functional consequences of alterations to hydrophobic amino

THE JOURNAL OF BIOLOGICAL CHEMISTRY IC 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 28, Issue of October 5, pp. 18839-18845, 1991 Printed in U.S.A.

Functional Consequences of Alterations to Hydrophobic Amino Acids Located at the M4S4 Boundary of the Ca2+-ATPase of Sarcoplasmic Reticulum*

(Received for publication, May 20, 1991)

Bente Vilsen and Jens Peter Andersen From the Danish Biomembrane Research Centre, Institute of Physiology, University of Aarhus, DK-8000Aarhus C, Denmark

David H. MacLennan From the Banting and Best Deaartment of Medical Research, C. H. Best Institute, University of Toronto, Toronto, Ontario M5G lG, Canada

Site-specific mutagenesis was used to investigate the functional roles of amino acids in the relatively hydrophobic sequence Ile-Thr-Thr-Cys-Leu-Ala-320, located at the MsS4 boundary of the sarcoplasmic reticulum Ca”+-ATPase. Each of the residues was replaced with either a less hydrophobic, a polar, or a charged residue. Mutants Ile-315 + Arg and Leu-319 * Arg were devoid of any Ca2+ transport function or ATPase activity, while the mutant Thr-317 + Asp retained about 5 and 7% of the wild-type Ca2+ transport and ATPase activities, respectively. These three mutants were able to form the ADP-sensitive phosphoenzyme intermediate (EIP) by reaction with ATP, but this intermediate decayed very slowly to the ADP-insensitive phosphoenzyme intermediate (E2P). In the mutants Ile-315 + Arg and Leu-319 * Arg, the level of E2P formed in the backward reaction with inorganic phosphate was extremely low, but hydrolysis of E2P occurred at a normal rate. These mutants, in addition, displayed a higher apparent affinity for Ca2+ than the wild-type enzyme. In the mutants Ile-315 + Ser and Ile-315 + Asp, the Ca2+ transport and ATPase activities were moderately reduced to 30-40% of the wild- type activities, but normal affinities for Ca2+, Pi, and ATP were retained, as was the low affinity modulatory effect of ATP. Mutation of Thr-316 to Asp, Thr-317 to Ala, Cys-318 to Ala and Ala-320 to Arg had little or no effect on Ca2+ transport or ATPase activities. Intro- duction of two negative and one positive charge by triple mutation of the Ile-Thr-Thr-317 sequence cre- ated a mutant enzyme that, although completely inac- tive, was inserted into the membrane, consistent with a location of these residues on the cytoplasmic side of the M4S4 interface.

Our findings suggest that the amphipathic character of the S4 helix and/or the distribution of charges in S4 is important for the stability of the E2P intermediate.

The membrane-bound Ca”-ATPase of sarcoplasmic reticulum is responsible for the ATP-driven active uptake of

* This research was supported by grants (to B. V. and J. P. A,) from the Danish Biomembrane Research Centre, the Danish Medical Research Council, the NOVO Foundation and the Carlsberg Foun- dation (Denmark), and by a grant (to D. H. M.) from the National Institutes of Health. 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.

calcium ions from the myoplasm during muscle relaxation (Hasselbach and Makinose, 1961). In its purified form, the Ca2+-ATPase protein consists of a single type of polypeptide chain with Mr 110,000 (MacLennan, 1970). Each Ca2+-ATP- ase monomer binds two calcium ions with high affinity and accepts the y-phosphate of ATP, forming a covalent bond with Asp-351 (Inesi, 1985). A structural model for the Ca2+- ATPase was developed from analysis of the amino acid sequence deduced from the sequence of cloned cDNA (Mac- Lennan et al., 1985; Brand1 et al., 1986). Functional analysis of site mutants (Maruyama and MacLennan, 1988; Maruyama et al., 1989; Clarke et al., 1989a, 1989b, 1990a, 1990b, 199Oc; Andersen et al., 1989; Vilsen et al., 1989;) has led to models of structure-function relationships in the Ca2+-ATPase in which the binding sites for ATP and phosphate are located in two globular cytoplasmic domains, whereas the calcium li- gands are contributed by residues located in the transmembrane region. The cytoplasmic and the transmembrane domains are connected through a pentahelical stalk (segments S1-S,), identified in negatively stained specimens (Taylor et al., 1986). Ca2+ translocation seems to be accomplished through conformational changes in the phosphorylated enzyme, whereby the Ca2+-binding sites lose their high affinity and the calcium ions gain access to the luminal membrane surface (de Meis, 1981; Andersen, 1989; Inesi et al., 1990; MacLennan, 1990). The structural changes in the Ca2+ sites are closely associated with the transformation of the aspartylphosphate bond from a “high energy” state (E,P),’ reactive with ADP, to a “low energy” state (E2P), unreactive with ADP, but reactive with water. I t is unknown how this energy coupling is brought about. On the basis of studies of the influence of organic solvent on the intermediary reaction steps, de Meis (de Meis et al., 1980; de Meis, 1981, 1989) has proposed that the decrease in the free energy of hydrolysis of the aspartylphosphate bond results from a hydrophilic-hydrophobic transition of the catalytic site. We have shown that, in E, forms, the AI tryptic fragment (residues 199-505) is less exposed to hydrophobic labeling from the membrane phase than in Ez forms (Andersen et al., 1986).

