Mechanism and regulation of the Two-component FMN-dependent monooxygenase ActVA-ActVB from Streptomyces coelicolor.

The ActVA-ActVB system from Streptomyces coelicolor is a two-component flavin-dependent monooxygenase involved

The flavin dependent monooxygenases are important enzymes that are involved in a wide variety of biological reactions.One of their fundamental functions uses their reduced flavins to activate molecular oxygen by generating flavin-hydroperoxide intermediates, the reactive species that oxygenate the substrate.Two different classes of flavindependent monooxygenases have been described.The first class comprises the wellknown single-component flavoprotein monooxygenases, in which within a single polypeptide chain, the flavin cofactor is reduced by NAD(P)H and reacts with O 2 to generate the flavin-hydroperoxide species.Some of these flavoproteins have been investigated in great detail, p-hydroxybenzoate hydroxylase being the prototype of this class of flavoprotein monooxygenase (1).More recently, several two-component flavindependent monooxygenases, present in prokaryotic cells and involved in a broad range of oxygenation reactions, have been reported.These systems use two different proteins, a reductase for reducing the flavin and an oxygenase for binding the reduced flavin and catalyzing the reaction with O 2 .Examples of two-component monooxygenases that use reduced FAD (FAD red ) as a cofactor are 4hydroxyphenylacetate monooxygenase (HpaB) from Escherichia coli (2), phenol hydroxylase (PheA1) from Bacillus thermoglucosidaflurescens (3) and styrene monooxygenase (StyA) from Pseudomonas fluorescens (4,5).Two-component flavindependent monooxygenases that use reduced FMN (FMN red ) include enzymes involved in the biosynthesis of the antibiotic pristinamycin in Streptomyces pristinaespiralis (6,7), in the utilization of sulfur from aliphatic sulfonates in Escherichia coli (8) and in the desulfurization of fossil fuels by Rhodococcus species (9), as well as the hydroxylation of 4hydroxyphenylacetate in Acinetobacter baumannii (10).We have been investigating the mechanism of the FMN-dependent twocomponent enzyme system, ActVA-ActVB, which participates in the last steps of the biosynthesis of the antibiotic actinorhodin in Streptomyces coelicolor (11)(12)(13) (Scheme 1).
The first enzyme of the two-component flavin-dependent oxygenases is a NAD(P)H:flavin oxidoreductase that catalyzes the reduction of free oxidized flavins (FAD and/or FMN) by the reduced pyridine nucleotides, NADPH or NADH (3,(14)(15)(16).The second enzyme is an oxygenase that binds the resulting free reduced flavin and promotes its reaction with O 2 to hydroxylate the substrate (2,3,(5)(6)(7)(8)(9)(10)17).Therefore, in contrast to the single-component flavin hydroxylases, the two-component systems catalyze the reduction of the flavin and the hydroxylation of the substrate on separate polypeptides.In the biosynthesis of actinorhodin the hydroxylase system consists of the reductase, ActVB, and the oxygenase, ActVA (14,15,17).There is now general agreement that the hydroxylation steps of flavin-dependent hydroxylases proceed through the reaction of the reduced flavin with molecular oxygen to generate a flavin C(4a)-hydroperoxide intermediate (18).This is followed by the transfer of an oxygen atom from the peroxide to the substrate.
Catalysis by the two-component flavindependent monooxygenases requires that the reduced flavin be transferred from the reductase to the monooxygenase without being oxidized by molecular O 2 .This would appear to be a very challenging process because free reduced flavin becomes oxidized quite rapidly in the presence of O 2 .Such an oxidation must be avoided or the reduction of the flavin and the ensuing hydroxylation reaction would effectively be uncoupled and, instead, NAD(P)H would be consumed to produce H 2 O 2 and superoxide.We have previously demonstrated that with the ActVA-ActVB system from S. coelicolor the transfer of FMN red from the reductase to the oxygenase is thermodynamically favorable (17).Indeed the monooxygenase ActVA binds FMN red with a K d value of 0.39 µM and FMN ox with a K d value of about 20 µM, while the reductase ActVB binds FMN ox more tightly than FMN red.There is also no evidence for specific binding interactions between the two components that might facilitate a channeling mechanism allowing the flavin to travel from one protein to another within a protein complex.These data suggest that in spite of the drawbacks associated with the simultaneous presence of free FMN red and O 2 , FMN red is probably transferred between the two components by diffusion and rapid binding to the oxygenase during the hydroxylation reaction.
In this work, we have determined that FMN red can be transferred very efficiently from the reductase to the monooxygenase without the participation of any protein-protein complexes.In addition, our data show that in the S. coelicolor system the rate of transfer of FMN red from the reductase, ActVB to the hydroxylase, ActVA, is controlled by the release of NAD + from ActVB, and that the concentration of NAD + strongly affects this rate.This is manifest by the formation of a strong charge-transfer complex between NAD + and FMN red in the active site of ActVB.
Anaerobic procedures and rapid reaction experiments.All reactions were performed in 50 mM Tris-HCl buffer, pH 7.4, at 4°C.Solutions were made anaerobic in glass tonometers by ~20 cycles of vacuum and equilibration with Ar 2 gas that was depleted from oxygen contamination with an Oxiclear oxygen removal column (Labclear) as described previously (22).To minimize oxygen contamination of reaction mixtures due to diffusion through valves, etc, all anaerobic solutions contained an oxygen-scrubbing solution consisting in 12.5 to 25 µM protocatechuate (PCA, final concentrations) and ~0.06 units/ml of protocatechuate dioxygenase (PCD, final concentration).This mixture has been shown to maintain anaerobic conditions efficiently (23).
