Proc. Nati. Acad. Sci. USA Vol. 77, No. 1, pp. 249-252, January 1980 Biochemistry

Physical interaction and activity coupling between two induced by immobilization of one (bacterial //flavin reductase/flavoproteins/monooxygenase) SHIAO-CHUN Tu* AND J. WOODLAND HASTINGS The Biological Laboratories, 16 Divinity Avenue, Harvard University, Cambridge, Massachusetts 02138 Communicated by E. R. Blout, October 15, 1979

ABSTRACT Flavin reductase and bacterial luciferase are to provide a partially purified flavin reductase, which exhibited believed to be coupled in the in vivo light emitting reaction. In an activity of 0.21 unit/mg; 1 unit is defined as the oxidation extracts, however, they are both soluble enzymes that exhibit little or no association. Immobilized luciferase, covalently at- of 1 ,mol of NADH at 23°C in 1 ml of 0.05 M phosphate, pH tached to Sepharose, was found to bind the soluble reductase 7/50 jAM FMN/0.2 mM NADH. Cyanogen bromide and and to exhibit activity in the coupled reaction with an enhanced Sepharose 6B were products of Eastman and Pharmacia, re- efficiency of electron transfer. spectively. FMN and NADH were obtained from Sigma. Luciferase activity was determined in a standard assay Bacterial luciferase utilizes reduced (nonturnover) as the initial maximum light intensity upon the (FMNH2) as its , a compound that is highly autoxi- injection of 1 ml of 50 pM FMNH2 (catalytically reduced) into dizable (1, 2). In cell extracts there are enzymes (reductases) 1 ml of 0.05 M phosphate (pH 7) containing 0.2% bovine serum capable of producing FMNH2 from NAD(P)H In flavin reductase albumin, 0.001% decanal, and luciferase. this nonturnover NAD(P)H + H+ + FMN - NAD(P)+ + FMNH2 assay, all of the luciferase-flavin intermediate is formed within the first second, at which time the bioluminescence intensity FMNH2 + 02 + CH3(CH2)nCHO (light-emitting rate), in quanta sec-1, reaches a maximum (IO). luciferase Light intensity then decreases exponentially; the first-order rate -* light + FMN + H20 + CH3(CH2)nCOOH constant of this decay provides a measure of the catalytic rate. The reduced flavin produced in vitro by the reductase is free In the coupled assay, the bioluminescence is initiated by the and subject to autoxidation (3, 4). But it has always seemed to addition of flavin reductase to 1 ml of 0.05 M phosphate (pH us that in vivo there must be some complex between the re- 7) containing luciferase, 0.2% bovine serum albumin, 0.001% ductase and luciferase such that the cellular supply of reduced decanal, 50 ,uM FMN, and 0.25 mM NADH. In the coupled pyridine nucleotide would not be subject to uncontrolled dis- assay, both enzymes turn over. Luminescence was measured sipation via a shunt leading to freely autoxidizable flavin and at 23°C with a calibrated photomultiplier photometer (11). H202 production (5). One report suggested the existence of a Luciferase was immobilized on Sepharose 6B by using the luciferase-flavin reductase complex, but there is still little ev- method of cyanogen bromide activation (12). Sepharose 6B gel, idence for specific interaction between these two enzymes about 5 ml of bed volume, was washed by filtration with 200 (5-7). ml of water and suspended with constant stirring in an equal Although both luciferase and reductase in cell lysates are volume of water; 1.25 g of cyanogen bromide was added over soluble and not associated with a particulate fraction, there is a period of 20-30 min. During this step, the temperature was evidence suggesting that luciferase functions in vivo in associ- kept between 18 and 22°C by the addition of ice and the pH ation with membrane proteins (8). As a model for the mem- was kept at 10.5-11.5 by the addition of 2.5 M NaOH. The brane-bound system, we covalently attached luciferase activation of gel is about complete when no more base is needed to Sepharose 6B and examined the interaction of the immo- to maintain the desired pH. The suspension was immediately bilized luciferase with soluble flavin reductase. The results in- cooled to 4°C, filtered, washed with 200 ml of precooled 0.1 dicate that both binding and electron transfer between the two M phosphate (pH 7.5), and suspended in 5 ml of the same proteins are enhanced with the immobilized luciferase. precooled buffer containing 10 mg of luciferase. After gentle MATERIALS AND METHODS shaking at 4°C for about 18 hr, the suspension was washed by centrifugation at 0°C five times in 10 ml of 0.1 M phosphate, Luciferase was purified from Beneckea harveyi strain 392 (9) pH 7. After each centrifugation, the supernatant was collected by the method of Gunsalus-Miguel et al. (10) to a specific ac- for determination of luciferase activity, and protein content was tivity of 1.7 X 1014 quanta/sec per mg of protein determined measured by the method of Lowry et al. (13). This allowed us at 230C by the standard FMNH2-initiated assay (see below). to the amount of bound A NADH-dependent flavin reductase activity was resolved calculate by difference luciferase; this from luciferase activity at the stage of DEAE-Sephadex column value is used in estimating the specific activity of the immo- chromatography. The peak activity fractions were combined bilized luciferase. For experimental studies, the thoroughly washed gel was suspended in an equal volume of 0.1 M phos- The publication costs of this article were defrayed in part by page phate, pH 7/0.1 mM dithiothreitol and stored at 0°C. charge payment. This article must therefore be hereby marked "ad- vertisement" in accordance with 18 U. S. C. §1734 solely to indicate * Present address: Department of Biophysical Sciences, University of this fact. Houston, Houston, TX 77004. 249 Downloaded by guest on September 26, 2021 250 Biochemistry: Tu and Hastings Proc. Natl. Acad. Sci. USA 77 (1980) Table 1. Properties of soluble and immobilized its catalytic rate with the 10-carbon aldehyde (decanal) was somewhat (20%) slower. Parameter* Soluble Immobilized Thus, the decreased activity cannot be attributed to de- Relative specific activity, % 100 4 creased catalytic efficiency. The lower activity might be due Catalytic ratet min- to quenching of excited states produced in the reaction. Alter- Octanal 2.4 2.3 natively, 95% of the attached molecules might be catalytically Decanal 17.8 14.1 inactive. Luciferase is a heterodimer designated af3 (15, 16); Dodecanal 2.4 2.2 a is involved in , whereas the specific function of 3 is Denaturation temperatures 0C not known (4). Because luciferase activity is much more sensi- 0.05 M phosphate 44 48 tive to modifications in a than to those in : (17-20), inactive 0.75 M phosphate 55 57 attached luciferase might be attributable to the site on the Energy of activation for thermal protein at which the covalent link is made. inactivation, kcal-mol'1 In addition to its decreased activity, the immobilized lucif- 0.05 M phosphate 64 28 erase is also different from the soluble enzyme with regard to 0.75 M phosphate 89 24 its thermal stability. Heat denaturation of immobilized lucif- * All activity measurements were made by using the FMNH2-initiated erase in phosphate buffer required a higher temperature and standard assay at pH 7 and 230C. t Luminescence decay rate at 230C in the nonturnover standard exhibited a much lower energy of activation (Table 1). As assay. shown, both of these parameters were sensitive to ionic strength Temperature at which half the activity is lost in 5 min. or phosphate concentration, but the immobilized luciferase was distinctly different from the soluble form in its stability prop- erties under all conditions examined. RESULTS AND DISCUSSION When partially purified flavin reductase (NaDodSO4 gel, The properties of immobilized luciferase are shown in Table see Fig. 1A) was mixed for 5 min at 00C in 0.05 M phosphate 1. Its apparent specific activity was lower than that of the sol- (pH 7) with the luciferase immobilized on Sepharose 6B, about uble enzyme by a factor of about 20, but its kinetic properties, 75% of the reductase activity was bound, as determined by the as judged by the decay rate in the nonturnover assay, were not reductase activity remaining in the five washes. Comparison greatly altered. Depending upon the chain length of the alde- of the patterns of NaDodSO4 gel electrophoresis of the material hyde used in the reaction, there are considerable differences solubilized from the Sepharose 6B-luciferase before (Fig. 1B) in turnover time with soluble luciferase (14, 15). Similar dif- and after (Fig. 1C) complexing with flavin reductase shows that ferences were exhibited by the immobilized luciferase, although relatively few of the proteins from the array of those present