By site-directed mutagenesis, residues essential to the E,P- E,P transition have been identified in the putative transmem-

The abbreviations used are: E,P, ADP-sensitive phosphoenzyme intermediate; E2P, ADP-insensitive phosphoenzyme intermediate: EGTA, [ethylenebis(oxyethylene-nitri1o)ltetraacetic acid; MES, 2- (N-morpho1ino)ethanesulfonic acid; TES, N-tris[hydroxymethyl] methyl-2-aminoethanesulfonic acid; MOPS, 4-morpholinepropane- sulfonic acid; SDS, sodium dodecyl sulfate.

18839

18840 Amino Acid Alterations at the M4S4 Boundary

brane sequence, Mq, as well as in the cytoplasmic sectors (Vilsen et al., 1989; Andersen et al., 1989; Clarke et al., 1990b, 1990~). These observations suggest that the M4 helix with adjoining segments may be engaged in the transmission of conformational changes from the phosphorylation site down to the Ca2+ sites in the membrane (Vilsen et al., 1989). In the present study we have used mutagenesis to analyze the func- tions of residues located at the boundary between M4 and the 4'h stalk segment, Sq, forming the direct physical link between M4 and the phosphorylated aspartyl residue. Some of the residues at the M4S4 boundary are highly conserved within the family of cation-transporting ATPases with a phosphorylated intermediate (Shull and Lingrel, 1985). A characteristic feature of the NH2-terminal end of the S, segment is its relatively high content of hydrophobic residues. If an a-helical structure is assumed, the hydrophobic side chains would be located predominantly a t one surface, making the helix amphipathic (Fig. 1). We have replaced 6 hydrophobic residues located at the M4S4 boundary with less hydrophobic, polar, and charged residues. In three cases in which Ile-315, Thr- 317, and Leu-319 were replaced with charged residues, we observed a strong inhibition of the activity, which could be ascribed to a reduced formation of E2P, whereas less pronounced or no inhibition of activity was observed with the other mutants studied.

EXPERIMENTAL PROCEDURES

The methods employed in this study have been described in detail elsewhere (Maruyama and MacLennan, 1988; Clarke et al., 198913; Andersen et al., 1989; Vilsen et al., 1989, 1991). A summary of the methods is as follows. Mutations were introduced into the rabbit fast twitch muscle Ca'+-ATPase cDNA using the site-specific mutagenesis method of Kunkel (1985). The presence of the correct mutation was confirmed by nucleotide sequencing according to Sanger et al. (1977). The entire Ca"-ATPase cDNA containing the desired mutation was cloned into the EcoRI site of vector p91023 (Kaufman et al., 1989) for expression in COS-1 cells (Gluzman, 1981). Microsomes were prepared from transfected cells and assayed for Ca'+ transport as described by Vilsen et al. (1989). Calcium-activated ATPase activity was measured as described by Mdler et al. (1980) and Vilsen et al. (1991). Studies of phosphoenzyme intermediates were carried out as described by Andersen et al. (1989) and Vilsen et al. (1991) with modifications as detailed in the legends to figures and tables. Quan- titation of radioactivity on the dried gels was obtained by autoradiography followed by densitometric analysis, using a LKB 2202 U1- troscan Laser Densitometer. An enzyme-linked immunosorbent assay was used to quantify the expressed Ca2+-ATPase in each microsomal preparation as described by Clarke et al. (1989b), using monoclonal antibody A52 (Zubrzycka-Gaarn et al., 1984) and a polyclonal goat anti-Ca"-ATPase antibody kindly provided by Dr. H. Brogren, The National Food Agency of Denmark.

RESULTS

Mutations in the M4S4 Segment-Fig. 1 shows an a-helical representation of the M4S4 segment with indication of the individual amino acid residues and the substitutions carried out in this study. Mutations were carried out in a relatively hydrophobic cluster of 6 juxtaposed residues Ile-315, Thr-316, Thr-317, Cys-318, Leu-319, and Ala-320, which may be cytoplasmic or membrane buried, depending on the exact location of the border between S4 and M4 (MacLennan et al., 1985; Shull et al., 1985; Clarke et at., 1989). In order to introduce shifts in hydropathy at various points, each of the residues indicated, was replaced with a residue which was either less hydrophobic, polar, or charged. In this way hydropathy changes ranging between -0.7 and -9.0 on the scale of Kyte and Doolittle (1982) were achieved (Table I). Thr-317 was mutated not only to Asp possessing a lower hydropathic index, but in addition to Ala, representing an increase in the hydropathic index of +2.5. Ile-315 was mutated to 3 different

Ala A S P , Ala I s4

M4

FIG. 1. Amino acid residues and mutations of the M4S4 segment. The amino acid residues of the wild-type enzyme (one-letter code) are drawn on an a-helical net, with indication of the substitutions carried out in the present study (three-letter code). M, and S, constitute the transmembrane and cytoplasmic parts of the helix, respectively. The indicated borderline between them is suggested on the basis of secondary structure prediction and homology studies of the family of cation-transporting ATPases.'

TABLE I Hydropathy change, relative expression, calcium transport rates, and

calcium affinities of mutants

Mutation Hydropathy Relative Specific rate Calcium change" expressionb Of

affinityd

Wild type Ala-320 -+ Arg Leu-319 + Arg Cys-318 -+ Ala Thr-317 -+ Ala

Thr-316 + Asp Ile-315 -+ Arg

-+ Asp

-+ Asp -+ Ser

Ile-Thr-Thr-317 -+ Arg-Asp- ASP

0 -6.3 -8.3 -0.7 +2.5 -2.8 -2.8 -9.0 -8.0 -5.3

% 100 95

107 82 75 71 92 93 86

101

30

% K ~ P M 100 0.25 70 0

0.18

78 0.25 84 0.28

100 0.17

37 0.25 34 0.16

5

0

0

Calculated according to the hydropathy scale suggested by Kyte and Doolittle (1982).