Rapid kinetics reactions were performed with a Hi-Tech Scientific stopped-flow Model SF-61DX instrument operating either in single or in double mixing mode, and using either a diode array spectrophotometer or a photomultiplier detector.Kinetic traces were analysed and fitted with KinetAsyst 3 software (Hi-Tech Scientific, Salisbury, UK).Before each experiment, the stopped-flow instrument flow system was made anaerobic by incubating the flow system overnight with an oxygen scrubbing solution consisting of 100 mM PCA and 0.06 units/ml PCD in potassium phosphate buffer, pH 7.
Reaction of ActVA•FMN red with O 2 in the presence or in the absence of the pyronaphthoquinone substrate DHK red .A solution containing ActVA (50 µM), FMN ox (20 µM), deazaflavin (60 nM), EDTA (10 mM) and catalase (50 nM) in 50 mM Tris-HCl buffer, pH 7.4, was made anaerobic in a glass tonometer as described above.FMN ox was then photoreduced for 5 to 10 sec using a commercial 1000 W tungsten-halogen lamp placed ~10 cm away from the solution.The redox status of FMN was monitored by UVvisible spectrophotometry.When experiments were performed in the presence of DHK, the same reaction mixture as indicated above was used, but included the presence of 80 µM DHK.Both FMN ox and oxidized DHK (DHK ox ) were photoreduced together by the procedure described above.This solution was then mixed in the stopped-flow instrument with an equal volume of a solution containing various concentrations of oxygen, and the reaction was followed spectrophotometrically at various wavelengths, or by spectrofluorimetry with excitation at 390 nm and emission at wavelengths > 530 nm.Oxygenated solutions were prepared by equilibrating 50 mM Tris-HCl buffer for 30 min with air (21% oxygen) or with certified oxygen/nitrogen gas mixtures (5, 10, 50, 100 % oxygen).
Investigation of the kinetics of FMN red transfer.To study the transfer of FMN red from ActVB to the monooxygenase (ActVA or HPAH-C 2 ), the ActVB•FMN red complex was prepared.A solution containing ActVB (85 µM) and FMN ox (20 µM) in Tris-HCl buffer, pH 7.4, in a glass tonometer equipped with a cuvette was made anaerobic as described above.The excess ActVB was to assure that most of the FMN was bound.The complex was then reduced by titration with an anaerobic solution of NADH delivered by a gastight Hamilton syringe connected to the tonometer via an Airless Ware fitting.Approximately 20 µM of NADH was sufficient to fully reduce the ActVB•FMN complex.This complex was then mixed in the stopped-flow instrument with an equal volume of Tris-HCl buffer containing ActVA (45 µM) and O 2 (600 µM), HPAH-C 2 from A. baumanii (45 µM) and O 2 (600 µM), oxidized DCPIP (2,6dichlorophenolindophenol, 20 µM), or oxidized menadione (40 µM).The concentrations given are those in the syringes before mixing.The inhibition of FMN red transfer from ActVB to HPAH-C 2 by NAD + was investigated by a double mixing stoppedflow technique.The ActVB•FMN red •NAD + complex prepared as described above (235, 40 and 40 µM of each component respectively), was first mixed into an aging chamber with a solution containing various concentrations of NAD + (from 0 to 5 mM).After a 5 s delay to assure that the binding of NAD + had come to equilibrium, the solution from the first mix was mixed with an equal volume of 50 mM Tris-HCl buffer containing HPAH-C 2 (50 µM) and O 2 (255 µM), and the FMN red transfer to HPAH-C 2 was followed spectrophotometrically at 386 and 700 nm to observe the formation of the C(4a)-FMN-OOH intermediate and the disappearance of the FMN red -to-NAD + charge-transfer complex.

Reaction of ActVA-FMN red with O 2 in the absence of the pyronaphthoquinone substrate. Formation and decay of the C(4a)-FMN-OOH intermediate. It has been previously shown that
ActVA is able to stabilize a C(4a)-FMN-OOH species within its active site, but the kinetics of its formation and decay were not determined (17,24).Here, we have investigated the kinetics of the formation and decay of the C(4a)-FMN-OOH intermediate by stoppedflow spectrophotometry at 4°C, as described in Materials and Methods.An anaerobic solution of FMN red (20 µM), containing an excess of ActVA (50 µM) to promote full complex formation (K d FMNred = 0.39 µM, (17)), was mixed with equal volumes of Tris-HCl buffer solution containing various concentrations of O 2 .The reaction was monitored by UV-visible spectrophotometry.As shown in Fig. 1A (thick line) the first spectrum recorded at 5 s ([O 2 ] = 595 µM after mixing) has a main absorption band at 386 nm.This spectrum is characteristic of a C(4a)-FMN-OOH species, such as those found in other single-and two-component flavin-dependent hydroxylases, including 4hydroxyphenylacetate 3-hydroxylase (HPAH-C 2 ) from A. baumannii (22) and phydroxybenzoate hydroxylase (PHBH) from Pseudomonas aeruginosa (25,26).Reaction traces monitored at 386 nm were fit to a single exponential model and the rate constants for formation of the absorbing species were found to be proportional to the O 2 concentration (Fig. 1B).The linear plot of k obs at 386 nm versus oxygen concentration yielded a second-order rate constant of k = 2.97 ± 0.06 x 10 4 M -1 s -1 (Fig. 1C).These data are consistent with a bimolecular reaction between the reduced flavin bound to ActVA and oxygen to form the C(4a)-FMN-OOH intermediate.
The C(4a)-FMN-OOH then slowly converted to FMN ox with a half-life of about 1400 s and was complete by ~1 hour, as seen from the appearance of the two characteristic bands of the oxidized flavin at 370 and 445 nm (Fig. 1A).The absorbance changes at 445 nm (Fig. 1B) were found to be independent of oxygen concentration.These data, which are consistent with our previous results from experiments at 25 °C, clearly show that ActVA strongly stabilizes the C(4a)-FMN-OOH intermediate (17,24).At the end of this slow decay of the intermediate, 10 µM FMN ox was detected (ε 445nm = 13 mM -1 cm -1 ), indicating that all of the initial FMN red was quantitatively oxidized to FMN ox by O 2 .An ε value at 386 nm of 9.6 mM -1 cm -1 for the C(4a)-FMN-OOH species could then be calculated.
