Rubidium Occlusion within Tryptic Peptides of the H,K-ATPase*

T H EJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 268, No. of 11, Issue

April 15,pp. 8012-8018,1993 Printed in U.S.A.

Rubidium Occlusion within Tryptic Peptides of the H,K-ATPase* (Received for publication, September 21, 1992)

Edd C. RabonS, Kent Smillie, VenitaSeru, andRobin Rabon From the Center for Ulcer Research and Education, Veterans Administration Center, Wadsworth Division, Los Angeles, California 90073 and the DeDartment of Medicine. Universityof California at Los Angeles, School of Medicine, Los Angeles, California90024

"Rb+ binding to the H,K-ATPase was measured in cations are near or within the membrane and are inthe M8+-vanadate-inhibitedenzyme at 4 "C. The con- accessible totrypsinattack.The membrane-bound centration dependence of esRb+binding in detergent- fragments of the H,K-ATPase were enriched ina pepfree preparations exhibited twocomponents, one sat- tide at molecular mass = 20 kDa and a second band urable witha &&(Rb+) of 0.76 0.3 mM and a binding near the dye front below the molecular mass = 14.4 capacity of 2626 f 690 pmol of Rb+/mgof protein and kDa standard.Sequence analysis indicatedthat theNthe second nonsaturable, but linearly dependent,upon terminal residueof the molecular mass = 20 kDa pepthe "Rb+ concentration. The concentration dependthe third tide began at Asne47~eus64 withinC-terminal ence of "Rb+ binding was unaffectedbydigitonin of the molecule, treatment with a Ko.6(Rb+)of 0.63 f 0.09 mM and a binding capacity of 2824 f 152 pmol of Rb+/mg of protein, but the amplitudeof the nonsaturablecompoThe gastric H,K-ATPase is a member of the P type ion nent was eliminated. The level of "Rb+ binding was optimized by vanadate and decreasedby ADP and translocating ATPase family. It is responsible for the active ATP, suggestingthat cation binding is stabilized inthe countertransport of 2H+ for 2K+ (1-3) and the passive exEz-like conformation and antagonized in the El con- change of K+ for K+ (4-6) across the secretory membrane of formation. The Rb+-dependent stabilization of the Ez the parietal cell. Like other members of this ATPase family, enzyme conformationwas confirmed fromthe fluores- a variety of data indicate that ion translocation utilizes a cent quench responseof the fluorescein isothiocyanate cyclical series of conformational changes generally described (F1TC)-labeled enzyme, where ssRb+ bound tothe using an E1/E2 nomenclature (7-15). FITC-labeled enzymewith a KO, = 0.85 f 0.3 mM and Although reciprocal conformational changes within the a saturable binding capacity of 2121 pmol of 86Rb+/mg cytoplasmic nucleotide and extracytoplasmic cation binding a domains are predicted for all P type ion pumps, only recently of protein and quenched the FITC fluorescence with Ko.a(Rb+) of3.6 f 0.3 mM. The K+-competitive inhibi- has it been shown that ATP-dependent phosphorylation intor, 1-(2-methylphenyl)-4-methylamino-6-methyl-fluences the conformation of the cation binding domain of 2,3-dihydropyrrolo[3,2-c]quinoline (MDPQ), also the H,K-ATPase (16). The identification of this cation bindquenchedFITCfluorescence with a Ko.6(MDPQ) of ing site is a central problem in the elucidation of structure/ 24.5 f 0.6 p~ and competitively inhibited "Rb+ bind- function relationships of the H,K-ATPase and may provide = 35.8p~ (MDPQ). The MDPQ-induced evidence of the basis of cation selectivity within the P type ing with a quench of FITC fluorescence at Lys"" within the cytoplasmic M4/M5 nucleotide domainand displacement family of pump molecules. Recently, Munsen et al. (17, 18) have shown that Meof "Rb+ from a functionally defined extracytoplasmic DAZIP,' a photolyzable K+ competitive inhibitor of the H,Kbinding domain indicate that structural determinants of the Ez conformational state exist within both cyto- ATPase, labels extracytosolic peptide residues delimiting a plasmic andextracytoplasmicdomains of the H,K- loop bounded by membrane spanning domains Ml/M2. In an ATPase and thus provide evidence of concerted con- alternative approach to implicate residues involved in cation binding, Karlish (19,20) has obtainedprotease digested Na,Kformational changes between the nucleotide and cation binding domains within the FITC-labeled H,K-ATP- ATPase capable of Rb+ occlusion. The trypsin digested memase. Membrane-bound fragments of the H,K-ATPase brane preparationof the Na+ pump includes a molecular mass were prepared by tryptic hydrolysis in KC1 medium. = 19 kDa peptide beginning with Amw1 and smaller peptides Tryptic digestion rapidly inactivated the phosphoen- corresponding to eight putativemembrane spanning domains. zyme site with a time course where k = 0.25 2 0.04 These results, albeit in difference P type ATPases, suggest min" but both "Rb+ binding and MDPQ fluorescence that residues from distant primary sequence and membrane responses were retained. The concentration dependspanning domains may participate in cation binding within ence of "'Rb+ binding and Rb+-dependentMDPQ fluo- these pumps. rescence responses in the trypsin-digestedmembranes A significant uncertainty in utilizing this analogy between were described by a single classof binding sites where the K+-competitive inhibitor site of the H,K-ATPase and the = 1.2 f 0.3 and 0.73 f 0.09 mM, respectively. The cation binding site of the Na,K-ATPase is that SCH 28080 is stability of the Rb+ and MDPQ sites suggest their lo-

*

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Supported by National Institutes of Health GrantAM34286 and by the United States Veterans Administration.

The abbreviations used are: MeDAZIP, 2,3-dimethyl-8-[(4-azidophenyl)methoxy]imidazo[l,2a]pyridine; CDTA, 1,2-cyclohexanediaminetetraacetic acid; FITC, fluorescein isothiocyanate; MDPQ/ SK & F 96356, l-(2-methylphenyl)-4-methylamino-6-methyl-2,3dihydropyrrolo[3,2-c]quinoline; Pipes, 1,4-piperazinediethanesulfonic acid; PVDF, polyvinylidine difluoride membrane; SCH 28080,2methyl-8-(phenylmethoxy)imidazo[l,2a]pyridine-3-acetonitrile.

