3.A.3 The P-type ATPase (P-ATPase) Superfamily
Nearly all of the members of this superfamily, found in bacteria, archaea and eukaryotes, catalyze cation uptake and/or efflux driven by ATP hydrolysis. Clustering on the phylogenetic tree is usually in accordance with specificity for the transported ion(s). Many of these protein complexes are multisubunit with a large subunit serving the primary ATPase and ion translocation functions. In eukaryotes, they are present in the plasma membranes or endoplasmic reticular membranes. In prokaryotes, they are localized to the cytoplasmic membranes. Gastric H+-translocating ATPases (ouabain-insensitive) comprise a subgroup of the larger and more diverse Na+/K+ ATPase family (ouabain-sensitive) (Family 1). The ouabain-binding pocket can be functionally reconstituted in the gastric H+,K+ ATPase by substitution of only seven residues (Qiu et al., 2005). Ca2+ ATPases of prokaryotes and eukaryotes comprise a very diverse family (Family 2) including in eukaryotes plasma membrane, golgi, and sarcoplasmic reticular types. Sarcoplasmic reticular Ca2+ ATPases (SERCA) in brown adipose tissue can uncouple ATP hydrolysis from Ca2+ transport and be thermogenic (de Meis, 2003). H+-translocating P-type ATPases of plants and fungi comprise their own family (Family 3). Plant P-ATPases have been reviewed by Wdowikowska and Klobus, (2011). Distinct bacterial enzymes specific for K+ (Family 7; only in prokaryotes) or Mg2+ (Family 4, mostly in prokaryotes; uptake), Cu2+, Ag+, Zn2+, Co2+, Pb2+, Ni2+, and/or Cd2+ (Family 6; efflux) and Cu2+ or Cu+ (Family 5; uptake or efflux, depending on the system) have been characterized, and each of these types of enzymes comprises a distinct family. Cu2+ or Cu+-translocating ATPases from bacteria, archaea and animals cluster together, and at least some of these also transport Ag+. The Cu+/Ag+ (Family 5 and heavy metal (Family 6)) ATPases have an 8 TMS topology (Mandal et al., 2002). A cys-pro-cys motif in CopA of E. coli (TC #3.A.3.5.5) is essential for Cu+/Ag+ efflux and phosphoenzyme formation (Fan and Rosen, 2002).
P-type ATPases play essential roles in numerous processes, which in humans include nerve impulse propagation, relaxation of muscle fibers, secretion and absorption in the kidney, acidification of the stomach and nutrient absorption in the intestine. Published evidence suggests that uncharacterized families of P-type ATPases with novel specificities exist. Thever & Saier (2009) analyzed the fully sequenced genomes of 26 eukaryotes including animals, plants, fungi and unicellular eukaryotes for P-type ATPases. They reported the organismal distributions, phylogenetic relationships, probable topologies and conserved motifs of nine functionally characterized families and 13 uncharacterized families of these enzyme transporters. Family 9 Na+- or K+-ATPases can be found in fungi, plants bryophytes and protozoa (Rodríguez-Navarro and Benito, 2010).
Chan et al. (2010) analyzed P-type ATPases in all major prokaryotic phyla for which complete genome sequence data were available and compared the results with those for eukaryotic P-type ATPases. Topological type I (heavy metal) P-type ATPases predominate in prokaryotes (approx. tenfold) while type II ATPases (specific for Na+,K+, H+ Ca2+, Mg2+ and phospholipids) predominate in eukaryotes (approx. twofold). Many P-type ATPase families are found exclusively in prokaryotes (e.g. Kdp-type K+ uptake ATPases (type III) and all ten prokaryotic functionally uncharacterized P-type ATPase (FUPA) familes), while others are restricted to eukaryotes (e.g. phospholipid flippases and all 13 eukaryotic FUPA families) (Thever and Saier, 2009). Horizontal gene transfer has occurred frequently among bacteria and archaea, which have similar distributions of these enzymes, but rarely between most eukaryotic kingdoms, and even more rarely between eukaryotes and prokaryotes. In some bacterial phyla (e.g. Bacteroidetes, Flavobacteria and Fusobacteria), ATPase gene gain and loss as well as horizontal transfer occurred seldom in contrast to most other bacterial phyla. Some families (i.e., Kdp-type ATPases) underwent far less horizontal gene transfer than other prokaryotic families, possibly due to their multisubunit characteristics. Functional motifs are better conserved across family lines than across organismal lines, and these motifs can be family specific, facilitating functional predictions. In some cases, gene fusion events created P-type ATPases covalently linked to regulatory catalytic enzymes. In one family (FUPA Family 24), a type I ATPase gene (N-terminal) is fused to a type II ATPase gene (C-terminal) with retention of function only for the latter. Several pseudogene-encoded nonfunctional ATPases were identified. Genome minimalization led to preferential loss of P-type ATPase genes. Chan et al. (2010) suggested that in prokaryotes and some unicellular eukaryotes, the primary function of P-type ATPases is protection from extreme environmental stress conditions. The classification of P-type ATPases of unknown function into phylogenetic families provides guides for future molecular biological studies (Chan et al., 2010).
