The Mechanism of Action of the Flavoprotein Melilotate Hydroxylase*

reaction mechanism of melilotate hydroxylase of two ternary complexes


VINCENT MASSEY
From the Department of Biological Chemistry, The University of Michigan, Ann Arbor, Michigan 4810.4

SUMMARY
The reaction mechanism of melilotate hydroxylase has been investigated by a variety of kinetic methods. A steady state analysis has indicated that the enzyme has a mechanism involving two ternary complexes but not a quatemary complex of the enzyme and the three substrates.
The reduction of the enzyme-melilotate complex by NADH has been studied by measuring fluorescence changes in the stopped flow apparatus.
This reaction is second order and very rapid.
The product of this reaction, observed by stopped-flow spectrophotometry, is believed to be a chargetransfer complex between the reduced FAD of the enzyme and NADf.
It has a broad, long wave length band centered at 750 nm and little absorption at 450 nm. This long wave length band disappears with a rate independent of the NADH concentration; this is presumably due to the dissociation of NAD+ from the complex.
A charge-transfer complex can also be formed by anaerobic titration of reduced enzyme with NAD+ or 3-acetylpyridine-NADf.
The charge-transfer band is sensitive to the nature of the pyridine nucleotide.
The spectrum of the complex between reduced enzyme and NAD+ appears to be identical with the spectrum observed transiently during the reduction of the enzyme by NADH in the stopped flow apparatus.
The reaction of reduced melilotate hydroxylase with molecular oxygen in the absence of melilotate is a second order reaction.
The rate of this reaction is enhanced more than lo-fold by the presence of melilotate, and the reaction profile becomes more complex. This complexity is due to the formation of an intermediate, which is believed to be an adduct of molecular oxygen to the reduced FAD of the enzyme.
The spectrum of this intermediate has been determined by analog computer simulation using experimentally determined rate constants for its formation and decay, as well as the known extinction coefficients for reduced and oxidized enzyme.
A reaction mechanism is proposed based on the steady state analysis.
This analysis enabled the kinetic constants for the reaction to be determined.
These kinetic constants have also been predicted from individually determined rate constants assuming the proposed mechanism.
The values * This work was supported by Grant GM11106 from the United States Public Health Service.
$ Recipient of a National Science Foundation Predoctoral Fellowship.
obtained by these two methods are in excellent agreement. They also correlate with the &iss previously determined for the dissociation of melilotate from the enzyme-substrate complex.
Melilotate hydroxylase is a flavoprotein which catalyzes the conversion of melilotate (2-hydroxyphenylpropionate) to 2,3dihydroxyphenylpropionate (1). The enzyme isolated from Pseudomonas has many properties in common with other bacterial flavoprotein hydroxylases.
The most noticeable among these similarities is that melilotate forms a 1:l complex with the enzyme and that the rate of reduction of this complex by NADH is approximately 10S-fold faster than the reduction of free enzyme (2).
In the past several years, a number of kinetic investigations of flavoprotein hydroxylases has been published. The reaction of p-hydroxybenzoate hydroxylase (Pseudomonas Jlwrescens) with its effector 6-hydroxynicotinate has been the subject of a detailed steady state analysis (3), and this enzyme has also been studied extensively by stopped flow spectrophotometry (3)(4)(5). Likewise, p-hydroxybenzoate hydroxylase from Pseudomonas desmolytica (6) and salicylate hydroxylase from Pseudomonas putidu (7) have been investigated by steady state and rapid reaction techniques.
Finally White-Stevens et al. have made extensive use of stopped flow spectrophotometry to investigate salicylate hydroxylase obtained from an unidentified soil organism (8).
The present paper reports a kinetic investigation of melilotate hydroxylase by steady state and rapid reaction techniques.
A reaction mechanism is postulated from a steady state analysis and corroborated by measurement of the individual rate constants in the stopped flow spectrophotometer.
In addition, these rapid reaction studies have revealed the presence of two intermediates in the catalytic cycle. The first appears to be a charge-transfer complex between reduced melilotate hydroxylase and NAD+. The second is postulated to be an oxygenated flavin intermediate composed of enzyme, melilotate, and oxygen, similar to the one proposed for p-hydroxybenzoate hydroxylase (4,5). The reoxidation studies have further shown that the presence of melilotate increases the rate of reaction of the reduced enzyme with molecular oxygen and that its presence is necessary for the formation of the oxygenated intermediate.