A

.1

0.2 -

"To

0.1 -

1

I ] 0.5 RF FIG. 1. NaDodSO4 gel electrophoresis ofthe partially purified soluble flavin reductase (A), proteins solubilized from Sepharose 6B-immobilized luciferase (B), and the immobilized luciferase complexed with flavin reductase (C). (A) Flavin reductase (50 ,g) was incubated in 0.1 ml of 0.01 M sodium phosphate, pH 7/1% NaDodSO4/1% 2-mercaptoethanol at 371C for 2 hr and then applied to a gel for electrophoresis. (B) Sepharose 6B-immobilized luciferase (0.4 ml gel volume) was similarly incubated in 0.4 ml ofthe buffer described above. (C) Immobilized luciferase complexed with flavin reductase and incubated as in B. After centrifugation, 0.1 ml of each supernatant was used for electrophoresis. Bands I, and 12 (B and C) were impurities in the luciferase sample used for the immobilization. Because luciferase is a dimeric (af) protein, the a and the : bands observed in B and C must be derived from those enzyme molecules in which only one ofthe two subunits was covalently attached. The amounts of the NaDodSO4-dissociable a and f3 range from 3 to 5% of the total attached protein. Downloaded by guest on September 26, 2021 Biochemistry: Tu and Hastings Proc. Natl. Acad. Sci. USA 77 (1980) 251