Calculated from the yields of mutant and wild-type Ca"-ATPase protein/mg of total microsomal protein in the preparation.

e Measured at 27 "C, in the presence of 5 mM MgATP, 100 mM K+, 5 mM oxalate, pH 6.8, and 10 PM Ca2' as previously described (Vilsen et al., 1989).

Determined by measurement of Ca'+ uptake rates a t various Ca2+ concentrations set with EGTA as previously described (Vilsen et al., 1989).

e Transport activity too low for determination of calcium affinity.

residues, Ser, Asp, and Arg, representing hydropathy changes of -5.3, -8.0, and -9.0, respectively.

Calcium Transport and Expression-The mutants Thr-316 -+ Asp, Thr-317 + Ala, Cys-318 + Ala, and Ala-320 + Arg displayed maximum specific Ca2+-uptake rates differing little or insignificantly from the wild-type maximum specific rate (Table I). The mutants Ile-315 + Asp and Ile-315 - Ser displayed moderate reductions of the Ca2+ transport activities to 37 and 34% of the wild-type activity, respectively. In the Thr-317 + Asp mutant, the specific Ca2+ transport rate was reduced to 5% of the wild-type rate, and in the Ile-315 + Arg and Leu-319 + Arg mutants, the Ca2+ transport rates were immeasurably low (Table I). The latter three mutants, as well

' N. M. Green, personal communication.

Amino Acid Alterations at the M4S4 Boundary 1884 1

as the other mutants with single amino acid replacements indicated in Fig. 1, were well expressed by the COS-1 cells, to levels between 70 and 110% of the wild type (Table I), indicating that proper insertion in the microsomal membrane occurred, despite the substitution of charged and polar residues for hydrophobic. In addition, we performed a triple mutation of Ile-Thr-Thr-317 to Arg-Asp-Asp. This latter mutant did not transport Ca'+ a t a measurable rate. Its expression level was lower than observed for any of the mutants with single amino acid replacements, but still significant, corresponding to 30% of the wild type (Table I).

For those mutants, which retained sufficient Ca2+ transport activity, we determined the apparent Caz+ affinities through measurement of the Ca" dependence of Ca2+ transport. All of the transport-active mutants showed an apparent affinity for Ca", which was indistinguishable from that of the wild type (Table I ) .

ATPase Activity-The Ca2+-activated specific ATPase activities were measured at ATP concentrations of 100 PM and 5 mM, in the presence of the ionophore A23187 to prevent back inhibition, due to accumulation of Ca2+ in the microsomes (Table 11). It can be seen that the relative Ca2+-ATPase activities a t 5 mM ATP agreed well with the relative Ca'+ uptake activities measured at the same ATP concentration (compare Tables I and 11), indicating that neither the mutants with reduced Ca'+ uptake activity, nor the mutants unable to transport Cap+, hydrolyzed ATP through an uncoupled path- way. We have previously demonstrated that the rate of ATP hydrolysis in the wild-type Ca2+-ATPase in COS-1 cell microsomes displays a secondary rise in the millimolar ATP-concentration range, identical to that observed for the Ca'+- ATPase in sarcoplasmic reticulum (Vilsen et al., 1991). The

TABLE I1 Specific Ca"-A TPaw activity and phosphorylation

from ATP and P, in wild t v m and mutants Specific Cn-ATl'ase

Mutation activit.y" Specific phosphorylationh

100 p~ ATP 5 mM ATP ATP (2 p ~ ) P, (500 BM)

%<

Wild-type 43 100 100 100 Ala-320 + Arg 27 68 >90 >90 I,eu-319 + Arg 0 0 >90 c 1 0 Cys-318 + Ala 38 95 >90 >90 Thr-317 -+ Ala 41 94 >90 >90 + Asp 0 7 >90 >90

Thr-316 + Asp 32 86 >90 >90 Ile-315 Arg 0 0 >90 4 0 + Asp 14 43 >90 >90 + Ser 11 33 >90 >90

Ile-Thr-Thr-317 0 0 0 0 + Arg-Asp- Asp

" Calcium-activated ATPase activity was measured on 0.5-2 pg of Ca"-ATl'ase a t 37 "C by the NADH-coupled assay described by Mdler el al. (1980). The assay medium contained 20 mM TES, pH 7.5, 0.1 M K', 1 mM Mg'+, 0.1 mM Ca", 0.15 mM NADH, 1 mM I,hosphoenolpyruvate, lactate dehydrogenase (-30 IU), pyruvate ki- nase (-30 IU), 2 p~ of the calcium ionophore A23187, together with 100 pM or 5 mM MgATP, as indicated.

" Phosphorylat.ion from ATP was carried out a t 0 "C for 15 s in the presence of 20 mM MOPS buffer, pH 7.0, 80 mM K', 5 mM Mg', 0.1 mM Ca')+, 2 p~ [-y-'"l']ATP. Phosphorylation from Pi was carried out for 2 min at 20 "C in the presence of 50 mM MES, pH 6.0, 5 mM Mi"', 0.5 mM :"Pi, 2 mM EGTA, and 20% (v/v) dimethyl sulfoxide.

' The specific calcium act,ivated ATPase activity is shown relative to that. of the wild-t.ype enzyme measured a t 5 mM MgATP. The specific phosphorylat.ion levels are also shown relative to t,hat of t,he wild-t.ype enzyme.

biphasic ATP concentration dependence reflects the dual role of ATP as a substrate and as an allosteric modulator (for a review, see Andersen, 1989). In the active mutants, the specific Ca"-ATPase activity increased by a factor identical to that observed for the wild-type enzyme, when the ATP concentration was increased from 100 PM to 5 mM. Thus, the modulation by ATP bound with low affinity was intact in these mutants.