PROCEDURE
J~aterials-i\elilotate and melilotate hydrosylase (2) and glucose oxidase (9) were prepared as previously described. NADH (Grade III) was purchased from Sigma; catalase (B grade) from Calbiochem; and 3-acetylpyridine-NAD from P-L Biochemicals. Steady State Experiments-The steady state experiments were done by two methods, both at 1-1.5".
Experiments in which the concentration of oxygen was held constant were done with a Cary 17 recording spectrophotometer at 340 nm. All reactants except the enzyme were added to the cuvette, which was then incubated at 1" in the sample compartment of the spectrophotometer for 5 min. The reaction was started by the addition of enzyme.
The sample compartment was constantly flushed with dry air to prevent condensation of water on the optical surfaces. Those experiments in which the concentration of oxygen was varied were done with the stopped-flow apparatus of Gibson and Mimes (10). The electronics of the instrument were replaced with a solid state circuit which includes an internally calibrated, logarithmic operational amplifier.' This circuit produces an output directly proportional to absorbance. The enzyme solution, air-saturated at l", was placed in one syringe of the stopped flow apparatus.
Buffer, containing the appropriate amount of melilotate and NADH, was equilibrated with one atmosphere of 5% oxygen (95% nitrogen), 10% oxygen (90% nitrogen), 21 "/o oxygen (air), or 100% oxygen in a tonomet.er at 25". This solution was transferred to the stopped flow apparatus and equilibrated for 23 min at 1". The final concentration of oxygen was calculated from the known solubilities of oxygen in water at the appropriate temperatures. The concentration of enzyme was chosen so that the reaction took place over approximately 10 to 60 s. The concentrations of NADH, melilotate, and melilotate hydroxylase were determined spectrophotometrically using their respective extinction coefficients: NADH, 6220 M-' cm-1 at 340 nm (11) ; melilotate, 1910 M-' cm-l at 271 cm (1) ; and melilotate hydroxylase, 11,300 Me' en-i at 450 run (2). The steady state results are expressed as turnover numbers, i.e. moles of NADH oxidized per mm per mole of enzyme-bound FAD.
The initial rate equations were derived by the schematic method of King and Altman (12).
Stopped FZou, Experiments-All stopped flow experiments were done at l-1.5" with the unit described above. Anaerobic techniques were as previously described (2) except that the nitrogen gas was purified by storage over Fieser's solution (13). For stopped flow measurements at wave lengths longer than 600 mu, a Corning 3-70 glass filter was placed directly in front of the face of the photomultiplier tube to eliminate second order light from the monochromator.
The fluorescence experiments were done with the stopped flow apparatus using a fluorescence observation cell based on the design of Gibson et al. (14). The tungsten ribbon filament lamp was replaced with a 150~watt xenon arc powered by a Hewlett Packard 6267B power supply.
The starting voltage for the lamp was provided by a custom-made unit.' The xenon lamp was mounted on a table independent of the mixing block to reduce vibrational noise associated with the hydraulic driving mechanism.
For these experiments, excitation light was monochromated at 340 nm by the normal stopped flow monochromator; emitted light was filtered by a Corning 3-75 glass filter to eliminate contributions from the exciting light. To insure complete anaerobiosis in the st,opped flow apparatus, glucose and glucose oxidase were routinely added to the tonome-ter containing the enzyme solution.
In the tonometers containing other reactants, agitation was vigorous enough during the anaerobiosis procedure to make this unnecessary.
However, in the fluorescence experiments done at low concentrations of reactants, this oxygen scrubbing system was included in both tonometers.
The tonometer used for the enzyme solution in these experiments had a ground glass joint which allowed attachment of a spectrophotometer cell. This arrangement enabled anaerobic, static spectra to be recorded of the solution in the tonometer. For reoxidation studies, the enzyme was reduced either by a 2fold excess of NADH or by irradiation with visible light in the presence of EDTA (2). Identical results were obtained for 10th methods.
