HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 279, Issue 31, 32515-32523, July 30, 2004

This Article Abstract Full Text (PDF) All Versions of this Article: 279/31/32515    most recent M403242200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by McIntosh, D. B. Articles by Andersen, J. P. Search for Related Content PubMed PubMed Citation Articles by McIntosh, D. B. Articles by Andersen, J. P. Social Bookmarking                      What's this?

Roles of Conserved P Domain Residues and Mg²+ in ATP Binding in the Ground and Ca²+-activated States of Sarcoplasmic Reticulum Ca²+-ATPase^*

David B. McIntosh, Johannes D. Clausen¶, David G. Woolley, David H. MacLennan||, Bente Vilsen¶, and Jens Peter Andersen¶**

From the Chemical Pathology, Department of Clinical Laboratory Sciences, Faculty of Health Sciences, University of Cape Town, and National Health Laboratory Service, Groote Schuur Hospital, Cape Town 7925, South Africa, the ^||Banting and Best Department of Medical Research, University of Toronto, Toronto, Ontario M5G 1L6, Canada, and the ^¶Department of Physiology, University of Aarhus, DK-8000 Aarhus C, Denmark

Received for publication, March 23, 2004 , and in revised form, April 28, 2004.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Residues in conserved motifs ⁶²⁵TGD, ⁶⁷⁶FARXXPXXK, and ⁷⁰¹TGDGVND in domain P of sarcoplasmic reticulum Ca²⁺-ATPase, as well as in motifs ⁶⁰¹DPPR and ³⁵⁹NQR(/K)MSV in the hinge segments connecting domains N and P, were examined by mutagenesis to assess their roles in nucleotide and Mg²⁺ binding and stabilization of the Ca²⁺-activated transition state for phosphoryl transfer. In the absence of Mg²⁺, mutations removing the charges of domain P residues Asp⁶²⁷, Lys⁶⁸⁴, Asp⁷⁰³, and Asp⁷⁰⁷ increased the affinity for ATP and 2',3'-O-(2,4,6-trinitrophenyl)-8-azidoadenosine 5'-triphosphate. These mutations, as well as Gly⁶²⁶ Ala, were inhibitory for ATP binding in the presence of Mg²⁺ and for tight binding of the ,-bidentate chromium(III) complex of ATP. The hinge mutations had pronounced, but variable, effects on ATP binding only in the presence of Mg²⁺. The data demonstrate an unfavorable electrostatic environment for binding of negatively charged nucleotide in domain P and show that Mg²⁺ is required to anchor the phosphoryl group of ATP at the phosphorylation site. Mutants Gly⁶²⁶ Ala, Lys⁶⁸⁴ Met, Asp⁷⁰³ Ala/Ser/Cys, and mutants with alteration to Asp⁷⁰⁷ exhibited very slow or negligible phosphorylation, making it possible to measure ATP binding in the pseudo-transition state attained in the presence of both Mg²⁺ and Ca²⁺. Under these conditions, ATP binding was almost completely blocked in Gly⁶²⁶ Ala and occurred with 12- and 7-fold reduced affinities in Asp⁷⁰³ Ala and Asp⁷⁰⁷ Cys, respectively, relative to the situation in the presence of Mg²⁺ without Ca²⁺, whereas in Lys⁶⁸⁴ Met and Asp⁷⁰⁷ Ser/Asn the affinity was enhanced 14- and 3–5-fold, respectively. Hence, Gly⁶²⁶ and Asp⁷⁰³ seem particularly critical for mediating entry into the transition state for phosphoryl transfer upon Ca²⁺ binding at the transport sites.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
The sarcoplasmic reticulum Ca²⁺-ATPase¹ from rabbit muscle couples the hydrolysis of ATP to the transport of Ca²⁺ to the reticular lumen, and the lowering of the sarcoplasmic Ca²⁺ concentration leads to muscle relaxation. This ion pump is a well studied example of P-type ATPases, the catalytic cycle of which includes a phosphorylated aspartyl intermediate and progression through E1 and E2 states. ATP-dependent phosphorylation takes place upon Ca²⁺ binding in the E1 form, and hydrolysis of the aspartyl phosphoryl bond occurs in the E2P phosphoenzyme conformation following Ca²⁺ translocation across the membrane. Two atomic structures of the protein have been elucidated by x-ray crystallography, one in the presence of Ca²⁺ (designated E1(Ca₂)) and another in the absence of Ca²⁺ and presence of the specific inhibitor thapsigargin (E2(TG)) (1, 2). They reveal 10 membrane-spanning helices, connected through a stalk section to the phosphorylation (P) domain in association with nucleotide (N) and actuator (A) domains. The three head domains are loosely attached in E1(Ca₂) and closer together, forming a more compact entity, in E2(TG). For ATP-dependent phosphorylation of Asp³⁵¹ to occur in the Ca²⁺ bound E1 state, domain N must close over the phosphorylation site, such that the nucleotide site is brought close to Asp³⁵¹. ATP likely straddles both domains, the nucleoside portion anchored in domain N and the phosphates stretching into domain P (cf. Fig. 1A).

View larger version (34K): [in this window] [in a new window]   FIG. 1. Disposition of domain P residues in Ca2+-ATPase (A) and of equivalent residues in a pseudotransition state of phosphoserine phosphatase (B). In A, the structure of Ca2+-ATPase is that of the E2(TG) crystal structure of Ca2+-ATPase (Ref. 2, PDB accession no. 1IWO [PDB] ). Domain A has been omitted for clarity. A model of an ATP molecule, separated between the - and -phosphates (indicated by dashed line), is shown in yellow. The ATP was taken from an NMR structure of recombinant N domain of Na+,K+-ATPase complexed with ATP (Ref. 18, PDB accession no. 1MO8 [PDB] ). The AMP portion has been inserted in the Ca2+-ATPase in a position that approximates that of the NMR structure, and the pyrophosphate portion is positioned such that the -phosphate approximates the phosphoryl group of phosphoserine in a crystal structure of the homologous phosphoserine phosphatase with bound substrate (Ref. 5, PDB accession no. 1L7P [PDB] ). The green lines in A indicate distances that need to shorten by 4–5 Å to approach the structure of phosphoserine phosphatase complexed to AlF and Mg2+, which is presented in B (Ref. 5, PDB accession no. 1L7N [PDB] ). This structure is thought to mimic the transition state in the phosphoryl 3transfer reaction (5). The AlF (homolog of the phosphoryl) is shown in yellow and Mg2+ as a green ball. Three of the water molecules present in the crystal are highlighted as 3light blue balls. The green lines in B show distances of less than 3.2 Å. The equivalent residues of Ca2+-ATPase are indicated in parentheses. Both pictures were prepared by use of DS Viewer Pro (Discovery Studio, Accelrys Inc.).