Interestingly, when free FMN red was mixed with an air-saturated ActVA solution in the stopped-flow apparatus, kinetic traces were the same as those obtained when a preformed ActVA•FMN red complex was mixed with the same concentration of oxygen (data not shown).This implies that almost no spontaneous oxidation of free FMN red by O 2 occurred, even though autoxidation of free FMN red by O 2 is a reasonably fast process.For example, 40 µM FMN red is oxidized by 240 µM O 2 , final concentration, in an autocatalytic reaction with a t 1/2 ~0.5 s (27, unpublished data), and no detectable C(4a)-FMN-OOH species is produced (28).This result implies that binding of FMN red to ActVA preceding the formation of the C(4a)-FMN-OOH species is fast and precludes any significant oxidation of free FMN red by oxygen.These results explain and are in full agreement with our previous steady-state kinetics experiments (24).
Reaction of ActVA-FMN red with O 2 and pyronaphthohydroquinone substrate.The reaction of O 2 with FMN red bound to ActVA in the presence of its substrate, reduced dihydrokalafungin (DHK red ), was investigated by stopped-flow spectrophotometry.An anaerobic solution containing 80 µM of DHK red , 20 µM of FMN red and an excess of ActVA (50 µM, to ensure that most of the FMN red was in complex with ActVA) was mixed with equal volumes of oxygenated Tris-HCl buffer solution containing various amounts of O 2 .In Figure 2, the reported data were observed with a final O 2 concentration of 595 µM.The reaction was followed at : (i) 365 nm, a wavelength that is useful for monitoring the formation of the C(4a)-FMN-OOH intermediate species (22,29,30) and at which DHK red contributes to the absorbance minimally; (ii) 445 nm to monitor the formation of FMN ox ; (iii) 520 nm to monitor the formation of the hydroxylated DHK ox product (DHK ox -OH) (24).In addition, fluorescence emission at 530 nm (excitation at 390 nm) was recorded in order to follow the formation of C(4a)-FMN-OH (hydroxy-FMN) and FMN ox species.For many flavindependent hydroxylases, the C(4a)-FMN-OOH intermediate does not exhibit significant fluorescence emission upon excitation at 390 nm, whereas the C(4a)-FMN-OH species is often highly fluorescent (22,29,31).The absorbance values at the above mentioned wavelengths are plotted as a function of time in Fig. 2A.These data indicate that the reaction proceeds through at least four successive phases (Fig. 2A, Table 1).Each trace could be fitted by the sum of four exponential processes using a single set of rate constants for data at all wavelenghts.
The rate constant k 1 for phase 1 was directly dependent on O 2 concentration (data not shown), indicating a bimolecular reaction with O 2 .No FMN ox (445 nm trace) and almost no C(4a)-FMN-OH (fluorescence at 530 nm) were observed during this first phase.Thus, the bimolecular rate constant (3.96 ± 0.05 x 10 4 M - 1 s -1 ) determined from the increase in absorbance at 365 nm is clearly associated with the formation of the C(4a)-FMN-OOH intermediate.This rate constant is only slightly different from that determined in the absence of the DHK red substrate (Fig. 1C, 2.97 ± 0.06 x 10 4 M -1 s -1 ).In the subsequent phases, k 2 , k 3 , and k 4 , were found to be independent of O 2 concentration (data not shown), indicating that they did not involve direct reactions with O 2 .
In phase 2, the appearance of fluorescence at 530 nm (with excitation at 390 nm) reached a maximum at ~2 sec reaction time, with a rate constant of 1.48 ± 0.15 s -1 .Although there is only a small decrease in absorbance at 365 nm in this phase, the fluorescence data clearly reflect the hydroxylation reaction.Therefore, the above data are consistent with the formation of the C(4a)-FMN-OH intermediate species in phase 2, implying that hydroxylation of the DHK red substrate occurred during that phase.The trace at 520 nm does not show any increase in absorbance during this phase as might be expected for DHK ox -OH.The immediate product is therefore most likely DHK red -OH, which has very little absorbance or fluorescence (unpublished data), and is not expected to contribute significantly to the absorption and fluorescence data during phase 2.
Phase 3 is best indicated by the strong increase in absorbance at 445 nm characterized by a rate constant of 0.16 ± 0.02 s -1 .This was due to formation of FMN ox , which is expected to arise from dehydration of the C(4a)-FMN-OH intermediate.The small decrease of fluorescence with an identical rate constant (0.16 ± 0.02 s -1 ) likely reflects the smaller fluorescence quantum efficiency of FMN ox (with excitation at 390 nm) as compared to that of C(4a)-FMN-OH, as has been shown for other flavin-dependent hydroxylases (29,31).
Finally, phase 4 was mainly characterized by a slow increase of absorbance at 520 and 450 nm (0.013 ± 0.002 s -1 ), which most likely represents the oxidation by O 2 of the hydroquinone, DHK red -OH, into the quinone, DHK ox -OH.This conclusion was also consistent with the spectrum recorded 200 s after mixing (Fig. 2B) that has an additional band extending to 600 nm.This spectrum is similar to that of a mixture of ~10 µM FMN ox and ~10 µM DHK ox -OH; the latter spectrum was presented in our previous work (24).Oxidation of hydroquinones and their derivatives such as DHK red by O 2 appears to proceed through an autocatalytic mechanism with phases that are only partially dependent on the concentration of O 2 (28,32,33).This is also the case for the reaction of O 2 with DHK red , which becomes oxidized in an autocatalytic process with very little O 2dependence (data not shown).
