News Physiol Sci 18: 164-168, 2003;
doi:10.1152/nips.01441.2003
1548-9213/03 $5.00
News in Physiological Sciences, Vol. 18, No. 4, 164-168,
August 2003
© 2003 Int. Union Physiol. Sci./Am. Physiol. Soc.
Found: Na+ and K+ Binding Sites of the Sodium Pump
R. F. Rakowski1 and
S. Sagar2
1 Department of Biological Sciences, Ohio University, Athens, Ohio 45701; and
2 Department of Physiology and Biophysics, Finch University of Health Sciences, North Chicago, Illinois 60064
 |
Abstract
|
|---|
Homology modeling and valence mapping have been used to predict the location and structure of Na+ and K+ binding sites in the Na+-K+-ATPase on the basis of the known atomic resolution structure of SERCA. Additional sites are predicted that may be associated with intracellular access and extracellular egress pathways for Na+. The model predictions are in excellent agreement with previous structure-function and electrical studies.
|
Introduction
|
|---|
It has become clear that "despite the practically unlimited number of protein sequences, the number of . . . protein folds seems . . . to be relatively small with probably no more than 10,000 folds in existence" (2). This makes evolutionary sense. There is little doubt that protein families are monophyletic (have a common ancestor) whose folding pattern has been highly conserved, because the fold is essential for their biological function. The similarity within a protein family goes beyond preservation of the backbone structure and includes the structures essential for the particular biochemical reaction step, such as ATP binding or hydrolysis, that it catalyzes. It is not surprising that large proteins are assemblies of domains having distinct functional roles in the overall sequence of biochemical reaction steps catalyzed by the protein. An excellent example of an assembly of functional domains is shown in Fig. 1
, which is a representation of the structure of the Ca2+ pump of the sarcoplasmic reticulum (SERCA) determined at atomic resolution by X-ray crystallography by Toyoshima et al. (12). Domains for nucleotide binding (N), phosphorylation (P), and activation (A) are indicated. These intracellular domains regulate the accessibility of two Ca2+ binding sites within the membrane domain (gray spheres).
A single structure, even at atomic resolution, is helpful but does not provide all of the information required to obtain an atomic-level understanding of the mechanism of ion transport. Investigations of active transport of Ca2+ are now focused on the structural changes that occur during the cycle of ATP hydrolysis and ion translocation. A major step toward a detailed understanding of the mechanism of active transport of Ca2+ was made recently with the publication of a second crystal structure of SERCA at 3.1-Å resolution with no bound Ca2+ ions. The two structures of SERCA are likely to represent the postulated E1 (Ca2+-bound) and E2 (Ca2+-free) forms of the enzyme. These are the stable conformations of an alternating access mechanism of ion transport in which Ca2+ is bound at the myoplasmic face of the enzyme and released into the sarcoplasmic reticulum.
|
Homology modeling of the Na+ pump
|
|---|
Homology modeling can be used to infer detailed structural and mechanistic information about a closely related P-type ATPase, the electrogenic Na+ pump (Na+-K+-ATPase). Sweadner and Donnet (10) pointed out the close structural similarity of the Na+-K+-ATPase and SERCA. They used the gapped BLAST algorithm of the Cn3D utility program and threaded the Na+ pump sequence onto the known structure of SERCA to examine the topological location of specific residues. They concluded that the Na+ pump and SERCA share the same fold. The next logical step is to predict the detailed three-dimensional structure of the Na+-K+-ATPase. We have used the homology modeling and energy minimization capabilities of the utility program Swiss Protein Database Viewer (http://us.expasy.org/spdbv) to predict the structure of the 
1-subunit of Na+-K+-ATPase from Xenopus neural plate [Geninfo Identifier (GI): 1228150] based on its gapped BLAST alignment with SERCA using the DDV utility program of Cn3D (the same alignment program used by Sweadner and Donnet).
A valence-mapping program (VALE) was used to find the putative Na+ binding sites in the predicted Na+ pump structure (courtesy of Drs. Enrico Di Cera and Thierry Rose; Washington University, St. Louis, MO). The electrostatic valence is based on principles formulated by Pauling more than 60 years ago. An empirical relationship for bond strength, or valence (
) contributed by a given coordinating oxygen is given by = (R/R1)-N, where R is the distance from the oxygen to the cation, R1 is the value of R giving = 1, and N is an empirical constant. Values of R1 and N have been determined for various cation-oxygen pairs by analysis of crystals of metal oxides. The overall site valence is calculated by summing over all of the ligating oxygens. The VALE program calculated the theoretical valence on a three-dimensional grid of points spaced at 0.1-Å intervals superimposed on the predicted pump structure (3, 4). Five putative Na+ binding sites were identified in the transmembrane domain or at its margin that had a calculated site valence of >0.8. These five putative Na+ binding sites are shown in Fig. 2. Two loci having site valences of 1.21 and 0.92 were found that closely correspond to the two Ca2+ binding sites of SERCA. A third site was located between helical segments 6 and 7 and was found to closely correspond to the intracellular access site postulated by Shainskaya et al. (9). Two sites that may play a role in the release of Na+ to the extracellular medium were also identified.