Phosphorylation from ATP and P,-The phosphoenzyme intermediates of the mutants were studied in the series of experiments described below. All the mutants with single amino acid replacements were able to form an acid-stable phosphoenzyme with a steady-state concentration similar to that of the wild type, when phosphorylated in the presence of 2 PM [y-'"P]ATP a t 0.1 mM Ca'+ (Table I1 and lowerpanelof the inset of Fig. 3). Since the wild-type enzyme displayed an apparent KO., around 0.5 PM under these conditions, any significant reduction of the apparent affinity for ATP, in- duced by the mutation, should have been revealed as a decrease in the relative phosphorylation level. For the mutants with severely reduced turnover rates, Ile-315 - Arg, Thr-317 + Asp and Leu-319 + Arg, we performed a titration of the ATP concentration dependence of phosphorylation at [y-"'PI ATP concentrations varying between 0.1 and 10 PM. The apparent ATP affinities determined in this way did not differ from the apparent affinity of the wild-type enzyme (not shown). In Fig. 2, the Ca'+ affinities of the transport-defective mutants were examined by Ca:!+ titration of the phosphorylation from ATP. The apparent Ca" affinity of the Thr-317 + Asp mutant was indistinguishable from that of the wild type (compare with Fig. 5 of Vilsen et al., 1989), whereas the Ca2+ dependencies of the Ile-315 + Arg and IRU-319 + Arg mutants were slightly shifted toward a higher affinity.

Fig. 3 and the last column of Table I1 show the phosphorylation levels obtained in the backward reaction with inorganic phosphate in the absence of Ca:". The mutants He-315 +

Arg, Thr-317 + Asp and Leu419 - Arg were defective in carrying out the reaction with P,. The Thr-317 + Asp mutant displayed an apparent affinity for P,, which was slightly lower than that of the wild-type enzyme, by a factor of 1.8, whereas the maximum level of phosphorylation was normal, as judEed

Ileu3 15 4 Arg

Thr317 --. Asp

Leu319 -* Arg

pCa 4.0 5.0 5.5 6.0 6.2 6.5 7.0 8.0 FIG. 2. Calcium dependence of phosphorylation of the Ile-

315 4 Arg. Thr-317 4 Asp, and Leu-319 -+ A r g mutants from ATP. Phosphorylation o f th r C:n.'-A'I'l'ase in mirrtrsomal fractions isolated from COS-1 cells transfertrd with mutant (h."- ATPase cDNA was carried out at 0 "C for 15 s in the prrsencc of 20 mM MOPS huffer, pH 7.0. 80 mM K'. 5 mM MC', 2 pM [y-"l']A'I'l', and various concentrations of free Cay* srt with 14X;TA. The nrid- quenched samples containing an equivalrnt amount of exprrsqed ATPase were suhjected to SDS-polyncrylarnidr gel elrctrophorrsis nt pH 6.0, and radioactivity was detected hv nutoradiography.

18842 Amino Acid Alterations at the M4S, Boundary

4 20 100 500 2500

Pi (PM) FIG. 3 . Phosphorylation of the Ile-315 + Arg, Thr-317 +

Asp, and Leu-319 + Arg mutants and the wild-type Ca'+- ATPase from inorganic phosphate. Phosphorylation was carried out at various concentrations of.'.'I', for 2 min at 20 "C in the presence 01' 50 mM MES, pH 6.0, 5 mM Mi", 2 mM EGTA, and 20% (v/v) dimethyl sulfoxide. The acid-quenched samples were subjected to SL)S-polyacrvlamide gel electrophoresis at pH 6.0 followed by autoradiography and quantitation of radioactivity by densitometry. The specific phosphorylation levels (calculated per milligram of Ca'+- ATPase protein applied to the gel) are indicated relative to the phosphorylation level of the wild-t?ipe enzyme at 2.6 mM Pi. All data points represent the average values of three to five determinations. Sote the logarithmic scale on the abscissa. 0, wild t.ype; X, Ile-315 -+

Arg; A, Thr-317 -* Asp; 0, Leu-319 -+ Arg. The upprr panel of the insct shows examples of the autoradiograms from which the graphs were derived (P, concentration 100 p ~ ) . The lower panel of the inset shows ATP phosphorylations obtained with the same samples (conditions as in Fig. 2. at 0.1 mM Ca"). The following samples are shown in the insrt (lanes 1-4 from left to right): Ile-315-+ Arg mutant, Leu- :<I9 + Arg mutant, Thr-317 + Asp mutant, wild-type Ca'+-ATJ'ase. Equivalent amounts of expressed ATPase were applied to the lanes.

from the titration data in Fig. 3. In the two mutants, Ile-315 -P Arg and Leu-319 + Arg, the levels of phosphorylation from P, were less than 20% a t a Pi concentration of 2.5 mM, which corresponds to 50 times the concentration required to obtain half-saturation of the wild-type enzyme under identical conditions. We were unable to obtain reliable results a t higher P, concentrations, due to background phosphorylation of non- Ca"-ATPase proteins present in the crude microsomal preparation. It was therefore not possible to decide whether the low phosphorylation levels of these mutants were referable to a reduction of the affinity or of the maximum phosphorylation level, or of both. The inset of Fig. 3 shows an example of the autoradiographic data from which the graphs were derived. The complete phosphorylation obtained with ATP in the same samples is also shown in this inset (lower panel), to document that the lack of phosphorylation from Pi was due to specific effects of the mutations and not to nonspecific denaturation of the mutant enzymes in the particular prepa- rations under study.