The kinetic path of hydroxylation of DHK red catalyzed by ActVA is summarized in Scheme 2. One interesting feature of this reaction is that the formation of the C(4a)-FMN-OOH intermediate, which in the presence of O 2 forms quickly after FMN red is transferred to the monooxygenase, has very little dependence on the presence of the DHK red substrate.This indicates that the substrate may bind after the hydroperoxide intermediate is formed, as observed with HPAH-C 2 (22), cyclohexanone monooxygenase (34), and microsomal flavin monooxygenase (35).
FMN red transfer from ActVB to ActVA.Our previous investigations have shown that transfer of FMN red from the reductase ActVB to the monooxygenase ActVA is thermodynamically favorable (17).To study the kinetics of this transfer reaction, we took advantage of the UV-visible signature of the ActVB FMN red NAD + complex, which has a broad charge-transfer (C-T) band between 550 and 800 nm, as well as the formation of the C(a)-FMN-OOH intermediate that can be monitored at 380 nm.The C-T band originates from the interaction between NAD + and FMN red in the ActVB active site (14).Our previous studies showed that, under anaerobic conditions, the addition of 1 equivalent of ActVA to the ActVB FMN red NAD + complex resulted in total disappearance of the C-T band, implying that FMN red was transferred from ActVB to ActVA (17).
In Fig. 3, an ActVB•FMN red •NAD + complex (85, 20, and 20 µM, of each component, respectively) was anaerobically prepared by adding one equivalent of NADH to FMN ox in the presence of ActVB.Then, the complex was mixed at 4°C with an equal volume of solution containing 45 µM ActVA and 600 µM O 2 in a stopped-flow spectrophotometer.The UV-visible difference spectra (spectra at various times minus the initial spectrum of the ActVB•FMN red •NAD + complex; this mode of display makes it easier to view the overall changes) were recorded as a function of time from 7 to 150 ms after mixing.Addition of the aerobic solution of ActVA to the anaerobic solution containing ActVB•FMN red •NAD + resulted in total disappearance of the broad C-T band within ~1 s, shown in Fig. 3A by the loss of absorbance between 440 and 700 nm.Simultaneously, an absorbance band centered at ~386 nm appeared, due to formation of the C(4a)-FMN-OOH species.The presence of an isobestic point at about 435 nm indicated an apparent direct conversion of the ActVB•FMN red •NAD + complex to the C(4a)-FMN-OOH-ActVA complex.
Fig. 3B shows the kinetics of the absorbance changes at 386 nm (formation of C(4a)-FMN-OOH) and at 700 nm (disappearance of the C-T complex ActVB•NAD + •FMN red ).Both traces could be fitted to a single exponential model with almost identical rate constants (k 386nm = 4.6 ± 0.1 s -1 and k 700nm = 5.0 ± 0.4 s -1 ).It can be noted that this rate constant is smaller than that for the formation of the C(4a)-FMN-OOH•ActVA complex from the reaction of free FMN red with O 2 in the presence of ActVA (k 386nm = 10.6 ± 0.1 s -1 for the reaction of 10 µM free FMN red mixed with 22.5 µM ActVA and 300 µM O 2 final concentrations, Fig. 1 and data not shown).Although these results suggest that under the experimental conditions of Fig. 3 formation of the C(4a)-FMN-OOH intermediate is rate-limited by the release of FMN red from ActVB, the relatively small difference in rates of these two processes tempers confidence in this conclusion.In order to further clarify the kinetics of dissociation of FMN red from ActVB, we performed the experiments described below.
The HPAH-C 2 monooxygenase from A. baumanii is part of a two-component FMNdependent monooxygenase similar to ActVA.Recently, it has been shown that under similar conditions to those used for ActVA in the experiment shown in Fig. 1, HPAH-C 2 monooxygenase binds free FMN red and reacts with O 2 to form a C(4a)-FMN-OOH adduct very rapidly.The rate constant for binding FMN red to HPAH-C 2 has been estimated to be > 10 7 M -1 s -1 and the ensuing reaction with O 2 occurs with a second-order rate constant of 1.1 10 6 M -1 s -1 (22).Thus, if HPAH-C 2 is reacted with free FMN red in the presence of 300 µM O 2 , the formation of the C(4a)-FMN-OOH intermediate will occur with a rate constant of ~330 s -1 , a value about 30 times greater than that reported for ActVA.Substitution of HPAH-C 2 for ActVA in an experiment such as that in Fig. 3 would clearly distinguish whether the release of FMN red from ActVB is rate limiting.The experiment shown in Fig. 3 was repeated in the presence of HPAH-C 2 monooxygenase (25 µM, final concentration) in place of ActVA.As reported for the experiment with ActVA (Fig. 3), the ActVB•FMN red •NAD + complex disappeared in parallel with the formation of the C(4a)-FMN-OOH intermediate (data not shown).The kinetics of the absorbance changes at 386 nm (formation of the C(4a)-FMN-OOH species) and at 700 nm (disappearance of the C-T complex ActVB•NAD + •FMN red ) could be fitted to a single exponential model, with rate constants (Table 2, k 386nm = 4.7 ± 0.1 s -1 and k 700nm = 3.8 ± 0.1 s -1 ) nearly identical to those obtained with ActVA (Table 2).These data are consistent with the limiting step for the transfer of FMN red from ActVB to the monooxygenase component being the release of FMN red from the ActVB•NAD + •FMN red complex.