0U 10-2 a)~~~~~~~~

1 l3 Time, min 10-3~~~~~ FIG. 3. Effect of dilution on the kinetics of the coupled biolum- inescence with the soluble luciferase plus flavin reductase (Upper) or the Sepharose 6B-luciferase-flavin reductase complex (Lower). (Upper) Bioluminescence kinetics of a sample containing 50 ,g of luciferase and 100 jtg of flavin reductase (-) are compared with that 10-4~~~~~ observed with samples diluted 1:5 (---) and 1:50 (.....). (Lower) Similar kinetic comparisons are shown for the complex containing 1 1lo- 10-2 103 50 ,g of immobilized luciferase and the adsorbed flavin reductase Relative concentration (-), and the 1:5 (---) and 1:50 (----) diluted samples. The coupled FIG. 2. Effect of dilution on the initial velocity of the coupled bioluminescence was initiated with 25MgM NADH, as described in Fig. reductase-luciferase reaction, comparing bound and soluble forms 2. The light intensities ofdifferent samples are normalized to the same of the enzymes. The initial maximal intensity (IO) of the coupled value at the peaks. bioluminescence was determined at 231C in 1 ml of 0.02 M phosphate (pH 7) containing 0.2% bovine serum albumin, 0.001% decanal, 50,4M bo'bnd complex this is not so (Fig. 3). This is especially signifi- FMN, 0.25 mM (0 and !) or 25MM (0 and 0) NADH, and soluble cant with regard to the kinetics for attaining the steady state, luciferase plus flavin reductase (0 and 0) or the reductase-luciferase which, unlike the soluble system, are about the same at all complex (0 and *). The total amount of luciferase plus reductase was varied by dilution, so that the relative amounts of the two enzymes concentrations for the immobilized enzymes. With regard to were constant. For samples of soluble luciferase plus reductase, an decay kinetics, the reaction rate falls off rapidly over the first arbitrary unit of 1 is equivalent to 0.12 mg of luciferase plus 0.24 mg 2 min at higher concentrations in the soluble system. This de- of flavin reductase per ml in total enzyme concentration and 4.5 X 1012 crease may be attributed to the fact that the free reductase (-) or 3.4 X 1012 (0) quanta/sec in bioluminescence intensity. For rapidly shunts electrons from reduced pyridine nucleotide to samples of the reductase-luciferase complex an arbitrary unit of 1 is so that within 2 min the concentration of the equivalent to 3.1 X 1011 (-) or 1.2 X 1011 (o) quanta/sec in biolumi- reduced nescence and 50 Mig ofimmobilized luciferaseplus the adsorbed flavin pyridine nucleotide substrate has decreased markedly. In the reductase per ml of assay solution in total enzyme concentration. coupled system, however, the turnover is necessarily very slow because the luciferase molecule itself turns over about only once in the partially purified reductase preparation (Fig. IA) were every 10 sec. picked up by the gel. The two new bands (1 and 2), lying be- With the soluble luciferase, the maximum initial rate with tween the : subunit of the luciferase and an impurity, are likely NADH in the coupled reaction (at saturating flavin reductase to be the reductase(s), which have been variously reported to concentration) is about 75% of that obtained by initiating the have molecular weights between 24,000 and 40,000 (5, 21, 22). reaction with FMNH2 and luciferase alone. With the immo- The two other noncovalently bound proteins (bands 3 and 4) bilized luciferase and saturating reductase, an initial intensity have a considerably lower apparent molecular weight. in the NADH-initiated reaction can be obtained that is 3 times The immobilized luciferase-flavin reductase complex is greater than that resulting from the reaction initiated with active in the NADH-initiated coupled bioluminescence reac- FMNH2 and immobilized luciferase alone. tion. A key feature of the activity of this protein-protein Prior studies involving the immobilization of multienzyme complex is that FMNH2, the of the first enzyme, is systems have been carried out by immobilizing both (or all) of evidently passed to the luciferase with an enhanced efficiency. the different enzymes on the same particle, thereby creating As reported (7), the activity in the coupled assay with mixtures an "artificial" multistep enzyme system (23, 24). It is expected of the two enzymes in the soluble state decreases 1:100 when that in such systems there will be higher local concentrations the mixture is diluted 1:10 (Fig. 2). With the immobilized lu- of the intermediate substrates within the microenvironment ciferase-reductase complex a similar dilution results in only a of the enzymes, thereby establishing a more favorable operating 1:10 decrease in the activity (Fig. 2). This can be explained by condition for the second (next) enzyme in the sequence. Ex- assuming that the two enzymes are functionally coupled. The periments have confirmed this interpretation: the observed rate FMNH2 formed by the reductase is effectively transferred to of the overall reaction catalyzed by a set of enzymes immobil- the luciferase. ized on a single particle reached steady state more rapidly than The two systems differ kinetically as a function of dilution the soluble set, and enzymes immobilized to separate particles in a way that can be interpreted similarly. In the soluble system behaved like the soluble system (23). In our experiments we the kinetics differ at different concentrations, whereas in the presume that similar considerations apply, but even more so; Downloaded by guest on September 26, 2021 252 Biochemistry: Tu and Hastings Proc. Natl. Acad. Sci. USA 77 (1980)

