The inhibition of bacterial luciferase by mixed function oxidase inhibitors.

Abstract On the hypothesis that bacterial luciferase may be classed as a mixed function oxidase, compounds reported to be specific inhibitors for such enzymes were tested. Inhibitors designated as SKF (2-diethylaminoethyl-2,2-diphenyl valerate), DPDA (N,N-diethyl-2,4-dichloro(6-phenylphenoxy)-ethylamine), and DPEA (2,3-dichloro(6-phenylphenoxy)ethylamine) were effective in blocking the in vitro reaction of pure luciferase at inhibitor concentrations between 10-5 and 10-4 m. The mode of action of the two different types of compounds has been found to be distinctly different. One (SKF) interferes with the reaction of one of the substrates, namely FMNH2, while the other (DPEA) blocks the site for aldehyde, the second substrate. Both compounds are effective inhibitors of bacterial growth and luminescence, affecting both functions in similar although not identical degrees.


The Inhibition of Bacterial Luciferase by Mixed Function Oxidase
Inhibitors* (Received for p\lblication, September 23, 1971) SUMMARY On the hypothesis that bacterial luciferase may be classed as a mixed function oxidase, compounds reported to be specific inhibitors for such enzymes were tested. Inhibitors designated as SKF (Z-diethylaminoethyl-2,2-diphenyl valerate), DPDA (N,iV-diethyl-2,4-dichloro(6-phenylphenoxy)ethylamine), and DPEA (2,3-dichloro(6-phenylphenoxy)ethylamine) were effective in blocking the in uifro reaction of pure luciferase at inhibitor concentrations between lo-" and 10e4 M. The mode of action of the two different types of compounds has been found to be distinctly different. One (SKF) interferes with the reaction of one of the substrates, namely FMNH2, while the other (DPEA) blocks the site for aldehyde, the second substrate.
Both compounds are effective inhibitors of bacterial growth and luminescence, affecting both functions in similar although not identical degrees.
Most, possibly all, bioluminescent. reactions require molecular oxygen as a substrate (1). In the reactions of bot,h the firefly (2, 3) and the crustacean Cypridina (4), the stoichiornetry involves the utilization of 1 molecule of oxygen per molecule of substrate and the incorporation of both atoms of oxygen. In both cases the luciferin molecule is split, yielding ground state CO2 as one product and an aromatic species in the excited state as the other. In the formal sense these luciferases may be classed as oxygenases (5,6). In the firefly one of the oxygen atoms appears in t.he aromatic product and the second in water (7), while in Cypridina one of the oxygen atoms has beeu reported to occur in the CO2 (8).
In the bioluminescent reaction catalyzed by bacterial luciferase the reaction may be more strictly classed as a nlonooxygenase, or mixed function oxidase (1). In addition to oxygen and FMNHz the reaction requires a long chain aldehyde ( >8 carbon atoms), the over-all enzymatic reaction having beeu postulated to occur as shown, where FMN* is oxidized flavin in an electronically excited state. This scheme is only a postulate, neither the stoichiometr\ 11w the products having been experimentally established. Nevertheless, the availability of compounds reported to be specific inhibitors for reactions of the mixed function osidase type was of interest, especially with regard to establishing the irltel,mediate steps in the luminescent reaction and possibly more general enzymatic features of mixed function oxidases. Compounds designated as SKF' 525-A, DPDA, and DPEA ( Fig. 1) have been reported to be effective inhibitors of the I'-450 mixed function oxidase system from liver microsomes (9). Sl,udies have also indicated the efficacy and specificity of these inhibitors in a bacterial system (10). III the present study we describe a11 inhibitory effect of these compounds on bacterial bioluminescence, both in vivo and in vitro. Iprom an analysis of their inhibitory effects in the in vi&o system, with pure luciferase, the mode of action of the two different types of compounds has been found to be distinctly different. One (XI(F) interferes with the binding (or reaction, or both) of reduced flavin mononucleotide (FIZIXHZ), while the other (I)PEA) apparently reacts with the luciferase at the binding site for aldehyde. MATERIAT In addition to MAV and Pf, several other strains of luminous bacteria were used for in uivo experiments.