View larger version (47K):
[in this window]
[in a new window]
|
FIGURE 2. Putative Na+ access and egress sites in the transmembrane domain of the Xenopus Na+-K+-ATPase. In addition to putative Na+ binding sites 1 and 2, three additional loci were identified in the transmembrane domain of the homology model of the Xenopus Na+-K+-ATPase that may play a role in Na+ translocation. Water molecules were not included in the analysis. A putative binding site that may play a role in Na+ binding at the intracellular membrane border was predicted to have a site valence of 0.84. The putative access site is located at model coordinates x = 52.4, y = 27.0, z = 19.0 and makes contact with Asn771-OD1, Glu825-O, and Gln826-O. The model coordinate axes were chosen to have the same origin and orientation as that in the unit cell of SERCA (Swiss Protein Database accession no. 1EUL). Two sites at the extracellular face of the predicted structure are postulated to play a role in the release of Na+. The innermost of these two putative egress sites has a site valence of 0.90, is located at model coordinates x = 71.6, y = 15.2, z = -13.0, and makes contact with Asn129-OD1, Met974-O, Asp975-O, and Val976-O. The outermost site has a site valence of 0.94, is located at model coordinates x = 60.8, y = -0.8, z = -26.8, and makes contact with Trp894-O, Thr895-O, Asn896-O, and Asp897-OD2.
|
|
|
Predicted location of the three Na+ and two K+ binding sites in the Na+ pump
|
|---|
Independent of our work, Ogawa and Toyoshima (5) achieved the remarkable success of predicting the location and detailed coordination geometry of all three Na+ and both K+ binding sites in the human Na+-K+-ATPase. Two additional pieces of information allowed them to achieve this goal. By introducing water molecules into the cavities of the protein by using the WHATIF utility program, they predicted a third Na+ binding site in close proximity to Na+ sites 1 and 2. Two water molecules are required to complete the coordination shell of site 3. These increased the calculated site valence to 0.96. The availability of a three-dimensional structure for SERCA without bound Ca2+ (11) allowed prediction of the location of the two K+ binding sites. The oxygen atoms that are predicted to coordinate Na+ in the human and Xenopus Na+-K+-ATPase are listed in Table 1. It is remarkable how well these predictions agree with previous site-directed mutagenesis studies of ion binding by the Na+ pump. Ogawa and Toyoshima (5) provide an excellent discussion of this agreement with previous work. There is little doubt that their predictions are correct in their essential features. The details of the molecular architecture, however, remain to be verified by other methods.
|
Qualitative agreement of the predicted structure with electrical measurements
|
|---|
In addition to being consistent with various results from studies of the effects of mutations on ion binding, the predicted location of the Na+ and K+ binding sites is consistent with electrical measurements of the dielectric distances for ion binding and release. The dielectric distance corresponds to the fraction of the membrane potential that is sensed by the ion at its binding site. This provides a rough estimate of the fractional physical distance at which the binding site is located. Various estimates of the dielectric distance for release of Na+ to the extracellular medium give a value of ~0.65 (6). The dielectric distance for extracellular K+ binding is smaller and has a value of ~0.27 (7). This is in general agreement with the structural predictions of Ogawa and Toyoshima that place the Na+ binding sites one helical turn deeper within the pump molecule than those for K+ binding. The putative intracellular Na+ access site proposed based on the Xenopus structure at the internal junction of M6 and M7 is consistent with a relatively shallow depth for intracellular Na+ binding (dielectric coefficient ~ 0.25) (1). Much more detailed calculations of the profile of the electrical field within the membrane domain will be required to make a more quantitative comparison of the structural predictions and electrical results.