Menadione and DCPIP are two redox dyes known to react very rapidly with reduced free flavin (36).When free FMN red is mixed with two equivalents of either one of these oxidized dyes, the FMN red becomes nearly fully oxidized during the dead time of the stoppedflow experiment (implying that k ox > 300 s -1 , Table 2).However, when ActVB•FMN red •NAD + was mixed with either DCPIP or menadione under anaerobic conditions, the FMN red was oxidized with rate constants of 6.2 ± 0.1 s -1 and 6.1 ± 0.1 s -1 for DCPIP and menadione, respectively (Table 2).In addition, the C-T complex (measured at 700 nm) disappeared with rate constants (6.5 ± 0.1 s -1 and 6.3 ± 0.1 s -1 for DCPIP and menadione respectively, Table 2).These rate constants are more than 50-fold smaller than those obtained when free FMN red reacted with these dyes (Table 2).As shown in Table 2, the rates of disruption of the charge-transfer complex obtained in the presence of the monooxygenases ActVA or HPAH-C 2 and with the two dyes were almost identical.These results suggest that the rate of oxidation of the ActVB•NAD + •FMN red complex by the dyes as well as by the oxygenases is likely to be primarily limited by the release of FMN red from the complex.The slightly greater rates obtained with the dyes compared to those obtained with the two oxygenases can be explained by the dyes reacting with the FMN red before it is fully released from ActVB, whereas the C(4a)-FMN-OOH species can only form after FMN red is completely released and transferred to the oxygenases.
Altogether, these data demonstrate that the formation of the C(4a)-FMN-OOH intermediate with the monooxygenases ActVA and HPAH-C 2 is limited by the release of FMN red from the ActVB•NAD + •FMN red complex followed by its transfer to the monooxygenase.
Inhibition of FMN red transfer by NAD + .Previous steady-state kinetics studies had shown that during the ActVB-catalyzed reaction (FMN ox + NADH → FMN red + NAD + ) the release of the NAD + product occurs before that of FMN red (14).We tested whether the presence of NAD + influences the transfer of FMN red from the ActVB•NAD + •FMN red complex to the monooxygenase by carrying out experiments similar to those described in the previous section with various concentrations of NAD + .In these experiments HPAH-C 2 was used in place of ActVA because, as mentioned earlier, HPAH-C 2 forms C(4a)-FMN-OOH very rapidly (22).Therefore, under these conditions, the formation of the C(4a)-FMN-OOH species provides a convenient and reliable coupled assay for monitoring the release of FMN red from ActVB.
As shown in Fig. 4, the ActVB FMN red NAD + complex (235, 40, and 40 µM of the respective components), was anaerobically mixed with equal volumes of solutions containing various amounts of NAD + .After 5 s (to allow the NAD + binding to come to equilibrium), this solution was mixed with an equal volume of an aerobic solution of HPAH-C 2 (50 µM HPAH-C 2 and 255 µM O 2 ), and the absorbance values at 386 and 700 nm of the final mixture were followed as a function of time (Figs.4A and 4B respectively).Addition of increasing amounts of NAD + resulted in decreases of the rates of both the loss of the C-T complex (700 nm) and the formation of C(4a)-FMN-OOH (386 nm).For a given NAD + concentration, traces at both 386 and 700 nm could be fitted with a single exponential function with almost identical rate constants (Figs.4A and 4B).As shown in Fig. 4C, a plot of the rate constants obtained from the traces at 386 and 700 nm versus NAD + concentration indicated that the presence of an excess of NAD + could totally inhibit the transfer of FMN red from ActVB to HPAH-C 2 as well as the formation of the C(4a)-FMN-OOH species.The IC 50 concentration for inhibition by NAD + was determined to be ~40 µM.

DISCUSSION
In this work, we have carried out a detailed kinetic study of the reaction of the twocomponent FMN-dependent monooxygenase, ActVA-ActVB from Streptomyces coelicolor, by stopped-flow spectrophotometric methods in order to test our earlier propositions on the reaction mechanism of ActVA-ActVB system (17,24) and to gain insights into its mechanism of regulation.Recent characterization of several members of the two-component flavin-dependent monooxygenase family has revealed similarities to but also some specific differences from the well-studied single component flavoprotein monooxygenase family, such as p-hydroxybenzoate hydroxylase.In both types of systems, the monooxygenase activates the oxygen molecule by forming a C(4a)-flavin hydroperoxide intermediate that delivers an oxygen atom to the substrate.In contrast to the single component hydroxylases, the reduction of the flavin in the two-component flavin-dependent monooxygenase systems is catalyzed by a separate NADH oxidoreductase and the resultant reduced flavin must then be delivered to the oxygenase.Thus, the mechanism by which the reduced flavin cofactor is transferred from the reductase to the monooxygenase in the presence of O 2 represents one of the most challenging aspects of these systems.Here we have presented several experiments that elucidate the essential features of the mechanism of transfer of FMN red between ActVB and ActVA.Compared with the few studies recently published on other twocomponent FMN-dependent monooxygenase systems, our results highlight a unique property of the ActVA-ActVB system, namely that the overall hydroxylase reaction is partially limited by the release of FMN red from the reductase, and this release is regulated by the dissociation rate of the charge transfer complex between NAD + and FMN red .Large concentrations of NAD + , a product of the flavin reductase reaction, can greatly reduce the rate of FMN red release.Moreover, we show that no complexes between ActVB and ActVA are required for efficient transfer of FMN red between the two.
At first, we studied the monooxygenasedependent reaction in the absence of the reductase component ActVB and under conditions where the ActVA•FMN red •DHK red complex was formed before it reacts with O 2 .We showed that the hydroxylation mechanism proceeds with the participation of two different reaction intermediates, a C(4a)-FMN-OOH that precedes, and a C(4a)-FMN-OH species that accompanies the formation of the reaction product, DHK red -OH (Fig. 2).The C(4a)-FMN-OOH intermediate, characterized by a typical single absorbance band centered at 386 nm, results from the bimolecular reaction of the ActVA•FMN red complex with O 2 .This step occurs with a rate constant of ~4 x 10 4 M -1 s -1 , essentially independent of the presence of the DHK red substrate.We note that in the monooxygenase HPAH-C 2 component from A baumannii, when C 2 in complex with FMN red and its substrate, 4-hydroxyphenylacetate (at 2 mM), is mixed with O 2 , the rate of formation of the C(4a)-FMN-OOH intermediate is significantly smaller than when the substrate is absent or when the HPAH-C 2 -FMN red complex is mixed with a solution containing 4hydroxyphenylacetate and O 2 .Thus, it could be concluded that normally FMN red binds first, then O 2 reacts to form the C(4a)-FMN-OOH, and finally, 4-hydroxyphenylacetic acid binds before hydroxylation takes place (22).For ActVA, with the substrate concentrations used in this study (0, 10 and 40 µM DHK red , Fig. 1C, data not shown and Fig. 2A, respectively), no effect of DHK red on the kinetics of the reaction could be observed.However, because of the limited availability of the DHK substrate (it is not commercially available and has limited solubility), we could not determine if such an effect occurs with higher concentrations of DHK red , and therefore we do not know the preferred order of binding.