The present study expands on our previous studies of domain P mutations, which investigated the role of the conserved phosphorylation loop ³⁵¹DKTGTLT in MgATP binding (6, 7). We now focus on the other conserved segments of the P domain mentioned above. We have introduced new mutations of some of the conserved residues in these segments, including the putative Mg²⁺ binding residue, Asp⁷⁰³, as well as Lys⁶⁸⁴ and Asp⁷⁰⁷, and we also reexamine previously studied mutants (8–10), using our recently developed techniques for measurement of nucleotide binding and kinetics of phosphorylation, to assess the roles of these residues in nucleotide and Mg²⁺ binding and stabilization of the Ca²⁺-activated transition state for phosphoryl transfer. Residues in the highly conserved hinge segments, ⁶⁰¹DPPR and ³⁵⁹NQ(R/K)MSV, which link the N and P domains, have been studied to probe interdomain movements and the role of the hinge. Because some of the mutants exhibited very slow or negligible phosphorylation, it was possible to measure ATP binding in the presence of both Mg²⁺ and Ca²⁺, thereby providing insight in the interaction with ATP in the transition state for phosphoryl transfer.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Mutagenesis and Expression—Oligonucleotide-directed mutagenesis of cDNA encoding the rabbit fast twitch muscle Ca²⁺-ATPase (SERCA1a isoform) was carried out as described previously (9). For expression, the cDNA in pMT2 vector (11) was transfected into COS-1 cells by calcium phosphate precipitation (12). The microsomal fraction containing expressed wild-type or mutant Ca²⁺-ATPase was recovered by differential centrifugation (13). The concentration of expressed protein was quantified by a specific enzyme-linked immunosorbent assay (9).

Phosphorylation—Transient kinetic experiments at 25 °C were performed using a Bio-Logic quenched flow module QFM-5 (Bio-Logic Science Instruments, Claix, France) as described (14). The reaction was started by the addition of ATP to the Ca²⁺-bound Ca²⁺-ATPase. Other phosphorylation and dephosphorylation experiments were carried out by manual mixing techniques (7). Acid quenching was performed with 0.5–2 volumes of 25% (w/v) trichloroacetic acid containing 100 mM H₃PO₄. The acid-precipitated protein was washed by centrifugation and subjected to SDS-polyacrylamide gel electrophoresis in a 7% polyacrylamide gel at pH 6.0 (15), and the radioactivity associated with the separated Ca²⁺-ATPase band was quantified by imaging, using a Packard Cyclone^TM Storage Phosphor System. The background phosphorylation level was determined in parallel experiments with control microsomes isolated from mock-transfected COS-1 cells.

Assays for Nucleotide Binding—The synthesis of [-³²P]TNP-8N₃-ATP, the photolabeling of COS-1 cell microsomes containing wild-type or mutant Ca²⁺-ATPase (at room temperature), the inhibition by ATP, and the quantification of labeled bands by radio-imaging following SDS-polyacrylamide gel electrophoresis were carried out as described previously (16). The concentration of [-³²P]TNP-8N₃-ATP was 3 x K_0.5 in the inhibition experiments with ATP.

Tight ,-Bidentate Cr(III)ATP (CrATP) Binding—The assay consists of preincubating the Ca²⁺-ATPase with CrATP at 37 °C and then determining the fraction of enzyme not tightly complexed with CrATP by photolabeling with [-³²P]TNP-8N₃-ATP in excess EDTA following a large dilution, as described previously (6) and also in the legend to the figure. ,-Bidentate CrATP was prepared as described by Dunaway-Mariano and Cleland in method B (17).

Calculations and Data Analysis—The phosphorylation data and the [-³²P]TNP-8N₃-ATP labeling data were analyzed as detailed previously (14, 16). The "true" dissociation constant for ATP and MgATP binding was calculated using the validated equation for competitive inhibition of the [-³²P]TNP-8N₃-ATP labeling (16).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
The mutants examined are listed in Table I. To follow the results, it is helpful to know the disposition of these residues in the phosphorylation site and roughly how ATP and Mg²⁺ may bind. A model based on the crystal structure of Ca²⁺-ATPase in the E2(TG) conformation (2), the atomic structure of the homologous phosphoserine phosphatase with bound phosphoserine (5), and an NMR structure of ATP bound to domain N of Na⁺,K⁺-ATPase (18) is presented in Fig. 1A. ATP is shown in yellow. The AMP portion of ATP is lodged in domain N, and the - and -phosphates in domain P in a position close to Asp³⁵¹, the residue receiving the -phosphate from ATP. The ATP molecule is stretched between the - and -phosphates (indicated by dashed line) to reach both sites, and it is apparent that further movement of the two domains is needed for the ATP to straddle both binding sites. This problem is even more pronounced in the E1(Ca₂) Ca²⁺-ATPase crystal structure (1), which led us to prefer the E2(TG) structure rather than E1(Ca₂) for modeling the ATP binding site. In a transition state-like structure of phosphoserine phosphatase, where AlF₃ is bound as a phosphoryl transition state analog, Mg²⁺ is coordinated by a fluoride atom, the residues equivalent to Asp³⁵¹ and Asp⁷⁰³, the main chain carbonyl of the equivalent of Thr³⁵³, and two water molecules (Fig. 1B). AlF₃ is ligated to main chain amides of residues equivalent to Lys³⁵², Thr³⁵³, and Gly⁶²⁶, as well as to side chains of the equivalents of Thr⁶²⁵ and Lys⁶⁸⁴ (Ref. 5 and Fig. 1B). Hence, these residues are candidates as ligands of Mg²⁺ and -phosphate, respectively, in the transition state of the Ca²⁺-ATPase.