These strains are Issue of li'ebruary 10,1972 K. H. ~Vmlson atrd .I. IV. Jlastiugs 889 from diverse origins; they include both symbiotic and free living forms. The strain numbers used here are those which will be used in a later publication describing the origins and properties of these strains.
In the medium used for growth the NaCl concentration was changed from 3 to 1% in some experiments.
The medium consisted of 30 g (or 10 g) of NaCI, 7.0 g of NazHP0J.7Hz0, 1 g of KH2P04, 0.5 g of (NH&PO+ 0.1 g of MgS04, 3 ml of glycerol, 0.5 g of Difco yeast extract, and 5 g of Difco Bacto Tryptone per liter of distilled water.
For t,he "flavin" assay, stock O.lyO v/v aldehyde solutions were prepared as suspensions by sonication of 10 ~1 of aldehyde in 10 ml of distilled water.
Although molarities after dilution in the reaction mixtures were calculated assuming that all of the aldehyde was in solution, this is probably not so. Thus K, values given may not be exactly correct, but the relative values obtained in the presence and absence of inhibitors were reproducible and are considered to be significant.
Aldehydes prepared as saturated aqueous solut)ions might have been used instead, but these are difficult to prepare and suffer from capricious autoxidation (13). Light-measuring apparatus consisted of a light-tight chamber, designed to hold a standard scintillation vial exposed by means of a shutter mechanism to a photomultiplier tube (RCA -lP21). After appropriate amplification the output was monitored on an Esterline Angus Speed Servo recorder and expressed in quanta see-', established by the standards of Hastings and Weber (14).
Reduced flavin-initiated in vitro assays (flavin assays) were initiated by the injection via a syringe of 1 ml of 5 X 1O-5 M reduced FMNHS into the sample vial containing the luciferase in 1 ml of assay buffer (0.01 M phosphate buffer, pH 7.0, 0.1% BSA) plus 50 ~1 of the appropriate stock aldehyde (except decanal with MAV, which was 10 ~1).
Oxygen-initiated in vitro assays (15), referred to as dithionitc assays, were accomplished by the injection of 1 ml of an airsaturated dodecanal solution (5 ml of stock aldehyde in 100 ml of distilled water) into a reaction vial containing 10 ~1 of luciferase in 1 ml of assay buffer, 1 ml of 5 X 10m5 M FMN, and about 5 mg of solid dithionite (Na&O.J, adequate to deplete the oxygen and fully reduce the flavin. Light-initiated bioluminescence was measured by the method of Hastings and Gibson (16). Pure light-induced protein, 10 ~1 (17, 18), was mixed with 2 ml of 0.05 M phosphate buffer in a Vycor syringe and irradiated by flash discharge (19). The irradiated sample was then injected into a sample vial positioned above the phototube containing 20 ~1 of dodecanal stock in 1 ml of 0.05 M phosphate buffer. In all assays, the activity was measured by recording the initial maximum light intensity (IO). In both the flavin and dithionite assays the initial intensities and kinetics are identical when identical amounts of luciferase are used.
Flavin-binding experiments were done with the dithionite assay, because at low FMNHz concentrations ( <lop7 M), the amount of autoxidation in the syringe is so significant that the injection method of the flavin assay is neither accurate nor reproducible (15). Dilutions of the stock flavin solution were used and the total volume before aldehyde addition was adjusted to 2 ml with distilled water.
Cell growth and in vivo luminescence were monitored as previously described (20) MAV cells from a late exponential phase culture were diluted 1 to 20 into either 3%, 2y,, or 1% NaCl (indicated by 3,d, and 1, respectively). After 10 min, 10 ~1 of the indicated inhibitor were.added and luminescence wasmonitored.
in viva luminescence was measured with 50 ~1 of growing cells suspended in 1 ml of an NaCl solution (1, 2, or 3%) and adding the inhibitors to this volume.