Ogawa and Toyoshima’s homology model represents a milestone in our progress toward understanding the molecular mechanism of Na+ pump operation. Its value is the ability to make detailed atomic-level hypotheses about the pump mechanism directly from the postulated structure. These predictions can be tested by site-directed mutagenesis and augmented by other modeling techniques such as molecular dynamic simulation of partial reaction steps. A comparison of the structural prediction of Ogawa and Toyoshima with ours serves to illustrate both how robust these modeling results are and how they can be used to make further detailed predictions of molecular function.
|
Robustness of homology model predictions
|
|---|
A critical step in homology modeling is the alignment of residues in the unknown and known structures. This can be done with gapped BLAST utility programs, with or without subsequent manual adjustments. Ogawa and Toyoshima chose to manually adjust their initial alignment of the human Na+-K+-ATPase 1-subunit obtained from the program DIALIGN using information derived from comparison of known Na+ pump sequences in the transmembrane helical segments that varied principally in the lipid-facing part of the packed helical structure. Their final alignment, however, was identical to that used by Sweadner and Donnet (10), who did not attempt this manual adjustment. The success of alignment algorithms depends strongly on the percent identity of the two sequences to be aligned. Although the overall sequence identity between SERCA and the Na+ pump is ~27%, the identity is much higher in highly conserved regions of the molecule and in the -helical transmembrane segments where the ion binding sites are located. Figure 3 shows a comparison between the alignment we used for SERCA and the 1-subunit of the Xenopus Na+ pump and that used by Ogawa and Toyoshima to align the human 1-subunit sequence of the Na+ pump with SERCA.
Note that we have chosen a numbering convention that conveniently makes corresponding residues in the Xenopus and human sequence identical in the regions of interest. In M4, the residues that bind Ca2+ in four positions of SERCA (indicated by #) correspond precisely to four positions that bind Na+ in the Xenopus (x) and human (*) Na+ pump. Note the change, however, from Ile309 in SERCA to Val332 in the Na+ pump sequences.
Although our alignment of Xenopus with SERCA predicted a gap of four residues opposite fkia, and that of Ogawa and Toyoshima does not, the subsequent alignment starting from IGII . . . is identical and the gap caused no difficulty in predicting the coordinating oxygens that straddle the Pro333 kink in the M4 helix. Note also that, although the automated alignment program has difficulties in M7 (a segment that contains no groups that are involved in ion binding), subsequent regions of high sequence identity with SERCA bring the Xenopus and human sequences back in register.
|
Effect of including water in the homology model
|
|---|
Although Ogawa and Toyoshima chose to include water molecules (somewhat arbitrarily located) in cavities within their model, the absence of water molecules in our homology model has some interesting consequences. For example, site 1 in the human Na+ pump is predicted to involve coordination by one water molecule. In the Xenopus model, however, this is replaced by coordination with the main chain oxygen of Ala330 (see Table 1
). Since Ala330 also was found to coordinate the Na+ ion at site 2, we suggest that it may have a dual role in the sequence of transitions that are involved in the occupancy of Na+ sites as water and ions pass this locus during the cycle of ion translocation. Other differences between the predictions obtained in the presence and absence of water (Gln930/Glu786 at site 1, Thr814/Asp815 at site 2, and the participation of Asp811 at site 2) may also have functional significance.
The coordination geometry of the two Na+ sites predicted for Xenopus is shown in Fig. 4. Na+ site 1 is shown in Fig. 4A. Ala330 takes the place of a water molecule in site 1 of the Ogawa and Toyoshima model. The distance from the center of the Na+ ion to the center of the two coordinating oxygens of Glu786 is rather long, but Glu930, proposed by Ogawa and Toyoshima to coordinate Na+ at this site, is even farther away in our predicted structure (12.5 Å; not shown). Asp815 is clearly an important residue for ion binding. Both side chain oxygens are predicted to coordinate Na+ at site 1 of the Xenopus structure (but a side chain oxygen of Tyr814 is proposed for the human site 1). Asp815 is predicted to coordinate Na+ at site 2 (Fig. 4B) in both the human and Xenopus Na+ pump.
|
Intracellular access and extracellular egress sites
|
|---|
Shainskaya et al. (9) concluded that charged residues in the intracellular loop between M6 and M7 "contribute to the initial recognition of transported cations which then move into the occlusion sites within the membrane." The putative intracellular access site predicted for Xenopus includes Asp770 located near the intracellular border of M5 and Tyr824 and Glu825 in the intracellular loop between M6 and M7, in good agreement with the conclusion of Shainskaya et al. that this is the site to which Na+ ions bind before moving to sites deeper in the transmembrane domain, where they become occluded.
Schneider and Sheiner-Bobis (8) pointed out the homology between the sequence that occurs in the extracellular loop between M7 and M8 and the P-loop of Na+ channels. Mutations introduced in the motif 891DDRW894 (human numbering) resulted in decreased ouabain binding affinity. They postulated that this region is involved in the "acceptance and/or release of Na+ ions coming from the cytosolic regions of the protein." Consistent with this hypothesis, we find that the outer egress site in Xenopus involves coordination by Trp894, Thr895, Asn896, and Asp897 with a site valence of 0.94.