Anaerobic Titrations-Anaerobic titrations were done in the unit described by Foust et al. (15). Previously described techniques were used (2). NllDf and AP-NAD+2 were dissolved in 0.1 M KP;, pH 7.3, and were standardized using extinction coefficients of 17,800 M-I cm-l at 259 nm and 16,400 M-' cnTr at 260 nm, respectively (16). The enzyme solution was placed in the cuvette assembly, made anaerobic, and photoreduced in the presence of EDTA.
When reduction was complete, the titration was begun by adding pyridine nucleotide.
Analog Computer Simulations-Computer simulations were done on an Applied Dynamics AD2-64PD analog computer console. The reaction scheme simulated was: where E is melilotate hydroxylase.
The appropriate values for the rate constants lc, and ks are shown in Table II. Previously determined extinction coefficients for reduced and oxidized melilotate hydroxylase were used (2).
Spectrophotometric Jlfethods-Routine enzyme assays were performed as previously described (2) with a Gilford recording spectrophotometer.
Static spectra were taken with a Cary 17 recording spectrophotometer.
For wave lengths longer than 700 nm, the infrared detector and multipotentiometers were used.

Steady
Initial rate equations derived for other mechanisms may be considered special cases of this general form. The kinetic parameters, #JO, 4a, etc., can be evaluated by plotting the reciprocal initial velocity versus the reciprocal concentration of each substrate at fixed concentrations of the other two (17).
Lineweaver-Burk plots (18) in which the concentrations of melilotate and NADH were varied at a fixed concentration of oxygen are shown in Fig. 1 The arrow on the main figure shows the calculated Ediss of melilotate from the enzyme.
concentration of NADH (Fig. 2), and (b) the concentrations of NADH and oxygen were varied at a fixed concentration of melilotate (Fig. 3). Both of these latter two series yielded sets of parallel lines.
The initial rate equation which fits these results has the form: A reaction scheme which agrees with this initial rate equation must include a ternary complex of enzyme, NADH, and melilotate (17). One mechanism which is consistent with this analysis, written in the shorthand notation of Cleland (19) This type of reaction is called either a concerted-substitution, type II, b mechanism (17)) or a Bi Uni Uni Bi Ping Pong mechanism (19). A concerted reaction of the enzyme with two of the substrates occurs which forms an altered form of the enzyme, and this altered form then reacts with the third substrate in a separate step.
The values of the kinetic constants for the reaction mechanism can be determined from secondary plots of the initial rate data (17). Plots of this nature are shown in the insets to Figs. 1, 2, and 3. The kinetic constants so determined are summarized in Table I.
For the Lineweaver-Burk plots shown in Fig. 1  plex. The value determined for this Kdiss from the data in Fig. 1 is 3.6 X lop5 M; and Kdiss measured by a spectrophotometric titration of the enzyme with melilotate is 3.8 X low5 M. This relationship holds only if melilotate binds to the enzyme prior to NADH in an obligatory order mechanism (20) or if melilotate and NADH bind in random order to the enzyme (21). The possibility of a random addition of the first two substrates can be ruled out by replotting the data of Fig. 1 as reciprocal turnover number uersus reciprocal NADH concentration at several melilotate concentrations.
These plots will of course also give a set of converging lines. If the mechanism is truly random order for the first two substrates, however, the negative reciprocal of the horizontal coordinate of the point of intersection in this plot should now be equal to the Kdiss for the NADH-enzyme complex (21). The value determined from a replot of the data of Fig. 1 as described is 6.3 x 10m5 M. This is not in agreement with the measured Kdiss for the NADH-enzyme complex, 3.8 X 10P4 M (2). Thus, it appears that the substrate addition occurs by an obligatory order mechanism, with melilotate binding first.

Reduction of Melilotate
Hydroxylase by NADH-The reduction of melilotate hydroxylase by NADH is a relatively slow process in the absence of melilotate, having an extrapolated rate constant of 1.4 min-1 at lo (2). In the presence of melilotate, the reduction is extremely rapid and was impossible to follow in the stopped flow spectrophotometer at the concentrations normally employed.
However, by lowering the concentrations of the reactants and observing changes in fluorescence rather than absorbance, it was possible to study this reaction in two ways. In the first experiments, NADH and melilotate hydroxylase were reacted at equal concentration in the presence of melilotate. The disappearance of the NADH fluorescence obeys the kinetics for an irreversible, second order reaction as shown in Fig. 4A. The second order rate constant determined for this process is 1.4 X lo* ~-1 min+.