The binding parameters obtained with wild type and the mutants are presented in Table I, and examples of the underlying experimental data are shown in Fig. 2. In the table, values of K_{0.5(TNP-8N3-ATP)} and K_D(ATP) more than 2-fold different from wild type are shown in bold type, and, of these, those showing an increase in affinity are in italics. For side chains contributing to ligation of the nucleotide, a reduced affinity is expected upon mutation. For charged side chains, removal of the charge by mutation may lead to an increase of the affinity for the nucleotide, if there is electrostatic repulsion in the wild type. Examining domain P mutations in the absence of Mg²⁺ first, it can be seen that most charge-removing mutations increased the affinity for TNP-8N₃-ATP and ATP. This was especially pronounced for Asp⁷⁰⁷ Asn, where the affinity was increased 33- and 14-fold, respectively. Asp⁷⁰⁷ Cys hardly changed the affinity for ATP, which may be taken as an indication that the sulfhydryl group is ionized, thereby maintaining the same electrostatic environment as the aspartate. Rather surprisingly considering their conservation in related enzymes, but in keeping with what has been found for Asp³⁵¹ (6), replacement of Asp⁶²⁷ and Asp⁷⁰³ with the longer glutamate side chain had no effect on binding of TNP-8N₃-ATP or ATP, suggesting there are no steric clashes. We conclude that in wild type none of the residues appear to be involved in ligation of the phosphates of TNP-8N₃-ATP or ATP in the absence of Mg²⁺, and mostly electrostatic repulsive forces prevail.

View larger version (25K): [in this window] [in a new window]   FIG. 2. TNP-8N3-ATP photolabeling (upper panels) and ATP inhibition (lower panels). Photolabeling was performed with COS-1 cell microsomes in 25 mM EPPS/TMAH, pH 8.5, 20% (w/v) glycerol, variable concentrations of [-32P]TNP-8N3-ATP, and 2 mM EDTA (closed circles), 1 mM MgCl2 + 0.5 mM EGTA (open circles), or 1 mM MgCl2 + 0.05 mM CaCl2 (closed triangles). The samples were subjected to SDS-polyacrylamide gel electrophoresis and the radioactivity associated with the Ca2+-ATPase band quantified. The sum of a hyperbolic function and a linear component was fitted to the data. In the presented data and curves, the linear component has been subtracted. In the ATP inhibition experiments, the concentration of [-32P]TNP-8N3-ATP was close to 3 x K0.5 (see Table I), and ATP was included at the concentrations indicated. A simple binding function with an offset representing nonspecific labeling was fitted to the data. The presented data and curves have the offset subtracted. All data sets are the average of two or more experiments. The arrow indicates the difference induced by Ca2+.

Asn³⁵⁹ and Asp⁶⁰¹ are located on the flexible hinge segments linking domains N and P (cf. Fig. 1A). Mutation of these two residues had little or no effect on ATP binding in the absence of Mg²⁺, but in the presence of Mg²⁺, Asn³⁵⁹ Ala and Asp⁶⁰¹ Asn caused pronounced increases in affinity for both nucleotides, whereas Asp⁶⁰¹ Glu was detrimental.

Phosphorylation from [-³²P]ATP and Phosphoenzyme Turnover—In the presence of Ca²⁺, the interaction with MgATP normally leads to phosphoryl transfer to Asp³⁵¹ with resulting formation of the ADP-sensitive phosphoenzyme intermediate, E1P. It was previously reported that Asp⁶⁰¹ Glu, Gly⁶²⁶ Ala, Asp⁷⁰³ Ala, and Asp⁷⁰⁷ Asn are unable to form significant amounts of phosphoenzyme upon incubation with 2 µM [-³²P]ATP and Ca²⁺ for 10 s at 0 °C, whereas Asp⁶⁰¹ Asn, Asp⁶²⁷ Glu, Lys⁶⁸⁴ Arg, Asp⁷⁰³ Glu, and Asp⁷⁰³ Asn form a phosphoenzyme under these conditions (8–10). We were able to confirm these findings, except for Asp⁶⁰¹ Glu, which phosphorylated to a wild type-like level in our hands. With respect to those mutants that have not been previously studied, we observed wild type-like accumulation of phosphoenzyme in the case of Asn³⁵⁹ Ala and Asp⁶²⁷ Asn, whereas Lys⁶⁸⁴ Met, Asp⁷⁰³ Cys, Asp⁷⁰³ Ser, Asp⁷⁰⁷ Cys, and Asp⁷⁰⁷ Ser were unable to form a significant amount of phosphoenzyme when incubated with 2 µM [-³²P]ATP and Ca²⁺ at 0 °C. We then proceeded to study the rate of phosphorylation at 25 °C, using a quenched flow procedure (14). At 25 °C, all the mutants, except Lys⁶⁸⁴ Met and the three mutants with alteration to Asp⁷⁰⁷, were capable of undergoing measurable phosphorylation from [-³²P]ATP in the presence of Mg²⁺ and Ca²⁺ (E1 E1P transition). Examples of data obtained under these conditions are shown in Fig. 3 (A and B), and all the observed rate constants are listed in Table I. It should be understood that, because the MgATP concentration applied (5 µM) is close to the K_m for phosphorylation of wild type, changes in the affinity for MgATP as well as the V_max for phosphorylation are being reflected in the phosphorylation rate.

View larger version (12K): [in this window] [in a new window]   FIG. 3. Phosphoenzyme formed from [-32P]ATP. A, phosphorylation of mutants Asp601 Glu (squares), Asp627 Asn (solid circles), Asp703 Asn (triangles pointing upward), Asp703 Glu (triangles pointing downward) and wild type (open circles) was carried out at 25 °C in a medium containing 40 mM MOPS/Tris (pH 7.0), 80 mM KCl, 100 µM CaCl2, 5 mM MgCl2, and 5 µM [-32P]ATP. The samples were acid quenched at serial time intervals, using a Bio-Logic QFM-5 quench-flow module with mixing protocol as previously described (14), and further processed as described under "Experimental Procedures." The data for the wild type and each mutant were normalized separately, taking the maximum level of phosphorylation as 100%. For the mutants, the lines show the best fits of a monoexponential function to the data, the rate constants being listed in Table I. B, phosphorylation and data analysis were carried out as described above, except that a manual mixing technique was used for mutant Asp703 Ala (diamonds); note the time scale. The circles reproduce the wild-type data from panel A for direct comparison. C, E1P E2P transition. Phosphorylation was carried out for 15 s at 0 °C (wild type, Asn359 Ala, Asp703 Asn, and Asp703 Glu) or for 30 s at 25 °C followed by cooling on ice for 30 s (Asp703 Ala, because of its inability to phosphorylate at 0 °C) in 40 mM MOPS/Tris (pH 7.0), 80 mM KCl, 5 mM MgCl2, 1 mM EGTA, 0.955 mM CaCl2 (resulting in 10 µM free Ca2+), 2 µM calcium ionophore A23187 [GenBank] , and 5 µM [-32P]ATP. To follow dephosphorylation, the phosphoenzyme was chased at 0 °C by addition of 6.7 mM EGTA with either 1 mM nonradioactive ATP (open symbols) or 1 mM ADP (closed symbols), and acid quenching was performed at the indicated time intervals. The lines show the best fits of a monoexponential decay function, giving the rate constants listed in Table I. Symbols are as for panels A and B, except that squares represent Asn359 Ala.