Dephosphorylation Experiments-Fig. 4 depicts the results of kinetic experiments, in which we compared the decay rates of the ADP-sensitive phosphoenzyme intermediates (EIP) of the wild-type enzyme and the Ile-315 + Arg, Thr-317 + Asp and Leu-319 + Arg mutants at 0 "C. It can be seen that the dephosphorylation was blocked to various extents in the mutants. In the Ile-315 + Arg and Leu-319 + Arg mutants, the dephosphorylation rate constant was strongly reduced, to 10 and 396, respectively, of that of the wild type. A less pronounced inhibition of dephosphorylation, to about 20%, was

0 10 20 30 40 50 Time (9 )

FIG. 4. Dephosphorylation of the EIP intermediate of the Ile-315 + Arg, Thr-317 + Asp, and Leu-319 + Arg mutants and the wild-type Ca"+-ATPase a t 0 OC. The ADP-sensitive E I J ' phosphoenzyme intermediate was formed by phosphorylation with ATP at 0 "C in t.he presence of 20 mM MOPS buffer, pH 7.0, 80 mM K', 5 mM Mi'+, 2 pM [-y-:"P]ATP, 0.1 mM Ca", and 2 pM of the calcium ionophore A23187. Aft.er phosphorylation for 16 s, 1 mM EGTA was added, followed by acid quenching a t serial time intervals as indicated. The acid-quenched phosphoprotein was quantified as descrihed for Fig. 3. The phosphorylation levels are indicated relative to the phosphorylation measured after direct quenching without prior addition of EGTA. All data points represent the average values of three to five determinations. 0, wild type; 0, IIe-315 + Arg; A, Thr- 317 -+ Asp; 0, Leu-319 + Arg. In a parallel set. of experiments the ADP sensitivity o f the phosphoenzyme of the wild type and mutants was demonstrated by inclusion of 1 mM ADP with EGTA, resulting in complete dephosphorylation within 5 s (not shown).

Wild type Ileu3 1 S+Arg Thr3 17-Asp Leu3 19-Arg

?

n- n ri - +ADP - +ADP - +ADP - +ADP FIG. 5. Accumulation of the&P intermediate in the Ile-315

+ Arg, Thr-317 + Asp, and Leu-319 + Arg mutants and the wild-type Ca"+-ATPase. In order to accumulate the ADP-insensi-

ence of 2 p~ (-y-?']ATP, 100 mM TES/Tris, pH 8.36, 10 mM Mi", tive E,P intermediate, phosphorylation was carried out in the pres-

50 p M &'+, 2 p M of the calcium ionophore A23187, for 15 s at 0 "C. The phosphorylated sample was acid quenched directly (-), or the ADP sensitivity was tested by addition of 1 mM ADP with 1 mM EGTA, followed by acid quenching 6 s later (+ A U P ) . The acid- quenched samples containing equivalent amounts of expressed ATP- ase were subjected to SDS-polyacrylamide gel electrophoresis a t pH 6.0, and radioactivity was detected by autoradiography.

observed in the Thr-317 + Asp mutant. Dephosphorylation experiments similar to those presented

in Fig. 4 were also performed with the Ile-315 + Ser and Ile- 315 + Asp mutants which displayed steady-state Ca'+ uptake activities in the range between 30 and 40% of the wild-type activity, but we were unable to demonstrate any difference between the dephosphorylation rates of these mutants and that of the wild-type enzyme (not shown).

The decay of EIP normally occurs through the conversion of this phosphoenzyme intermediate to the hydrolyzable ErP intermediate. In Fig. 5 we studied the steady-state level of E2P, during phosphorylation from ATP a t alkaline pH and low alkali metal concentration. Under these conditions the rate of hydrolysis of E,P is much lower than the rate of the E,P-EzP interconversion, in the wild-type enzyme, and ErP therefore accumulated, as demonstrated by the low reactivity with ADP (lane 2 of Fig.5). With the Ile-315 + Argand Leu- 319 + Arg mutants, however, the observed steady-state levels of E,P were almost zero. With the Thr-317 + Asp mutant, an intermediate level of ErP was observed. These data suggested that the E,P-E,P interconversion was inhibited in the

Amino Acid Alterations at th.c M,,S, Boundary 18843

mutants, consistent with the low dephosphorylation rates ohserved in Fig. 4.

In Fig. 6 we examined whether there was any inhihition of the catalysis of E,P dephosphorylation. In Fig. 6A, the E,P intermediate was formed from inorganic phosphate. The rates of phosphoenzyme decay ohserved after dilution of the mutant &P intermediates in a medium containing K' a t neutral pH were much higher than the decay rates observed for the E,P intermediates, and indistinguishahle from the decay rate of the wild-type &P, at the time resolution of our experimental set up. With the Ile-315 + Arg and Leu-319 + Arg mutants, we performed supplementary experiments in which the decay of E,P, formed in the forward reaction with ATP, was meas-

A

100

h

x v 80 d

4 60 .- 0

4

h I+ 2 40 a VI 0

a c 20

0

B

I I

0 5 l o Time ( s )

lleu315 - Arg

CI "