In the presence of DHK red , the C(4a)-FMN-OOH intermediate in the ActVA active site converted to a C(4a)-FMN-OH species, with a rate constant of 1.48 s -1 .Formation of the C(4a)-FMN-OH species results from transfer of the distal oxygen of C(4a)-FMN-OOH to the substrate, so that hydroxylation of DHK red to give DHK red -OH almost certainly occurred during that phase.In the subsequent step, the dehydration of the C(4a)-FMN-OH intermediate to form FMN ox was observed at a rate of 0.16 s -1 (Fig. 2 and Scheme 2).
The mechanism of DHK red hydroxylation catalyzed by ActVA with participation of the two reaction intermediates demonstrated in this study (Scheme 2) appears to be quite similar to that recently determined for the HPAH-C 2 component from A. baumannii (22), as well as that proposed for several single component hydroxylases (1,18,25,28,34,35).These results reveal that, despite a rather low sequence identity between ActVA and HPAH-C 2 (23 % identities), the monooxygenase components in the two-component flavindependent monooxygenase family use similar chemical mechanisms of hydroxylation.
In the absence of the substrate, the C(4a)-FMN-OOH intermediate forms rapidly (~3 x 10 4 M -1 s -1 ) and is remarkably stable, as previously observed (17,24), and thus slowly converts to FMN ox (and probably H 2 O 2 ) with a half-life of ~1400 s.Formation of the hydroperoxide intermediate in the absence of substrate has also been observed for other members of the two-component flavindependant monooxygenase family, such as styrene monooxygenase from Pseudomonas Putida S12 (37), the HPAH-C 2 monooxygenase from A. baumannii (22) and bacterial luciferase (38).The recently solved X-ray structure of the HPAH-C 2 monooxygenase from A. baumannii in complex with FMN red has revealed the presence of a cavity adjacent to the isoalloxazine ring of the flavin molecule that is appropriate for harbouring the oxygen atoms of the C(4a)-FMN-OOH intermediate (39).It was proposed that this cavity provides a solventfree environment that prevents rapid breakdown of the peroxide intermediate, thus accounting for its stability.In ActVA, one might expect an even more sequestered environment around the C(4a)-position of the flavin, because the C(4a)-FMN-OOH intermediate is about 75-fold more stable than in HPAH-C 2 (half-life of 1400 s for ActVA and 18 s for HPAH-C 2 , this work and ( 22)).A sequence alignment between these two proteins shows that in HPAH-C 2 the aminoacid residues involved in the formation of the cavity near the C(4a)-position of the flavin are not totally conserved in ActVA.Presumably, these differences are responsible for the variations in stability of the peroxide intermediates.
Because in two-component FMNdependent monooxygenases the reduction of the flavin takes place in a protein separate from the monooxygenase, FMN red must be transferred from the reductase to the monooxygenase in the presence of O 2 without being oxidized unproductively.This represents a rather challenging process.Autooxidation of FMN red can be toxic to the cell because it produces the harmful O 2 •-and H 2 O 2 species, and in addition, is wasteful of reducing equivalents by effectively decoupling the reduction of the flavin by NADH and the hydroxylation reaction.We have previously shown that in the case of the ActVA-ActVB system such decoupling does not occur to any measurable extent (17,24), but the mechanism by which it is avoided was not documented.The results of the present study now provide some insight into this mechanism.
i) First, we showed that ActVA binds free FMN red rapidly enough to prevent autooxidation of free FMN red .Bruice and colleagues have determined that the initial rate of reaction of O 2 with free FMN red occurs at 250 M -1 s -1 (40).Thus, in the presence of 200 µM O 2 , this initial rate should be 0.05 s -1 and, even considering the ensuing autocatalytic oxidation process known for such reactions (28,32,40), very little net oxidation of free FMN red will occur within the first 100 ms (27).The formation of the C(4a)-FMN-OOH intermediate (Fig. 1C) occurs at a rate of 5.9 s -1 (t 1/2 ~100 ms) whether FMN red is premixed with ActVA before or after mixing with O 2 .Because the binding of FMN red to ActVA does not limit the rate of formation of the rate of formation of the C(4a)-FMN-OOH, binding must occur at a rate ≥ 60 s -1 (t 1/2 ≤ 10 ms), and thus will be complete before any appreciable autooxidation of flavin happens.
This fast binding of FMN red to ActVA represents one of the key catalytic features of these systems where reduction of the flavin and activation of O 2 occur in two different polypeptides.In addition to avoid uncoupling reactions during FMN red transfer, this fast binding of FMN red to ActVA also allows the rationalization of why the formation of a complex between the reductase and the monooxygenase is not required for efficient oxygenation reactions.
ii) Second, we demonstrated that dissociation of NAD + from ActVB limits the release and transfer of FMN red from the reductase ActVB to the monooxygenase ActVA.We showed that the dissociation of FMN red from ActVB is tightly regulated by NAD + , which must dissociate before FMN red can be released in the rate limiting step.A stable C-T complex between FMN red and NAD + within the active site of ActVB limits the dissociation rate of NAD + .Accordingly, when the kinetics of the FMN red transfer between ActVB and ActVA was studied in the presence of increasing amounts of NAD + , a strong inhibition of the transfer of FMN red to ActVA was observed.An IC 50 for NAD + of 40 µM was determined, and the inhibition of the FMN red transfer was almost complete in the presence of 1 mM NAD + .Therefore in S. coelicolor, ActVB, the reductase component can act as a regulatory component of the monooxygenase activity by controlling the transfer of the FMN red cofactor to the monooxygenase.