In these experiments, the possibility existed that the disappearance of the NADH fluorescence was associated with XADH binding to the enzyme, rather than the actual oxidationreduction reaction.
Therefore, an experiment using concentrations of the two reactants identical with those described for Fig. 4A  and 450 nm was observed.
The changes in absorbance at these two wave lengths.are identical in rate and order with the changes in fluorescence.
Thus, this fluorescence decrease is associated with the oxidation of NADH and the reduction of the enzyme flavin.
Another series of experiments, also using the stopped flow fluorescence technique, was performed in which the NADH concentration was held constant and the enzyme concentration varied.
The enzyme concentration was always at least five times as large as the NADH concentration (pseudo-first order conditions).
The enzyme was used as the reactant in excess to avoid the background fluorescence that would be associated with excess NADH.
A plot of the reciprocal of the observed first order rate constant versus the reciprocal of the enzyme concentration (22, 23) was linear and passed through the origin. This is further evidence that the reaction is second order. These experiments yielded a rate constant of 1.4 x lo8 M? min-1, identical with the value determined under second order conditions (Fig. 4A).
In order to correlate the stopped flow results with the steady state results (see "Discussion"), it was necessary to evaluate the magnitude of the rate constant for the reverse reaction.
il convenient graphical method for this purpose is as follows.
Consider a reversible, second order reaction: where 121' g kl [B] (pseudo-first order approximation) (7) If one assumes that the concentration of C at the beginning of the reaction is zero, the rate equation can be easily integrated and simplified by introducing the equilibrium condition (24). The resulting integrated rate equation is: In (A -A,) = -(k,' + k2)t + 111 (A,, -A,) where A0 is the initial concentration of A, A is the concentration at any time t, and A, is the concentration at equilibrium. Thus, the approach to equilibrium is a first order process, and the observed rate constant from a plot of ln A versus t will be the sum of the rate constants for the forward and reverse reactions. k ohs = k,' + kz (9) or k,bs = h[Bl + k, Therefore, a plot of kobs versus [B] should be linear with a slope of kl and a y intercept of kz. Such a plot for the reduction experiments performed under pseudo-first order conditions is shown in Fig. 4B. The least squares regression analysis estimate of the y intercept of this plot is -0.20 f 0.28 (95% confidence interval, Reference 25). Thus, within the limits of error in these experiments, the rate constant for the reverse reaction can be considered zero (kr, Fig. 10). When the reduction of melilotate hydroxylase by NADH is studied by the usual stopped flow spectrophotometric techniques, approximately 90% of the absorption at 450 nm associated with oxidized enzyme is lost in the 3-ms dead time of the instrument. However, a new absorption band at longer wave lengths appears during the dead time and then disappears more slowly.
The spectrum of the intermediate which is formed in the first 3 ms after mixing was determined as described previously (4) I  I  I  I  I  I  I  I  I  I  I  I I  I  I  I  I  I  350  400  450  500  550  600  650  700  750  800  850  900  950  shown in Fig. 5. The intermediate has a lower extinction coefficient, at 450 nm than fully reduced enzyme and has a very broad long wave length band centered at about 750 nm. The rate of disappearance of the long wave length band is 1300 min-l and is independent of NADH concentration over the range tested (0.25 to 1 .O mM). It seemed probable that this absorption band is associated with a charge-transfer interaction between the reduced FAD of the enzyme and NAD+.
The slower decay of the absorption band could then be attributed to the dissociation of N,4D+ from the complex.
The inset to Fig. 5 shows two similar experiments in which the long wave length absorption band was observed.
In initially (until a steady state level is reached) at the same rate as anaerobically. This implies that during catalysis oxygen reacts with the reduced enzyme after the dissociation of NAD+ has already occurred.
Charge-Transfer Inkm&n between Reduced Melilotate Hydroxylase and NAD+-If the long wave length band observed in the stopped flow apparatus is due to a charge-transfer interaction as postulated, it should be possible to form this species in static experiments.