Many eukaryotic P-type ATPases are monomeric or homodimeric enzymes of the catalytic subunit that hydrolyzes ATP. They contain the aspartyl phosphorylation site and catalyzes ion transport. The Na+,K+-ATPases, the Ca2+-ATPases and the (fungal) H+-ATPases of higher organisms exhibit 10 transmembrane α helical spanners (TMSs), some of them highly tilted. Additional subunits that appear to lack catalytic activity may be present in the ATPase complex. For example, the 10 TMS catalytic α-subunit of the Na+,K+-ATPase of animals is tightly complexed to the 1 TMS β-subunit and the tissue-specific, regulatory, 1 TMS γ-subunit. The β-subunit, which may influence the activity of the α-subunit, probably functions to facilitate proper insertion of the α-subunit into the membrane, to allow proper targeting to a subcellular membrane site in post-translational processing, and to stabilize the catalytic subunit. The β-subunit can therefore be considered to be an auxiliary protein of the Na+,K+-ATPase catalytic subunit. The γ-subunit of the Na+,K+-ATPase has been reported to influence kinetic parameters and is homologous to a family of pore-forming peptides, the peptides of the phospholemman family (TC #1.A.27), and the C-subunits of V-type ATPases (TC #3.A.2). This γ-subunit is induced under stress conditions and modulates Na+,K+-ATPase activity and cell growth (Wetzel et al., 2004). The Na+, K+-ATPase can serve as a steroid hormone receptor (Schoner, 2002). Several other P-type ATPases also depend on small proteolipids, the functions of which are uncertain.
The annular lipid-protein stoichiometry in a native pig kidney Na+/K+ -ATPase preparation has been studied by [125I]TID-PC/16 labeling, giving results that indicated that the transmembrane domain of the Na+/K+ -ATPase in the E1 state is less exposed to the lipids than in the E2 state, i.e., the conformational transitions are accompanied by changes in the numbers of annular lipids but not in the affinity of these lipids for the protein (Mangialavori et al. 2011). The lipid-protein stoichiometry was 23 ± 2 (α subunit) and 5.0 ± 0.4 (β subunit) in the E1 conformation and 32 ± 2 (α subunit) and 7 ± 1 (β subunit) in the E2 conformation.
The stoichiometries of transport are sometimes known and complex. In the case of the Na+,K+-ATPases, 3 Na+ are exchanged for 2 K+ per ATP molecule hydrolyzed. The gastric H+-translocating ATPases replace H+ for K+ but with an H+/K+ stoichiometry of 2:2. Thus, although these two enzymes are ~65% identical, the Na+,K+-ATPases are electrogenic while the H+,K+-ATPases are electroneutral. Gastric H+, K+-ATPase transports 2 moles of H+ together with two H2O (two H3O+) per mole of ATP hydrolyzed in isolated hog gastric vesicles. Protons are charge-transferred from the cytosolic side to H2O in sites 2 and 1, the H2O coming from the cytosol, and H3O+ in these sites are transported into the lumen during the conformational transition from E1P to E2P (Morii et al., 2008). Ca2+ ATPases may catalyze Ca2+/K+ or Ca2+/H+ antiport. A single organism often possesses multiple isoforms of these enzymes.
