Inhibition of vacuolar H (+)-ATPase by disulfide bond formation ...

2 downloads 0 Views 3MB Size Report
We have previously demonstrated that the coated vesicle vacuolar H+-ATPase (V-ATPase) can be inacti- vated by formation of intramolecular disulfide bonds.
Vol. 269,No. 18,Issue of May 6 , pp. 13224-13230,

THE JOURNAL OF B I O ~ ~ ; I C CHEMISTRY AL 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

1994

Printed in U.S.A.

Inhibition of Vacuolar H+-ATPase by Disulfide Bond Formation between Cysteine 254 and Cysteine 532 in SubunitA* (Received for publication, December 13, 1993, and in revised form, March 3, 1994)

Yu Fen& and Michael Forgacs From the Department of Cellular andMolecular Physiology, nLfts University School of Medicine, Boston, Massachusetts 02111

73-kDa A subunit (2). This cysteine residue is conserved in all We havepreviouslydemonstratedthatthecoated vesiclevacuolarH+-ATPase(V-ATPase)canbeinactiV-ATPases for which sequence is available (3-7). The region vated by formation of intramolecular disulfide bonds that contains cysteine 254 has a glycine-rich sequence motif, (Feng,Y., and Forgac,M. (1992)J. BioZ. Chern. 267,19769- GtXI4GKT,which is thought, from the crystal structure of the 19772). The disulfide bond responsible for inactivation Ha-ras oncogene product p21and adenylate kinase, to bind the can be distinguished from other disulfide bonds that phosphates of the nucleotide(8,9).This region is conservedin form by the fact that formation of the inactivating di- many nucleotide binding proteins (10, 11). By sulfide bondis blocked by ATP or high ionic strength. Cysteine 254of subunit A is the most reactive cysteineresitaking advantage of these properties, we selectivelyla- due in t h e V-ATPase complex. When clathrin-coated vesicles beled the ATPase at the relevant cysteine residues with are treated with cystine, cysteine 254is the only cysteine resifluorescein maleimide. After analyzing the proteolytic due able to form a disulfide bond with the cysteine moiety of fragments that contain the labeled cysteine residues, we cystine througha thio-disulfide exchange reaction. The resultfound that cysteine254 and cysteine 532 in subunit A of the bovine V-ATPase are the residues thatform the di- ing V-ATPase becomes inactive upon disulfide bond formation (2). Inactivation of the V-ATPase can alsobe caused by formasulfide bond resulting in inactivation of the enzyme. Cysteine 254 and cysteine 532 correspond to 2 of the 3 tion of intramolecular disulfide bonds.We have shownthat the cysteine residues that are conserved in all available V- V-ATPase solubilized from clathrin-coated vesicles and recontends t o become inactive upon ATPase A subunit sequences. Cysteine254 is located in stituted into artificial liposomes the medium (12). removal of sulfhydryl-reducing reagents from the consensus motif, G(X),GKT, corresponding to resiThe resulting inactive ATPase is insensitive toNEM, presumdues 250-257, which is conserved in many nucleotide binding proteins. Cysteine 532 is located in a region notably due to formation of a disulfide bond between cysteine 254 previously shown to be in proximity to the nucleotide and another unidentified cysteine residue that protects cysbinding site. Modification of cysteine 254 by disulfide teine 254 from reacting with NEM. This disulfide bond also bond formation with cysteine 532 or thio-disulfide ex- appears to form under in vivo conditions. We have shownthat, change with cystine does not impair binding of 2-azidoin bovine brain clathrin-coated vesicles, there are reversibly [s8P]ATPto the A subunit. The inhibition is therefore inactivated and NEM-insensitive V-ATPases presentunder likely caused by disruption of the catalytic function of conditions where the oxidative state of cysteine 254 has not the ATPase on formation of the disulfide bond. Apossible been altered (12). role in regulating intracellular acidification by reversiIn this paper, we report identification of cysteine 254 and ble sulfhydryl oxidation and reduction is discussed. cysteine 532 of subunit A as the residues that form a disulfide bond resulting in inactivation of the V-ATPase. Inactivation is caused by disruption of the catalytic function of t h e V-ATPase The vacuolar H+-ATPase (V-ATPase),’ which acidifies intra- by the disulfide bond. It is possible that reversible oxidative cellular compartments,is composed of nine different subunits, inactivation of t h e V-ATPase may play an important role in which are divided into two distinct sectors. The peripheral V, regulating intracellular acidification. sector, which contains subunitsA (73 m a ) , B (58 m a ) , 4 0 , 3 4 , EXPERIMENTALPROCEDURES and 33 kDa, possesses the nucleotide binding sites of the VMaterials-Calf brains were obtained from a local slaughterhouse. ATPase. The integral V, sector, which consists of subunit 100, Phosphatidylcholineand phosphatidylserine were purchased as chloro38,19, and 17 kDa, contains the pathway for proton conduction form solution fromAvantipolar Lipids, Inc. and stored at -20 “C. across the membrane (1). V-ATPases are characterized by their 9-Amino-6-chloro-2-methylacridine and fluoresceinmaleimidewere sensitivity to sulfhydryl reagents, such as N-ethylmaleimide NEM inhibition of (NEM). The cysteine residue responsible for the bovine brain-coated vesicleV-ATPase is cysteine 254of the