Indeed, when an anaerobic solution of reduced melilotate hydroxylase (produced by EDTA-light irradiation) is titrated with NAD+, a new absorption band does appear (Fig.  6). The similarity of the spectrum of this species to that observed in the stopped flow apparatus (Fig. 5)  .4 complex similar in nature to the above is formed when AP-NADf is added to an anaerobic solution of reduced enzyme. III this case, some reoxidation of the enzyme flavin occurs due to the fact that AP-NAD+ had a higher oxidation-reduction potential than NAD+ (27). The broad, long wave length band in t.his case is centered at about 630 nm. The dissociation constant of this pyridine nucleotide from the reduced enzyme was not determined, but from the failure to approach full formation at the concentrations used is clearly larger than that for NAD+.
Oxidation of Reduced Melilotate Hydroxylase-Melilotate hydroxylase which has been reduced by NADH or by irradiation in the presence of EDTA is reoxidized by molecular oxygen in a second order fashion in the absence of melilotate (Fig. 7A) The calculated second order rate constant is 9.7 X lo5 M-' min-I.
In the presence of melilotate, the reoxidation reaction is corlsiderably more complicated (Fig. 8). At 450 nm, a lag phase is observed followed by a process whose rate is independent of the osygen concentration.
The first order rate constant for this process is 1400 min-l.
At 405 nm, a rapid increase in absorbance is followed by a slight decrease. The rapid phase shows a direct dependence on oxygen concentration and has a second order It is possible, however, to obtain the spectrum of the intermediate by another method.
For the reoxidation process described above, the intermediate formed is essentially isosbestic with the oxidized enzyme at 415 nm. Therefore, the rate constant for the formation of the intermediate from reduced enzyme can be obtained from observations at this wave length (1.6 x lo7 M-' min-l).
The rate constant for the decay of the intermediate to oxidized enzyme can likewise be determined from observations at 490 nm, a region where the reduced enzyme is isosbestic with the intermediate (1400 min-') . The known parameters of the reoxidation reaction are thus the absorption spectrum of the initial and final enzyme, and the rate constants for both the formation and decay of the intermediate.
(At different oxygen concentrations, it is necessary to change the rate constant for the formation of the intermediate; Fig. 7B). By analog computer simulation, it is then possible to obtain an accurate representation of the absorption spectrum of the intermediate.
The assumed extinction coefficient of the intermediate at each wave length was varied until the computer simulation matched the actual display observed on the oscilloscope of t.he stopped-flow apparatus.
An excellent fit of the simulated data to the real data could be obtained at all wave lengths by varying the extinction coefficient of the intermediate. Examples of this agreement are shown in Fig. 8. The calculated absorption spectrum of the intermediate has a maximum at 405 nm and an extinction coefficient at that wave length of 8500 UISCUSSION The steady state results reported in this paper suggest that melilotate hydroxylase has the reaction mechanism shown in Fig. 10. This mechanism involves two different ternary complexes but not a quaternary complex of the enzyme and its three substrates.
There is an alternate mechanism which would fit the initial rate data. However, this mechanism is inconsistent with other known facts about the enzyme. Thus, Dalziel's concertedsubstitut,iou II, c mechanism (Cleland's Bi Bi Uni Uni Ping Pony; mechanism) would have both products from the first ternary complex released before the addition of the third substratc.
This clearly cannot be the case with melilotate hydroxylase. hlelilotate must be bound in a ternary complex with oxygen since the source of the oxygen atom in the incorporated hydrosyl group is molecular oxygen (1).
Previously, the order of substrate addition has been investi- The order of product release (for DiOH and H20) is an arbitrary assignment. gated for p-hydroxybenzoate hydroxylase. Nakamura et al. (6) reported that the addition of p-hydroxybenzoate and NADPH to this enzyme is ordered, with the aromatic substrate binding first. However, this conclusion was based on the fact that the enzyme-p-hydroxybenzoate complex is reduced much more rapidly than the free enzyme, which is not conclusive evidence. For example, Howell et al. (3) have shown by steady state analysis that with the nonsubstrate effector 6-hydroxynicotinate, which does enhance the rate of reduction of the enzyme by NADPH, the addition of the effector and pyridine nucleotide is random.
The present work documents that with melilotate hydroxylase the substrate melilotate binds first in an ordered addition process.
The reduction of melilotate hydroxylase by NADH (k~) appears to be effectively a second order, irreversible reaction. Rigorously speaking, a complex of oxidized enzyme and NADH must exist.