Considerable evidence is available showing that animals have Cl- translocating, Cl- stimulating P-type ATPases. Although extensive biochemical data are available, the protein sequence of any one such Cl- ATPase has not yet been determined (Gerencser, 1993; Inagaki et al., 1996; Zeng et al., 1999). Evidence for mammalian iron-inducible, iron-transporting ATPases is also available (Baranano et al., 2000). Finally bacterial Na+-transporting P-type ATPases probably exist (Ueno et al., 2000). Evidence for a Na+-P-type ATPase has been obtained for the halotolerant cyanobacterium, Aphanothece halophytica (Wiangon et al., 2007). Thus the breadth of substrates transported by P-type ATPases is likely to be much greater than currently recognized.
The Na+,K+-ATPase acts as a signal transducer and transcription activator, modulating cell growth, apoptosis, and cell motility. A prominent binding motif linking the Na+,K+-ATPase to intracellular signaling effectors is the N-terminal tail of the Na+,K+-ATPase catalytic α-subunit which binds directly to the N-terminus of the inositol 1,4,5-trisphosphate receptor (Zhang et al., 2006). Three amino acyl residues, LKK, conserved in most species and most α-isoforms, are essential for binding. In wild-type cells, low concentrations of ouabain trigger low frequency calcium oscillations that activate NF-κB and protect from apoptosis. Thus, the LKK motif binds the inositol 1,4,5-trisphosphate receptor and triggers an anti-apoptotic calcium signal. However, the N-terminal hydrophilic region in front of the first TMS does not interact with the transported cation (Na+, K+, or Ca2+) although this first TMS does (Einholm et al., 2007). Of the P-type ATPases, only Na+, K+-ATPases are receptors that respond to endogenous cardiotonic steroids such as ouabain and marinobufagenin (steroid 'hormones') which regulate Na+ excretion and blood pressure (Liu and Xie, 2010).
P-type ATPases provide a polar transmembrane pathway, to which access is strictly controlled by coupled gates that are constrained to open alternately, thereby enabling thermodynamically uphill ion transport. Reyes and Gadsby (2006) have examined the ion pathway through the N+, K+-ATPase, a representative P-type pump, after uncoupling its extra- and intracellular gates with the marine toxin palytoxin. They found a wide outer vestibule penetrating deep into the Na+, K+-ATPase, where the pathway narrows and leads to a charge-selectivity filter. Acidic residues in this region, which are conserved to coordinate pumped ions, allow the approach of cations but exclude anions. Reversing the charge at just one of those positions converts the pathway from cation selective to anion selective. Cysteine scans from TM1 to TM6 in the Na+, K+-ATPase revealed a single unbroken cation pathway that traverses palytoxin-bound Na+,K+-pump-channels from one side of the membrane to the other (Takeuchi et al. 2008). This pathway comprises residues from TM1, TM2, TM4 and TM6, passes through ion-binding site II, and is probably conserved in structurally and evolutionarily related P-type pumps. Close structural homology among the catalytic subunits of Ca2+-, Na+, K+- and H+, K+-ATPases argues that their extracytosolic cation exchange pathways all share these physical characteristics (Reyes and Gadsby, 2006). The mechanistic details of type II ATPases, notably those for which 3-d structures are available (Na, K+-, gastri H+, K+-, Ca2+ and H+, K+-ATPase); TC Families 1,2 and 3, respectively, as well as prepared cation translocation pathways, have been discussed by Bublitz et al. (2010).
The X-ray crystal structure at 3.5 Å resolution of the pig renal Na+,K+-ATPase has been determined with two rubidium ions bound in an occluded state in the transmembrane part of the α-subunit (Morth et al., 2007). Several of the residues forming the cavity for rubidium/potassium occlusion in the Na+,K+-ATPase are homologous to those binding calcium in the Ca2+-ATPase of the sarco(endo)plasmic reticulum. The β- and γ-subunits specific to the Na+,K+-ATPase are associated with transmembrane helices αM7/αM10 and αM9, respectively. The γ-subunit corresponds to a fragment of the V-type ATPase c subunit. The carboxy terminus of the α-subunit is contained within a pocket between transmembrane helices and seems to be a novel regulatory element controlling sodium affinity, possibly influenced by the membrane potential.