8012

Cation Occlusion in the H,K-ATPase not a K+-competitive inhibitor of the Na,K-ATPase nor has Rb+ occlusion been demonstrated in the H,K-ATPase. To address this issue, we have developed a method of measuring Rb+ occlusion within the H,K-ATPase and have evaluated its a fluorescent K+-competitiveinhibitor, interactionwith MDPQ. We have obtained evidence that Rb+ occlusion occurs within the H,K-ATPase and this site(s) is maintained in the trypsin digestedenzyme.Although tryptic digestion selectively eliminates the active site of the H,K-ATPase, it spares and in fact enhancesboth Rb+ occlusion and MDPQ fluoresH,K-ATPase. The results provide cence responses within the functional evidence that cation binding and K+-competitive inhibitor binding are closely associated and suggest that residues labeled by K+-competitive inhibitors and those associa t e d with =Rb+/MDPQ binding withinthe H,K-ATPase may contribute to the identity of the cation binding site of the H,K-ATPase. MATERIALS AND METHODS Preparation of Membrane-bound H,K-ATPase-Microsomal H,KATPase was prepared by a previously described methods where the purified microsomes were retrieved from a 250 mM sucrose, 7% Ficoll interface following 90-min centrifugation a t 59,000 rpm in aBeckman 2-60 zonal rotor (21). This gradient fraction was diluted &fold into 10 mM Pipes, Tris, pH 6.7, collected by centrifugation at lo5 X g for 60 min, and resuspended in the same buffer. The procedure was modified to resuspend the zonal purified fractions by two 10-s bursts ina polytron homogenizer setto medium speed ratherthan by homogenization ina Teflon/glass homogenizer. This protein was lyophilized, resuspended in 1 mM EDTA and 40 mM Pipes/Tris, pH 6.7 and stored in aliquots a t -80 'C. Measurement of@Rb+Occlusion-The H,K-ATPase was incubated in a reaction solution containing 100 mM choline chloride, 2 mM CDTA, 0.3-1 mM =RbCl ( ~ 1 . 5X 106cpm/60 pgof protein) and 40 mM Pipes/Tris, pH 6.7 for 5 min a t room temperature and 30 min a t 4 "C. Ligands and inhibitors were added after the 5-min room temperature incubation preceding the 30-min incubation a t 4 "C. Digitonin (detergent/protein = 2.0 (w/w)) was added by gentle vortex to the equilibrated preparation and incubated for 5 min at 4 "C. Digitonintreatment decreased =Rb+ binding approximately 40% but increased the fractional contributionof the nucleotide-sensitive component to suggest that change was due to theloss of a residual space within the lyophilized preparation. Protein exposure to digitonin was limited to 5 min to avoid inactivation of the enzyme occurring with a monoexponential time course with t 1 / 2 = 35 min at 4 "C. Free =RbCI wasseparated from that bound to theenzyme by rapid column filtration a t 4 'C using Dowex 50x8-200 cation exchange columns constructed as described by Shani et al. (22). Typically, a 60-pl aliquot containing 40-70 pg of protein was diluted 1:lO (v/v) into anice-cold stop solution containing 100 mM choline chloride, 2.0 m M MgC12,0.2mM vanadate, and 40 mM Pipes/Tris, pH 6.7, previously layered on the Dowex column. Vacuum pressure was applied to obtain a flow rate of 0.2 ml s-'. Each column was washed with 1.5 ml of ice-cold sucrose medium containing 300 mM sucrose and 40 mM Pipes/Tris, pH 6.7. Column blanks obtainedby Dowexcolumn filtration were 250 cpm/l.5 X lO'cpm of =Rb+ applied. Data Analysis-Data were analyzed by nonlinear regression analysis using P. Fit software purchased from Biosoft. Saturation of =Rb+ binding was obtained from the fit to equation,

Y = Vm,

X

X/Ko,5

+ X + Nsp X X

(Eq. 1)

where KOAis the Michaelis constant for =Rb+ binding, V,, is the maximum binding capacity, and X is the quantityof =Rb+ recovered from the Dowex columns per mg of protein. Nsp is a measure of the unsaturable component of =Rb' recovered at each Rb+ concentration in the absence of digitonin. By its detergent sensitivity, this component most likely representsa residual trapped space within the lyophilized preparation. Enzyme PhsphoryIation-100-pg aliquots of the H,K-ATPase were phosphorylated with 25 p~ [y3*P]MgATP (100 cpm/pmol) using previously described methods (23).Eachmeasurement was reproduced in 8-11 samples and adjusted for a nonspecific blank