.. M

a b c d e f

FIG. 6. Dephosphorylation of the B2P intermediate of the Ile-315 + Arg, Thr-817 4 Asp, and Leu-SIR .-, Arg mutants and the wild-type Ca"-ATPaseat 0 O C . /'and A , E;21'was formed I)y phtrsphnrvlat ion in the presence of.500 p~ "I1, under the contlit ions tlescriht~d for Fig. 3 . Following cooling of the sample to 0 "C, dephosphorylation was initiated hv 20-fold dilution of an aliquot into an ice- c ~ ) l t l medium contnining 20 mM MOPS hrtffer pH 7.0, 80 mM K', 5 mM Mg:", 0.1 mM Ca"' , and acid quenching was perlhrmed at serial time intervals as indicated. The acid-quenched phosphoprotein was quantified as descrihetl for Fig. 3. The phosphorylation levels are indicated relative to the wild-t,ype level ohserved hy quenching hefore clilut ion of the sample. 0, wild t.ype; 0, Ile-315 + Arg; A, Thr-317 - Asp: 0, Leu419 - Arg. I'anrl 13, E?I' was accumulated hv phos- phory1:ltion in the presence of 2 pM [y-."'P]ATI', 100 mM TI<S/Tris, p H 8.:15. 1 0 mM Mf ' , 50 pM Ca:", 2 p M o f the calcium ionophore A'L3187, for 15 s at 0 "C, followed hy incnhation for an additional 60 s in the presence of 1 mM I*X'I'A. An aliqout of 100-pI sample was then trrinsl'ered t o 2.0 ml of' an ice-cold medium, containing 20 mM MOPS buffer, pH 7.0, 80 mM K', 5 mM Mp:", 1 mM EGTA without (lnncs a-c) o r with (/ants d - f ) 1 mM AIIP. Acid quenching was performed simlrltaneously with (lanrs a and d ) , and 10 s (Iancs h and 0 ) or 20 s (Ianrs c and f ) after the transfer. The acid-quenched s;lmples containing equivalent amounts of expressed ATPase were suhjccted to SDS-polyacrylamide gel elect rophoresis at. pH 6.0, and radioactivity was detected hv autoradiography.

ured (Fig. 6 H ) . The phosphorylation with ATP was carried out for 15 s under the same conditions as described for Fig. 3, followed hy addition of EGTA and incubation for another 60 s t o allow accumulation of as much E,P as possible. An aliquot was then diluted into the pH 7.0,O.I M K' medium to ohserve dephosphorylation after 10 and 20 s. As can he seen in Fig. f H (lanrs h and c ) , only part of the phosphoenzvme decayed rapidly. This contrasted with the wild-t.ype Ca."-ATl'ase, in which all the phosphoenzyme disappeared within 10 s under similar conditions (Vilsen rt a/ . , 1991 ). I'arallel experiments in which ADP was added, however, showed that the mutant phosphoenzvme remaining after 10 s was A D P sensitive (Fig. 6H, lanes P and f ) and, therefore, represented the part of P;,I' not being converted to within the 60-s incuhation with EGTA. These results are consistent with a normal rate of hydrolysis of E,P in the IIe-315 -. Arg and I,eu-319 -. Arg mutants.

DlSClJSSlON

In the present study we have characterized a numher of Ca- ATPase mutants with replacements of hvdrophohic amino acids located at the M.,S, houndary. The mutations Ile-31.5 -. Arg, and Leu-319 + Arg resulted in complete loss of Ca"+ uptake and ATPase activity, while the Thr-317". Asp mutant retained 5 and 7 % of the wild-type activities, respectively. On the hasis of our kinetic analvsis of these three mutants , the turnover inhihition can he ascribed to a hlock of the enerkT- transducing F,'IP-R,P interconversion of phosphoenzvme intermediates. This is deduced from the low dephosphorylat ion rate of the ADP-sensitive E 1 J ' phosphoenzvmc intermediate, in conjunction with the finding that dephosphorylation from the ADP-insensitive I$P intermediate occurred at a normal rate. Moreover, the steady-state concentrat ion of the F,'.P intermediate was very low during phosphorylation with ATI', under conditions which producerl !*,':I' accumrtlntion in the wild-type enzyme.

The inhihition of the EIP-E,P transit inn was most pronounced in the He-31.5 + Arg and I,eu-319 -. Arg mutants, with which we also observed a dramatically r e d r ~ c c d level of E2P, formed in the hackward reaction with inorganic phos- phat.e. Even a t a I', concentration 50 times higher than that required for half-saturation of the wild-typ,r~ enzyme, the phosphorylat,ion levels of these mutants still constituted less than 20% of the wild-t.ype phosphorvlation level (Fig. 3 ) .

Previously, we ohserved a diminished phosphorylation from P, with mutants in which a conserved glycine ((;ly-2:%3) in tha t par t of t.he 13-strand sector adjoining the memhrane had heen changed to either a glutamic acid, a \valine. o r an arginine (Andersen rt a/. , 1989), and with a mutant in which the conserved Lvs-684 of the "hinge domain" had heen changed to an arginine (Vilsen P I nl., 1991). In these mutants t h e w was a concomitant hlock of the E I P - K 2 I ) transition in the forward reaction and, accordingly, the functional tlisturhnnce closely resemhled that of the lle-315 -. Arg and l,eu-:ll9 -. Arg mutants. Moreover, in the Ile-315 -. Arg and L e u - 3 1 9 +

Arg mutants, as well as in the GIV-233 mutants, the apparent affinity for Ca:", measured in the phosphorylation reaction, was found to he higher than that ofthe wild type, an ohser- vation which mav be explained by the kinetic effect of accumulation of EIP (Andersen ct al., 1941)).