It is remarkable that phosphoenzyme did accumulate in Asp⁷⁰³ Ala/Cys/Ser, despite the low rate of phosphorylation, suggesting that a step in the reaction sequence leading to dephosphorylation is also blocked. This was further examined for Asp⁷⁰³ Ala, and, as seen in Fig. 3C, the dephosphorylation observed upon chase of the phosphoenzyme with nonradioactive ATP was unusually slow. A rapid dephosphorylation was, on the other hand, observed upon addition of ADP to reverse the phosphorylation step, indicating that the accumulated phosphoenzyme intermediate was ADP-sensitive E1P and that the slow step in the forward dephosphorylation is the E1P E2P conformational transition.

Among the mutants studied, only Asn³⁵⁹ Ala and Asp⁷⁰³ Glu showed Ca²⁺ transport activity higher than 10% of wild type in the presence of a high MgATP concentration of 5 mM at 37 °C (50 and 30%, respectively; data not shown; assay carried out as described in Ref. 20). A block or slowing of the E1P E2P transition has previously been described for mutants Asp⁶⁰¹ Asn, Lys⁶⁸⁴ Arg, Asp⁷⁰³ Asn, and Asp⁶²⁷ Glu, on the basis of measurements similar to those shown in Fig. 3C (8, 9). We confirmed these results and, furthermore, observed a marked inhibition of the E1P E2P transition in Asp⁶⁰¹ Glu and Asp⁶²⁷ Asn, and a less pronounced slowing in Asn³⁵⁹ Ala and Asp⁷⁰³ Glu (Fig. 3C and Table I). Because E1P E2P normally is the rate-limiting step in the transport cycle at saturating ATP and Ca²⁺ concentrations, the inhibition of this step must contribute, besides a reduced phosphorylation rate, to reduce the Ca²⁺ transport activity in the mutants.

Nucleotide Affinity in the Presence of Ca²⁺—In the presence of both Mg²⁺ and Ca²⁺, TNP-8N₃-ATP is a substrate of the Ca²⁺-ATPase (19). Hence, it is normally not possible to measure TNP-8N₃-ATP or ATP binding by the photolabeling assay under these conditions, because the -phosphoryl group would be cleaved off, both from TNP-8N₃-ATP and from the ATP used to competitively inhibit the photolabeling. However, for the mutants exhibiting no or very low phosphorylation activity, the assay could be employed with both Mg²⁺ and Ca²⁺ present, i.e. the conditions where the transition state in the phosphoryl transfer reaction normally would be attained. Results are shown in Fig. 2 and Table I. With respect to TNP-8N₃-ATP binding, a pronounced Ca²⁺ effect was noticed only for Lys⁶⁸⁴ Met, where the binding affinity was lowered 10-fold relative to the situation without Ca²⁺. There were slight increases in affinity for the Asp⁷⁰³ and Asp⁷⁰⁷ mutants, the 4-fold increase in affinity seen for Asp⁷⁰⁷ Ser clearly being significant. In the case of ATP, the Ca²⁺-induced changes were quite dramatic and surprising for some of the mutants. In Fig. 2, the arrow indicates the effect of Ca²⁺. In Gly⁶²⁶ Ala, the ATP affinity was very low in the presence of Ca²⁺, 100-fold reduced relative to the situation with Mg²⁺ alone. Ca²⁺ lowered the ATP affinity 12-fold in Asp⁷⁰³ Ala. In contrast, Ca²⁺ increased the ATP affinity 14-fold in Lys⁶⁸⁴ Met. Similarly, 3- and 5-fold increases in ATP affinity were observed for Asp⁷⁰⁷ Ser and Asp⁷⁰⁷ Asn, respectively (Table I). Mutant Asp⁷⁰⁷ Cys was again the exception, Ca²⁺ lowering the affinity 7-fold. Ca²⁺ had little effect on Asp⁷⁰³ Ser. In Asp⁷⁰³ Cys, Ca²⁺ produced a biphasic inhibition curve, which is difficult to interpret (data not shown). It should be noted that in these experiments the Mg²⁺ concentration was 20-fold higher than the Ca²⁺ concentration, thus ensuring that the major part of the nucleotide was still present as the Mg²⁺ complex. Hence, the changes induced by Ca²⁺ are caused by Ca²⁺ binding at the transport sites.

Tight Binding of CrATP—The roles of these residues in nucleotide binding were further investigated by determining the rate at which CrATP becomes tightly bound at the active site in the presence of Ca²⁺ (21, 22). We have previously demonstrated that it is possible to determine the time course of tight CrATP binding in an assay in which the enzyme is preincubated with CrATP at 37 °C for various times followed by dilution into medium containing TNP-8N₃-ATP, which photolabels the enzyme fraction not having CrATP tightly bound (6). Results obtained using this assay are shown in Fig. 4 and summarized in Table I. Each of the panels shows photolabeling following preincubation for various times in the absence (open circles) and presence (closed circles) of CrATP. In the presence of CrATP, the disappearance of the ability to become photolabeled (i.e. the formation of the complex with CrATP) took place with a t of 10 min for wild-type enzyme, whereas in the absence of CrATP the photolabeling level remained steady over the time course. For Gly⁶²⁶ Ala, there was little effect of CrATP, indicating that no tight complex with CrATP was produced. Asp⁷⁰³ Ala/Cys/Asn/Ser and Asp⁷⁰⁷ Cys/Ser showed some effect of CrATP, but clearly the formation of the tight complex was slower in these mutants compared with wild type. In the hinge, the two mutations of Asp⁶⁰¹ moderately slowed the formation of the tight complex, and Asn³⁵⁹ Ala was like wild type. For Asp⁶²⁷ Asn, Lys⁶⁸⁴ Met, and Asp⁷⁰⁷ Asn, a fall off in labeling with time was noted even in the absence of CrATP. The presence of CrATP did not enhance this decline; therefore, none of these mutants appeared to produce a tight complex with CrATP.