8013

obtained by acid precipitation of the enzyme prior to addition of [y32P]ATP. MDPQ Fluorescence-All measurements were performed on an SLM-Aminco SPF 500" spectrofluorimeter, fitted with a magnetic device for stirring the cuvette's contents and water temperature control. Steady-state fluorescence was determined at room temperature in a 2.0-ml solution containing 50 mM Pipes/Tris, pH 7.4, 2 mM MgCI2, 30 pg of protein, and 0.3 p~ MDPQ at the wavelength pair e ~ c i t a t i o n ~ ~ ~ . , / e m i s s inm o nand ~ ~ ~slit widths of 2.5 nm (excitation) and 4.0 nm (emission). Dilutions of MDPQ were obtained daily from a 10 mM stock prepared in dimethyl sulfoxide. Fluorescence responses were corrected for photobleaching and dilution artifacts using synchronized responses of equivolume isotonic additions of choline chloride. FITC Labeling and Fluorescence-FITC was bound to the membrane-bound H,K-ATPase essentially as described by Jackson et al. (12) and frozen a t -80 "C until use. Steady-state fluorescence was determined at room temperature in a 2-ml solution containing 100 mM choline chloride, 50 mM Pipes/Tris, pH 7.4, 2 mM MgClz, and 40-100 pg of FITC-labeled H,K-ATPase. FITCfluorescence intensity was optimized by measurement the at wavelength pair excitationlg6.,/Emission~t~nm with slit widths of 7.5 nm (excitation) and 2.5 nm (emission). Tryptic Hydrolysis of the H,K-ATPase-The H,K-ATPase was suspended in medium at 37 "C composed of 30 mM KC1, 100 mM imidazole, pH 7.0, 3 mM EDTA, and 2 mg/ml protein. Trypsin was added at a protein/trypsin ratio of 20 (w/w). Aliquots were removed at timed intervals, stopped by addition of trypsin inhibitor protein (trypsin inhibitor protein/trypsin = 6 (w/w)) and placed on ice. For SDS-polyacrylamide gel electrophoresis, the membrane-bound protein was collected by centrifugation at 28 p.s.i. for 30 min in a Beckman Alrfuge, resuspended, and resolved on a Laemelli minigel (15%). For =Rb+ binding and Rb+ titrations of MDPQ fluorescence, the medium was replaced by centrifugation through Sephadex G-50 equilibrated in the KC1-free buffer composed of 50 mM Pipes/Tris, pH 7.4, and 2 mMMgC12. Sequence Analysis-Following trypsin digestion as outlined above, the membrane-bound protein was washed three times utilizing a wash cycle composed of a 10-min room temperature incubation, collection by 1-h centrifugation at lo5 x g, and resuspension by Teflon/glass homogenization. The molecular mass = 20 kDa peptide was resolved on a Laemelli gel (15%)pre-run for 5 min with 0.1 mM thioglycolate. The molecular mass = 20 kDa peptide was transferred overnight (150 mA constant current) to PVDF in transfer buffer composed of 149 mM glycine, 41 mM Tris, pH 8.4,20% MeOH, and 0.48 mM 1,4dithiothreitol. The Coomassie-stained band at molecular mass = 20 kDa was cut from the PVDF paper and analyzed at theUCLA Protein Microsequencing Facility using an Applied Biosystems model 475A sequencing system composed of a 470A Protein Sequencer, a 120A phenylthiohydantoin analyzer, and a 900A data module. The peptide sequence was identified by comparing the amino acid sequence elevated above background in each sequence cycle. The preparation contained amajor sequence as well as minor sequences. For the major sequence presented for the molecular mass = 20 kDa peptide fragment, 18 of 20 residues were correctly identified using the elevated sequence comparisons obtained from three separate samples. RESULTS

=Rb+ Binding to the Vanadate-inhibited H,K-ATPase-The concentration dependence of ffiRb+binding at 4 "C t o t h e vanadate-inhibited H,K-ATPase in the presenceand absence of digitonin is shown in Fig. 1.In the absenceof digitonin ( A ) data from two experiments were combined and the individual data points fitted to a model incorporating a saturable and a nonsaturablecomponent.Theuppertracedescribesthis model with Ko,(Rb+) = 0.76 f 0.3 mM and a maximal binding of protein capacity of 2626 f 690 pmolof =Rb+ bound per mg plus a concentration-dependentnonsaturablecomponent. The lower trace ofthe same panel presents the calculatedfit to the kinetic parameters limited to the derived saturable component. To evaluate whether a residual trapped space within the lyophilized H,K-ATPase preparation could account for the nonsaturable component,the measurement was repeated with

Cation Occlusion i~'2. the H,K-ATPase

8014

TABLEI Ligand dependence of =Rb+ binding t o the H,K-ATPase The H,K-ATPase was equilibrated for 5 min at room temperature in medium composed of 100 mM choline chloride, 40 mM Pipes/Tris, pH 6.8, 1 mM =RbCl (30 cpm pmol") and 1mg ml" protein. Various ligands were added to aliquots of this suspension and equilibrated for an additional 30 min at 4 "C. Digitonin was added at a digitonin/ protein ratio = 2 (w/w) and the suspension incubated 5 min a t 4 "C. Aliquots containing 70 pg of the protein were applied to 530 p1 of an ice-cold isotonic choline chloride stop solution layered directly on Dowex 50x8-200 columns. __ ~~

0

1

2

RbCl

3

4

0

1

2

3

4

(mM)

RbCl (mM) FIG. 1. "Rb+ binding in the H,K-ATPase. The H,K-ATPase was equilibrated with %Rb+during incubation periods of 5 min a t room temperature and 3 h at 4 "C in solutions composed of 100 mM vanadate, 40 mM Pipes/Tris, choline chloride, 2 mMMgC12,0.2mM pH 6.7, =RbCl ranging from 0.05 to 5.0 mM (specific activity, 600 to 6 cpm/pmol) and 1 mg ml" protein. Data points combined from two experiments were obtained from aliquots containing50 pg of protein. A , detergent-free enzyme (A); the upper trace was drawn tothe X X/(K0,5+ X ) Nsp X X,where = 0.76 equation; Y = V, mM, V,, = 2626 pmol of =Rb+/mg of protein, and Nsp = 708. The lower trace is a line drawn tothe calculated parameters of the digitonin saturable component only. B, digitonin-treated enzyme (0); was added (digitonin/protein = 2.0 (w/w)) to the protein 5 min prior to assay. The trace was drawn to the equation; Y V, X XI(K0.5 X), where K0.5 = 0.63 mM and V, = 2824 pmol of =Rb+/mg of protein.