The long term effect was determined by growing the cells in medium containing various amounts of the inhibitors.
The effect of the inhibitors on the in vitro activity was determined by mixing the inhibitor with the assay buffer-enzyme solution before addition of the flavin (or aldehyde in the dithionite assay). No difference was seen if the inhibitor was injected simultaneously with the FMNH2.

RESULTS
Freshly grown cells, harvested during the exponential period of growth and resuspended in a nonnutrient 3.0% NaCl buffer will continue to emit light without a significant increase or diminution in intensity for periods of 20 to 40 min. The addition of DPEA to such a cell suspension results in a prompt decrease in light emission (Fig. 2). After the addition of SKF, there is a transitory and unexplained stimulation, followed by an inhibition, the degree of inhibition being greater at lower salt concen- to produce an intermediate la, which after reaction with oxygen has properties similar to the analogous chemically generated intermediate II.

trations.
The lower osmolarity apparently facilitates inhibitor penetration, an effect previously'reported for these cells (12). A number of different strains of luminous bacteria were tested in both 1 and 37, NaCl for their response to these inhibitors and found to differ considerably in sensitivities, probably due to differences in permeability ( Table I). The fact that the luminescence of at least one strain was blocked in 3% NaCl by each of the inhibitors indicates that permeability is involved, not an inactivation of the inhibitor by high salt. In addition, nearly every strain tested was more sensitive in 1% than in 3% NaCl.
In order to determine whether or not the inhibition of the in viva luminescence by these compounds could be attributed to their action on the luminescent reaction itself, studies were carried out with the in v&o system, with pure luciferase. TWO distinctly different types of luciferases were used (Pf and XIAV, see "LIaterials and Methods") and while some differences (as noted below) were found, the effects were similar in most respects. With DPEX and DPDA the inhibition was found to be attributable to reaction, or competition, or both, at the aldehydebinding site. SKF, on the other hand, apparently interacts with the luciferase at or near the reduced flavin-binding site, thereby causing an inhibition.
This difference and the experiments which indicate it can be best presented in terms of the previously postulated reaction scheme (1, 21). As shown in Fig. 3, luciferase (E) reacts with the first substrate (FMNHa), then oxygen, and finally (but optionally) aldehyde.
The nonobligatory nature of the aldehyde step is particularly significant, so that the reaction up to Intermediate II proceeds in essentially the same way whether aldehyde is present or not. Beyond that, the reaction may proceed either with or without aldehyde, the clear difference being that the photon yield is high in the presence of aldehyde and low in its One unit = 2 X 1Or0 quanta see-l. FMNHt concentrations given are before mixing; final concentrations are one-half of these absence. The experiments of Fig. 8, which will be presented in a later section, may be referred to in order to clarify this point.
The difference between the inhibitory character of the two compounds may be seen in Fig. 4. With DPEA the initial intensity is less, but the rate of decay is also slower, so that the total amount of light obtained is similar.
In fact, the yield is, within error, the same for the two (see inset, Fig. 4). Referring to the reaction scheme (Fig. 3)) we can conclude that with DPEA the same amount of Intermediate II must have been formed in the first instance, and that while it was discharged over a longer time period, the relative amount reacting via the aldehyde (high quantum yield) pathway was unchanged.
On the other hand, when the luminescent reaction is inhibited with SKF, the rate of decay is not significantly different, so that a lower yield accompanies an inhibited reaction. This can be explained by assuming that the compound inhibits the actual formation of intermediate, possibly from the very outset, but that the intermediate which is formed possesses essentially unaltered kinetic properties.
The inhibitor thus might act by blocking at the initial stage of reaction between FMNHz and luciferase.
Similar experiments are illustrated in Fig. 5, plotted on a logarithmic scale to facilitate comparison of the rates of decay of luminescence in the different cases. In this figure we have included as well experiments carried out at a loo-fold lower concentration of FiMNH2 which indicate that SKF, but not DPEA, is competitive with FMNHZ.