Earlier steady-state kinetics studies showing that during ActVB catalysis, after the reduction of FMN ox by NADH via hydride transfer, the NAD + product is released before FMN red from the active site, in an ordered sequential mechanism (14).This confirms the notion that FMN red cannot dissociate from the ActVB active site when NAD + is bound.Then, according to Le Chatelier's principle, an increase of NAD + concentration is likely to promote the formation of ActVB•FMN red •NAD + complex and thereby prevent the release of free FMN red .This is consistent with the experiments presented above showing strong inhibition of the transfer of FMN red from ActVB to ActVA by NAD + .
Control of the monooxygenase activity by NAD + is unprecedented in the two-component flavin dependent monooxygenase family.However, in most studies, high concentrations of NAD + were not present in reaction mixtures.Furthermore, few systems share with ActVB the observable rather stable flavin red •NAD + charge-transfer complex.PheA2, the reductase of a two-component phenol hydroxylase from Bacillus thermoglucosidasius, is one such example (3), and it is possible that similar control of the activity by NAD + also occurs in that case.Crystal structures show that NAD + binds directly over the FAD red in a position that is ideal for charge-transfer interactions (41).Recently, it was proposed that the activity of the monooxygenase component in the A. baumannii system was limited by the transfer of FMN red from the reductase, HPAH-C 1 , to the monooxygenase, HPAH-C 2 (27).Furthermore, HPA, the monooxygenase substrate, was shown to function as an effector of HPAH-C 1 for both FMN reduction and FMN red release (19,27).Thus, as in the case of the ActVA-ActVB system, HPAH-C 1 , the reductase component, is likely to regulate the overall monooxygenase activity.However, in that case the effector is the substrate of the monooxygenase, HPA, whereas in ActVA-ActVB system it is NAD + , the product of the reductase.
Interestingly, it was also previously suggested that the oxygenases of the ActVA-ActVB system (24) and the 4hydroxyphenylacetate hydroxylase (42) systems could be endowed with a regulatory function with respect to their reductases.In the absence of substrate, ActVA, traps free FMN red and converts it to a stable C(4a)-FMN-OOH intermediate.The consequence is that the cellular content of free FMN can become depleted, resulting in a shut-down of the reductase component activity.As highlighted by this work, the regulatory role of ActVB that is dependent on the intracellular concentration of NAD + adds another important piece to the puzzle of the dynamic regulation of the ActVA-ActVB monooxygenase activity in a cellular context.It might allow coordination of the production of the DHK red -OH (and consequently the formation of the actinorhodin) to the energetic state of the cell, as determined by the NADH/NAD + ratio.When this ratio is low in cells depleted of NADH, the FMN red would remain bound to the reductase component, preventing further oxidation of NADH and leading to the arrest of the hydroxylase activity.
It therefore seems that for the twocomponent monooxygenase systems there is a mutual regulation mechanism allowing the interpretation of different cellular signals for rapidly tuning the activities of each partner according to the demand and the status of the cell.In the case of the ActVA-ActVB system, ActVA could be perceived as the sensor of the intracellular concentration of DHK, whereas ActVB could be a sensor of the cellular energetic level indicated the NADH/NAD + ratio.
Finally, the results on the two-component flavin-dependent systems from A. baumannii (27) and S. coelicolor (17, 24 and this work) reveal that free diffusion of FMN red between two components in the presence of O 2 can be very efficient for supporting monooxygenation reactions.It is clear that single component flavin-dependent monooxygenases can coordinate and carry out both reduction of the flavin and hydroxylation of substrates.Therefore, the advantage of a system using separate reductases and a monooxygenases might be questioned.Because flavin reduction and oxygen-dependent reactions have very different requirements that are difficult to fulfil with a single polypeptide, dividing the tasks by using two different enzymes is a reasonable strategy to alleviate this challenge.In the case of the single component flavoprotein hydroxylases, coordinated dynamics between two different conformations that effectively provides two "active sites" has been shown to be important for dealing with this challenge (1,18).This coordination is usually linked to the substrate of the enzyme, so that adaptation to new substrates to trigger this dynamic dance of catalysis can be quite involved and complicated.Therefore, the two-component monooxygenase organization might represent an evolutionary advantage for quickly adapting to the metabolism of new substrates.The reductase and the monooxygenase polypeptides can acquire new complementary properties individually without modifying the structural requirements of the other component.This strategy more readily allows for variations to adapt to the hydroxylation of a broad range of substrates, as well as to different regulatory profiles.As pointed out above, even though the hydroxylation reactions occur with similar reaction mechanisms, the regulatory processes of the monooxygenase activity at the level of the reductase component in the case of the A. baumannii and S. coelicolor appear to be very different.Although such regulatory processes have not yet been explored for other two-component flavin monooxygenase systems, one might expect to find similar as well as new regulatory schemes for other systems.Consistent with this notion is the realization that two-component systems are continually being shown to be involved in the oxidation of an ever broader range of substrates.