View larger version (26K): [in this window] [in a new window]   FIG. 4. Tight CrATP binding. The microsomes were incubated at 37 °C in 50 mM MOPS/TMAH, pH 7.0, 100 mM NaCl, 2 mM MgCl2, 0.2 mM CaCl2, with (solid circles, test) or without (open circles, control) 0.84 mM CrATP. At the indicated times, aliquots were diluted 15.6-fold into ice-cold irradiation medium containing 25 mM EPPS/TMAH, pH 8.5, 2 mM EDTA, 20% (w/v) glycerol, [-32P]TNP-8N3-ATP at a concentration of 3 x K0.5, and either no CrATP (test) or 0.054 µM CrATP (control) and irradiated immediately. The samples were then subjected to SDS-polyacrylamide gel electrophoresis, and the radioactivity associated with the Ca2+-ATPase band quantified.

View larger version (21K): [in this window] [in a new window]   FIG. 5. Effect of Asp703 mutations on Pi and Mg2+ concentration dependence values of phosphorylation from 32Pi. The reaction was performed for 10 min at 25 °C in 100 mM MES/Tris (pH 6.0), 2 mM EGTA, 30% (v/v) dimethyl sulfoxide, and either (A) 10 mM MgCl2 and varying concentrations of 32Pi or (B) 0.5 mM 32Pi and varying concentrations of MgCl2. The lines show the best fits of the equation, EP = EPmax·[Pi]n/(K0.5 + [Pi]n) to the data, giving the K0.5 values indicated in parentheses: circles, wild type (K0.5(Pi) 11 µM, K0.5 (Mg2+) 58 µM); squares, Asp703 Glu (K0.5(Pi) 40 µM, K0.5 (Mg2+) 805 µM); triangles, Asp703 Asn (no significant phosphorylation).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Electrostatic Effects and Role of Mg²⁺ in Nucleotide Binding—A large difference was found between mutational effects on nucleotide binding in the absence and presence of Mg²⁺. In the absence of Mg²⁺, mutations removing the charges of domain P residues Asp⁶²⁷, Lys⁶⁸⁴, Asp⁷⁰³, and Asp⁷⁰⁷ mostly enhanced TNP-8N₃-ATP and ATP binding, suggesting proximity of the phosphate chain to these residues in the wild type, but not ligation. In the presence of Mg²⁺, these mutations, as well as Gly⁶²⁶ Ala, were inhibitory for ATP binding, suggesting that in the wild type there is close favorable interaction of these residues with MgATP, i.e. ligation. For TNP-8N₃-ATP, only mutations of the ⁶²⁵TGD loop were inhibitory in the presence of Mg²⁺, placing the phosphate chain of the photolabel at some distance from the other P domain residues under these conditions. Mutations of the hinge residues, Asn³⁵⁹ and Asp⁶⁰¹, had quite large effects on ATP binding in the presence of Mg²⁺, and only smaller effects on binding in the absence of Mg²⁺.

On the basis of the atomic structure of phosphoserine phosphatase in a transition state-like configuration (Ref. 5 and Fig. 1B), domain P of Ca²⁺-ATPase must undergo substantial rearrangements from the known E1(Ca₂) and E2(TG) atomic structures for phosphorylation to occur. Hence, the distances between the backbone carbonyl of Thr³⁵³ and the carboxyls of Asp⁷⁰³ and Asp⁷⁰⁷, and that between the amide of Gly⁶²⁶ and the amino group of Lys⁶⁸⁴ need to shorten by 4–5 Å, if they are to approach the transition state configuration of the phosphatase (see green lines in Fig. 1A). A further necessary adjustment involves Asp³⁵¹. The carboxyl is turned "downward," away from the expected approach of the -phosphoryl group of ATP, and needs to rotate upward for an in line attack and also for ligation of Mg²⁺, as in the transition state of the phosphatase (Fig. 1, compare A and B). Part of the difficulty in gathering these residues closer is the strong negative electrostatic potential; three aspartates, Asp³⁵¹, Asp⁷⁰³, and Asp⁷⁰⁷, surround one lysine, Lys⁶⁸⁴. On the other side of the phosphorylation well are Lys³⁵², Asp⁶⁰¹, and Asp⁶²⁷. The phosphate chain of ATP, if placed in the vicinity, would obviously exacerbate the electrostatic imbalance. Mutagenesis of these conserved domain P residues should be interpreted against the background of strong electrostatic effects within a network of interactions. Eliminating the charge on one residue is bound to influence neighbors and possibly beyond. Lys⁶⁸⁴ is positioned midway between Asp³⁵¹ and Asp⁷⁰⁷ in the crystal structures, and eliminating the charge on the lysine can be expected to change the pK_a of the carboxyls so that one (or both) could become protonated and lose the charge. Eliminating the charge on one of the aspartates may do the same to the lysine. In line with this, there are similarities between the effects of replacement of either Lys⁶⁸⁴ or Asp⁷⁰⁷.

The present data together with those obtained previously for mutations of the phosphorylation loop, ³⁵¹DKTGTLT (6, 7), suggest that the main gathering movements of conserved domain P residues is a cooperative effect of ATP and Mg²⁺ binding, and that Ca²⁺ binding at the transport sites provides further structural adjustments to reach the tight ligation needed in the transition state. Without ATP, Mg²⁺ and Ca²⁺ do not achieve the structural changes alluded to above in domain P, as the crystal structures show (1, 2). ATP alone seems to mainly experience electrostatic repulsion, as expected with the existence of a strong negative electrostatic potential at the phosphorylation site. The electrostatic effects on ATP binding in the absence of Mg²⁺ and Ca²⁺ are quite large, with more than 10-fold increases in affinity for mutations Asp³⁵¹ Asn (6) and Asp⁷⁰⁷ Asn (Table I). The effect with the latter mutation was even larger for the TNP nucleotide (33-fold change). Even mutation Lys⁶⁸⁴ Met increased the affinity for both ATP and TNP nucleotide, and this may be the result of linked partial protonation of neighboring Asp³⁵¹ and Asp⁷⁰⁷ as mentioned above. Electrostatic effects seem to extend to Asp⁶²⁷, a bit out of the phosphorylation well itself, where a 5-fold increase of affinity was found upon elimination of the charge. Thus, in the absence of Mg²⁺, the phosphate chain of both TNP-8N₃-ATP and ATP seems to come into proximity of these aspartates, but repulsive effects prevent ligation. The phosphates likely interact at Lys⁴⁹² and Arg⁴⁸⁹ in domain N, according to our previous mutational analysis (20), and at neighboring Arg⁶⁷⁸ stretching from domain P (23).

