constant, k,,, was determined to be -600 Me1 min+.
On the assumption that koff is equal to the rate of formation of the intermediate in the presence of phosphate (7.2 mini), Kd is therefore calculated on kinetic grounds to be 1.2 X lo-' M, in good agreement with the values obtained by spectrophotometric titration and from inhibition studies. While not investigated in the same detail, analogous results have been found with a variety of inorganic and organic anions which have been found to be inhibitors of lactate oxidase. These all produce detectable, and sometimes marked perturbations of the visible absorption spectrum of the enzyme. Values of dissociation constants, obtained in the same way as shown in Fig. 2, are given in Table II for some of these compounds.
Chloride ion is also found to bind weakly to lactate oxidase. For this reason we have worked at a low concentration of imidazole-HCl buffer (0.01 M), in order to minimize Cl-inhibition, while at the same time buffering the system.
Demonstration that Intermediate IS Complex of Reduced Enzyme and, Pyruvate-The identification of the nature of the long wave length-absorbing intermediate observed transiently on anaerobic reduction of the enzyme by L-lactate is of obvious importance, since on kinetic grounds it is the only observable species which could be an intermediate in catalysis. The spectrum of the enzyme found after disappearance of the intermediate is typical of the anionic form of fully reduced flavins (17) and is quantitatively indistinguishable from that found on reductive titration with dithionite or by anaerobic reduction with NaBH4 (18).  (14) to determine the dissociation constant of the reduced enzyme-pyruvate complex, as well as its maximal absorbance (obtained by extrapolation of the Benesi-Hildebrand plot to the ordinate).
The dashed line (Fig. 5, Curve 7) shows the extrapolated spectrum of the fully formed reduced enzyme-pyruvate complex.
At selected wave lengths are also shown the absorbance values observed in stopped flow studies at, the end of the fast stage of the reduction process. It is evident that the transient intermediate observed in anaerobic reduction with lactate is the same as that formed in an equilibrium reaction between reduced enzyme and pyruvate.
That this formulation is correct is supported by several lines of From the data of Fig. 5, the Kd value for the reduced enzymeevidence.
In Fig. 5  alternative way of estimating Kd would bc by following the kinetics of the spectral changes illustrated in Fig. 5. The results of such a study, carried out at a slightly lower temperature, 20", are shown in Fig. 6. At this temperature, the rate of decay of the intermediate is 2 min-I. When reduced enzyme and pyruvate were mixed anaerobically and the increase in absorbance at 540 nm monitored in the stopped flow apparatus, apparent first order rate constants were obtained.
When correction was made for the real first order process (ka = 2 min-I), involved in the approach to equilibrium, it is apparent that a second order reaction was occurring between reduced enzyme and pyruvate, with a rate constant, k-3 of lo3 M-I min-I.
Thus, the value of Kd determined kinetically, 2 X 1OV M, is very similar to that obtained in the static equilibrium tit,ration experiment of Fig. 5. A striking difference in the rate of reaction with O2 is displayed by free reduced enzyme and reduced enzyme in complex form with pyruvate. This is illustrated in Fig. 7, in which reduced enzyme titrated with various concentrations of pyruvate was mixed with air-equilibrated buffer and reoxidation monitored in the stopped flow apparatus at 460 nm. The reoxidation is markedly biphasic, the extent of the reaction occurring in the rapid phase of the reaction being markedly influenced by the pyruvate concentration.
On varying the 02 concentration it was found that the reaction with 02 is second order for both the fast and slow phases. The reduced enzyme in complex form with pyruvate was found to react with O2 approximately 200 times faster than the uncomplexed reduced enzyme. At pH 7.0, 25", the second order rate constants were fomld to be 1.1 x lo8 M-l min-l and 5.4 X lo5 M-l min-', respectively. If the proportion of the reaction occurring in the fast phase is assumed to represent the proportion of reduced enzyme in complex with pyruvate, the data of Fig. 7 can also be used to calculate the Kd for the reduced enzyme-pyruvate complex. Treatment of the data in this way is shown in the inset of Fig. 7; the Kd value so obtained is 2.2 x 1OP M, in good agreement with the results already described, and obtained by different experimenbal methods.