+

+

digitonin present in the assay buffer. The concentration dependence of =Rb+ binding in the digitonin-treated enzyme is shown in B of Fig. 1.The best fit of the combined data from two experiments incorporated a single saturable component with a &,5(Rb+) of 0.63 k 0.09 mM and a maximal binding capacity of 2824 k 152 pmol of =Rb+/mg of protein. To determine the relationship between the number of =Rb+ binding sites and the number of active catalytic sites in the H,K-ATPase, both thenative and digitonin-treated preparations were phosphorylated with [y3'P]ATP. The stoichiometry of the phosphoprotein inthe control and digitonin-treated preparations ranged from 1109 f 59 ( n = 11) to 902 f 24 ( n = 8) pmol/mg, respectively, whereas the measured amount of =Rb+ bound to the enzyme remained constant or increased slightly. This range in phosphorylation capacity introduced a variation of the relative =Rb+/EP stoichiometry from 2.4 f 0.6 in the control to 3.1 +- 0.2 in the digitonin-treated preparation. Ligand Sensitivities of =Rb+ Associated with theH,K-ATPase"There is considerable evidence that ligands of the H,KATPase influence the conformational equilibrium and stabilize specific conformational states. Table I presents several measurements of the amount of =Rb+ bound inthe presence of other ligands predicted to either stabilize El, EP, or to initiate catalytic turnover.In general, the level of =Rb+ binding was greatest inthe presence of the inhibitor vanadate and least in the presence of nucleotides. Pump turnover was not a factor mediating the level of cation bindingwhere ATP and ADP in the presence or absence of Mg2+ displaced approximately 70% of the bound =Rb+. Intermediate levels of =Rb+ remained bound to the Mg' or M$+-free enzyme. Although thesemeasurements show that cation binding occurs in the vanadate-inhibited enzyme, it was not known if the bound cation remained accessible to themedium. A selectivity series was derived from the vanadate-inhibited enzyme using 20 mM cation to competitively displace 0.3 mM =Rb+. =Rb+ was displaced with a selectivity where K+ (1)> Rb' (0.8) > Cs+ (0.4) >> Na+ (0.04)> Li+ (0). The numbers in parentheses reflect the normalized amount of =Rb+ displaced following competition with the indicated cation for 1 min at

Ligand

0.2 mM vanadate + 2.0 mMM&12 5.0 mM CDTA 2.0 mM MgCL 5.0 mM ATP + 5.0 mM CDTA 5.0 mM ADP + 5.0 mM CDTA 5.0 mM ATP + 2.0 mM M&lz 5.0 mM ADP + 2.0 mMMgC1,

Control

100 f 2% ( n = 9) 65 f 2% (n = 6) 56 f 2% ( n = 6) 26 2 1% ( n= 9) 33 f 10% ( n = 3) 32 f 1% ( n= 6) 28 f 5% ( n = 3)

4 "C. K+ was most effective, displacing approximately 40% of the associated "Rb+. These experiments indicate that =Rb+ bound within the digitonin-treatedvanadate-inhibited enzyme is displaced with a selectivity sequence similar to that shown previously to stabilize the E2 conformation (14) and stimulate turnover of the H,K-ATPase(24). Conformational Changes Caused by =Rb+ and K+ Competitive Znhibitor Binding--It is thought thation translocation in the H,K-ATPaserequires a concerted conformational change where the cationbinds to an extracytoplasmic domain to produce a change within the cytoplasmic nucleotide domain. To obtain evidence of cation binding and concerted conformational change, both =Rb+ binding and FITC fluorescence responses to Rb+ were measured in theFITC-labeled enzyme. Preliminary measurements showed that =Rb+ saturated the vanadate-inhibited enzyme with a KOa(Rb+)= 1.22 k 0.2 mM and a capacity of 2916 f 134 pmol of =Rb+/mg of protein, whereas in the absence of vanadate the FITC fluorescence response was saturated with a Ko,5(Rb+) of 3.6 k 0.3 mM. The lower trucein Fig. 2A shows that MDPQ binding also quenches FITC fluorescence in a bindingprocess represented by a single relatively low affinity class of sites where Ko,5(MDPQ)= 24.5 f 0.6 PM. Thus the FITC-modified enzyme is capable of binding =Rb+ and MDPQ and in each case exhibits afluorescence response like that observed with K+ binding. Although it has been shown previously that MDPQ competitively inhibits the K+-stimulatedATPase activity by stabilization of the E2P-I intermediate (16), it is not known if the competition occurs at the level of cation binding or if MDPQ binds to the dephosphoenzyme to mimic K' sufficiently to stabilize the E2-Ienzyme conformation. To determine whether =Rb+ and MDPQ binding are competitive, =Rb+ binding was measured in the presence of MDPQ. In the experiment shown in Fig. 2 A , the preparation was initially equilibrated in 0.3 mM =Rh+ prior tothe addition of MDPQ. The inhibitor was added 30 min later anddigitonin 5 min prior to assay. The MDPQ-dependent displacement of =Rb+ exhibited a hyperbolic dependence on MDPQ concentration with a Ko.5 of45.2 f 12 pM. Ifwe assume simple competitive behavior, a replot of this data can be used to obtain the inhibitory constant, 'Ko.b(MDPQ), for MDPQ binding (22). The linear transform shown in Fig. 2B presents the MDPQ-sensitive fraction ERb,/ERb, + MDPQ as a function of MDPQ concentration. From this plot a linear slope, + (Rb+)] x 1/'K0.5(MDPQ), is equal to [Ko.5(Rb+)/Ko.5(Rb+) expected for competitive behavior. Over a large range of MDPQ concentrations, the data conform to thisexpectation. 'K0,(MDPQ), the inhibitory constant for MDPQ in the ab-