(20 µM), DHK ox (80 µM), in 50 mM Tris-HCl buffer, pH 7.4, containing 10 % glycerol, was incubated for 20 minutes in the dark in the presence of PCA (50 µM), PCD (0.12 unit.mL),EDTA (10 mM), and 5-deazaflavin (60 nM).The mixture was illuminated by white light until the FMN and DHK were fully reduced (5-10 s), and then was mixed at 4°C in the stopped-flow instrument with an equal volume of the same buffer containing oxygen (595 µM O 2 ). A. Traces of fluorescence (λ exc =390 nm and λ em > 530 nm) ( ) and of absorbance at 365 ( ), 445 () and 520 nm (∇) vs. time (dotted lines).The four reaction phases are demarcated by dashed vertical lines.The dependence of each individual phase on oxygen is indicated at the top of the figure.Solid lines are fits for multi-exponential models, with k 1 = 3.96 ± 0.05 x 10 4 M -1 s -1 , k 2 = 1.48 ± 0.15 s -1 , k 3 = 0.25 ± 0.03 s -1 for the 365 nm trace, k 3 = 0.16 ± 0.02 s -1 and k 4 = 0.013 ± 0.002 s -1 for the 445 and 520 nm traces and k 2 = 1.48 ± 0.15 s -1 , k 3 = 0.16 ± 0.02 s -1 , and k 4 = 0.013 ± 0.002 s -1 for the fluorescence trace.For the sake of clarity, all the traces are shifted to be zero intensity at 5 ms.B. Same protocol, but monitored with a diode array spectrophotometer.Spectra shown are recorded vs. anaerobic buffer (), and at 150 ms ( ), 2 s ( ), 200 s ( ) after the addition of O 2 .(20 µM) or solely FMN red (20 µM), as well as PCA (25 µM), and PCD (0.12 unit.mL -1 ) in 50 mM Tris-HCl buffer, pH 7.4, containing 10 % glycerol, were mixed with an equal volume of the same buffer containing either ActVA (45 µM) and O 2 (600 µM), HPAH-C 2 from Acinetobacter baumannii (45 µM), and O 2 (600 µM), oxidized DCPIP (20 µM), or menadione (40 µM).All reactions were carried out at 4°C and the concentrations given are those after mixing.The absorbance traces at 386 and 700 nm were recorded as a function of time (as in Fig. 3) and were fit with a single exponential model.The rate constants (s -1 ) obtained from the traces at 386 (a), 450 (b) and 700 nm (c) are given below.

Figure 1 .
Figure 1.Reaction of ActVA•FMN red with O 2 at pH 7.4, 4°C.A solution containing ActVA (50 µM) and FMN red(20 µM)  in anaerobic 50 mM Tris-HCl buffer, was mixed in the stopped-flow instrument with an equal volume of oxygenated 50 mM Tris-HCl buffer, (O 2 was 595 µM after mixing). A. UVvisible spectra of the mixture (from the bottom to the top at 450 nm) recorded before and 5, 62, 180, 300, 480, 660, 930, 1200, 1860, 2400, and 4320 s after mixing.B. Same protocol, but performed in the presence of various concentrations of O 2 (60, 125, 299 and 595 µM after mixing).The absorbance traces at 386 nm as a function of time (dotted lines) are shown for each O 2 concentration.Solid lines are fits to a single exponential model.(♦) is the absorbance trace at 445 nm in the presence of 595 µM O 2 , indicating the appearance of oxidized FMN. C. Plot of the pseudo first-order rate constants determined from the traces recorded at 386 nm shown in Fig.1Bas a function of O 2 concentration.A second-order rate constant of 2.97 ± 0.06 10 4 M -1 s -1 for the formation of the ActVA•C(4a)-FMN-OOH species was calculated from the slope of the straight line.

Figure 3 .
Figure 3. FMN red transfer from ActVB to ActVA and formation of C(4a)-FMN-OOH.An anaerobic solution containing ActVB (85 µM), NAD + (20 µM), FMN red(20 µM), PCA (50 µM), and PCD (0.12 unit.mL) was mixed in the stopped-flow instrument with an equal volume of oxygenated buffer containing ActVA (45 µM) and O 2 (600 µM) at 4°C.Both solutions were in 50 mM Tris-HCl buffer, pH 7.4, containing 10 % glycerol.Concentrations given are those before mixing.A. UV-visible difference spectra of the mixture (spectrum at a given time minus the initial spectrum) recorded from 7 ms to 1.5 s with a diode array spectrophotometer.The arrows indicate the direction of spectral evolution as a function of time.B. Absorbance traces (solid line) at 386 nm and 700 nm as a function of time.The smooth lines are fits to a single exponential model.Rate constants: k 386nm = 4.6 ± 0.1 s -1 and k 700nm = 5.0 ± 0.4 s -1 were determined.

Figure 4 .
Figure 4. Inhibition of FMN red transfer from ActVB to HPAH-C 2 by NAD + .An anaerobic buffered solution containing ActVB (235 µM), NAD + (40 µM), FMN red (40 µM), PCA (50 µM), PCD (0.12 unit.mL), was first mixed with an equal volume of the same buffer containing various concentrations of free NAD + (0 to 5000 µM).After 5 s this solution was mixed with an equal volume of the same buffer solution containing HPAH-C 2 (50 µM) and O 2 (255 µM).All concentrations given are those before mixing.The buffer used was 50 mM Tris-HCl, pH 7.4, containing 10 % glycerol, 4°C.The absorbance traces at 700 nm (A) and 386 nm (B), for each concentration of NAD + (10, 40, 125, 325, 625 and 1250 µM, concentrations after the second mix) are plotted as a function of time.The solid and smooth lines are fits for a single exponential model.C. The rate constants k obs 386 nm (○) and k obs 700 nm ( ) obtained from Figs. 4A and 4B are plotted as a function of NAD + concentration.The concentration of NAD + required to inhibit the rate of FMN red transfer from ActVB to HPAH-C 2 by 50 % is 40 µM.

Table 1 .
Rate constants determined from multi-exponential fits for the 365, 445, and 520 nm absorbance and fluorescence traces of Fig.2A.The k 1 (M -1 s -1 ) value is directly proportional to O 2 concentration, whereas k 2 , k 3 , and k 4 (s -1 ) values are independent of O 2 concentration.