The experiments shown in Fig. 6 provide additional evidence &pj-X Id6 CM-') 2 FIG. 6. Reciprocal plot of the relationship between FMNHz concentration and initial light intensity, in the presence and absence of inhibitors.
The data for the control (X) and with SKF (0) are plotted on a linear scale in the inset. Reaction mixtures included 10 ~1 of stock luciferase and, when used, 5 ~1 of inhibitor. In order to illustrate the lack of effect of DPEA upon the Km, the data was normalized to the V,,, obtained in the control and SKFinhibited reactions.
Ordinate, reciprocal of velocity of reaction, calculated in intensity units, where 1 unit = 2 X lOlo quanta se@. The data for the control (X) and with DPEA (0) are plotted on a linear scale in the inset. Reaction mixtures and ordinates as given in legend to Fig. 6. V,,, is not reached with the inhibitor competitive with aldehyde because the reaction with MAV luciferase is inhibited at high aldehyde concentrations (II), as is evident from the control.
The V,,, of the control was used in plotting the data. The data for the SKF-inhibited reaction was also normalized to the control V,,, as described in Fig. 6. for the competitive nature of the interaction of SKF at the flavin site for luciferase.
Without inhibitor present the apparent Michaelis constant for the binding of reduced flavin is 7 x lo-' M; in t,he presence of 1.7 X 1O-4 M SKF the Km is 4 times greater (2.7 x low6 M).
In the case of reactions inhibited by 1.1 X lop4 M DPEA the Km for FMNHZ is not greatly altered.
DPEA is, however, competitive with aldehyde, while SKF is not (Fig. 7). The value obtained for the uninhibited enzyme An additional and very significant feature of the inhibition by DPEA is illustrated in Fig. 8. In the earlier experiments it was shown that the lifetime of Intermediate II is considerably extended in the presence of the inhibitor, while at the same time its quantum yield is unaffected. The question as to whether or not the presence of aldehyde is required for DPEA to have this effect on the lifetime of the intermediate was examined in the experiments of Fig. 8. Reactions were initiated without aldehyde and without inhibitor in one series (A), and without aldehyde but with inhibitor in the second group (B). Aldehyde was then added secondarily at several later times. The partial time course for each of these reactions is shown. By connecting the initial maximum intensity for each of these (dotted line), one can deduce the lifetime of the intermediate. The time course for a reaction in which both aldehyde and inhibitor were present from the beginning is also shown, illustrating that the lifetime of the intermediate in the presence of inhibitor is extended in a similar way, both with and without aldehyde. The action of the inhibitors when added secondarily (after the reaction has been initiated) provides an interesting method for examining the effect of these inhibitors on the previously formed intermediates during the decay phase of the reaction. SKF has no effect on the luminescence when added secondarily. This is expected since its action appears to involve the reduced flavin site, where reaction is essentially complete within 1 set after FMNHz addition. Since excess FMNH? is oxidized nonenzymatically (Fig. 3), no further substrate is available to the luciferase after the first exposure.
DPEA, on the other hand, exerts a marked effect when added secondarily (Fig. 9). This effect is quite different with different aldehydes, being most marked with hexanal and octanal, but less with decanal and absent with dodecanal. There is also some difference with the other type of luciferase, Pf, for only a slight effect of secondary addition of DPEA occurs with hexanal, and there is no effect when any of the other aldehydes are used. This marked difference between Pf and MAV luciferases regarding the effect of sec0ndar.y addition of DPEA is the only major difference between the two luciferases found with respect to their interactions with these inhibitors.
An alternative and novel method for initiating in vitro bacterial bioluminescence was reported some years ago by Gibson et al. (19). Luciferase preparations exposed to an intense brief light from a flash discharge lamp were found to subsequently emit a kinetically characteristic aldehyde-stimulatable bioluminescence. The only differences are that this light-induced bioluminescence involves modified luciferase (17) and it neither requires nor is it stimulated by added FMN; the pathway was therefore postulated as shown in Fig. 3, entering at the stage of Intermediate II (16). A strong prediction of the present studies with inhibitors is that the DPEA should inhibit this light induced bioluminescence and SKF should not. This was indeed found to be the case (Table  III).