H,K-ATPase Cation Occlusion the in

8015

necessary to maintain the respective binding domains, the ATP-dependent phosphoenzyme, %Rb+binding, and MDPQ fluorescence responses were measured during the timecourse of tryptic hydrolysis. Fig. 3shows theextent of peptide cleavage obtained during thehydrolysis of the KC1-stabilized enzyme. The catalytic monomer, observed as the major band at molecular mass = 94 kDa in lanes 2-9, is initially cleaved to transient peptide fragments at molecular mass 55 kDa (arrow 1 ) and 43 kDa (arrow 2 ) . These peptides accumulate during the intermediate stages of proteolysis but are subsequently cleaved and are almost absent in the latter stagesof O.* 0.0 the digestion. The final tryptic peptides obtained under these experimental conditions arevisualized as an intensely stained bandat molecular mass = 20 kDa(arrow 3 ) and aless intensely stained band migrating near the dye front below the molecular mass = 14.4 kDa standard (arrow 4 ) . The difference in the MDPQ fluorescent intensity of the phosphorylated and unphosphorylated enzymespecies was initially used toprovide a measure of each inhibitor environment within the trypsin-treated preparation. Fig. 4A provides a single measure of each response over a 30-min period. The w n-.-n ATP-dependent fluorescence response was rapidly lost in a 0 15 30 45 60 monoexponential time course, where k = 0.13 f 0.01 min". MDPQ (pM) In contrast, the K+-dependent fluorescence response actually increased during the time course of trypsin treatment. FIG. 2. Competition between MDPQ and "Rb+ in the H,KSeparate measurements of the level of the phosphoenzyme ATPase. The H,K-ATPase was equilibrated in solutions composed of 100 mM choline chloride, 2 mM MgCl,, 0.2 mM vanadate, 40 mM during the timecourse of trypsin digestion indicated that the Pipes/Tris, pH 6.7, 0.3 mM 8BRbCl (specific activity,110 cpm/pmol) phosphoenzyme rapidly decreased during tryptic digestion in and 1 mg ml" protein. MDPQ, rangingfrom a final concentrationof a monoexponential manner,where k = 0.25 f 0.04 min" (data 7-50 p ~ was , then added and incubated for 30 min at 4 "C. 5 min in time prior to assay, digitonin was added (digitoninlprotein = 2 (w/w)) to not shown). The rapid inactivation and similarity the H,K-ATPase at course for the ATP-dependent MDPQ fluorescence response the enzyme suspension.A , =Rb+ associated with each MDPQ concentration (0).The line was fitted to Y = 1 - [ V,. andthephosphoenzyme suggested that tryptic hydrolysis X X/(Ko.6+ X ) ] ,where = 45 & 12 PM MDPQ, V,, = 0.93 (A) selectively inactivated the phosphorylation site. Thisis comFITC fluorescence at each MDPQ concentration. The FITC-labeled patible with the placementof the MDPQ site within an area H,K-ATPase (0.03 mg/ml) was equilibrated at room temperature in of restricted trypsin access, possibly near the cation binding a solution composed of 100 mM choline chloride, 2 mM MgC12, and 40 mM Pipes/Tris, pH 7.4. The MDPQ-dependent quench of FITC site. T o test the prediction that the functional Rb+ binding site fluorescence was monitoredwithaddition of MDPQ and a trace drawn to the equation Y = 1 - [V,, X X/(K0.5+ X ) ] ,where = was also present in tryptic peptides of the H,K-ATPase, %Rb+ 24.5 & 0.6 mM MDPQ, V,, = 129 arbitrary fluorescence units. B, binding was measured in the trypsin-treated preparation.As linear transformof the data obtained in A (0)with slope = 0.0189. shown in Fig. 4B, the amount of %Rb+bound to theenzyme, like the MDPQfluorescence response, increased over the time sence of Rb+ was obtained from this plotby substituting the course of tryptic hydrolysis. Since a subsaturating concentraKo,5(Rb+)= 0.63 mM derived from the experiments in Fig. 2 tion of %Rb+ (0.3 mM) was used in this experiment, it was in the absence of MDPQ and the measured slope of 0.0189 possible that the increased level of %Rb+ bound to enzyme the obtained from the plot. From these values, 'K0.5(MDPQ) = might be derived from a decrease in the KO.,for =Rb+ rather 35.8 PM MDPQ. than an increase in the cation bindingcapacity. The saturaThe Cation DomainwithinTryptic Digests of the H,K- tion kinetics describing=Rb+ binding in the trypsindigested ATPase-Theprevious experiments show thatffiRb+and enzyme are shown in Fig. 5. The Ko.5(Rb+),1.2 f 0.3 mM, is MDPQ binding are competitive and stabilize a similar con- within the rangeof the control preparation measuredearlier. formation within theFITCnuCl,ide domain. These similarities A second measure of cation binding, the Rb+ quench of MDPQ show that thetwo binding sites are functionally closely asso- fluorescence, was also obtained in the trypsin-treated prepaciated and possibly reside near each other in the molecule. ration andis shown in the inset of Fig. 5. The kinetic constant On the assumption that residues comprising the MDPQ and describing the concentration dependenceof the Rb+-dependcation binding sites arelocated within membrane or inacces- ent fluorescence quench was K0.5 = 0.73 f 0.09 compared with sible extracytoplasmic domains, the H,K-ATPase was treated that in the undigested preparation (not shown) where Ko.5; with sufficient trypsin to extensively cleave the cytoplasmic 0.94 f 0.08. Both observations confirm the presenceof both domain. Although less severe conditions of tryptic digestion Rb' and MDPQ binding sitesin the trypsin-treated preparahave beenshown to rapidly inactivatetheATPase as a tion and argues that the affinity of the Rb+ siteis unchanged. consequence of peptide bond cleavage within cytoplasmically Sequence Analysis of the Molecular Mass = 20 kDa Pepexposed domains (10, 25), the inactivation could be due to tide-The molecular mass = 20 kDa peptide represents the loss of the active phosphorylation site, the cation binding largest peptide within the trypsin digested K+-stabilizedprepdomain, or an essential peptide linkage needed to transfer aration. It was therefore of interest to determine itsposition conformational change from one domain to another. To de- in the primary structure of the H,K-ATPase. The peptides termine if functional elementsof different structural domains produced by tryptic digestion were resolved by SDS-polyacrylare selectively maintained following proteolysis and if so t o amide gel electrophoresis utilizinga 15% Laemlli gel and provideevidence of the minimal size of peptide fragment transferred to PVDF paper. The molecular mass = 20 kDa

A

L

Cation Occlusion the in

8016

197" i

FIG.3. Tryptic peptides in the KC1-stabilizedH,K-ATPase. Trypsin was added to the H,K-ATPase equilibrated a t 37 "C in medium containing 30 mMKC1. 100 mM imidazole. DH 7.0. 3 mM EDTA, 2 mg/ml protein it an ATPase/trypsin ratio = 20 (w/w). At the indicated times, aliquots containing 23 pg of protein were stopped by addition of trypsin inhibitor protein a t trypsin in= 6 (w/w). The hibitor protein/trypsin samples were diluted by 150-pl addition of the imidazole buffer and collected by centrifugation. The samples were resuspended, solubilized,and applied to a 15% Laemlli gel. The gel lnnes 2-9 represent aliquota obtained a t digestion intervals of 0, 1, 3, 7, 13, 20, 28, and 35 min.