Mg²⁺ creates favorable interactions for ATP binding in domain P, because the mutations here were consistently detrimental in the presence of Mg²⁺ and absence of Ca²⁺, apart from Asp⁷⁰⁷ Cys (the cysteine may be ionized) and charge conserving mutations of the other two aspartates to glutamate. Replacement of Lys⁶⁸⁴ with methionine was severely detrimental for MgATP binding in the absence of Ca²⁺ (22-fold reduction of the affinity, Table I), compatible with Lys⁶⁸⁴ bonding the -phosphoryl group in the wild-type Ca²⁺-ATPase as in phosphoserine phosphatase (cf. Fig. 1B). Of particular interest is also the inhibition of MgATP binding caused by most mutations of Asp⁷⁰³, which is considered a putative Mg²⁺ coordinating residue on the basis of the phosphoserine phosphatase model (Ref. 5 and Fig. 1B). In the absence of Ca²⁺, MgATP affinity was reduced 5–8-fold by replacement of Asp⁷⁰³ with alanine, cysteine, asparagine, or serine, whereas ATP affinity was not reduced. Because the replacement with asparagine produced a change in MgATP affinity similar to that seen with alanine, it appears that the negatively charged carboxylate group of the Asp⁷⁰³ side chain is required for proper MgATP ligation in the absence of Ca²⁺. Replacement of Asp⁷⁰³ with glutamate resulted in little inhibition of MgATP binding, indicating little or no steric constraint around this residue in the E1 form, which is compatible with the relative isolation of this side chain in the atomic structures and its exposure to the aqueous medium in the homologous phosphoserine phosphatase structures. The phosphorylation of the E2 form with inorganic phosphate was very sensitive to mutations of Asp⁷⁰³ (Fig. 5), and even the replacement with glutamate induced a large (14-fold) reduction of the apparent affinity for Mg²⁺ as activator of phosphorylation by P_i, consistent with a critical role of Asp⁷⁰³ in Mg²⁺ ligation in the E2P phosphoenzyme intermediate. For Asp⁷⁰⁷ Asn/Ser and Asp⁶²⁷ Asn, 3–8-fold reductions of MgATP affinity, relative to wild type, were observed in the absence of Ca²⁺. In these mutants, as well as most of the Asp⁷⁰³ mutants, the increased affinity for ATP and the reduced affinity for MgATP translate into little or no difference between the affinities for ATP and MgATP. Hence, the binding data do not permit distinction between the roles of Asp⁷⁰³, Asp⁷⁰⁷, and Asp⁶²⁷ in MgATP binding. In all three cases, perturbation of favorable interaction may be masked by the simultaneous reduction of repulsive forces resulting from the elimination of the negative charge. In phosphoserine phosphatase, Asp⁶²⁷ is replaced by glycine (cf. Fig. 1B). A role for the aspartate in P-type ATPases might be to help fix domains N and P together, and at the same time this residue could interact with the -phosphate of bound ATP (cf. Fig. 1A). In phosphoserine phosphatase, the aspartate corresponding to Asp⁷⁰⁷ is within hydrogen bonding distance to a water molecule coordinating Mg²⁺ and, moreover, appears to stabilize the backbone in a conformation allowing the equivalent of Asp⁷⁰³ to coordinate Mg²⁺ (cf. Fig. 1B). A similar scenario in Ca²⁺-ATPase would be consistent with the present finding of rather similar reductions in MgATP binding affinity for Asp⁷⁰⁷ and Asp⁷⁰³ mutations removing the negative charge.

In the case of TNP-8N₃-MgATP, mutations Asp⁷⁰⁷ Asn/Ser enhanced binding, which not only shows that electrostatic repulsive effects need to be considered, but also demonstrates that Asp⁷⁰⁷ does not contribute to coordination of the Mg²⁺ ion of TNP-8N₃-MgATP. Mutation Lys⁶⁸⁴ Met also enhanced TNP-8N₃-MgATP binding, in line with the hypothesis that the -phosphoryl of this nucleotide does not reach this far into the well. Not even mutation of Asp⁷⁰³ perturbed the binding of TNP-8N₃-MgATP. It appears that, in contrast to the situation with ATP, Mg²⁺ does not anchor the phosphate chain of TNP-8N₃-ATP, thus confirming the notion that these two nucleotides are bound differently, although at overlapping sites (20).

CrATP, on the other hand, appears to bind much like MgATP, because the P domain mutations found to inhibit MgATP binding also eliminated or slowed formation of the tight enzyme complex with CrATP. Presumably, the chromium(III) ion takes a position similar to that of the Mg²⁺ of MgATP, with the phosphates reaching into the phosphorylation well. This would also be consistent with our previous finding that mutation of Asp³⁵¹ eliminates tight CrATP binding (6). It should, however, be emphasized that the -phosphoryl group of CrATP is not transferred to Asp³⁵¹ in the wild-type enzyme, even though the enzyme-CrATP complex, just as the E1P phosphoenzyme, has Ca²⁺ bound in an occluded state at the transport sites (21, 22), thus indicating that with chromium(III) replacing Mg²⁺ the transition state for phosphoryl transfer cannot be fully reached. The tight binding of CrATP likely results from exchange of one or more of the intramolecular ligands associated with the chromium(III) ion (four water molecules and the - and -phosphates) with oxygen or nitrogen functionalities of the protein, derived from Asp³⁵¹ (6) and, probably, from some of the P domain residues studied here.

The Ca²⁺-activated Transition State—The negligible phosphorylation activity of several of the mutants provided an opportunity to investigate the energetics in the pseudo-transition state reached in the presence of both Mg²⁺ and Ca²⁺. Ca²⁺ binding produced large and contrasting changes in MgATP affinity for some of the mutants, compatible with a closer approach of the residues around the phosphates and Mg²⁺.

Previously, we found that mutant Asp³⁵¹ Asn, which is also inactive in phosphorylation, exhibits a very high affinity for ATP in the presence of Mg²⁺ (K_D = 6.5 nM), and the addition of Ca²⁺ lowers the affinity (K_D = 28 nM) (6). For Asp³⁵¹ Ala, the reverse is true, the affinity is not so high with Mg²⁺ alone (K_D = 25 nM) and increases with both Ca²⁺ and Mg²⁺ present (K_D = 1.1 nM) (6). These changes are instructive for what is happening at the phosphorylation site in the transition state and for interpretation of the present results. Evidently, the amino group of the asparagine is less perturbing compared with alanine in the ground state, but the larger side chain is more damaging in the tight activated state. Thus, rather minor differences in side chain size can have quite large effects on MgATP binding in a compacted transition state. Even though the Asp³⁵¹ side chain receives the phosphoryl group from ATP, this residue does not appear to be required for the structural rearrangement and favorable interaction with MgATP involved in stabilization of the transition state for phosphoryl transfer, because a large increase of nucleotide affinity occurred upon Ca²⁺ binding in the Asp³⁵¹ Ala mutant. It seems likely that a similar increase of favorable interaction with MgATP occurs in the transition state in wild-type enzyme, although this cannot be measured because of the activation of phosphoryl transfer in the wild type. By contrast, Ca²⁺ is expected to lower the MgATP binding affinity in mutants with alterations to residues participating in favorable interaction with the nucleotide in the transition state.