Results from Stopped Flow Turnover ~.~perit.r~enfs-~~~l~n lactate oxidasc is mixed with excess reducing substrate in the presence of 02, a steady state level of oxidized enzyme and intermediate is rapidly established, which persists for varying lengths of time depending on the initial concentrations of 02 and reducing substrate used. Fig. 8 shows the results of such an experiment with lactate as substrate, at an initial 02 concentration of 2.6 x lo-4 M, and monitoring the absorbance at 460 nm. When in the same experiments the absorbance changes were monitored at 540 nm very little intermediate was detected in the steady state, but gradually accumulated as the 02 was depleted.
It is evident from Fig. 8 that most of the enzyme is in the oxidized form during the steady state. This is in agreement with the findings already documented, that the reaction of intermediate with OS (see previous section) is much faster than the rate of formation of the intermediate from oxidized enzyme and lactate (see Fig. 1). Following the principle enunciated by Chance (8) and as detailed by Gibson et al. (9), the data of Fig. 8 Fig. 8 yields the series of parallel Lineweaver-Burk plots shown in Fig. 9. A secondary plot of the intercepts at infinite O2 concentration (from Fig. 9) is given in Fig. 1 that produced by lactate, and may be generated in static equilibrium by addition of ac-keto-j%methylvalerate to reduced enzyme.) The phenomenon of the turnover numbers being consistently lower than the corresponding rates of formation of the intermediate, and the parallel nature of the reciprocal plots, is well illustrated with this substrate.
A striking feature of the turnover results with this substrate, compared to lactate, is that although it clearly leads to intermediate formation at a much slower rate, nevertheless, less enzyme is in the oxidized form in the steady state than when lactate is employed as substrate (compare Figs. 8 and 10). This would imply that the intermediate in this case must react with 02 considerably more sluggishly than does the reduced enzymepyruvate complex.
Indeed this was found to be the case. 10. Stopped flow turnover studies with n-a-hydroxy-Pmethylvalerste.
The conditions were identical with those described for Fig. 8, except for the different substrate.
The initial A460 value of oxidized enzyme was 0.069.  Fig. 9). The other points (0) were turnover numbers obtained by extrapolation to infinite oxygen concentration, employing the conventional 02.electrode assay for activny.
When reduced enzyme was titrated with a-keto-fi-methylvalerate and reacted with O2 at 25", the second order rate constant was found to be 2.7 x lo6 Me' min-i. This value should be compared to that for the reduced enzyme-pyruvate complex of 1.1 X lo8 M-l min-l.
The finding of parallel Lineweaver-Burk plots such as shown in Fig. 9 (and found for all substrates) is that expected for a binary complex mechanism (19,20) in which the first product dissociates from the reduced enzyme before reaction of the latter with the second substrate (OS). However, as pointed out previously, there are limiting cases of ternary complex mechanisms which also yield parallel Lineweaver-Burk plots (21,22). The results obtained here are very reminiscent of those obtained previously with n-amino acid oxidase, which was also shown to operate by a ternary complex mechanism in which dissociation of oxidized product from the reoxidized enzyme was the ratelimiting step in catalysis (5,22). In the present study the reaction of O2 with intermediate has been shown to be second order. The minimal (and undoubtedly oversimplified) mechanism consistent with the observed results is therefore While kq is different depending on whether lactate or cu-hydroxy-P-methylvalerate is employed as substrate, the observed turnover numbers cannot be explained by the rate-limiting step being either kz or kd; in both cases the expected turnover numbers would be practically indistinguishable from the observed rates of formation of intermediate.
The results can be explained very satisfactorily, however, by assuming that the rate-limiting step is ks. For the above reaction sequence, the initial rate equation following Palmer and Massey (22)  The conditions for obtaining parallel Lineweaver-Burk plots (such as shown in Fig. 9, and found for all substrates) is that the last term in the denominator of the initial rate Equation 7 be negligibly small. This is entirely in keeping with the large value of kd compared t'o the other known or calculated rate constants which go to make up K (in view of evident difficulty of reversing the anaerobic reduction sequence (cf. Fig. 5) it is also reasonable to assume that kWz has a very small value).
The above mechanism also accounts very satisfactorily for the observation that with all substrates, the reciprocal plot of catalytic turnover number and reducing substrate concentration is parallel to the similar reciprocal plot of observed rate of reduction to the intermediate. This is due to the dissociation of product from the reoxidized enzyme (kJ being at least partly ratelimiting.
On the assumption that the above mechanism is correct, kh may be calculated from the known V,,, and the determined value of kz by the use of Equation 8. For lactate as substrate, the value of kj calculated in this way is 11,300 min?. This calculated value may then be used in Equations 9 and 10 to predict values for KR CHOH oOOH and Ko2.
