macy, 428 Church St., University of Michigan, Ann Arbor, MI 48109-. 1065. ... LKB Biotechnology Inc. Electrophoresis reagents, protein dye re- agent, and ...
THE JOURNAL OF BIOLOGICAL CHEMISTRY
Communication
(0 1991 by
Vol. 266, No. 34, Issue of December 5, pp. 22789-22791, 1991 The American Society for Biochemistry and Molecular Biology, Inc. Printed in (1. S.A .
Purification of the Inducible Murine MacrophageNitric Oxide Synthase
kines such as interferon-y (Stuehr and Marletta,1985; Iyengar et al., 1987). ‘ N O is reactiveinoxygenatedaqueous solution and decomposes to NO, and NO;. The expression of ‘NO synthase activity in macrophages is coincident with their cytotoxic/cytostatic properties, and the ‘NO generated appears to play a role in augmenting the cell-mediated imIDENTIFICATIONAS A FLAVOPROTEIN* mune response by causing the degradation of critical iron(Received for publication, July 24, 1991) sulfur centersin the targetcells (Hibbs et al., 1988; Lancaster Joan M. Hevel$, Kimberly A. White$, and and Hibbs, 1990). Endothelial cell synthesis of ‘NO causes Michael A. MarlettaSBT smooth muscle relaxation, and most of the accumulated eviFrom the $Interdepartmental Program in Medicinal dence supports the proposal that endothelium-derivedrelaxChemistry, College of Pharmacy and the §Department of ing factor is in fact ’NO (Palmer et al., 1987; Palmer et al., Biological Chemistry, School of Medicine, University of 1988). The mechanism of smooth muscle relaxation is cGMPMichigan, A n n Arbor, Michigan48109-1065 dependent and involves the stimulation of the enzyme guanylate cyclase in the smooth muscle. Another signaling role The synthesis of nitric oxide (‘NO) from L-arginine for ‘NO has been foundin neurons where the stimulation of has been demonstrated in a number of cell types and functions either as a cell signaling agent or as a key NMDA receptorsby the neurotransmitter glutamate activates component of the cell-mediated immune response. Both the synthesis of ’NO. Excitatory amino acid-mediated transconstitutive and inducible activitieshave been de- mission is accomplished when ’NO signals the adjacent neuscribed. Herein we report the purification of inducible rons through a cGMP-dependent mechanism (Garthwaite et ‘NO synthase (EC 1.14.23) from activated murine al., 1988; Bredt and Snyder, 1989). fall into two The ’NO synthasescharacterizedtodate macrophages usinga two-column procedure. Crude 100,000 X g supernatant was passed through a 2‘-5‘- general categories. The ‘NO synthaseinvolved in signaling is ADP-Sepharose 4B affinity column followed by a constitutively expressed and has beenpurified from rat cereDEAE-Bio-Gel A anion exchange column. The ‘NO bellum (Bredt and Snyder, 1990; Schmidt et al., 1991) and synthase ran as a band of M , = 130,000 on sodium porcine brain (Mayeret al., 1990). This ’NO synthase requires dodecyl sulfate-polyacrylamide gel electrophoresis. both Ca2+ and calmodulin. The ’NO synthase from murine Gel filtration experiments using a Superose 6 HR 101 macrophages is inducible and shows no requirement for Ca’+ 30 column estimated the native molecular weight to be and calmodulin (Marletta et al., 1988). Bothforms of the 260 f 30 kDa, indicating that the native enzyme exists .NO synthaseutilize the co-substrates L-arginine and molecas a dimer. Activity was dependent upon L-arginine ular oxygen and require NADPH(Marletta et al., 1988). ( K , = 16 f 1 PM at 37 “C and pH 7.5) and NADPH. Tetrahydrobiopterin has beenshown to be a required cofactor Both (6R)-tetrahydro-~-biopterin and FAD enhanced activity, whereas Mg2+ and FMN hadno effect on for ‘NO synthase activity derived from macrophages (Tayeh activity. Fluorescence studies demonstrated the pres- and Marletta, 1989; Kwon et al., 1989), and the same appears ence of