Bacterial Luciferase

Txm JOURNAL cm BIOLOGICAL CHEMI~TBY Vol. 240, No. 8, Isme of April 25. pp. 2762-2768, lQ75 Printed in U.S.A.

Bacterial BINDING

Luciferase

OF OXIDIZED

FLAVIN

MONONUCLEOTIDE* (Received

THOMAS 0. BALDWIN, MIRIAM

Z. NICOLI,

JAMES E. BECVAR,

for publication,

AND J. WOODLAND

August 23, 1974)

HASTINGS

From the Biological Laboratories, Harvard University, Cambridge, Massachusetts 0.2158

SUMMARY Bacterial luciferase catalyzes a bioluminescent oxidation of reduced flavin mononucleotide; the products include a photon and oxidized FMN. The experiments reported here show that luciferase binds oxidized flavin mononucleotide in a 1: 1 molar ratio with an apparent dissociation constant of 1.2 X lo-’ M at 3” in 0.05 M 2,2-bis(hydroxymethyl)-2,2’,2”nitriloethanol (bis-tris), pH 7.0. Analysis of the binding at temperatures between 3 and 30” indicates an enthalpy of binding (AEZ,) of -10.0 kcal per mol. The absorption spectrum of luciferase-bound FMN shows considerable alteration relative to that of free flavin. There is one major peak at 366 run, and the 445~nm band is resolved into two distinct peaks at 434 and 458 nm; this spectrum is indicative of binding in a nonpolar environment. The circular dichroism spectrum of FMN bound to luciferase has structure which correlates well with the optical absorption spectrum of the bound flavin. The detail in the spectra of the bound FMN probably reflects the resolution of vibrational structure which is blurred in polar environments. The optical activity shown by the CD spectrum presumably results from binding in an electronically asymmetric fashion. Although FMN free in solution is highly fluorescent, FMN bound to luciferase is nonfluorescent, thus indicating that the emitting species is not an excited state of product FMN located in the same site in which luciferase binds oxidized FMN.

(I)

FMNH,

+E -E-FH

FMN + H,O,

o2

2

E + FMN + H202 (dark)

* This work was sunported in part by National Science Foundation Research Grant GR 31977X ito J. w. H.) and by United States Public Health Service Postdoctoral Fellowshins (to T. 0. B. and J. E. B.). Preliminary reports of these results have appeared (1,2). 1 The abbreviations used are: FMNHz and FMN, reduced and 2,2-bis(hydroxyoxidized riboflavin 5’-phosphate; bis-tris, methyl)-2,2’,2”-nitriloethanol.

. 2

I + JI

L

1 E + FMN+ RCOOH +H20 + hv495 REACTION

Bacterial luciferase is a 79,000-dalton dimer composed of nonidentical subunits, (Y and /3 (3, 4), having different molecular weights and different primary structures (5). The enzyme catalyzes the oxidation of FMNHQ and a long chain aldehyde by molecular oxygen to yield the corresponding acid, HzO, FMN, N 495 run) (6-9). Hastings and Gibson (10) and light (X,,, deduced a reaction scheme (Reaction 1) in which FMNHz re-

,

1

Knowledge of the interaction of the product FMN with luciferase is of interest for several reasons. First, the identity of the emitting species is not yet known, but it most certainly involves a flavin of some sort bound to enzyme (14, 15). Secondly, the binding of FMN should provide an excellent probe of the luciferase active center, with regard to both the number of sites and the protein environment of the site(s) (16, 17). The present experiments confirm the existence of a single FMN binding site (18). Moreover, FMN bound to luciferase is altered significantly in its absorbance properties and is nonfluorescent. Since the bioluminescent quantum yield with respect to flavin is about 0.05 to 0.15 (12, 19), the results indicate that the emission in the bioluminescent reaction does not occur from the excited state of product FMN located in the same site in which luciferase binds oxidized FMN.