The mutants Thr-416 + Asp, Thr-317 -* Ala. Cys-318 -. Ala, and Ala-320 + Arg displayed maximum specific (*a"' uptake rates close to or equal to that o f t he wild type, whereas in t.he mutants Ile-315 -. Ser and He-315 -. Asp, the ('a *

transport activity was reduced to 34 and WY of the wild-type activity, respectivelv. It is noteworthy that almost the same

18844 Amino Acid Alterations ut the M4S4 Boundary

relative reduction was found for Ca2+ transport as for ATP hydrolysis, consistent with a tight coupling between these two activities. The apparent Ca2+ affinities of these mutants, measured in the Ca2+ transport assay, did not differ signifi- cantly from the apparent affinity of the wild type (Table I). Moreover, in the Ile-315 + Ser and Ile-315 + Asp mutants, the specific Ca”-ATPase activity increased by a factor identical to that observed for the wild-type enzyme when the ATP concentration was increased from 100 p M to 5 mM. The apparent affinities for ATP and Pi displayed in the phosphorylation reaction were normal as well, suggesting that neither the high affinity catalytic ATP site nor the low affinity modulatory site was preferentially affected. The observations with these mutants, therefore, seem to be consistent with inhibition of a reaction following the transfer of the y-phosphate of ATP to the enzyme. The E1P-E2P interconversion is a candidate for such a reaction, but a direct confirmation of a partial E1P-E2P block would require a higher time resolution in the kinetic measurements than was attainable in the present study.

The hydrophobic nature of the NH2-terminal part of the S4 segment makes it difficult to define the border between Mq and Sq, solely on the basis of secondary structure prediction (MacLennan et al., 1985). The finding that the 3 residues Ile- 315, Thr-316, and Thr-317 could be mutated to charged residues, both singly and as a triple mutation Ile-Thr-Thr- 317 * Arg-Asp-Asp, with maintenance of the insertion of the protein in the microsomal membrane, seems to be consistent with a location of the 3 residues at the cytoplasmic side of the S4M4 interface (Fig. l), at least in one of the major conformational states of the protein. During enzyme turnover, hydrophobic residues of Sq may form transient contacts with the lipid phase and with the transmembrane segments (Brand1 et al., 1985). A deeper immersion of the S4 helix in the lipid bilayer in the E,P form would be consistent with the present observation of an EIP-E,P block after replacing some of the hydrophobic Sq residues with charged amino acids, and with our previous findings with a hydrophobic photoactivatable reagent 3-trifluoromethyl-3-(iodophenyl)diazirine, incorpo- rated in the lipid phase (Andersen et al., 1986). With this reagent we demonstrated an increased binding to the Ca2+- ATPase in the E,P form, relative to EIP, and the preferential hydrophobic labeling of E,P was located on the A1 tryptic fragment, part of which is constituted by the S4 segment (Andersen et al., 1986). Disruption of the high affinity Ca2+- binding sites and release of occluded Ca2+ to the luminal surface, in relation to the EIP-E,P transition, may involve a rotation or tilting of the Mq helix (Vilsen et al., 1989; Mac- Lennan, 1990). The present results would support a crucial role of Ile-315, Thr-317, and Leu-319 in the coupling of such conformational changes with changes in the cytoplasmic domains.

Studies of the effects of organic solvents on the reactivity of the Ca”-ATPase with Pi, and on the synthesis of ATP from the phosphorylated enzyme in the presence of ADP, have shown that energy transduction in the pump may be explained on the basis of a hydrophilic-hydrophobic transition at the catalytic site, occurring in relation to the E1P-E2P conformational change. By making the catalytic site in E2P hydrophobic, and thereby similar to a gas-phase, the conformational change would reduce the free energy of hydrolysis of the aspartylphosphate bond (de Meis et al., 1980; de Meis, 1981, 1989). It is, therefore, of interest to note that the Ile- 315 + Arg and Leu-319 + Arg mutations, which affected the formation of the E2P intermediate most profoundly, also brought about the largest reduction of hydropathy and that

the degree of turnover inhibition seemed to increase with decreasing hydropathy in the cases where two or more different substitutions of the same residue were carried out (Table I). One possibilty is that the conformational rearrangements involved in the E1P-E2P transition lead to a relatively close interaction between the critical S4 residues and the covalently bound phosphate. Thus the phosphate would rest on top of the headpiece of the protein in the EIP form, whereas it would be closer to the membrane in the E2P form.

It may also be suggested that the bulky and highly charged guanidinio group of the arginine side chain in the Ile-315 + Arg and Leu-319 + Arg mutants might interfere with inter- domain movements which close the catalytic cleft in connec- tion with the formation of E2P (Ross and McIntosh, 1987).

In the light of the pronounced inhibition of E2P formation observed with the Leu-319 +Argand Ile-315 --., Arg mutants, it is remarkable that Ala-320 could be substituted with arginine without a major effect. A possible explanation is that Leu-319 and Ile-315 would be located at the same surface of the putative helix formed by S4 (Fig. l ) , whereas Ala-320, juxtaposed with Leu-319 in the primary sequence, would be located on the opposite surface. Thus, the amphipathic character of the helix would be destroyed by the Ile-315 + Arg and Leu-319 + Arg mutations, but not by the Ala-320 + Arg mutation. The difference between the respective effects of substituting Asp for Thr-316 and Thr-317, may likewise be explained by reference to the possible locations of these residues on opposite surfaces of a putative helix structure. Since Thr-317 is located on that surface of the helix which is already polar in the wild-type enzyme, it is possible that the negative charge was responsible for the ElP-E2P block in the Thr-317 + Asp mutant. In the wild-type enzyme, the polar helix surface contains only positive charges (Fig. 1).