one bound FAD and one bound FMN per sub- to be true for theneuronal enzyme (Mayer et al., 1990). Additional reports of ‘ N O synthase purifications include a unit. calmodulin-independent form of the enzyme from rat neutrophils (Yui et al., 1991a) and very recently an inducible form of the synthase from rat macrophages (Yui et al., 1991b). We report here the purification of the inducible form of the Intense interest hasfocused on nitricoxide ( ‘ NO)’ because ’NO synthasefrom murine macrophages. This ‘NO synthase of the critical role it plays as a cell signaling agent and its function in cell-mediated immunity. The synthesis of ’NO by is distinct from all others characterized so far. Additionally ‘NO synthase (EC1.14.23) (see Scheme 1)occurs in a number we show here that the murinemacrophage ‘NO synthaseis a of mammalian cell types (Marletta et al., 1990). The pathway flavoprotein containing both FAD and FMN. was first identified in the macrophage as a reaction involving EXPERIMENTAL PROCEDURES the formation of citrulline, NO;, and NO; from L-arginine Materials-Hepes, glycerol, L-arginine, NADPH, NADP+, L-malic following stimulation with bacterial endotoxin and/or cyto*This research was supported by UnitedStatesPublicHealth Service Grant CA 50414. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ll To whom correspondence should be addressed: College of Pharmacy, 428 Church St., University of Michigan, Ann Arbor, MI 481091065. ’ The abbreviations used are: ’NO, nitric oxide; BH,, (6R)-tetrahydro-L-biopterin; cGMP,guanosine 3’,5’-cyclic monophosphate; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonicacid; PMN, polymorphonuclear neutrophils; NMDA, N-methyl-D-aspartate; NO,, nitrite; NO;, nitrate;SDS-PAGE, sodiumdodecyl sulfatepolyacrylamide gel electrophoresis.
acid, FAD, FMN, Naja naja snake venom,magnesium acetate, bovine serum albumin, dihydropteridine reductase, andhemoglobin (human, 2 X crystallized) were purchased from Sigma. Potassium phosphate (dibasic, enzyme grade) was purchased from Fisher Scientific (Pittsburgh, PA). Z’B’-ADP-Sepharose 4B was purchased from Pharmacia LKB Biotechnology Inc.Electrophoresisreagents,protein dye reagent, and DEAE-Bio-GelA were purchased from Bio-Rad. BH, was purchased from Dr. B. Schircks Laboratories (Jona, Switzerland). Enzyme Assay- ’ NO synthase activitywas measured as described previously (Olken et al., 1991). Briefly, ‘NO formation was measured at 37 “C by observing the time-dependent increase in absorbance a t 401 nm associated with the formation of methemoglobin from oxyhemogblobin. Final concentrations of the components were: L-arginine, 1 mM; NADPH, 100 p M ; Mg(OAc),, 1 mM; HbO,, 6 ,AM;BH,, 12 pM; 0.17 mM dithiothreitol. The K,, for arginine was determined
22789
W I N Y O
I
NNH D
I
+
.NO
i n the al)srnce o f SIg'* and \vas calculated using a nonlinear least squares program t o I'it the data. /'rf>tf*;rf /)c,l~'rn1ir7rrlion-'~heconrentration of proteininmacroI)haeesupernatantandpurifiedf'ractions was determined hy t h e l3radf'ord protein assay o r Hratlfordproteinmicroassay(Bin-Rad) using hovine serum a l h m i n a s a standard. Dilute protein samples \vere first concentrated in Centricon40 microconcentrators (Amicon, 1)anvers. 51.4) or in a Speed-Yac concentrator (Savant Instruments. 200 Inc.. Farmingdale. S Y )and then analyzed using the microassay. -. .__ 116.25 I'urificntion of ',\'(I .S~~nthnsc~-Cellculture procedures, induction 97.4 01' ' S O synthase activity. and the preparation o f macrophage 100.000 X ,c supernatant tvere carried out as described previously ('l'ayeh and -, 66.2 \larletta. l9S9). A l l purification steps werecarriedout at 4 "C. 1 0 0 . 0 0 0 X g supernatant (=25 mg of protein) was applied to a 2'-5'.AI)I'-Sepharose ,113 column (4.5 X 1 cm)equilibratedwith 10 mM K2Hf'0.: containing 10';. glycerol and 0.5 11131 I.-arginine a t pH 7.5 (l3ul'f'erA ) . The columnwas successively washed with 1 0 ml of Huffer !\ containing O . . i 11 SaCl. I O ml o f Huffer A. 15 mlof Hul'fer A containing :1 t n malate ~ and 0.5 NADP'. and 20 ml of' Huffer A. 1 2 3 'SO synthase activity was eluted with 11 ml 01 Huf'fer A containing FIG. 1. SDS-PAGE of macrophage 'NO synthase. Shown is 1 m11 S.4L)I'H and 2 pts1 RH.,. ' N O synthase-containingfractions a IOc; gel that was silver-stained. I,nrw 1, NADPH eluate from the \vere pooled and immediately applied to a DEAE-Hio-Gel A column ~'-:i'-AI)I'-Sepharose 413 column: /nnr 2. Hio-(;el A eluate containin:. 1.-) X 0.7 cm) equilil)rated w i t h Huffer A. The column was washed \vith :I m l of' Ruffer A followed hy 10 ml of Huf'f'er A containing 80 the purified enzyme: lonc ;I. molecular weight markers o f ovallwmin (66.200). phosphorylase h (97,.400). 11111 SaCI. 'SO synthase activity was eluted with Hul'fer A containing (45.000). hovine serum alhumin !$galactosidase ( I l6,250), and myosin ( 2 0 0 , 0 0 0 ) . 120 11111 S a C l and 2 p\t HH.,. \\'hen the K,,, for[.-arginine was determined. the I~ul'fersIhr the last t w o steps on the anion exchange column did not contain 1.-arginine. \\'hen the protein concentration curve from theNojn nnjn-treated standard curve.'I'he resulting curve l'or the anion exchange step was determined. glycerol was omitted indicated the amount o f FAD present in each instance. f'rom all hfl'ers due to its interference with the Hradford microassay. \\'hen the flavin conrentrations were determined. HH, was omitted RESULTS ANDDISCUSSION its interferencein f'rom theanionexchangeelutionhullerdueto fluorescence measurements. The purification of the inducible murine macrophage 'NO .SI)S-I'~~(;~:"Electrophoresis was performedusing a Rio-Rad synthase was accomplished ina two-step procedure involving \lini-I'HOTEr\S 11 dual slah cellaccording to the manufacturer's an affinity column and an anion exchange column. The results instructions. a discontinuous hult'er system (1,aemmli. 1970) and a of this purification are summarized in Table I. Others have 10"; separating gel. Gels were silver-stained using the sil\rer staining used similar chromatographic steps in the isolation of the kit from Hio-Rad. .Ynticc .tlo/ccu/or Wright I~ctcrrnir~ntion by ( ; r / Fi/lmlion-Analyt brain ' N O synthase (Hredt and Snyder, 1990; Schmidt ~t ai., ical gel filtration chromatography was carried nut on a Superose 6 1991) and a constitutive PMN-derived ' N O synthase (Yui et H I { 10/:10 column(Pharmacia I.KR) equi1il)ratedwithHuffer A ai., 1991a). We found that the use of a number of affinity containing 0.15 51 NaCl at a flow rate of O.:iml/min. Standards and a number of anion resins that recognize NADP+/NADPH and 100.000 X g supernatant or purified ' N O synthase were applied to exchange resins all resulted in the co-purification of another the column in 200-pl aliquots. Standards were monitored at 2880 nm. protein of M , = 64,000 on SDS-PAGE that demonstrated and ' S O synthase was detected enzymatically as described ahove. malic enzyme activity. Attempts to elute the activity with 3 The standards (Sigma)used were thyroglohulin. 669.000: apoferritin. 44:i.OOO: ,j-amylase, 200.000: hovine serum allxlmin, 66.000: and carmM r,-malate alone were not. successful; however, elution with h n i c anhydrase. 29.000, malate and 0.5 mM NADP+ provided a malic enzyme prepaI.'/n~,in Idcntificntionand Qunntitntion-The presence of flavin i n rationapproachinghomogeneity"withoutany loss of ' N O purified ' S O synthase was determined hy tluorescence performed on synthaseactivity.Increasedstability of t h e ' N O synthase a ratio spectrolluorolneter constructed hy Gordon Ford andDr. David IMlou (Department of Hiological Chemistry. IJniversity of Michigan activity during the purification was achieved with 10% glyc5~ledical School) previously described elsewhere (Moore cl d . , 1978). erol. While no apparent loss in activity was observed when ' N O synthase was stored overnight 50% in glycerol at -80 "C, The amnunt of flavin present was quantitated in the following manthe half-life of the enzyme purified in the absence of glycerol ner. Standard solutions containing equimolar concentrations o f FAD and F \ l S wereprepared. The excitation spectra were taken from wasapproximately 2 h a t 4 "C andcouldnothefrozen. :100-500 nm (A$.,,, = 520 n m ) of each concentration of the mixed flavin Therefore, t,he specific activity and yield of the purified enstandards and a denatured (hoiletl) ' N O synthase sample. Similarly. zyme in Tahle I is probably lower than that obtained with the emission spectra from -150-600 nm (Ac, = ,150 nm) were taken o f glycerol. the standards and the sample. Each standard (containing hoth FAD : t n d F5IS) and the denatured ' N O synthase sample were then treated Purified ' N O synthaseran as a dimerhaving a native 260 k 30 kDa as determined by gel lvith .Ynjn nnjn venom phosphodiesterase. tvhich cleaved the free FAD molecularweightof filtration. On SDS-PAGE ' N O synthase ran as a prominent t o F51S.and the fluorescence spectra recorded again. The amount of F51S was determinedusingtheuntreated(no Nnjnnnjn venom ;tdded) standard curve.anti the amount of FAD was calculated using a difference standard curve calculated hy subtracting the untreated
-
-
~
-
.NO Synthase
Purification of Macrophage Murine
22791
band of M , = 130,000 (Fig. 1) with an occasional one or two intermediate in the reaction (Stuehr et al., 1991)." The sequence homology to cytochrome P-450 reductasesuggests that faintbands withmolecularweightsslightly smallerthan 130,000 that appear to be degradationproducts. A recent reducing equivalents may be shuttled in to thehydroxylation report (Yui et al., 1991b) on the purification of macrophage site via the bound flavins. The flavin prosthetic group also 'NO synthase from rats also demonstrated the existence of a providesa reasonable moiety to catalyze theone-electron dimer under native conditions; however, the individual sub- chemistry that must atsome point be involved in the overall conversion. With the purification of this 'NO synthase and units had a M , = 150,000. The two reports of the rat cerebelthe characterizationof bound flavins, mechanistic studies can lum 'NOsynthasepurification differ with regard tothe be carried out t o probe the function of the various cofactors observed native molecular weight. Schmidt and co-workers and prostheticgroups. report a dimeric protein ( M , = 155 kDa/subunit) (Schmidtet al., 1991) that appears to be identical to the 'NO synthase Acknowledgements-We wish to thank Kristin M. Rusche, Robert first described by Bredt and Snyder(1990) which they report A. Pufahl, Jr., and Bruce A. Palfey for their help with some of the exists asa monomer. A 'NO synthasefrom rat PMN hasalso studies. We also wish to thank Drs. David Ballou and Vince Massey for their help with theflavin studies andDr. Thomas Michel (Harvard been purified and is reported to have a M , = 150,000 and exist Medical School) for his suggestion that the 64-kDa protein might be as a monomer under native conditions (Yui et al., 1991a). malic enzyme. 