The Na+,K+-ATPase can be transformed into an ion channel using pharmacological agents. Palytoxin (PTX), produced by soft coral of the genus Palythoa, binds to the ATPase with a Kd of 20 pM and creates a monovalent cation-selective channel with a single channel conductance of 10 pS (Rossini and Bigiani, 2011). The presence of external Na+ seems to be essential for channel activation (Wu et al., 2003). When the N-terminal 35 residues are removed from the ATPase, the toxin-activated channel does not exhibit a time-dependent inactivation gating at positive potentials as is characteristic of the wild-type protein. The truncated pump exhibits no electrogenic current, and the ion stoichiometry for active transport is altered. Addition of the synthetic peptide restores activity towards wild type. The N-terminal peptide therefore appears to act as an inactivation gate (similar to Shaker B channels of the VIC family (TC #1.A.1)). It may also play a critical role in determining the ion stoichiometry (Wu et al., 2003). Fluorometric studies indicate that under normal conditions, α- and β-subunits move towards each other during the E2 to E1 transition (Dempski et al., 2006).
The structures of the sarcoplasmic reticular Ca2+-ATPase have been solved at 2.6 Å resolution for the complex to which 2 Ca2+ are bound, and at 3.1 Å resolution for the complex lacking Ca2+ (Toyoshima et al., 2000; Toyoshima and Nomura, 2002). A total of eight different states of the Ca2+ -ATPase, representing the many steps in the reaction cycle, have been visualized by high resolution x-ray crystallography (Toyoshima et al., 2007; Toyoshima, 2008). The two Ca2+ are located side by side, surrounded by 4 transmembrane helices, two of which are unwound for efficient coordination geometry. There are 3 cytoplasmic domains, one, the central catalytic domain, bearing the phosphorylation site, a second bearing the adenosine nucleotide binding site, and a third of unknown function. The central domain has the same fold as haloacid dehydrogenases (Aravind et al., 1998; Stokes and Green, 2000). The Ca2+-free form shows large conformational differences from the Ca2+-bound form with the three cytoplasmic domains tightly associated to form a single headpiece and six of the ten TMSs largely rearranged. These latter rearrangements guarantee the release of external Ca2+ and create a pathway for entry of Ca2+ from the cytoplasm. ATPase activity and Ca2+ binding are cooperatively interdependent, but the two processes can be separated by mutations (Zhang et al., 2002).
Structures are available for both the E1 and E2 states of the Ca2+ ATPase showing that Ca2+ binding induces major changes in all three cytoplasmic domains relative to each other (Xu et al., 2002). Xu et al. proposed how Ca2+ binding induces conformational changes in TMS4 and 5 in the membrane domain (M) that in turn induce rotation of the phosphorylation domain (P). The nucleotide binding (N) and β-sheet (β) domains are highly mobile, with N flexibly linked to P, and β flexibly linked to M. Modeling of the fungal H+ ATPase, based on the structures of the Ca2+ pump, suggested a comparable 70º rotation of N relative to P to deliver ATP to the phosphorylation site (Kühlbrandt et al., 2002). One report suggests that this S.R. Ca2+ ATPase is homodimeric (Ushimaru and Fukushima, 2008).
Crystal structures (Gadsby, 2007) have shown that the conserved TGES loop of the Ca2+-ATPase is isolated in the Ca2E1 state but becomes inserted in the catalytic site in E2 states. Anthonisen et al. (2006) characterized the kinetics of the partial reaction steps of the transport cycle and the binding of the phosphoryl analogs BeF, AlF, MgF, and vanadate in mutants with alterations to the TGES residues. The data provide functional evidence supporting a role of Glu183 in activating the water molecule involved in the E2P → E2 dephosphorylation and suggest a direct participation of the side chains of the TGES loop in the control and facilitation of the insertion of the loop in the catalytic site. The interactions of the TGES loop furthermore seem to facilitate its disengagement from the catalytic site during the E2 → Ca2E1 transition.