from MolecularProbes, Inc. (Eugene, OR), and L3H1NEMwas obtained from DuPont NEN. V8protease was purchased from Boehringer Mannheim. Immobilon was from Millipore Corp. 2-A~ido-[p,y-~*PlATP was a giftfromDr. Richard Cross. Other reagents were purchased from Sigma, and cholesterol was recrystallized before use. * This work was supported in part by National Institute of Health Selective Labeling of the Cysteine Residuesof Subunit A with FluoGrant GM-34478. The costs of publication of this article were defrayed rescein Muleimide or PHJNEM-Vacuolar proton ATPase was purified in part by the payment of page charges. This article must therefore be from calf brain as previously described (13). The purified ATPase was hereby marked “advertisement”in accordance with 18 U.S.C. Section concentrated 3-fold using an Amicon ultrafiltration cell with a YM 30 1734 solelyto indicate this fact. $ Supported through a postdoctoral fellowship from the American filter. The concentrated ATPase was then reconstituted into liposomes containing phosphatidylcholine,phosphatidylserine, and cholesterol by Heart Association, Massachusetts Affiliate, Inc. 5 To whom correspondence should be addressed. Tel.: 617-956-6939; 3 days of continuous dialysis as previously described(13)with the minor modification that 2 m~ mercaptoethanol was present throughout the Fax:617-956-0445. The abbreviations used are: V-ATPase, vacuolar proton-translocat- entire reconstitution. The reconstituted vacuolar H+-ATPasewas then ing adenosine triphosphate; D m , dithiothreitol; NEM, N-ethyl- dialyzed against 10% glycerol,0.2 m~ EGTA, and 20 m~ HEPES pH 7.0 with at least four changes to remove mercaptoethanol. Dialysis promaleimide; PAGE, polyacrylamide gel electrophoresis.