2763

Downloaded from www.jbc.org by guest, on July 10, 2011

versibly binds to the enzyme, forming an enzyme-reduced flavin complex (I). Any FMNHt not bound to the enzyme is autoxidized in less than 1 s (11). Intermediate I reacts irreversibly with O2 to form a long lived intermediate, II (tl,z N- 10 s at 22’), which recently has been isolated by chromatography at -20” and found to break down in the absence of aldehyde to release FMN and HzOz, producing little or no light (12). Aldehyde, when present, is reversibly bound to form IIA, which is also long lived (13). This intermediate can proceed to form an excited species (*) whose decay produces the light. The scheme in Reaction 1 is simplified; for example, it does not include the possible nonluminescent breakdown of intermediates such as HA and the excited species.

2764 EXPERIMENTAL

PROCEDURE

RESULTS

Spectra 01 Enzyme-bound FMN-The most striking feature of the optical absorption spectrum of FMN bound to bacterial luciferase is its highly resolved structure (Fig. la). The bound flavin shows well defined maxima at 458, 434, and 366 nm with distinct shoulders at about 490,410, and 355 nm, and minima at Wavelength (nm)

I.. 20,ocO

!500

450 ,‘-‘~I’~

I,,>,, 25,000 Wovenumber km-‘) Wavelength (nm) 400

30,000

7

350 I

7

1

1

3.0 -

-2.o-

-3o-

-401,

, 20,000

,

,

,

,

,

25,000 Wavenumber km-‘)

,

,

,

,

j’

30,OOC

FIG. 1. a, optical absorption spectra of free (- - -) and luciferase-bound (-) FMN at 2 f 2”. The difference between the free and bound flavin spectra is plotted in the top panel. Compartment A contained luciferase with flavin, Compartment B contained buffer, Compartment C contained luciferase without flavin, and Compartment D contained the flavin-buffer solution against which the protein in Compartment A had been dialyzed. Distinct absorption maxima for the bound species occur at 21,830, 23,040, and 27,326 cm-l. b, circular dichroism spectra of free (-.-) and luciferase-bound (-) FMN at 6 f 1”. The luciferase-bound FMN spectrum was obtained by correcting for contributions of protein (- - -). Positive maxima for the bound species occur in the CD absorbance at 21,830 and 22,990 cm-r, and negative maxima at 20,410 and 27,860 cm-*.