However, if the collapse of this oxidized enzyme complex to the reduced enzyme complex is very rapid compared to its formation, an over-all second order process will be observed. The complex of oxidized enzyme and NADH will then be kinetically invisible.
The magnitude of the reverse rate constant kd (Fig. 10) was zero within the limits of detection reported here. In any case, it is clear from the 95% confidence interval of the y intercept of Fig. 4B that led can be no larger than approximately 30 to 40 min-I.
The intermediate species observed during the reduction of melilotate hydroxylase by NADH (Fig. 5) is certainly a complex between reduced enzyme and NAD+.
This conclusion is based on the observed absorption at 456 nm and on the fact that an identical species is formed upon anaerobically mixing reduced enzyme and NAD+ (Fig. 6). To prove that this interaction is of the charge-transfer type, several conditions must be met. First and foremost, the complex must have an absorption band not present in either the donor (FADHJ or acceptor (NAD+) molecule (28). The very broad absorption band centered at 750 nm for this complex obviously satisfies this requirement. Another good test for charge-transfer bands is that they should by guest on March 23, 2020 http://www.jbc.org/ Downloaded from be sensitive to substituents on the donor and acceptor molecules (28). The absorption band for the complex between reduced enzyme and AP-NAD+ has a wave length maximum of 630 nm, a shift of 120 nm. At first glance, the fact that AP-NAD+ shifts the absorption band to the blue may appear contradictory to the hypothesis.
The reduction potential (30", pH 7.0) for the NAD+ system is -320 mvolts (29); for the AP-NAD+ system, it is -248 mvolts (27). Due to its more positive reduction potential, one might expect the charge-transfer interaction between AP-NAD+ and reduced flavin to be facilitated and the new absorption band to have a maximum at a longer wave length than for NAD+.
However, the origin of the charge-transfer band must be kept in mind.
For weak charge-transfer interactions, the absorption arises from a transition which is effectively a one-electron transfer from the donor to the acceptor molecule (30). Therefore, the energy required for this transition will be sensitive not only to the reduction potentials of the donor and acceptor but also to the geometry of binding.
It may be in this case that AP-NAD+ is bound to the enzyme in a less favorable conformation than NAD' for partial electron transfer to occur. Thus, even though AP-NAD+ has a more favorable reduction potential for the process, more energy may be required.
The crucial fact is that the nature of the new absorption band is sensitive in some manner to changes in the acceptor molecule.
It should be noted that the spectrum reported here for the reduced enzyme-NAD+ complex is very similar to the spectrum of a complex between reduced lipoyl dehydrogenase and NAD+, in which evidence for a charge-transfer interaction was also obtained (31).
The reoxidation studies show that melilotate affects the rate of reoxidation of melilotate hydroxylase as well as the rate of reduction (2). In the presence of melilotate, the rate of reaction of the reduced enzyme with oxygen was increased by more than lo-fold.
Furthermore, in the absence of melilotate, no intermediate was seen during the reoxidation.
The intermediate observed during the reaction of reduced enzyme with oxygen in the presence of melilot.ate is particularly interesting.
This is the third case in which such an intermediate has been found for a flavoprotein hydroxylase (4,5). The spectrum of the intermediate in the present work was obtained by analog simulation.
This spectrum can only be as accurate as the rate constants and extinction coefficients also used in the simulation.
All of these values are known to at least 10%. Both of the previous observations of an intermediate in the reoxidation of a reduced flavoprotein hydroxylase were with p-hydroxybenzoate hydroxylase (4,5). In one case, using 2,4dihydroxybenzoate as the substrate (4), the intermediate had a spectrum somewhat similar to the one reported here but had much higher extinction coefficients (e.g. at 410 nm, e -13,000 ~-1 cm-l).
With p-hydroxybenzoate as the substrate (5), the intermediate had a spectrum with a maximum at 380 nm and an extinction coefficient of about 8,000 Me1 cm-l.
The spectrum of the intermediate using 2,4-dihydroxybenzoate must be considered the better determination.
In that case, the relative rates of formation and decay of the intermediate allowed essentially its quantitative production for observation in the stopped flow apparatus.
The question as to the molecular structure of the intermediate remains unanswered.
The fact that the rate of formation of the intermediate is directly dependent on the oxygen concentration while its decay is independent implies that it is some type of oxygen adduct.