In the case of mutation Lys⁶⁸⁴ Met, the affinity for MgATP increased 14-fold on the addition of Ca²⁺, which is rather similar to the change mentioned above for Asp³⁵¹ Ala (23-fold). Thus, even though the affinities for MgATP are very different in the Asp³⁵¹ and Lys⁶⁸⁴ mutants, both express a large affinity increase upon Ca²⁺ binding, and neither of these residues, therefore, seems required for the favorable interaction with nucleotide induced by Ca²⁺ in the transition state. With neighboring Asp⁷⁰⁷, the serine and alanine mutants showed only 3- and 5-fold increases of the affinity for MgATP, respectively, on the addition of Ca²⁺, indicating partial tightening, but possibly interference with and, thus, involvement of Asp⁷⁰⁷ in transition state stabilization. The involvement of Asp⁷⁰⁷ is also indicated by the 7-fold reduction of MgATP affinity seen for Asp⁷⁰⁷ Cys upon Ca²⁺ binding.

A spectacular 12-fold decrease of MgATP affinity was found upon addition of Ca²⁺ in Asp⁷⁰³ Ala, strongly suggesting that the presence of Asp⁷⁰³ is required for the Ca²⁺-induced increase of MgATP affinity described above. Hence, Asp⁷⁰³ appears to play a more critical role in MgATP binding in the transition state for phosphoryl transfer reached in the presence of Ca²⁺ than in the enzyme complex with MgATP formed in the absence of Ca²⁺. In the wild type, Ca²⁺ binding at the transport sites may facilitate very tight Mg²⁺ coordination by Asp⁷⁰³ in the presence of nucleotide, thereby activating the phosphoryl transfer. Interestingly, the replacement of Asp⁷⁰³ with asparagine only moderately slowed phosphorylation, in contrast to the almost complete block of phosphorylation seen for Asp⁷⁰³ Ala/Ser/Cys, indicating that a carboxamide group can partially substitute for carboxylate in the Ca²⁺-activated state. It thus seems likely that in the transition state the side chain of Asp⁷⁰³ only contributes one oxygen atom directly to Mg²⁺ coordination, as in phosphoserine phosphatase (cf. Fig. 1B).

The consequences of mutation Gly⁶²⁶ Ala are especially instructive in evaluating the effects of Mg²⁺ and Ca²⁺ on phosphate anchoring. The interpretation of the results of replacement of this residue is not complicated by electrostatic effects, and the mutation hardly affected nucleotide binding in the absence of Mg²⁺, indicating weak or no interaction with the phosphate chain, again suggesting that the latter is not in the phosphorylation well under these conditions. Addition of Mg²⁺ caused the mutation to become strongly perturbing (28-fold decrease in affinity for ATP), and even more significant is the finding that Ca²⁺ binding at the transport sites virtually prevented ATP binding to the Gly⁶²⁶ Ala mutant. Thus, in agreement with the model in Fig. 1A and the phosphatase transition state structure in Fig. 1B, our data are consistent with the phosphate chain interacting with the amide of Gly⁶²⁶ in the presence of Mg²⁺ and Ca²⁺. Evidently, Ca²⁺ binding in the presence of nucleotide induces quite a dramatic change in the binding site in the vicinity of Gly⁶²⁶ that is not apparent by comparison of the E1(Ca₂) and E2(TG) crystal structures. In this connection it is interesting to note that activation of CheY, a response regulator with active site configuration similar to Ca²⁺-ATPase, likewise involves a significant conformational change here with the residue equivalent to Thr⁶²⁵ repositioned to ligate the phosphate or analogue (24).

Roles of P Domain Residues in E1P E2P Transition—It is striking that mutation of any of the charged residues in domain P that were found important for MgATP or CrATP binding also affected the E1P E2P transition (stabilized E1P). This finding suggests that the unfavorable electrostatic environment in domain P alluded to above plays a role in the dissociation of ADP and destabilization of E1P conformation causing the relaxation to E2P (release of domain N and rebinding in a different orientation). In line with this hypothesis, the acceleration of the E1P E2P transition induced by secondary binding of ATP with low affinity to the phosphoenzyme ("modulatory effect") may be the result of unfavorable electrostatic interactions associated with packing an extra phosphate group into the phosphorylated catalytic site in E1P after the dissociation of ADP (20). The mutational effects on the E1P E2P transition could also be related to changes in Mg²⁺ coordination, as Mg²⁺ binding is required for a normal rate of the E1P E2P transition (25). Again the difference between the Asp⁷⁰³ Ala and Asp⁷⁰³ Asn mutants is noteworthy, the latter showing a 20-fold higher rate of E1P E2P transition than the former, thus indicating that in the E1P phosphoenzyme, as in the transition state for phosphoryl transfer, a carboxamide group can partially substitute for the carboxylate.

The Hinge Residues—Asp⁶⁰¹ and Asn³⁵⁹ are located in the hinge segments and are at the fulcrum of N-P interdomain movements. Mutations here should impact on partial reactions of the catalytic cycle in which such movements are critical and rate-limiting. We found that the binding of either TNP-8N₃-ATP or ATP was only affected with Mg²⁺ present, suggesting that Mg²⁺ is required for fastening of domain N over domain P. This result fits very well with Mg²⁺ inducing favorable interactions of the phosphate chain with critical domain P residues. Both mutations Asn³⁵⁹ Ala and Asp⁶⁰¹ Asn enhanced MgATP binding, indicating that some impediment to binding exists in the wild type in this hinge region. The strain could be important for release of domain N and ADP dissociation following phosphorylation, which would be consistent with the reduced rate of the E1P E2P transition in these mutants. Mutation Asp⁶⁰¹ Glu was quite perturbing for both TNP-8N₃-ATP and ATP binding in the presence of Mg²⁺, as well as for the E1P E2P transition, showing that steric factors become important at the fulcrum of the hinge when the domains fasten together in connection with nucleotide binding and in the ensuing conformational rearrangement. Despite their opposite effects on MgATP affinity, both mutations of Asp⁶⁰¹ slowed the rate of phosphorylation from ATP, and mutation Asn³⁵⁹ Ala had no effect on phosphorylation, despite the increased affinity for MgATP in this mutant, thus adding to the evidence that further Ca²⁺-induced rearrangement of domains N and P beyond that involved in the anchoring of the phosphate chain is required to reach the transition state in the subsequent phosphorylation step.