Materials-FMN and bis-tris were products of Sigma. The FMN, stored desiccated at -2O”, was estimated from thin layer chromatography in a variety of solvent systems to be more than 90% pure. This level of purity is confirmed by the homogeneity of binding to luciferase observed in these experiments. FMN concentrations were determined spectrophotometrically on the basis of a molar absorption coefficient of 12,200 M-l cm-l at 450 nm (20). All binding studies were done in 0.05 M bis-tris-HCl, pH 7.0. Buffers for use in equilibrium dialysis contained, in addition, 0.2 M NaCl. Luciferase Purification-Luciferase was purified from Beneckea harveyi (21), previously designated MAV (4)) by a procedure modified from that of Gunsalus-Miguel et al. (22). Cells were grown in a 250-liter New Brunswick Fermacell Fermentor, harvested, and stored at -20”. The cell paste was allowed to thaw overnight at 4”, mixed 1:l (w/v) with 30/, NaCI, and lysed osmotically, 100 g cells per liter of low ionic strength buffer (8 X 10e3 M Tris base, 5 X 10-e M dithiothreitol). After 30 min, DEAE-cellulose in the phosphate form (Whatman DE32) was added directly in a ratio of 100 ml of settled volume of DEAE-cellulose per liter of lysate. After adjusting the pH to 6.7 f 0.5 and allowing 30 min for adsorption, the material was vacuum filtered using Miracloth (Calbiochem) on a large Buchner funnel, rinsed twice with 100 ml of HzO, and five times with lOO-ml aliquots of 0.15 M phosphate, pH 7.0. Luciferase then was eluted batchwise with five lOO- to 200-ml aliquots of 0.35 M phosphate buffer, pH 7.0, and precipitated with ammonium sulfate between 40 and 75% of saturation. The luciferase then was suspended in a minimum volume of 0.25 M phosphate, 5 X lo-’ M dithiothreitol, 1 X 10m3M EDTA, pH 7.0, and dialyzed against the same buffer, in preparation for DEAE-Sephadex A-50 column chromatography (22). All of the above procedures were carried out at O-4”. The DEAE-cellulose was regenerated by washing with 0.2 M NaOH, H20 to neutrality, 0.2 M HaPO+ HzO, and 0.1 M phosphate buffer, pH 7.0, and stored in H20. New batches of DEAE-cellulose were carried through a regeneration cycle prior to use. We recently have found that this luciferase is highly susceptible to inactivation by proteases (23, 24), but is effectively protected from proteolysis by phosphate. The routine use of phosphate buffers may thereby contribute to the stability of the luciferase in extracts. For use in these experiments, the Beneckea harveyi luciferase was rechromatographed through a second column of DEAE-Sephadex A-50 to reduce the amount of the light-inducible form of luciferase (25). Luciferase concentrations were determined spectrophotometrically based on a molecular weight of 79,000 and an absorption coefficient of 0.94 (O.lyo, 1 cm) at 280 nm. Optical Absorption Spectroscopy-Absorption spectra were obtained with a Cary 15 spectrophotometer using two methods. TO determine the spectrum of enzyme-bound FMN, tandem cuvettes were used, with samples prepared by equilibrium dialysis at concentrations ranging from 1 to 6 X lo-’ M luciferase and 2.5 X 10es to 2.5 X 10-b M FMN. The spectrum of the flavin-protein mixture inside the bag was measured using the free flavin taken from outside the bag in one of the reference cell compartments. Since the free FMN concentration should be the same on both sides of the dialysis membrane, the observed spectrum should be that of the enzyme-bound FMN only. To compensate for luciferase absorbance and light scattering, free enzyme alone was placed in the second compartment of the tandem reference cuvette and buffer in the second sample cuvette compartment. To determine the isosbestic points, the spectrum of concentrated luciferase containing 7 X 10-K M FMN first was measured versus the free enzyme in the reference cell. The luciferase with FMN then was diluted with buffer containing an equivalent total FMN concentration and the spectrum was determined versus the reference sample of enzyme diluted in an equivalent fashion with buffer alone. Each dilution decreased the concentration of enzymebound favin by decreasing the total enzyme concentration while maintaining a constant total flavin concentration, thereby revealing an isosbestic point at 492 nm. The spectrum of luciferase-bound flavin obtained as described above then was normalized to the 492nm absorbance of free FMN to determine the molar absorption coefficient of luciferase-bound FMN at all wavelengths. Circular Dichroism Spectroscopy-CD spectra were taken with a Jasco J-20 spectropolarimeter, using a simple titration technique.

The reactants were delivered directly to the cuvette with microliter pipettes to yield the various luciferase-flavin mixtures. The path length of the sample solution was chosen to maintain an absorbance value below 2 at wavelengths above 300 nm. Fluorescence Spectroscopy-Fluorescence measurements were made with a Hitachi Perkin-Elmer MPF-PA fluorescence spectrophotometer.

2765

J-

[Ed

Wavelength 600 l”I’I

(nm)

550 “1

Wavenumber

I

km-’

500 I

@

’

)

3. Fluorescence emission spectra of FMN in the presence and absence of luciferase. Samples (8.9 X 10-6~ FMN f 4.9 X lo-’ M luciferase) in 0.05 M bis-tris, pH 7.0, at 3”, were excited at 450 nm. FMN alone, -; FMN + luciferase, - - -; FMN + luciferase, normalized for comparison with free FMN, 0. FIG.

446 and 392 run. The free FMN spectrum shows maxima at 444 and 373 nm with a shoulder at about 466 nm, and a minimum at 399 run. The difference spectrum of free FMN versus luciferasebound FMN shows large negative peaks at 476,445, and 385 nm, and positive peaks at 503 and 342 nm. Binding of oxidized flavin mononucleotide to bacterial luciferase generates substantial optical activity not exhibited by the unbound flavin (Fig. lb). The circular dichroism spectrum shows positive peaks at 458 and 435 m-n, a positive shoulder at about 405 run, and negative peaks at 490 and 359 nm, correlating well with the structure observed in the optical absorption spectrum. Stoichiom&y of FMN Binding-The stoichiometry of FMN bmding to luciferase was determined by the method of continuous

x 10-4

(Id-‘)

FIG. 4. Quenching of the fluorescence of 8.9 X 10-a M FMN as a function of luciferase concentration. Samples in 0.05 M bis-tris, pH 7.0, were excited at 450 nm and fluorescence (F&s) was monitored at 525 nm. Kd for each temperature was calculated from the slope of the line (Equation 1) and is shown in Table I.