Ogawa and Toyoshima also describe a site (S) near Glu847 that is isolated from the other Na+ binding sites but that is near the boundary of the hydrophobic core of the lipid bilayer. Although they dismiss this site as being unlikely to represent the third Na+ binding site, it may play a transient role in ion permeation. Since its site valence is >0.9, its precise role deserves further investigation.
|
References
|
|---|
- Heyse S, Wuddel I, Apell H-J, and Sturmer W. Partial reactions of the Na,K-ATPase: determination of rate constants. J Gen Physiol 104: 197–240, 1994.[Abstract/Free Full Text]
- Koonin EV, Wolf Yi, and Karev GP. The structure of the protein universe and genome evolution. Nature 420: 218–223, 2002.[Medline]
- Nayal M and Di Cera E. Predicting Ca2+ binding sites in proteins. Proc Natl Acad Sci USA 91: 817–821, 1994.[Abstract/Free Full Text]
- Nayal M and Di Cera E. Valence screening of water in protein crystals reveals potential Na+ binding sites. J Mol Biol 256: 228–234, 1996.[Web of Science][Medline]
- Ogawa H and Toyoshima C. Homology modeling of the cation binding sites of Na+ K+-ATPase. Proc Natl Acad Sci USA 99: 15977–15982, 2002.[Abstract/Free Full Text]
- Rakowski RF, Gadsby DC, and De Weer P. Voltage dependence of the Na/K pump. J Membr Biol 155: 105–112, 1997.[Web of Science][Medline]
- Rakowski RF, Vasilets LA, LaTona J, and Schwarz W. A negative slope in the current-voltage relationship of the Na+/K+ pump in Xenopus oocytes produced by reduction of external [K+]. J Membr Biol 121: 177–187, 1991.[Web of Science][Medline]
- Schneider H and Scheiner-Bobis G. Involvement of the M7/M8 extracellular loop of the sodium pump subunit in ion transport. J Biol Chem 272: 16158–16165, 1997.[Abstract/Free Full Text]
- Shainskaya A, Schneeberger A, Apell H-J, and Karlish SJD. Entrance port for Na+ and K+ ions on Na+, K+-ATPase in the cytoplasmic loop between trans-membrane segments M6 and M7 of the subunit. J Biol Chem 275: 2019–2028, 2000.[Abstract/Free Full Text]
- Sweadner KJ and Donnet C. Structural similarities of Na,K-ATPase and SERCA, the Ca2+-ATPase of the sarcoplasmic reticulum. Biochem J 356: 685–704, 2001.[Web of Science][Medline]
- Toyoshima C and Nomura H. Structural changes in the calcium pump accompanying the dissociation of calcium. Nature 418: 605–611, 2002.[Medline]
- Toyoshima C, Nakasako M, Nomura H, and Ogawa H. Crystal structure of the calcium pump of sarcoplasmic reticulum at 2.6 Å resolution. Nature 405: 647–655, 2000.[Medline]
This article has been cited by other articles:
|
|
|
|
 
D. C Holley and M. P Kavanaugh
Interactions of alkali cations with glutamate transporters
Phil Trans R Soc B,
January 27, 2009;
364(1514):
155 - 161.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. C Gadsby, A. Takeuchi, P. Artigas, and N. Reyes
Peering into an ATPase ion pump with single-channel recordings
Phil Trans R Soc B,
January 27, 2009;
364(1514):
229 - 238.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
O. Capendeguy, J. Iwaszkiewicz, O. Michielin, and J.-D. Horisberger
The Fourth Extracellular Loop of the {alpha} Subunit of Na,K-ATPase: FUNCTIONAL EVIDENCE FOR CLOSE PROXIMITY WITH THE SECOND EXTRACELLULAR LOOP
J. Biol. Chem.,
October 10, 2008;
283(41):
27850 - 27858.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R.F. Rakowski, P. Artigas, F. Palma, M. Holmgren, P. De Weer, and D. C. Gadsby
Sodium Flux Ratio in Na/K Pump-Channels Opened by Palytoxin
J. Gen. Physiol.,
July 1, 2007;
130(1):
41 - 54.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J.-D. Horisberger
Recent Insights into the Structure and Mechanism of the Sodium Pump
Physiology,
December 1, 2004;
19(6):
377 - 387.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. Artigas and D. C. Gadsby
Large Diameter of Palytoxin-induced Na/K Pump Channels and Modulation of Palytoxin Interaction by Na/K Pump Ligands
J. Gen. Physiol.,
March 29, 2004;
123(4):
357 - 376.
[Abstract]
[Full Text]
[PDF]
|
|
|
Copyright © 2003 by the Int. Union Physiol. Sci./Am. Physiol. Soc.
Top
Introduction
Homology modeling of the...