The present mutagenesis data may be compared with the conservation of the residues within the family of cation transporting ATPases with phosphorylated intermediates. At the positions corresponding to Ile-315 and Leu-319 in the Caz+- ATPase, the Na+,K+-ATPase, the H+,K+-ATPase, the H+- ATPase, and the plasma membrane Ca2+-ATPase, all contain 1 of the hydrophobic residues Ile, Leu, Val, or Met (Shull et al., 1985; Serrano et al., 1986; Shull and Lingrel, 1986; Shull and Greeb, 1988). At the position corresponding to Thr-317, the other ATPases contain either threonine or a more hydrophobic residue. At the positions corresponding to Thr-316 and Ala-320, these ATPases contain the identical residue or a more polar residue. This correlates well with our observation that polar residues can be introduced at positions 316 and 320 without a major functional disturbance, whereas normal function requires a hydrophobic residue at positions 315,317, and 319. On the other hand, Cys-318 is conserved among the sarcoplasmic reticulum Ca*+-ATPase, the Na+,K+-ATPase, and the H+,K+-ATPase, but could be mutated to Ala without a major effect.

Acknowledgments-We would like to thank Jytte Jflrgensen, Karin Kracht, and Janne Petersen for their expert and invaluable technical assistance, Dr. N. M. Green for discussion of the Ca”-ATPase structure, and Dr. R. J. Kaufman, Genetics Institute, Boston, for the gift of the expression vector p91023(B).

REFERENCES

Andersen, J. P. (1989) Biochim. Biophys. Acta 988,47-72 Andersen, J. P., Vilsen, B., Collins, J. H., and Jbrgensen, P. L. (1986)

Andersen, J. P., Vilsen, B., Leberer, E., and MacLennan, D. H. (1989)

Brandl, C. J., Green, N. M., Korczak, B., and MacLennan, D. H.

J. Membr. Bid. 93, 85-92

J. Biol. Chem. 264, 21018-21023

(1986) Cell 44, 597-607

Amino Acid Alterations at the M4S4 Boundary 18845

Clarke, D. M., Loo, T. W., Inesi, G., and MacLennan, D. H. (1989a) Nature 339,476-478

Clarke, D. M., Maruyama, K., Loo, T. W., Leberer, E., Inesi, G., and MacLennan, D. H. (198913) J. Biol. Chem. 264,11246-11251

Clarke, D. M., Loo, T. W., and MacLennan, D. H. (1990a) J. Biol. Chem. 265,6262-6267

Clarke, D. M., Loo, T. W., and MacLennan, D. H. (1990b) J. Biol. Chem. 265,14088-14092

Clarke, D. M., Loo, T. W., and MacLennan, D. H. (1990~) J. Biol. Chem. 265, 22223-22227

de Meis, L. (1981) in T h e Sarcoplasmic Reticulum: Transport and Energy Transduction (Bittar E., ed) Vol. 2, John Wiley & Sons, New York

de Meis, L. (1989) Biochim. Biophys. Acta 9 7 3 , 333-349 de Meis, L., Martins, 0. B., and Alves, E. W. (1980) Biochemistry

Gluzman, Y. (1981) Cell 2 3 , 175-182 Hasselbach, W., and Makinose, M. (1961) Biochem. Z. 333, 518-528 Inesi, G. (1985) Annu. Reu. Physiol. 47, 573-601 Inesi, G., Sumbilla, C., and Kirtley, M. E. (1990) Physiological Reu.

Kaufman, R. J., Davies, M. V., Pathak, V. K., and Hershey, J. W. B.

Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , 488-492 Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132 MacLennan, D. H. (1970) J. Biol. Chem. 245, 4508-4518 MacLennan, D. H. (1990) Biophys. J . 58, 1355-1365

19,4252-4261

70, 749-760

(1989) Mol. Cell. Biol. 9 , 946-958

MacLennan, D. H., Brandl, C. J., Korczak, B., and Green, N. M.

Maruyama, K., and MacLennan, D. H. (1988) Proc. Natl. Acad. Sci.

Maruyama, K., Clarke, D. M., Fujii, J., Inesi, G., Loo, T. W., and

Mbller, J. V., Lind, K. E., and Andersen, J . P. (1980) J. Biol. Chern.

Ross, D. C., and McIntosh, D. B. (1987) J. Biol. Chem. 262, 12977-

Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad.

Serrano, R., Kielland-Brandt, M. C., and Fink, G. R. (1986) Nature

Shull, G. E., and Lingrel, J. B. (1986) J. Biol. Chem. 261, 16788-

Shull, G. E., and Greeb, J. (1988) J. Biol. Chem. 263, 8646-8657 Shull, G. E., Schwartz, A., and Lingrel, J. B. (1985) Nature 316,

Taylor, K. A., Dux, L., and Martonosi, A. (1986) J. Mol. Biol. 187,

Vilsen, B., Andersen, J. P., Clarke, D. M., and MacLennan, D. H. (1989) J. Biol. Chem. 264, 21024-21030

Vilsen, B., Andersen, J. P., and MacLennan, D. H. (1991) J. Biol. Chem. 266, 16157-16164

Zubrzycka-Gaarn, E., MacDonald, G., Phillips, L., J@rgensen, A. O., and MacLennan, D. H. (1984) J. Bioenerg. Biomembr. 16,441-464

(1985) Nature 316, 696-700

U. S. A. 85,3314-3318

MacLennan, D. M. (1989) J. Biol. Chem. 264, 13038-13042

255,1912-1920

12983

Sci. U. S. A. 74, 5463-5467

319,689-693

16791

691-695

417-427

functional consequences of alterations to hydrophobic amino

Documents