For Equation 9 , absolute values for kl and Ll are not available, but their ratio is. Thus, the Kd for lactate (k-i/kl) from Fig. 1, is 5 X 1OP' M. The other rate constants in Equation 9 are available, either as direct'ly observed or as calculated values. If one then assumes various values for kl (and hence k-1) one can therefore determine the minimal values of kl and k-1 required to obtain a reasonable fit to the observed Michaelis constant, Klactate, of 2.22 x 10P2 3%. Values of lil of 106, 107, and lo8 nrP mine1 substituted in Equation  9 yield Klactate values of 2.85 x 10e2, 2.31 X 10M2, and 2.26 x lo+' M, respectively.
Thus the minimal value of kl required to fit the observed Klactate value is lo7 M-' mini, an entirely reasonable value for enzyme-substrate interaction and orders of magnitude lower than would be permitted in a diffusion-controlled process. A more telling test of the mechanism is the prediction of K,, from the kinetically determined rate constants. It should be remembered that the reaction of 02 with the intermediate appears t,o be second order, i.e. there is no experimental evidence for 0% actually forming a complex with the intermediate; if such a complex is formed it must have a very large dissociation constant.
Nevertheless a Michaelis constant for 02, Ko2, is readily measured, and with lactate as substrate, is determined as 7.1 x 10-j 31. All kinetic constants in Equation 10 are known, except for L.
If it is assumed that k-2 is small compared to the other rate constants, an assumption which has been discussed previously, then KOz can be calculated to be 5.5 X 10-S M. Considering that this value is derived from three independently meas-ured or calculated rate constants, the agreement between the calculated and experimental KOz is regarded as being very satisfactory. The calculation of Koz requires knowing k.+ The only other substrate for which kq has been determined is a-hydroxy-/% methylvalerate, where a much lower value was found than with lactate.
In this case the calculated value of Koz (8.1 x 10W5 M) was also in very good agreement with the experimentally determined value (8.0 X 10-j M).
The values of rate constants, experimentally determined or calculated, for the four substrates investigated in some detail, are shown in Table III.
It can be seen that in all cases the dissociation of product from the reoxidized enzyme, kg, is the step which largely controls the rate of over-all catalysis. with Oxygen-It is well known that Hz02 will decarboxylate keto acids in a nonenzymic reaction.
Hence the possibility existed that the primary products of the lactate oxidase reaction were free Hz02 and pyruvate and that these reacted independently of the enzyme to give the final products of acetate, COZ, and HzO. That this was not the case was shown in several ways. We tested the rate of reaction of HzOs with pyruvate under conditions even more favorable for nonenzymic reaction than those used for the enzyme-catalyzed reaction.
With the o-dianisidineperoxidase assay for H20z, we could detect no reaction in 30 min when 10e4 M H202 and 10P4 M pyruvate were incubated at 0" in 0.01 31 imidazole, pH 7.0, and only 30% destruction of Hz02 over 30 min with 10-4 M Hz02 and 1O-3 M pyruvate.
At 25", the reaction between 10-4 M H202 and 1OP M pyruvate was also very slow (13% destruction of I&O2 in 30 min).
In contrast, when 1O-5 M lactate oxidase is added to air-equilibrated 1OP M lactate, the reaction is complete within 10 s (to yield stoichiometric amounts of acetate and COZ). Clearly, the enzymecatalyzed reaction is much faster than the reaction of free pyruvate with free HzOz; therefore this pathway is eliminated as a significant route of product formation.
A second approach was to reduce enzyme with n-lactate under anaerobic conditions, wait a few minutes for the intermediate to disappear (i.e. allow time for pyruvate to dissociate from the reduced enzyme), then mix with air and assay for pyruvate and H202. The results, in Table IV, show that pyruvate and Hz02 are produced in amounts stoichiometric with the enzyme flavin. These results support the conclusion stated above and add further weight to the conclusions derived from kinetic experiments,   In a typical experiment, 1.18 X 1O-4 M enzyme was reduced anaerobically at 25" with a stoichiometric amount of n-lactate, then a lo-fold M excess of [i4C]pyruvate (labeled in the C-2 position and containing 1.22 x lo6 dpm) was added to produce the intermediate. Air was then admitted, and the entire volume of 2.6 ml immediately applied to a Dowex column as described under "Materials and Methods." The acet.ate neatly separated from other radioactive components with a total count of 74,500 dpm. From the results described, we feel that the following reactions as shown in Scheme 1 constitute a minimal mechanism for the reactions catalyzed by lactate oxidase. There are several lines + co 2 + Hz0 SCHEME I the data of Fig. 5 that actually 55% of the enzyme was in complex with pyruvate. This would correspond to an expected count of 60,000 dpm. The fact that more radioactivity is recovered in t,he acetate fraction than expected may be due to competition between 02 reacting with uncomplexed reduced enzyme and more pyruvate reacting with that reduced enzyme not previously bound.