0

10

20

Time (min)

30

H.K-ATPase

'I

66.2-

-4

31-

21 .b 14.-

0

1020

304050 Time (mh)

FIG.4. =Rb+/MDPQ fluorescence response%in tryptic peptides. A, MDPQ fluorescence; The H,K-ATPase was treated with trypsin under conditions described in Fig. 3. At the indicated times, aliquots containing 40 pg of protein were removed, stopped by addition of trypsin inhibitor proteina t a trypsin inhibitor protein/trypsin ratio = 6 (w/w), and immediately added to 2 mi of the fluorescence assay buffer containing 50 mM Pipes/Tris, pH 7.4, 2 mM MgCIz, and 0.3 p~ MDPQ. Each point represents a single measurement of the magnitude of the fluorescence responses to 0.5 mM ATP (0)or 20 mM K+ (0)over the time course of the trypsin digestion. The line fitted to the loss of the ATP-dependent fluorescence response was drawn to a single rate constant; k = 0.132 f 0.01 min-', whereas the line describing the KCI-enhanced fluorescence was drawn by eye. B, =Rb+ binding (0);aliquots containing 45 pg of protein were treated as in A and immediately applied to a Sephadex G-50 column equilibrated in 1 0 0 mM choline chloride, 40 mM Pipes/Tris, pH 6.8, 2 mM MgCI2, and 0.2 mM vanadate. The protein was equilibrated with 0.3 mM @Rb+(107cpm pmol-l) and 5 min prior to assay treated with digitonin (digitonin/protein = 2 (w/w)). Each data point represents a single observation. The trace was drawn by eye.

"

0

1

2

3

1

5

6

Rb+ (mM) FIG.5. The concentration dependence of "Rb*/MDPQ bind""Rb' binding to tryptic peptides. Tryptic ing intryptic peptides. 0, peptides were produced by 30-min incubation at 37 'C under conditions described in the legend to Fig. 3. The preparation was immediately applied to a Sephadex (3-50 column equilibrated in100 rnM choline chloride, 40 mM Pipes/Tris. pH 6.8.2 mM MgCI,. and 0.2 mM vanadate and equilibrated with 0.1-5.0 mM "Rh' (300 to 6 cpm pmol"). 5 rnin prior to assay, the samples were treated with di&onin (digitonin/protein = 2 (w/w)). Each data point represents a single observation with the line was drawn to Y = V,. X X / K o a+ X,where = 1.16 0.3 mM BBRb+ and V,. = 306 32 pmol/aliquot. Inset, MDPQ fluorescence (0);following trypsin treatment the sample was applied to a Sephadex G-50column equilibrated in a fluorescence assay buffer composed of 50 mM Pipes/Tris. pH 6.8, and 2 mM M&Iz. An aliquot containing 60 pg of protein was then added to 2.0 ml of fluorescence assay buffer containing 0.3 P M MDPQ and 4 pg of nigericin. Fluorescence was measured following sequential additions of RbCl ranging from 0.1 to 20 mM. The points are single measurements of the difference in fluorescence between equivalent additions ,V X X / of RbCl and choline chloride with the line fit to Y = 1 - [ . + X ) ] .Koa= 0.73 k 0.09mM RbCl and V,. = 26 k 0.6 arbitrary fluorescence units.

*

*

band wasvisualized by Coomassie Blue and removed for sequence analysis. Three determinations of sequence indicated that the peptide band contained a major sequence initiated at Asn"' or LeusM.Both residues place the peptide fragment in the C-terminal third of the molecule distal to the measurements ofRb' occlusion in the H,K-ATPase was recently suggested to arise from the relative rates of interconputative M6 pump segment. version of the EI/E2 enzyme conformations which yield an equilibrium constant 80-fold lower in the H,K-ATPase than DISCUSSION This study provided several insights into the nature of the in the Na,K-ATPase (14). This earlier study predicted that cation binding site within the H,K-ATPase and the confor- the enhanced rate of the reverse conformational transition, mational interactions occurring between the nucleotide and E2K to EIK, increased the rate of cation release from the cation binding domains. First, the dataindicate that BBRb+ is H.K-ATPase sufficient to preclude the accumulation of a occluded within the H,K-ATPase. Although it is not surpris- stable cation bound intermediate. Rb+ occlusionwas demoning that thisoccurs and indeed appears to be a property of all strated in the present study in the vanadate-inhibited enzyme the P type ATPases, it has not, until now, been demonstrated at 4 "C. Under these conditions, those which stabilize the El binding was obtained in the gastric H,K-ATPase. The difficulty in obtaining direct enzyme conformation, saturable BBRb+