Olesen et al. (2007) have described functional studies and three new crystal structures of the rabbit skeletal muscle Ca2+-ATPase. These represent the phosphoenzyme intermediates associated with (1) Ca2+ binding, (2) Ca2+ translocation and (3) dephosphorylation. They are based on complexes with a functional ATP analogue, beryllium fluoride or aluminium fluoride. The structures complete the cycle of nucleotide binding and cation transport of Ca2+-ATPase. Phosphorylation of the enzyme triggers a conformational change that leads to opening of a luminal exit pathway defined by the transmembrane segments M1 through M6. M1-M6 represent the canonical membrane domain of P-type pumps. Ca2+ release is promoted by translocation of the M4 helix, exposing Glu 309, Glu 771 and Asn 796 to the lumen. The mechanism explains how P-type ATPases are able to form the steep electrochemical gradients required for key functions in eukaryotic cells (Olesen et al., 2007). Moller et al. (2010) reveiwed structural studies of various conformers of the Ca2+ ATPase, (SERCA1a), present in skeletal muscle. The structures corresponding to the various intermediary states. They have been obtained after stabilization with structural analogues of ATP or metal fluorides as mimicks of inorganic phosphate. It is possible to provide a detailed structural description of both ATP hydrolysis and Ca2+ transport through the membrane.
The structure of a P-type proton pump was determined by X-ray crystallography by Pederson et al., (2007). Ten transmembrane helices and three cytoplasmic domains define the functional unit of ATP-coupled proton transport across the plasma membrane. The structure is locked in a functional state not previously observed in P-type ATPases. The transmembrane domain reveals a large cavity, which is likely to be filled with water, located near the middle of the membrane plane where it is lined by conserved hydrophilic and charged residues. Proton transport against a high membrane potential is readily explained by this structural arrangement.
As in other P-type ATPases, metal binding to transmembrane metal-binding sites (TM-MBS) in Cu+-ATPases is required for enzyme phosphorylation and subsequent transport. However, Cu+ does not access Cu+-ATPases in a free (hydrated) form but is bound to a chaperone protein. The delivery of Cu+ by Archaeoglobus fulgidus Cu+ chaperone CopZ to the corresponding Cu+-ATPase, CopA, has been studied (González-Guerrero and Argüello, 2008). CopZ interacted with and delivered the metal to the N-terminal metal binding domain(s) of CopA (MBDs). Cu+-loaded MBDs, acting as metal donors, were unable to activate CopA or a truncated CopA lacking MBDs. Conversely, Cu+-loaded CopZ activated the CopA ATPase and CopA constructs in which MBDs were rendered unable to bind Cu+. Furthermore, under nonturnover conditions, CopZ transferred Cu+ to the TM-MBS of a CopA lacking MBDs altogether. Thus, MBDs may serve a regulatory function without participating directly in metal transport, and the chaperone delivers Cu+ directly to transmembrane transport sites of Cu+-ATPases (González-Guerrero and Argüello, 2008). Wu et al (2008) have determined structures of two constructs of the Cu (CopA) pump from Archaeoglobus fulgidus by cryoelectron microscopy of tubular crystals, which revealed the overall architecture and domain organization of the molecule. They localized its N-terminal MBD within the cytoplasmic domains that use ATP hydrolysis to drive the transport cycle and built a pseudoatomic model by fitting existing crystallographic structures into the cryoelectron microscopy maps for CopA. The results also similiarly suggested a Cu-dependent regulatory role for the MBD.
In the Archaeoglobus fulgidus CopA (TC# 3.A.3.5.7), invariant residues in helixes 6, 7 and 8 form two transmembrane metal binding sites (TM-MBSs). These bind Cu+ with high affinity in a trigonal planar geometry. The cytoplasmic Cu+ chaperone CopZ transfers the metal directly to the TM-MBSs; however, loading both of the TM-MBSs requires binding of nucleotides to the enzyme. In agreement with the classical transport mechanism of P-type ATPases, occupancy of both transmembrane sites by cytoplasmic Cu+ is a requirement for enzyme phosphorylation and subsequent transport into the periplasmic or extracellular milieu. Transport studies have shown that most Cu+-ATPases drive cytoplasmic Cu+ efflux, albeit with quite different transport rates in tune with their various physiological roles. Archetypical Cu+-efflux pumps responsible for Cu+ tolerance, like the Escherichia coli CopA, have turnover rates ten times higher than those involved in cuproprotein assembly (or alternative functions). This explains the incapability of the latter group to significantly contribute to the metal efflux required for survival in high copper environments.