Hammond and White (10) reported that the carotenoid hydroxylation system of Staphylococcus aureus was more sensitive Ordinate, light intensity.
One unit = 2 X lOlo quanta se@. greater at any given cell density than the level of luminescence in the presence of inhibitor. Clearly no differential effect of this type was seen. In fact,, with I)l'EA a small but clear differential effect favoring luminescence was observed.
The luminescence of cells at mid-log phnhe was sli&t,ly greater than that of uninhibited cells at t,he same density (Fig. 10, Zest). With SKF a pronounced differential inhibition of in viva bioluminescence occurred, but only at the very early stages of growth (Fig. 10, right), at a time prior to the onset of the synthesis of new luciferase (20). The fact that this does not occur at later stages may be connected with the character of the SKF inhibition of growth, which involves a concentration-dependent lag, during which little growth occurs, followed by growth at a rate very similar to that of the uninhibited control (Fig. 10, right inset). It seems likely that the SKF is detoxified during this lag period, and that the apparent differential inhibition of luminescence may be only a reflection of the inhibit,ion of that luciferase which was present at the time of inoculation.
This pattern of growth inhibition is strikingly different from that which occurs with DPEA (Fig. 10, left inset) where the rate of growth is affected by and in proportion to the amount of the inhibitor present.  (7), Cypridina (4), and Renilla (23) might provide valuable contributions to our understanding of the more general aspects of the mechanism in bioluminescent systems, since the reactions in these differ considerably.
A prediction of the model proposed for DPEA inhibition is that the rate of luminescence decay, i.e. the apparent first order constant for the decay, should be related to the ratio of inhibitoraldehyde present, by virtue of the equilibria shown in Fig. 3 (ITaId e II * II DPEA). Decay rates intermediate between those shown in Fig. 5 (with and without DPEA) occur either with less inhibitor or with more aldehyde, independent of absolute concentrations. dcknowledgment-We are grateful to Miss Ellen Rothenberg for her assistance.

REFERENCES
Ijacterial luciferase provides an unusually rapid and sensitive method for the study of oxidase inhibitors.
The instantaneous reaction rate is given directly by the light intensity, eliminating the necessit,y for measuring the accumulation of a reaction product which in other systems may sometimes be remote from the point of inhibition.
Probably the most important advant'age of the luciferase system relates to favorable enzymology.
The enzyme is aT-ailable in the pure state, its subunit structure has been elucidated, and the substrates and certain aspects of the intermediate steps are known (11,22). In addition, the lifetime of the reaction intermediates is estraordinarily long (tens of seconds or minutes), making possible the analysis of the individual steps (21).
From the present study it is already clear that the two types of inhibitor,+ act at different sites on the lucifernse, each being competitive with a different one of the substrates.
The results are readily interpretable in terms of the intermediates and steps in the reaction 1)reviously postulated (Fig. 3). The specific and different effects of the two inhibitors is also of considerable interest in terms of the interpretation of the effect of these inhibitors ill other oxidase systems, where such distinctions have not yet been reported.
It seems reasonable that the site blocked by SKF, competitive with FX?L%, might be similar to analogous sites on other mixed function osidase enzymes, where flavins are in\rolved iu many instances (5). With DPEA and DPDS the more general basis for inhibition is less evident, since long chain nldehyde is not a substrate in most oxidase systems. It is clear that Dl'M inhibition does not involve the oxygen site as such, since secondary addition of DPEA inhibits, and it is known that the oxygen step occurs only at the very beginning (21). Indeed after initial reaction with oxygen the luminescence is independent of oxygen and secondary inhibition by DPEA occurs equally well with or without oxygen.
The fact that secondary inhibition by DPEA is ineffecti\-e with some aldehydes may relate to t,he relative strengths of binding of aldehyde and inhibitor.