Cation Occlusion H,K-ATPase in the

8017

enzyme. This close association of MDPQ to the Rb+ binding site provides an experimental rationale for utilizing MDPQ as a probe of the cation binding domain. The stability of the active site and the cation binding site to tryptic digestion vary dramatically. The proteolysis of the ATPase to peptides of molecular mass = 20 kDa and below molecular mass = 14.4 kDa eliminates the active phosphorylation site but appears to enhance cation and MDPQ binding to the dephosphoenzyme. The loss of theATP-dependent MDPQ response was attributed to the loss of the active site, kl kz E1Rb E,(Rb) e EZRb since the two measurements exhibit similar time courses of k-1 k-2 trypsin inactivation. =Rb+ binding retained in the trypsinREACTION 1 digested dephosphoenzyme displays saturation kinetics with respect to =Rb+ binding and the Rb' dependence of the =Rb+ is bound within astable conformation, E,,(Rb), in MDPQ fluorescence response. Interestingly, the increase in equilibrium with the EIRb and EzRb conformations. This =Rb+ binding could not be accounted for by a decrease in the intermediate is not kinetically favored at room temperature for =Rb+ and appears to be an almost %fold increase in but is stabilized at 4 "Cwhere kl k-9 must be greater than the cation binding capacity. The nature of this increase is a k-,+ kz. Presumably, vanadate binding further reduces kTl to favor the accumulation of the E,,(Rb) intermediate. The matter of speculation. Perhaps multiple cation binding sites so that only nucleotides inhibit =Rb+ binding by stabilizing the &-like exist within the H,K-ATPase but are constrained a fraction can be saturated in the intact enzyme. conformation defined structurally by the presentation of trypSome information relevant to the structuralassignment of tic-sensitive cleavage sites (10) and functionally by the ADP the cation binding site has been gained from measurements stimulation of an ATP/ADPexchange (29) orthe ADP stimulation of phosphoenzyme turnover (29, 30). The vanadate- within the tryptic peptide preparation. Cleavage of peptide stabilized E,,(Rb) conformation of the H,K-ATPase is de- bonds within the M4/M5 cytoplasmic loop rapidly inactivates fined as the low fluorescence state of the FITC-modified the active phosphorylation site at Asp3%. Thispreventsa enzyme (12) andan enhanced fluorescence state of the concerted conformational change between the active site and the MDPQ site necessary to obtain the high fluorescence MDPQ-inhibited enzyme (16). The =Rb+ occluded within the H,K-ATPase is likely in- MDPQ state but does not inhibit binding to the Rb+/MDPQ volved in cation transport,since the loss of =Rb+ binding due domains within the dephosphoenzyme. Thus, the minimal to competition with competing cationsexhibits the same peptide size essential for formation of the active phosphorylselectivity sequence as that of cation transport. In addition, ation site is larger than that needed for cation/MDPQ bindsince =Rb+ was displaced from the vanadate-stabilized en- ing. The enrichment of the molecular = 20 kDa peptide in zyme conformation, it is likely that competition occurs at the the trypsin-treated preparation establishes this as thelargest extracytosolic cation binding site assuggested previously from peptide size essential for cation occlusion. It is possible that the fluorescence measurements of the MDPQ binding site the cation/MDPQ binding domains are containedexclusively (16). within the molecular mass = 20 kDa peptide although this The stoichiometry obtained in this study suggests that two and other evidence does not eliminate the possibility that the or three esRb+ions may be bound per phosphoenzyme mole- binding domains could be comprised of more diverse pump cule. The upper range of this stoichiometry is difficult to elements also present as membrane-associated peptides, In reconcile with the transport stoichiometry for H+ transport the simplest case, the essential elements of the cation binding which is closer to 2H+ per ATP hydrolyzed (1, 3, 31). Active site would be present within the molecular mass = 20 kDa C=Rb+ transport stoichiometry has not been investigated ade- terminal peptide beginning with LeuS54.The circumstantial quately, but the electroneutral nature of net H+ transport 1) MeDAZIP, a evidence for diversity, however,is2-fold. requires a unitary exchange stoichiometry (24, 32) and thus photolyzable K+ competitive inhibitor of the H,K-ATPase 2H+/2Rb+. Possibly, the phosphoenzyme levels are undereslabels an N-terminal extracytosolic domain between M1 and timated due to the greater lability of this domain in comparM2 (18).If the MeDAZIP residue assignment had also been ison with that of the =Rb+ binding site or to impurities in commercially available [32P]ATP preparations (22). Further in the C-terminal portion of the H,K-ATPase, it would be investigations of the phosphoenzyme stoichiometry obtained plausible that the molecular mass = 20 kDa peptide could from 32Piand ADP binding as well as adequate measurements contain the intact cation and MDPQ sites. Since this is not the case, one could assume that the close association of the of =Rb+ transport are needed to resolve this issue. The experiments show that the MDPQ fluorescence site is MDPQ domain and thecation binding domain are due to the closely associated with the =Rb+ binding site. Functional alignment of residues produced by folding of various domains evidence such as the strictly competitive nature of MDPQ located predominantly within the membrane-spanning reinhibition of the ATPase and the ionophore requirement for gions of the pump. 2) The trypsin digests of the K+-protected both cation-stimulated ATPase and the K+ quench of the enzyme appear analogous to those obtained in theNa,KATP-dependent MDPQ fluorescence response implicates the ATPase, where a C-terminalmolecular mass = 19 kDa peptide extracytoplasmic side of the ATPase molecule (16, 33). The beginning with 831Asnis obtained. Further analysis of this experiments here extend these arguments of "close associa- preparation of the Na,K-ATPase and of a more extensively tion" to show that thecompetitive interaction between MDPQ digested trypsin preparationof the H,K-ATPase indicate that and the H,K-ATPase is demonstrable at the level of =Rb+ trypsin-digested preparations also contain smaller peptide binding to the vanadate-inhibited dephosphoenzyme. Inter- sequences corresponding to at least eight transmembrane estingly, MDPQ binding arguably at anextracytoplasmic site segments (20, 34). Thus, these experiments may establish a also produced a concerted E2-likeconformational change maximal size for apump segment containing the cation bindwithin the cytoplasmic nucleotide domain of the FITC-labeled ing site but do not permit the conclusion that theC-terminal in both detergent-free and digitonin-treated preparations. The ligand sensitivities of =Rb+ binding appear analogous to those of the closely related Na,K-ATPase where the nucleotides ATP and ADP reduce and vanadate stabilizes esRb+ binding (26-28). These ligand sensitivities demonstrate that the cation binding site is sensitive to enzyme conformation stabilized by the binding of ligands within the nucleotide domain. In reference to a simplified reaction mechanism,