Chintalapati et al. (2008) have characterized two copper-transporting ATPases, CtrA2 and CtrA3 from Aquifex aeolicus. CtrA2 has a CPC metal-binding sequence in TM6 and a CxxC metal-binding N-terminal domain, while CtrA3 has a CPH metal-binding motif in TM6 and a histidine-rich N-terminal metal-binding domain. CtrA2 is activated by Ag+ and Cu+ and presumably transports reduced Cu+, while CtrA3 is activated by, and presumably transports, the oxidized copper ion. Both CtrA2 and CtrA3 are thermophilic proteins with an activity maximum at 75 degrees C. Electron cryomicroscopy of two-dimensional crystals of CtrA3 yielded a projection map at approximately 7 A resolution with density peaks indicating eight membrane-spanning alpha-helices per monomer. A fit of the Ca-ATPase structure to the projection map indicates that the arrangement of the six central helices surrounding the ion-binding site in the membrane is conserved, and suggests the position of the two additional N-terminal transmembrane helices that are characteristic of the heavy metal, eight-helix P(1B)-type ATPases (Chintalapati et al., 2008).
The 8TMS CadA of Listeria monocytogenes (Family 6) confirs resistance to cadmium. Residues in TMS6 (Cys354 and Cys356), TMS8 (Asp692) and TMS3 (Met149) may bind Cd2+ (Wu et al., 2006). However, the two cysteine residues in the CPC motif act at different steps: Cys354 is involved in Cd2+ binding while Cys356 is involved in Cd2+ occlusion. The two equivalent cysteines in the yeast Cu2+ ATPase may also act at different steps. The conserved Glu164 may be required for Cd2+ release. Possibly two Cd2+ are involved in the reaction cycle of CadA (Wu et al., 2006). A hemerythrin-like two iron-binding domain in a P1B-type transport ATPase from Acidothermus cellulolyticus has been identified (Traverso et al., 2010).
The phospholipid translocating P-type ATPases (Family 8; TC #3.A.3.8) are found only in eukaryotes. They appear to function with β-subunits of about 400 aas with 2 TMSs. These have been functionally characterized from yeast and protozoans (TC #8.A.27; Perez-Victoria et al., 2006). These enzymes catalyze the ATP-dependent flipping of phospholipids and lysophospholipids from the outer leaflet of the cytoplasmic membrane to the inner leaflet. Residues defining phospholipid flippase substrate specificity have been identified (Baldridge and Graham, 2012). In the yeast Saccharomyces cerevisiae, Family 3.A.3.8 lipid flipping ATPases play a pivotal role in the biogenesis of intracellular transport vesicles, polarized protein transport and protein maturation. However, in mammals, these ATPases act in concert with members of the CDC50 protein family, putative beta-subunits for these ATPases, and many function as part of the vesicle-generating machinery (Paulusma and Oude Elferink, 2010). Family 8 ATPases may exert their cellular functions by combining enzymatic phospholipid translocation activity with an enzyme-independent action. The latter can involve the timely recruitment of proteins involved in cellular signalling, vesicle coat assembly and cytoskeleton regulation (van der Velden et al., 2010). The beta-subunit, CDC50A, allows the stable expression, assembly, subcellular localization, and lipid transport activity of the P4-ATPase ATP8A2 (Coleman and Molday, 2011). Residues in phospholipid specificity have been identified (Baldridge and Graham 2012).
Detailed high resolution x-ray structures of heavy metal P1B-type ATPases were not available prior to 2011 when Gourdon et al. (2011) reported the structure of CopA, a Cu+-ATPase from Legionella pneumophile at 3.2 Å resolution. The results provided the first in depth description of a heavy metal-translocatin P1B-type ATPase. A three-stage copper transport pathway involves several well conserved residues. A P1B-specific transmembrane helix kinks at a double-glycine motif displaying an amphipathic helix that lines a putative copper entry point at the intracellular interface. An ATPase-coupled copper release mechanism from the binding sites in the membrane via an extracellular exit site is probable (Gourdon et al. 2011).
The generalized reaction for P-type ATPases is:
nMe1 (out) + mMe2 (in) + ATP → nMe1 (in) + mMe2 (out) + ADP + Pi.