the product of the first enzyme may be fed rather directly to 8. Ne'eman, Z., Ulitzur, S., Branton, D. & Hastings, J. W. (1977) J. the second. This is presumably because the second enzyme is Biol. Chem. 252,5150-5154. not simply brought into proximity of the first; it binds in a way 9. Reichelt, J. L. & Baumann, P. (1973) Arch. Mikrobiol. 94, that may be physically and functionally closer to that of the in 283-330. vivo state. 10. Gunsalus-Miguel, A., Meighen, E. A., Nicoli, M. Z., Nealson, K. Protein-protein interaction has been of wide interest and H. & Hastings, J. W. (1972) J. Biol. Chem. 247,398-404. studied considerably, but in this instance, protein-protein 11. Mitchell, G. & Hastings, J. W. (1971) Anal. Biochem. 39, 243- binding has been found to be enhanced by the immobilization 250. of one of the two members of a pair. In the system studied here 12. Porath, J. & Axen, R. (1976) Methods Enzymol. 44, 19-45. immobilized luciferase might be used for the purification of 13. Lowry, 0. H., Rosebrough, N. H., Farr, A. L. & Randall, R. J. flavin reductase by utilizing either batch or column techniques (1951) J. Biol. Chem. 193, 265-275. 14. Hastings, J. W., Spudich, J. A. & Malnic, G. (1963) J. Biol. Chem. if suitable methods for eluting the reductase and for regener- 238,3100-3105. ating the luciferase can be developed. Moreover, the principle 15. Hastings, J. W., Weber, K., Friedland, J., Eberhard, A., Mitchell, should be more widely applicable to other protein pairs for G. W. & Gunsalus, A. (1969) Biochemistry 8, 4681-4689. which immobilization of one enhances binding to the other. 16. Friedland, J. M. & Hastings, J. W. (1967) Proc. Natl. Acad. Sci. This research was supported in part by Grant PCM 77-19997 from USA 58,2336-2342. the National Science Foundation. S-C.T. was a National Institutes of 17. Meighen, E. A., Nicoli, M. Z. & Hastings, J. W. (1971) Bio- Health Postdoctoral Fellow (5F32-GM00348). chemistry 10, 4069-4073. 18. Nicoli, M. Z., Meighen, E. A. & Hastings, J. W. (1974) J. Biol. 1. Gibson, Q. H. & Hastings, J. W. (1962) Biochem. J. 83, 368- 377. Chem. 249, 2385-2392. 2. Hastings, J. W. & Gibson, Q. H. (1963) J. Biol. Chem. 238, 19. Cline, T. W. & Hastings, J. W. (1972) Biochemistry 11, 3359- 2537-2554. 3370. 3. Gibson, Q. H., Hastings, J. W., Weber, G., Duane, W. & Massa, 20. Baldwin, T. O., Hastings, J. W. & Riley, P. L. (1978) J. Biol. J. (1966) in Flavins and Flavoproteins, ed. Slater, E. C. (Elsevier, Chem. 253, 5551-5554. Amsterdam), pp. 341-359. 21. Michaliszyn, G. A., Wing, S. S. & Meighen, E. A. (1977) J. Biol. 4. Hastings, J. W. & Nealson, K. H. (1977) Annu. Rev. Microbiol. Chem. 252, 7495-7499. 31,549-595. 22. Jablonski, E. & DeLuca, M. (1977) Biochemistry 16, 2932- 5. Duane, W. & Hastings, J. W. (1975) Mol. Cell. Biochem. 6, 2936. 53-64. 23. Mosbach, K. & Mattiasson, B. (1976) Methods Enzymol. 44, 6. Jablonski, E. & DeLuca, M. (1978) Biochemistry 17, 672-678. 453-478. 7. Hastings, J. W., Riley, W. H. & Massa, J. (1965) J. Biol. Chem. 24. Jablonski, E. & DeLuca, M. (1976) Proc. Natl. Acad. Sci. USA 240, 1473-1481. 73,3843851. Downloaded by guest on September 26, 2021