The possibility that it is an oxygen complex with the substrate cannot be excluded.
However. nerhans the most likely candidate is an adduct of molecular oxygen to reduced flavin. There are three positions on the isoalloxaxine ring system where this adduct might occur.
The positions C-4a and C-1Oa have previously been proposed as likely sites of oxygen attack (32). In fact, the spectrum of the intermediate reported here resembles that of a proposed C-1Oa methoxy adduct to lo-pentamethyl-l , 5-dihydro-isoalloxazine (33). The C-4a adduct might also be considered an attractive hypothesis in this case for several reasons. (a) Previously observed C-4a compounds have similar extinction coefficients to that of the intermediate reported here, although their absorption maxima are usually below 380 nm (34, 35). (b) C-1Oa addition to a reduced flavin blocks an amidine center, whereas C-4a addition blocks an azomethine center. In general, the C-4a addition might be considered more favorable (36). (c) Song8 has recently predicted from molecular orbital calculations that the C-4a oxygen adduct should have an absorption maximum close to 400 nm, compared to 445 nm for the C-1Oa species.
In addition to these two positions, N-5 must also be considered a possibility.
From theoretical considerations, it appears that N-5 is at least equally susceptible to attack by molecular oxygen (considered as an electrophile or radical) as C-4a and C-10a.3 Unfortunately, the unavailability of model flavins with an oxygen function at position N-5 prohibits further evaluation of this possibility.
The postulated reaction mechanism for melilotate hydroxylase is shown in Fig. 10, and the corresponding rate constants are tabulated in Table II. The mechanism shown is the one pre- 1.36 X 10-Q min 1.75 X lOeD M min 6.4 X 10-Q M min 6.7 X 10-Q M min 3.1 x 10-Q M 3.6 X 1O-6 Md -0 Assuming the mechanism shown in Fig. 10. h Calcu!ated from t,he data in Figs. 1, 2, and 3. c Calculated using the rate constants listed in Table II. d Calculated from the data of Fig. 1 (see text). 8 From a titration of melilotate hydroxylase with melilotate (2).
dieted on the basis of the steady state analysis.
Although this analysis appears conclusive, one must be cautious in assigning a mechanism on the basis of initial rate data alone. Correlation between a variety of kinetic methods is much firmer evidence for a postulated mechanism.
A comparison of the kinetic constants calculated from steady state and stopped flow results is shown in Table III.
The term $Q is the least meaningful of these comparisons, since the magnitude of kri is not known.
In Table III, i&i has been assumed to be large enough to make its reciprocal negligible in the expression for +Q. With this assumption the agreement between the two methods is good. It should be noted that kii is the release of one of the products from the oxidized enzyme, a reaction that could quite conceivably be fast enough to fit the above requirement.
For the calculation of &Ann, lcQ and kQ were measured directly with the stopped-flow technique.
Since kc is known to be near zero, however, the expression for q&rADH simply becomes l/k,.
The calculation of 40, is analogous to that for &.rADH, and in this case kr and kQ were measured directly.
Using similar arguments as those advanced for the estimation of kh, it is possible to assign a value of zero to kQ, so that +o, is equal to l/k,.
The agreement between &Ann and 4oQ obtained by the two methods is excellent (Table III).
It is important to note that this agreement is not sensitive to the abso1ut.e value of the estimated rate constants kc and ks. For example, if kq were actually 40 mind1 (a highly unlikely possibility in light of the data), the calculated value for &*nH would change only 3%. It should also be emphasized that the kinetic constants 40, +Mei, etc., are related to the more conventional kinetic constants listed in Table I ( (Table I) the agreement between the steady state and stopped flow results is equally as striking as that shown in Table III.
Accepting the proposed mechanism as valid allows determination of kl, which was not possible by the methods described here. Since her is equal to l/kl, kl should equal 5.7 x lo* M-i Inin-'.
Considering that this is for the reaction at l", the rate constant is similar in magnitude to those determined for the binding of salicylate to salicylate hydroxylase.
Finally, the kinetic analysis reported here clearly demonstrates the usefulness of probing an enzyme mechanism with a variety of kinetic techniques.
For melilotate hydroxylase, analysis of the reaction mechanism by steady state, rapid reaction, and t,itration methods gives complementary results. This is taken as extremely strong evidence that the proposed mechanism is correct.