13224

Inhibition of Vacuolar H+-ATPaseby Sulfhydryl Oxidation

13225

ceeded until the reconstituted ATPase completely lost ATP-dependent proton transport activity as a result of autoxidation. The inactive 100 ATPase was then dialyzed against a buffer containing 10% glycerol, 0.2 m EGTA, 30 m KCl, 50 r n NaCl, ~ and 20 m HEPES pH 7.0 for 24 h. The dialyzed ATPase was then incubated with 1 rn NEM for 1 h at room temperature to block reduced cysteine residues. Dithiothreitol (D’IT) (50 m) was then added to the ATPase solution to neutralize r 80 & excess NEM and reduce disulfide bondsbetween cysteine residues formed during oxidation. The reducedATPase was then dialyzed Iagainst 10% glycerol, 0.2m EGTA, 5 m ATP, and 20 m HEPES pH tam 7.0 for 2 days with four changes and then againsta thoroughly deoxyE e genated buffer containing 10% glycerol, 0.2 m EGTA, 100 m NaCl, & and 20 m HEPES pH 7.0 for 24 h to remove ATP. The dialyzed ATPase was subsequently incubated with 100 p~ fluorescein maleimide for 30 min or with 3 r n [3H]NEM ~ for 15 min at room temperature. The B labeling was stopped by adding 50 m dithiothreitol. Separation and Sequencing of Fluorescein Maleimide-labeled 20 Fragments-SDS (1%) was added to the fluorescein maleimide-labeled vacuolar H+-ATPase,which was prepared as describedabove. The ATPase was dialyzed against 10% glycerol, 0.1% SDS, and 50 rn TrisHCl, pH 6.8,for 12 h and then electrophoresed on a 10% acrylamide 0 SDS gel according to Fling and Gregerson (14).The fluorescein male0 1 2 3 4 5 imide-labeledsubunit A was cut out while illuminated by longW light. Days Subunit A was then electroeluted from the gel and digested byV8 protease as previously described (2) with minor modifications. Briefly, FIG.1.Oxidative inactivationof V-ATPase in the presence and the eluted subunit A was concentrated by lyophilization to about 200 absence of ATP and NaC1. The purified bovine V-ATPase was reconpg/ml, then dialyzed in 25,000 molecular weight cutoff dialysis tubing stituted into liposomes. Mercaptoethanol present during reconstitution against 150 m NaCl and 0.02% SDS for 3 days, and thenagainst 0.02% was then removed by dialyzing the reconstituted ATPase against 10% SDSfor 1 day at room temperature to reduceSDS in the solution glycerol, 0.2 m EGTA, 20 m HEPES pH 7.0 alone (0) or with the containing subunit A. EDTA (0.2 m),sodium phosphate (50 m, pH addition of either 5 m~ ATP (a)or 100 m NaCl(0). The proton trans7.8), KC1 (100 m),and V8 protease were then added to the dialyzed port activity of 2 pg of the reconstituted V-ATPase was assayed at the subunit A solution. KC1 precipitates residual SDS that may inhibit V8 indicated times as described under “Experimental Procedures.” protease a t a concentration over 0.2%. Digestion was continued for 1-3 days at room temperature with the digestion monitored by SDS-PAGE on a 20% acrylamide gel that was prepared according to Fling and bonds that lead to inactivation of the V-ATPase, it is necessary Gregerson (14)and subjected to preelectrophoresis in the presence of 10 to distinguish the relevant cysteine residues from the others. m reduced glutathione according to Moos et al. (15).The proteolytic We found that two distinct properties of formation of the difragments were then transferred from the gel to Immobilon and sub- sulfide bond that causes inactivation of the V-ATPase can be jected to amino-terminal sequencing as previously described (2). used to sort out the relevant cysteine residues. 2-Azido-f2PlATP Labeling-Labeling ofV-ATPaseby 2-azido-[p,yFormation of the inactivating disulfide bond can be inhibited 32PlATPwas carried out in a 200-9 reaction mixture containing 10 mg by ATP as well as high ionic strength. As shown in Fig. 1, the of reconstituted V-ATPase, 20 m 2-a~ido-[~~PRTP (specific activity, V-ATPase, which was reconstituted into liposomes, gradually 450-1, 200 cpdpmol), 10% glycerol, 50 m NaCl, 30 m KCl, 0.2 IIIM EGTA, 20 m HEPES pH 7.0. The mixture was incubated in the dark became inactivated upon removal of sulfhydryl-reducing refor 10 min and then placed a t 1 cm from a 254-nm wavelength W lamp. agents. The inactivation process is quite slow, taking 4-5 days After 15 min of illumination, 50 m DTT was added to the mixture to for complete inactivation. ATP at a concentration of 5 m~ or quench the reaction. SDS (1%) was added into the mixture followed by NaCl at a concentration of 100 m~ reduces the rateof oxidative dialysis of the mixture against 1% SDS, 10% glycerol, 2 rn mercaptoethanol, and 50 m Tris-HC1, pH 6.8, overnight to remove free 2-azido- inactivation. Taking advantage of these properties, we selectively labeled the relevant cysteine residues with L3H1NEMand ATP. fluorescein maleimideby the following procedure.The purified Purification and Reconstitution of Cystine-modified V-ATPase-VATPase in which cysteine 254 of subunit A was modified by cystine was V-ATPase was first reconstituted into liposomes and allowed to purified and reconstituted into liposomes as previously described (2). inactivate by removal of sulfhydryl-reducing reagents (Fig. 2). Other Analytic Methods-Proton transport activity was measured as The inactive enzyme was then dialyzed against a buffer conthe carbonyl cyanide m-chlorophenylhydrazone-sensitivefluorescent quenching using the pH-sensitive probe 9-amino-6-chloro-2-methylac- taining 50 mM NaCl and 30 m~ KC1 followed byincubation with ridine following the addition of MgATP to the assay mixture as previ- 1 mM NEM to block all accessible, reduced cysteine residues. ously described (2). ATPase activity was measured by a continuous The NEM-treated ATPase was then incubated with dithiospectrophotometric assay as previously described (2). Protein concen- threitol to reduce disulfide bonds. The resulting ATPase was tration was determined by the Lowry method (16). activated. As shown in Fig. 2, typically 50% of the original