+

+

Cation Occlusion in the H,K-ATPase

8018

portion of the pump is sufficient to comprise the cation binding site. In summary, these data have provided some understanding of the conformational coupling of catalytic and transport functions within the H,K-ATPase. They provide the initial demonstration of 86Rb+ occlusion within the enzyme and establish the competitive interaction between =Rb+ binding and the K+-competitive fluorescent inhibitor MDPQ. The study provides evidence of interactive conformational changes occurring on opposite faces of the H,K-ATPase and establishes that theloss of enzyme function due to trypsin digestion can be related to theloss of a single structural domain rather than the systemic destruction of all domains. Finally, these studies show that cation and MDPQ binding occur in membrane preparations containing a C-terminal molecular mass sz 20 kDa peptide. Thus the largest peptide capable of Rb+ occlusion can be molecular mass = 20 kDa. If the site is more complex, the resistance to trypsin attack suggests that other putative structuralelements related to these sites reside near or within the membrane. Acknowledgments-MDPQ/SK & F 96356 was kindly provided by Dr. C. A. Leach of the G.I. Chemistry Section of the Research and Development Division of SmithKline Beecham Pharmaceuticals. REFERENCES 1. Rabon, E. C., McFall, T. L., and Sachs, G. (1982)J. Biol. Chem. 267, 6296-6299 2. Skrabanja, A. T.P:, van,der Hijden, H. T. W. M., and De Pont, J.J. H.H. M. (1987)Biochzm. Bzophys. Acta 903,434-440 3. Skrahanja, A. T.P., De Pont, J. J. H. H. M., and Bonting, S. L. (1984) Biochim. Biophys.Acta 774,91-95 4. Soumarmon, A,, Rangachari, P. K., and Lewin, M. J. (1984)J. Biol. Chem. 259,11861-11867 5. Gunther, R. D., Bassilian, S., and Rabon, E. C. (1987)J. Biol. Chem. 262, 13966-13972 6. Schackmann, R., Schwartz, A,, Saccomani, G., and Sachs, G. (1977)J. Membr. Biol. 32,361-381

7. Stewart, B., Wallmark, B., and Sachs, G. (1981)J. BWL Chem. 256,26822690 8. Wallmark, B., Stewart, H. B., Rahon, E., Saccomani, G., and Sachs, G. (1980)J. Bwl.Chem. 255,5313-5319 9. Wallmark, B., and Mardh, S. (1979)J. BioL Chem. 264,11899-11902 10. Helmich-de Jong, M. L., van Emst-de Vries, S.E., and De Pont, J. J. H. H. M. (1987)Biochim. Biophys. Acta 905,35%370 11. Helmich-de Jong, M. L., van Duynhoven, J. P. M., Schuurmans Stekhoven, F. M. A. H., and De Pont, J. J. H. H.M. (1986)Biochim. Biophys. Acta 868,254-262 12. Jackson, R. J., Mendlein, J., and Sachs, G. (1983)Biochim. Biophys. Acta 731,9-15 13. Faller, L. D.(1989)Biochemistry 28,6771-6778 14. Rabon, E. C., Bassilian, S., Sachs, G., and Karlish, S. J. D. (1990)J. Biol. Chem. 265,19594-19599 15. Rabon, E. C., and Reuben, M.A. (1990)Annu Rev. Physiol. 52,321-344 16. Rabon, E., Sachs, G.,Bassilian, S., Leach, C., and Keeling, D. (1991)J. BioL Chem. 266,12395-12401 17. Munson, K. B., and Sachs, G. (1988)Biochemistry 27,3932-3938 18. Munson, K. B., Gutierrez, C., Balaji, V. N., Ramnarayan, K., and Sachs, G. (1991)J. Bwl. Chem. 266,18976-18988 19. Karlish, S. J. D., Goldshleger, R., and Stein,W. D. (1990)Proc. Natl. Acad. Sci. U.S. A. 87,4566-4570 20. Ca asso, J. M., Hoving, S. H., Tal, D. M., Goldshleger, R., and Karlish, S. D. (1992)J.Biol. Chem. 267, 1150-1158 21. Chang, H., Saccomani, E., Rabon, E., Schackmann, R., and Sachs, G. (1977)Biochim. Biophys. Acta 464,313-327 22. Shani. M.. Goldschleeer. R.. and Karlish. S. J. D. (1987) . . Biochim. Bio~hvs. . . Acta 904, 13-21 23. Rahon, E. C., Gunther, R. D., Bassilian, S., and Kempner, E. S. (1988)J. Bi01.Chem. 263,16189-16194 24. Sacha, G., Chang, H. H., Rabon, E., Schackman, R.,Lewin,M., and Saccomani, G. (1976)J. Biol. Chem. 261,7690-7698 25. Besancon, M., Shin, J. M., Mercier, F., Munson, K., Miller, M., Hersey, S., and Sachs, G. (1993)Biochemistry, in press 26. Glynn, I. M., and Richards, D. E. (1982)J.Physiol. ( L a n d . ) 330,17-43 27. Beauge, L. A., and Glynn, I. M. (1979)Nature 280,510-512 28. Glynn, I. M., Richards, D. E., and Hara, Y. (1985)in The Sodium Pump (Glynn, I. M., and Ellosy, J. C., eds) pp. 589-598,Company of Biologists, Cambri e United Kin dom 29. Rabon, E.%ichs, G., Mar& S., and Wallmark, B. (1982)Biochim. Biophys. Acta 688,515-524 30. Helmich-de Jong, M.L., van Emst-de, V. S., de Pont, J. J., Schuurmans S&khoven, F. M., and Bonting, S. L. (1985)Biochim. Biophys.Acta 821,

f.

I

.

.

3'1'1-383

31. Skrabania. A. T.P.. Astv. P..Soumarmon. A., De Pont, J. J. H. H. M., and LewiG M. J. M. (1986)Bibchim. Biophys. Acta 860.131-136 32. Rabon, E., Chang, H., and Sachs, G. (1978)Biochemistry 17,3345-3353 33. Ganser, A.L., and Forte, J. G. (1973)Biochim. Biophys. Acta 307, 169180 34. Sachs, G., Besancon, M., Shin, J. M., Mercier, F., Munson, K., and Hersey, S.(1992)J. Bioenerg. Bmomernbr. 24,301-308