Whatever the explanation, the tot,al counts recovered in the acetate fraction is still less than that corresponding to 1 mole per mole of total reduced enzyme originally present (108,000 dpm).
It is easy to discount the possibility that [14C]acetate was derived from equilibrium concentrations of [l*C]lactate from the possible equilibria, EFMNHz + pyruvate j intermediate $ EFMN j-lactate, since such L-of evidence which argue strongly that the long wave lengthabsorbing intermediate is the species of enzyme which reacts with O2 in the catalytic sequence. It is the only spectroscopically detectable intermediate produced sufficiently fast to be involved in cat.alysis. With all substrates, the rate of formation of free reduced enzyme (ka) is orders of magnitude slower than catalysis, making it unlikely that free reduced enzyme is involved in catalysis.
The catalytic noninvolvement of free reduced enzyme is also emphasized by its comparatively slow rate of reaction with 02, which is also too slow to account for catalysis. Furthermore, as shown by product analysis, reaction of this form of the enzyme with O2 results in stoichiometric formation of Hz02 rather than H,O, the normal product of catalysis. In contrast, the intermediate, formed by mixing reduced enzyme with excess pyruvate, reacts with 02 some 200-fold faster than the reduced enzyme, and acetate is the product of the reaction, together with Hz0 and COe. Thus a highly favorable situation exists in the enzyme for the normal catalytic reaction to occur, rather than the production of free keto acid and HzOz, since the intermediate, produced rapidly, is also capable of rapid reaction with OS, whereas free reduced enzyme and pyruvatc are formed only slowly, and free reduced enzyme, once formed, is capable of only slow reaction with OS.
The nature of the long wave length-absorbing intermediate is of obvious interest.
Its spectra.1 characteristics, with little absorbance at 450 nm, indicat'c that it is a form of reduced fiavin. That it can be produced in an equilibrium reaction with keto acid argues Ohat it may well be a charge transfer complex between reduced enzyme flavin and keto acid. Whatever its nature, it is evident that ket.0 acid is bound to the reduced enzyme in close juxtaposition to the FMNH2. Reaction of the latter with O2 to produce HsOz would then seem to be a reasonable postulate to account for catalysis, since the two products, Hz02 and keto acid, still bound to the enzyme, might be expected to react much more effectively than in free solution, to produce the oxidative decarboxylation products typical of this enzyme. With all substrates tested (cf. Table III) a step in catalysis which is at least partly rate determining, is the dissociation of products (k5). This step is responsible for the fact that in reciprocal plots such as illust.ratcd in Figs. 9 and 11 the plot for turnover number is parallel to that for the rate of production of the intermediate.
As discussed under 'LResults," the determined and derived kinetic constants are fully consistent with this mechanism.
The finding of distinctly different values of k5 depending on the substrate used argues strongly for kS being the rate constant associated with the release of R.COOH from the oxidized enzyme; if the release of COZ or Hz0 had been rate limiting, the same value of Icb would have been found for all substrates.
The finding of such a rate-limiting step is interesting in view of the similar phenomenon found previously with n-amino acid oxidase (5).
As stated above, the proposed scheme represents a minimal reaction pathway.
The mechanism of substrate activation by the enzyme in the early stages of the catalytic sequence is now under investigation in collaboration with Dr. R. H. Sbeles and his associates; in analogy to studies with D-and L-amino acid oxidases (23), Pmchlorolactate has been found t'o bc a substrate for lactate oxidase under both aerobic and anaerobic conditions. The products of the anaerobic reaction are pyruvate and Cl-; in the presence of O2 the products are H?O, CO*, and chloroacetate. Thus, as with D-and L-amino acid oxidases (23), it is probable that the primary step in the lactate oxidase reaction involves the abstraction of a proton from the substrate by the enzyme. This work will be reported in full in a separate communication.2