Concluding Comments—The present results have pinpointed the important roles of catalytic Mg²⁺ and conserved P domain and hinge residues in anchoring the phosphate chain of ATP. Ca²⁺ binding provokes marked changes in the interaction of Gly⁶²⁶ and Asp⁷⁰³ with MgATP, and structural rearrangements in domain P seem part of the cue for mediating entry into the transition state for phosphoryl transfer from ATP. The changes in domain P appear to be a cooperative effect of ATP, Mg²⁺, and Ca²⁺ binding. Coupling of binding of the transported ions to gaining the transition state at the active site is fundamental to P-type ATPases as transport proteins and contributes to distinguishing them from the catalytically related, soluble phosphotransferases, like phosphoserine phosphatase, where the active site rearrangements in the equivalent domain are much smaller and not linked to cation binding at distant transport sites. The present observations, showing that in the Ca²⁺-ATPase the close interaction of Gly⁶²⁶ and Asp⁷⁰³ with MgATP depends on Ca²⁺ binding at the transport sites, demonstrate an important difference from phosphoserine phosphatase, which may be fundamental to linking ion binding to phosphoryl transfer in the transport ATPases.

	FOOTNOTES

 
^* This work was supported in part by grants from the National Research Foundation of South Africa and University of Cape Town Research Committee (to D. B. M.), the Danish Medical Research Council, the Novo Nordisk Foundation, Denmark, the Lundbeck Foundation (Denmark), and the Research Foundation of Aarhus University (to J. P. A.), and by Canadian Institutes of Health Research Grant MT 12545 (to D. H. M.). 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.

To whom correspondence may be addressed: Chemical Pathology, Dept. of Clinical Laboratory Sciences, Faculty of Health Sciences, University of Cape Town, Cape Town 7925, South Africa. Fax: 27-21-4488150; E-mail: davidmci{at}chempath.uct.ac.za. ** To whom correspondence may be addressed: Dept. of Physiology, University of Aarhus, Ole Worms Allé 160, DK-8000 Aarhus C, Denmark. Fax: 45-86129065; E-mail: jpa{at}fi.au.dk.

¹ The abbreviations used are: Ca²⁺-ATPase, the sarco(endo)plasmic reticulum Ca²⁺-transporting adenosine triphosphatase (EC 3.6.1.38 [EC] ); CrATP, ,-bidentate chromium(III) complex of ATP; E1, enzyme form with cytoplasmically facing high affinity Ca²⁺ binding sites; E2, enzyme form with low affinity for Ca²⁺; E2(TG), E2 enzyme with bound thapsigargin; E1P, ADP-sensitive phosphoenzyme intermediate containing Ca²⁺ in the occluded state; E2P, ADP-insensitive phosphoenzyme intermediate with luminally facing low affinity Ca²⁺ binding sites; EPPS, N-2-hydroxyethylpiperazine-N'-3-propanesulfonic acid; K_0.5, ligand concentration giving half-maximum effect; MES, 2-[N-morpholino]eth-anesulfonic acid; MOPS, 3-[N-morpholino]propanesulfonic acid; TMAH, tetramethylammonium hydroxide; TNP-8N₃-ATP, 2',3'-O-(2,4,6-trinitrophenyl)-8-azidoadenosine 5'-triphosphate.

	ACKNOWLEDGMENTS

 
We thank Irene Mardarowicz and Joy Norman (University of Cape Town, Cape Town, South Africa) and Lene Jacobsen and Karin Kracht (University of Aarhus, Aarhus, Denmark) for expert technical assistance. We are indebted to Chikashi Toyoshima (University of Tokyo, Tokyo, Japan) for generously sharing information on unpublished work.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 

1. Toyoshima, C., Nakasako, M., Nomura, H., and Ogawa, H. (2000) Nature 405, 647–655[CrossRef][Medline] [Order article via Infotrieve]
2. Toyoshima, C., and Nomura, H. (2002) Nature 418, 605–611[CrossRef][Medline] [Order article via Infotrieve]
3. McIntosh, D. B. (1998) Adv. Mol. Cell. Biol. 23A, 33–99
4. Aravind, L., Galperin, M. Y., and Koonin, E. V. (1998) Trends Biochem. Sci. 23, 127–129[CrossRef][Medline] [Order article via Infotrieve]
5. Wang, W., Cho, H. S., Kim, R., Jancarik, J., Yokota, H., Nguyen, H. H., Grigoriev, I. V., Wemmer, D. E., and Kim, S.-H. (2002) J. Mol. Biol. 319, 421–431[CrossRef][Medline] [Order article via Infotrieve]
6. McIntosh, D. B., Woolley, D. G., MacLennan, D. H., Vilsen, B., and Andersen, J. P. (1999) J. Biol. Chem. 274, 25227–25236[Abstract/Free Full Text]
7. Clausen, J. D., McIntosh, D. B., Woolley, D. G., and Andersen, J. P. (2001) J. Biol. Chem. 276, 35741–35750[Abstract/Free Full Text]
8. Clarke, D. M., Loo, T. W., and MacLennan, D. H. (1990) J. Biol. Chem. 265, 22223–22227[Abstract/Free Full Text]
9. Vilsen, B., Andersen, J. P., and MacLennan, D. H. (1991) J. Biol. Chem. 266, 16157–16164[Abstract/Free Full Text]
10. Maruyama, K., Clarke, D. M., Fujii, J., Inesi, G., Loo, T. W., and MacLennan, D. H. (1989) J. Biol. Chem. 264, 13038–13042[Abstract/Free Full Text]
11. Kaufman, R. J., Davies, M. V., Pathak, V. K., and Hershey, J. W. (1989) Mol. Cell. Biol. 9, 946–958[Abstract/Free Full Text]
12. Chen, C., and Okayama, H. (1987) Mol. Cell. Biol. 7, 2745–2752[Abstract/Free Full Text]
13. Maruyama, K.