variation of Job (26, 27). The mole fractions of the two interacting species (FMN and luciferase) were varied and a parameter sensitive to the concentration of the complex, CD absorbance at 435 run, was monitored. At this wavelength, the CD intensity of the complex is maximal, while that of each free species is relatively small. The Job plot (Fig. 2) shows that the maximum amount of enzyme-flavin complex develops when the mole fraction of FMN is approximately 0.43, or the molar ratio of FMN to luciferase is about 0.8 : 1.0. This result confirms the observation of a single FMN binding site per enzyme molecule obtained by Scatchard analysis of data from equilibrium dialysis (18). The binding of FMN to luciferase causes dramatic quenching of the flavin fluorescence. This is shown in Fig. 3, where the fluorescence emission spectrum of FMN is shown in the presence and absence of 4.9 x 10e4 M luciferase. It may be seen that at 3’ this concentration of enzyme quenches the fluorescence of FMN about SO%, and uniformly so at all wavelengths. The decrease in flavin fluorescence upon binding to luciferase permits a sensitive measurement of the extent of flavin binding during the titration of a low concentration of FMN by increasing enzyme concentrations (28). The observed fluorescence at 525 run, Fobs, is related to &, the dissociation constant for the enzyme-flavin complex, and jFMN and js:vMN, the intrinsic fluorescence of free and bound FMN, respectively, by Equation 1 FFMN f f FMN FMN F

FMN

-F

Kd

ohs

-

fFMN - fE:FMN

[El

+

f-

- fEzFMN

(1)

where FFMN is the fluorescence of the flavin solution in the absence of enzyme. The total concentration of enzyme [El] is much greater than that of FMN, so [E] G [Et]. Linear plots of - Fobs) versus l/[Et] were obtained at each of FFMN/@FMN several temperatures between 3 and 30” (Fig. 4). The ordinate intercept (jvMn/ljv~N - jB:rMN]) in every cw is 1.0, indicating that the intrinsic fluorescence of the enzyme-flavin complex, jB:vMN, is zero ( . W., WEBER, G., DUANE, W., AND MASSA. J. (1966) in Flavins and Flavovroteins @LATER. E. C.. . ’ ’ ed) pp: 341-359; Elsevier, Amsterd& MEIGHEN, E. A., NICOLI, M. Z., AND HASTINGS, J. W. (1971) Biochemistry 10, 40624068 MEIOHEN, E. A.. NICOLI, M. Z., AND HASTINGS, J. W. (1971) Bioche&try lb, 40694073 CLINE. T. W.. AND HASTINGS. J. W. (1972) Biochemistru ” 11. I 335g3370 FRIEDLAND, J. M., AND HASTINGS, J. W. (1967) Biochemistry 8, 289342900 CLINE, T. W., AND HASTINQS, J. W. (1974) J. Biol. Chem. .

249, 46684669

I

35. MEIGHEN, E. A., AND HASTINGS, J. W. (1971) J. Biol. Chem. 248, 7666-7674 36. PALMER, G., AND MASSEY, V. (1968) in Biological Ozidations (SINGER, T. P., ed) pp. 263-300, Interscience, New York 37. KOZIOL, J. (1969) Photochem. Photobiol. 9, 45-53 38. SUN, M., MOORE, T. A., AND SONG, P.-S. (1972) J. Amer. Chem. Sot. 94, 1730-1740 39. WEBER, G., AND TEALE, F. W. J. (1957) Trans. Faraday Sot. 63, 646-655 40. LEE, J., AND MURPHY, C. L. (1973) Biophys. Sot. Annu. Meet. Abstr. 13, 274a 41. EHRENBERG, A., AND HEMMERICH, P. (1968) in Biological Ozidations (SINGER, T. P., ed) pp. 239-262, Interscience, New York 42. BEINERT, H. (1956) J. Amer. Chem. Sot. 78, 5323-5328 43. DUDLEY, K. H., EHRENBERG, A., HEMMERICH, P., AND MUELLER, F. (1964) Helv. Chim. Acta 4’7, 1354-1383 44. EBERHARD, A., AND HASTINGS, J. W. (1972) Biochem. Biophys. Res. Commun,. 47. 348-353