i

$40

proton transport activity and 30%of the original ATPase activity can be restored at this step. Poorer restoration was observed We have shown that thepurified V-ATPase, which has been if the inactivated V-ATPase was kept in a low ionic strength reconstituted into an artificial membrane consisting of phos- buffer when treated with NEM. phatidylcholine, phosphatidylserine, and cholesterol, can be reThe activated ATPase was then dialyzed against a low ionic versibly inactivated by removing sulfhydryl-reducingreagents, strength buffer containing ATP and a normal level of oxygenfor such as dithiothreitol and mercaptoethanol, from the medium 48 h. Under these conditions, disulfide bondsin the V-ATPase (12).The inactive V-ATPase becomes resistant to millimolar will form again except the inactivation-causing disulfide bond NEM, unlike the active enzyme that is inhibited by 10-100 whose formation is inhibited by ATP. The resulting V-ATPase micromolar NEM (13, 17). This indicates that a disulfide bond had slightly decreased activity due to the small portion of the must be formed in the inactive enzyme that protects cysteine V-ATPase,which undergoes disulfide bond formation at the 254 from reacting with NEM. There are at least five subunits of inactivating site. The V-ATPase was then dialyzed against a bovine V-ATPasecontaining cysteine residues that could poten- buffer containing 100 m~ NaCl and a low level of oxygen for12 and 33 kDa) h to remove ATP, which can partially prevent cysteine 254 from tially form disulfide bonds (subunits A, B, 40, 38, (3,18-21).To determine which cysteine residues form disulfide reacting with NEM and would therefore decrease the labeling RESULTS

-

Inhibition of Vacuolar H-ATPase by Sulfhydryl Oxidation

1

2

3

P

40k38k33k-

0

".e-

FIG.3. Selective labeling of the cysteine residues that form R disulfide bond causing inactivation of the hovine V-ATh" with [."HINEM.Samples of r c a n s t i t u t r d V-ATl'asr w r r r p r e p a n d a s follows hrforr I HINEM labeling, Imnr I. V-ATPasr ( 2 pg of protrln) wns rrconstituted into liposomes and then dialyzrd against n droxygrnatd dialysis huffcr containing 107 glycerol. 100mw SaCl. 0.2 msr E(:TA. 'LO msr HEPES pH 7.0 a t 4 "C for 10 h with several changrs tn rrmovr mercaptorthanol. Imnr 2. V-ATPase containing srlrctivrly rrducrd cy+ k i n e r r s i d u r s t h a t wouldform a disulfidca hnnd and Inartivntr thr V-ATPasr was prepared as drscrihed In t h r IeKrnd to Fig. 2 from s t r p s a./.L a n e 3 . V-ATPase was prepared as drscrihrd in the Irgrnd to FIR.2 from steps u-c followed by dialysis against thra h v r dialysls huffrr for 10hwithseveralchangesto rrmovr rxcess NE". In thls c a s r t h r oxidative inactivation-rcsponsihlecysteines remainrd disulfide-hondrd. V-ATPase was lahrlrd with 1 'HINEV and run on a 12'; acrylamidr grl followed by autnradiopaphy as drscrihrd under 'Exprrimrntal Prowdures." U