'NO synthase activity was dependent on L-arginine and REFERENCES NADPH and was enhanced up to 40% in the presence of 12 D. S., and Snyder, S. H. (1989) Proc. Natl. Acad. Sci. U. .'5 A . Bredt, p~ BH,. There was substantial activity in the absence of 86,9030-9033 added pterin suggesting that the pterinwas tightly bound. In Bredt, D. S., and Snyder, S.H. (1990) Proc. Natl. Acad. Sci. U. S. A . addition, at concentrationsof BH, less than 12 PM, dihydro87,682-685 pteridine reductase did not enhance activity (not shown) as Bredt, D. S., Hwang, P. M., Glatt, C. E., Lowenstein, C., Reed, R. R., and Snyder, S. H. (1991) Nature 351, 714-718 would be expected if the quinonoid tautomer of dihydrobioGarthwaite. J.. Charles. S. L.. and Chess-Williams. Nature , R.(1988) . pterin dissociatesfrom the enzyme. Further investigations 336,385-388 regarding the pterin and itsrole in the enzymatic mechanism Hibbs. J. B.. Jr.. Taintor. R. R.. Vavrin. Z.. and Rachlin.E. M. (1988) are under way. Although M$+ increases the activity in macBiochem. Biophys. Res: Commun. 157, k7-94 Iyengar, R., Stuehr, D. J., and Marletta, M. A. (1987) Proc. Natl. effect on the rophage 100,000 X g supernatant, it had no Acad. Sci. U. S. A. 84, 6368-6373 purified enzyme activity (not shown). The K, for arginine Kwon, N. S., Nathan, C. F., and Stuehr, D. S. (1989) J . Biol. Chem. and the maximum rate a t 37 "C and pH 7.5 were 16 f 1 p~ 264.20496-20501 and 26 k 2 nmol of 'NO h", respectively. Similar K,,, values Lancaster, J. R., Jr., and Hibbs, J. B., Jr. (1990) Proc. Natl. Acad. Sei. U. S. A. 87, 1223-1227 for the PMN-derived enzyme and the ratmacrophage enzyme Marletta, M. A,, Yoon, P. S., Iyengar, R., Leaf, C. D., and Wishnok, have been reported (Yui et al., 1991a, 1991b). J. S. (1988) Biochemistry 27,8706-8711 'NO synthase activity was increased in the presence of Marletta, M. A., Tayeh, M. A,, and Hevel, J. M. (1990) Biot'actors 2, 219-225 FAD but not FMN. However, the FAD enhancement was minimal and variable (10%). Fluorescence studies utilizing Mayer, B., John, M., and Bohme, E. (1990) FEBS Lett. 277, 215219 the purified enzyme demonstrated the presenceof both FAD Moore, E. G., Cardemil, E., Massey, V. (1978) J . Bid. Chem. 253, and FMN in a 1:l ratio (8 nM FADS nM FMN). By compar6413-6422 ison of the concentrationsof flavins to the protein concentra-Olken, N. M., Rusche, K. M., Richards, M. K., and Marletta, M. A. (1991) Biochem. Biophys. Res. Commun. 177, 828-833 tion of the purified 'NO synthase, it was determined that R. M. J., Ferrige, A. G., and Moncada,S. (1987) Nature 327, each subunit of ' N O synthase contained one FAD and one Palmer, 524-526 FMN. Although the protein concentration of the synthase Palmer, R. M. J., Ashton, D. S., and Moncada,S. (1988) Nature 333, 664-666 was at the lower limit of detection, this finding is consistent with the recent report by Snyder andco-workers (Bredt et al., Schmidt, H.H. W., Pollock, J. S., Nakane, M., Gorsky, L. D., Forstermann, U., andMurad,F. (1991) Proc. Natl.Acad. Sei. 1991) that the rat cerebellar form of 'NO synthase contains U. S. A. 88,365-369 both a FAD and a FMN binding site as deduced from the Stuehr, D. J., andMarletta, M. A. (1985) Proc. Natl. Acad. Sci. U. S. A . 82, 7738-7742 cDNA sequence and homology to cytochrome P-450 reducTayeh, M. A., and Marletta, M. A. (1989) J . Bid. Chem. 264, 19654tase. 19658 The overall conversion of L-arginine to 'NOinvolves a five- Yui, Y., Hattori, R., Kosuga, K., Eizawa, H., Hiki, K., Ohkawa, S., electron oxidation process. We had initially proposed that a Ohnishi, K., Terao, S., and Kawai, C. (1991a) J . Biol. Chem. 266, 3369-3371 likely first step in the pathway was an N-hydroxylation of arginine (Iyengar et al., 1987; Marletta et al., 1988) and the Yui, Y., Hattori, R., Kosuga, K., Eizawa, H., Hiki, K., and Kawai, C. (1991b) J . Biol. Chem. 266, 12544-12547 finding thatBH, was a required cofactor was consistent with this proposal (Tayeh and Marletta, 1989; Kwon et al., 1989). "R. A. Pufahl, Jr., P. G. Nanjappan, R. W. Woodard, and M. A. Indeed, N"-hydroxy-L-arginine hasnow been shown to be an Marletta, manuscript in preparation. ~~
~~~