h

c

d

'6

/

R

similar labeling pattern was also observed when fluorescein maleimide was used (Fig. 4). Thus the cysteine residues, which FIG.2. Proton transport and ATPase activities of V-ATPase at form a disulfide bond and inactivate the V-ATPase, are in subvarious steps during selective labeling of cysteine residues in- unit A. volved in oxidative inactivation. T h r purified hovine V-ATPase was Because there are 8 cysteine residues in subunit A ( 3 ) .we reconstituted Into liposomrs in the presence of mercaptoethanol ( 0 )and needed to determine which were involved in oxidativr inactithen oxidized hy the removal o f mercaptocthanol from the medium ( h ); the resulting oxidized ATPase was trrated with 1 mM NEM at room vation. For thispurpose,thefluoresceinmaleimide-labeled temperature for 1 h (c), and the NEM-treated ATPase was reducedby subunit A was elutedfollowing SDS-PAGE and digested hv V H incubation with 50 mM D7T at room temperature for 1 h ( d ) ;the reprotease, which cleaves a t the carbnxyl side of glutamate and duced ATPase was dialyzed against 5 msr ATP, 107 glycerol. 0.2 mM aspartate. When the proteolytic fragments were separated hy EGTA, 20 mM HEPES pH 7.0 for 2 days ( e ) and then further dialyzed against a deoxygrnated buffer containing 100 msr NaCI, 1 0 a glycerol, SDS-PAGE, two labeled fragments were observed. The larger 0.2 mM EGTA. 2 0 m u HEPES pH 7.0 for 24 h ( f ) ; the dialyzed ATPase fragment has an apparent molecular mass of 7.5 kDa. and the from step f was treated with 100 p NEM at room temperature for 30 small fragment has a molecular mass of 3.6 kDa (Fig. 4 ) . T h r min (g).The proton transport activity is shown in pnnrl A, and the two fragments (30 pmol) were sequenced after transfrr to ImATPase activity is shown in punel B . The activities of 2 mg o f the reconstituted V-ATPase were assayed as descrihed under "Experimen- mobilon. The10 amino acid residuesa t t h eNH, terminus of the tal Procedures." The initial ATPase activity is H mmol ATl'/min/mg 7.5-kDa fragment are VAKLIKDDFL. This fragment is thus protein. predictedtocorrespond to residues valine 511 to glutamate 572, which h a s a predicted molecular massof 7.2 kDa. Because efficiency in the next step (13). The resulting V-ATPase was fluorescein maleimide has a molecular weight of 427. the frag2). To identify the inac- ment containing the fluorescein maleimide-modified cysteine active and fully sensitive to NEM (Fig. tivation-relevant cysteine residues, this preparation was residue would have a predicted molecular mass of 7.6 kDa. treated with ['HINEM or fluorescein maleimide. The resulting which is quite close to the apparent molecular mass ohtained ["HINEM labeling patterns are shown in Fig. 3. When the V- fromSDS-PAGE.Because the region from valine 511 tn the ATPase was selectively labeled by ['HINEM at the inactivation- carboxyl terminus contains only 1 cysteine residue, cysteine relevant cysteine residues as described, the predominantly la- 532, this residue must participate in the formation of the dibeledprotein is subunit A. By contrast,thefullyreduced sulfide bond leading to inactivation of the V-ATPase. control vacuolar H+-ATPase was labeled with ['HHNEM in sevThe 10 amino acid residues a t t h e NHZterminus of the 3.6eral subunits including subunits A, R, 40, 38, and 33 kDa. A kDa fragment are KLPANHPLLT. Although we have not carI'rcparations of \'arunlar Proton ATPase

Inhibition of Vacuolar H-ATPase by Sulfhydryl Oxidation

A

B

1

2

I

.

.

.

: