The plasma membrane Ca2+ ATPase of animal cells: Structure, function and regulation

Available online at www.sciencedirect.com

ABB Archives of Biochemistry and Biophysics 476 (2008) 65–74 www.elsevier.com/locate/yabbi

Review

The plasma membrane Ca2+ ATPase of animal cells: Structure, function and regulation Francesca Di Leva a,b, Teuta Domi a,b, Laura Fedrizzi a,b, Dmitry Lim a,c, Ernesto Carafoli a,c,* b

a Department of Biochemistry, University of Padova, Viale G. Colombo, 3 35131 Padova, Italy Department of Experimental Veterinary Sciences, University of Padova, Viale dell’Universita`, 16 35020 Legnaro, Padova, Italy c Venetian Institute of Molecular Medicine (VIMM), Via Orus, 2 35129 Padova, Italy

Received 31 January 2008, and in revised form 15 February 2008 Available online 4 March 2008

Abstract Most important processes in cell life are regulated by calcium (Ca2+). A number of mechanisms have thus been developed to maintain the concentration of free Ca2+ inside cells at the level (100–200 nM) necessary for the optimal operation of the targets of its regulatory function. The systems that move Ca2+ back and forth across membranes are important actors in its control. The plasma membrane calcium ATPase (PMCA pump) which ejects Ca2+ from all eukaryotic cell types will be the topic of this contribution. The pump uses a molecule of ATP to transport one molecule of Ca2+ from the cytosol to the external environment. It is a P-type ATPase encoded by four genes (ATP2B1–4), the transcripts of which undergo different types of alternative splicing. Many pump variants thus exist. Their multiplicity is best explained by the specific Ca2+ demands in different cell types. In keeping with these demands, the isoforms are differently expressed in tissues and cell types and have differential Ca2+ extruding properties. At very low Ca2+ concentrations the PMCAs are nearly inactive. They must be activated by calmodulin, by acid phospholipids, by protein kinases, and by other means, e.g., a dimerization process. Other proteins interact with the PMCAs (i.e., MAGUK and NHERF at the PDZ domain and calcineurin A in the main intracellular domain) to sort them to specific regions of the cell membrane or to regulate their function. In some cases the interaction is isoform, or even splice variant specific. PMCAs knock out (KO) mice have been generated and have contributed information on the importance of PMCAs to cells and organisms. So far, only one human genetic disease, hearing loss, has been traced back to a PMCA defect. Ó 2008 Elsevier Inc. All rights reserved. Keywords: PMCA pump; Ca2+ homeostasis; Ca2+ signalling; PMCA KO; Hereditary deafness

General properties of PMCA pumps The plasma membrane Ca2+ pump belongs to the P-type pump family, which is characterized by the formation of a high-energy phosphorylated intermediate during the reaction cycle. The pump ejects Ca2+ from all eukaryotic cells, but PMCA-like pumps have also been described in intracellular membranes in yeasts [1]. Na+/Ca2+ exchangers, which * Corresponding author. Address: Venetian Institute of Molecular Medicine (VIMM), Via Orus, 2 35129 Padova, Italy. Fax: +39 049 8276125. E-mail address: [email protected] (E. Carafoli).

0003-9861/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2008.02.026

are particularly important in excitable cells, also eject Ca2+ and do so with higher transport capacity. Two conformational states of the phosphorylated PMCA have been described, E1 and E2 [2]. In the E1 state the enzyme is assumed to expose a high affinity Ca2+ binding site to the cytoplasmic side of the plasma membrane. The phosphorylation of an invariant aspartate by ATP (see Fig. 1) promotes a conformational change of the enzyme and the E1P ? E2-P transition. In the E2 conformation, the enzyme exposes the bound Ca2+ to the extracellular face and decreases the affinity of its binding site, liberating it outside the cell. After Ca2+ has been liberated, the E2-P intermediate is cleaved and the enzyme returns to the E1 conformation.

66

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

Fig. 1. A scheme of the reaction cycle of the PMCA pump. In the E1 conformation of the pump, Ca2+ is bound with high affinity at the cytoplasmic site of the plasma membrane. In the E2 configuration, the binding site exposes Ca2+ to the external site of the plasma membrane. In this configuration its lower affinity for Ca2+ favours its release.

At variance with the SERCA pump, which transport 2 Ca2+ ions per ATP hydrolyzed, the PMCA pump hydrolizes 1 ATP molecule per Ca2+ ion transported [3]. It thus only has one Ca2+ binding site, which is assumed to correspond to the site 2 now molecularly defined in this structure of the SERCA pump [4,5]. The PMCA exchanges 1 Ca2+ for 1 H+, i.e., the transport operation is partially electrogenic [6]. As all other P-type pumps, PMCAs are inhibited by orthovanadate (VO3(OH)2) and by La3+. Interestingly, La3+ increases the steady-state level of the phosphorylated intermediate as it slows down the E1P ? E2-P transition [7–9]. This is at variance with all other P-type pumps, where La3+ instead reduces the steady-state level of the phospho-intermediate. The property is useful in the identification of the phosphorylated PMCAs in preparations containing much higher amounts of other pumps, such as the SERCA pump: the amount of PMCA pumps usually does not exceed 0.1% of the total membrane proteins.

The structure In mammals, the PMCAs are the products of four distinct genes (ATP2B1–4), located on human chromosomal loci 12q21-q23, 3p25-p26, Xq28 and 1q25-q32 [10–13]. They are single polypeptide chains organized in 10 transmembrane domains (TM)1 and four main cytosolic domains: this topology scheme follows that experimentally established for the first time for the SERCA pump [4,5]. Fig. 2 shows a molecular modelling cartoon with the super1

Abbreviations used: TM, transmembrane domains; PL, phospholipid; CaM-BD, calmodulin-binding domain; PIP, phosphatidylinositol bisphosphate; InsP3, inositol triphosphate; DAG, diacylglycerol; PKA, protein kinase A; FAK, focal adhesion kinase; NOS-I, nitric oxide synthase I; CASK, calcium/calmodulin-dependent serine protein kinase; NMDA (N-methyl-D-aspartate); CaN, calcineurin; MET, mechano-electrical transduction; MECs, mammary epithelial cells.

Fig. 2. Structural superposition of the PMCA structure on the SERCA structural template (E1). Blue ribbon, PMCA pump; red ribbon, SERCA pump. Courtesy of Sergio Pantano (Montevideo, Uruguay).

position of the PMCA (isoform 2) on structural template of the SERCA pump. Following the nomenclature introduced for the SERCA pump, the cytosolic domains of the PMCA pump are an actuator domain, A, a phosphorylation domain, P, and a nucleotide-binding domain, N (see Fig. 3). The N-terminal cytoplasmic domain, encompassing the first 80–90 amino acids, is the most variable pump portion in the four basic isoforms. Together with the intracellular loop between TM2 and TM3, it forms the A-domain, which contains the TGE sequence motif, shown to be important in the SERCA pump during the phosphorylation process of the aspartate residue in the P-domain. This loop is predicted to be mainly composed of b-sheets, and contains a stretch of basic amino acids which is one of the two binding sites for activatory acidic phospholipid (PL) (see below). The intracellular loop between TM4 and TM5 contains the catalytic core of the pump, the P-domain. This domain contains the phosphorylated aspartate residue (D), whereas the N-domain contains the conserved lysine (K) which is part of the binding of ATP. The C-terminal region of the pump is longer than in the other Ca2+ ATPases: it is of particular interest since it contains a calmodulin-binding domain (CaM-BD), which also acts as an autoinhibitory sequence by binding to two sites in the main body of the pump. The binding with CaM is assumed to remove the CaM-BD from the inhibitory binding sites in the pump, freeing it from autoinhibition.

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

67

Fig. 3. A scheme of the membrane topology of the PMCA pump. The 10 transmembrane domains (TM) are symbolized by cylinders numbered 1–10. The three domains defined by the 3D structure of the SERCA pump are indicated by different colors. The autoinhibitory calmodulin-binding domain (CaMBD) is bound to receptor sites in the first (between TM2 and TM3) and the second (between TM4 and TM5) intracellular loops of the pump (ellipse). The A-domain contains the alternative splicing site A. Approximate positions of the phosphoryl-aspartate (D) in the catalytic site of the P-domain and of the signature lysine (K) in the ATP-binding site are shown. Site splice C is indicated in the C-terminal CaM-BD.

The C-terminal region also contains phosphorylation sites for protein kinase A and C (that for PKA only in isoform 1), and the binding sites for a number of other regulatory proteins (see below). Regulation of the PMCA activity Unstimulated PMCA pumps have poor affinity for Ca2+ (the Kd is in excess of 10 lM) and would therefore be essentially inactive at the physiological sub lM cytoplasmic Ca2+ concentrations (however, see below for PMCA2). A number of mechanisms increases the affinity of the pump for Ca2+, lowering its Kd to values as low as about 200 nM. The traditional activator is calmodulin: the interaction of PMCA with it was the first discovered mechanism for the regulation of PMCA pump activity [14,15]. Calmodulin interacts with a domain (CaM-BD) located in the C-terminal cytosolic tail of the pump (see Fig. 3). As mentioned, in the absence of calmodulin the CaM-BD is bound to an intramolecular receptor site composed of two portions: the first is close to the basic region that binds acidic phospholipids in the cytosolic loop between the second and third transmembrane domains (TM2 and TM3), the second is located next to the active site of the pump within the large cytosolic loop between TM4 and TM5 (see Fig. 3). The binding of calmodulin to the CaM-BD is assumed to displace the latter from the receptor sites, leading to the decrease of the Kd for Ca2+ from 10–

20 lM to <1 lM [16–19]. The solution of the three-dimensional structure of calmodulin complexed to a peptide corresponding to the N-terminal half of CaM-BD has revealed some unusual features of the complex [20]. The peptide was bound only to the C-terminal lobe of calmodulin, but, most significantly, calmodulin in the complex did not collapse to the hairpin conformation first seen with the myosin light chain binding peptide [21], but remained in the extended conformation. This was consistent with the observation that the pump, at variance with other calmodulin-dependent enzymes, can be activated by the separate C-terminal half of calmodulin [20,22]. This may be related to the observation that the full length pump (the isoform tested was PMCA3) and its truncated version that has lost the C-terminal half of the CaM-BD, have similar Ca2+ ejecting activities when overexpressed in model cells [23]. The observation may also indicate that PMCA pumps in vivo can be regulated by agents different from CaM [24]. The activity of PMCAs is influenced by the phospholipid composition in the surrounding plasma membrane [25,26]. As mentioned, acidic phospholipids and polyunsaturated fatty acids activate the pump by binding to two distinct regions: one is the CaM-BD itself [27], and the other is located in the first cytosolic loop of the pump [28] (see Fig. 3). The activating effect of phospholipids is even more pronounced than that of calmodulin, i.e., they decrease the Kd (Ca2+) to about 100 nM [16]. The concentration of phosphatidylserine in the plasma membrane presumably

68

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

surrounding the pump has been estimated to be adequate for about 50% of maximal activation [29]. Possibly then, the regulation of the PMCA by acidic phospholipids is important in vivo. Early studies have indicated phosphatidylinositol bisphosphate (PIP2) to be the most active in stimulating PMCA activity [30,31]. As the analysis of the fluorescence energy transfer between tryptophan residues of the PMCA and pyrene-labelled analogues of phospholipids has revealed a positive correlation between the strength of interaction and the number of negative charges, it is not surprising that PIP2 should be bound by the pump most tightly [32]. PIP2 is of particular interest as a potential modulator of PMCA activity since it is the only acidic phospholipid whose concentration in the plasma membrane may change rapidly in response to external stimulation. The products of its hydrolysis, inositol triphosphate (InsP3) and diacylglycerol (DAG) have no effect on the activity of the pump [33]. PMCAs can be phosphorylated both at serine/threonine and at tyrosine residues. Protein kinase A (PKA) phosphorylates a consensus site (KRNSS) only found in isoform 1 downstream of the CaM-BD. The phosphorylation increases both the Vmax and the Ca2+ affinity of the pump [34]. PKC phosphorylates PMCAs at a threonine residue directly within the CaM-BD. At variance with serine/threonine phosphorylation, tyrosine phosphorylation has been claimed to inhibit the activity of the pump [35,36]. The Src kinase phosphorylates the PMCA4 at tyrosine 1176 in vitro [35], and the focal adhesion kinase (FAK) has been proposed to phosphorylate the same residue in PMCA4 during platelet activation [36]. PMCAs are cleaved by calpain, just upstream of the CaM-BD. The cleavage produces a constitutively active fragment of about 124 kDa [34,37]. Somehow, the PMCA1 isoform is somewhat more sensitive to the cleavage by calpain than the other three basic isoforms [38]. The cleavage of PMCAs by calpain activates it irreversibly, and has been shown to play a role in human erythrocyte function [39] and in platelet activation [40]. Early radiation inactivation experiments on human red cell ghosts had indicated that the PMCA pump may exist in the membrane as a dimer [41,42]. Later on it was shown that the self-association is mediated by both the CaM-BD [43] and the CaM-BD-binding site [44]. It leads to an activation of the pump [45] which is alternative to that by calmodulin, as the PMCA dimers are no longer sensitive to the latter [46]. More recently, it has been proposed that oligomerization might represent a mechanism of self protection of PMCAs against spontaneous denaturation [47] since self-association in vitro requires a 10–20 lM concentration of the pump [46]. Since the amount of the PMCA pump in plasma membranes is only 0.1–0.01% of total membrane proteins, dimerization in vivo seems unlikely. However, PMCAs may not be uniformly distributed along the plasma membrane, one plausible site for their dimerization being the caveolae, where the pumps have been claimed to be concentrated [48,49]. Caveolae would repre-

sent a convenient site for PMCA regulation also because they are enriched in acidic phospholipids [50]. Interestingly, calmodulin has also been claimed to be associated with the caveolar membrane fraction, but not with non-caveolar fractions [51]. However, one obvious problem in suggesting the caveolae as sites of PMCA activation is their absence in a number of cells types. The pump isoforms and the alternative splicing process The complete sequence of the PMCA pump was obtained about 20 years ago [52,53]. The sequences of two rat brain isoforms contained 1176 and 1198 amino acids and were named PMCA1 and PMCA2, respectively. The sequence of the human brain pump shared more than 99% homology with rat PMCA1 in the first 1117 residues, but was significantly different in the C-terminal region downstream of it. The sequence discrepancy was found to be due to a complex alternative splicing process: detailed studies on it showed that it generated a wide range of splicing variants for all four basic isoforms of the pump that were eventually cloned [11,54]. The number and size of coding exons are conserved among the genes of four isoforms, although the size of the human genes ranges from 60 kb for PMCA3 to over 350 kb for PMCA2. The genes show different expression patterns, those of PMCA1 and PMCA4 being expressed ubiquitously, those of PMCA2 and PMCA3 being expressed in a tissue restricted way: in humans, the PMCA2 pump is a brain isoform [55], although is also expressed in the mammary gland [56], the PMCA3 pump is expressed in brain but in skeletal muscles as well [55,57,58]. PMCA1 is ubiquitous, but is expressed most abundantly in brain, lung and intestine [55,58–60], PMCA4 in kidney, erythrocytes, skeletal muscle, heart, stomach, intestine, brain and spermatozoa [55,60–62]. The multiplicity of PMCA isoforms is likely to respond to different demands of Ca2+ tuning in tissues and subcellular domains. The differential location, distribution and regulated expression of the pump isoforms add weight to the idea of the functional significance of diversity [63]. For instance, although all PMCAs respond to a number of activators, PMCA2 has the peculiar ability to operate efficiently also in their absence. This property may be important in the maintenance of a given level of non activated Ca2+ exporting activity in specialized cells, e.g., the hair cells of the inner ear [64]. As mentioned, all four basic isoform undergo alternative splicing, which occurs at site C in the C-terminal tail of the protein, and at site A in the first cytosolic loop (A-domain). Fig. 4 offers complete panorama of the splicing variants, and of their nomenclature. The number of splice variants at site A could be high, but only some variants have so far been detected for each pump isoform. The splicing involves a small exon of 36– 42 nucleotides (nt) present in all PMCA genes (see Fig. 4). In the PMCA1 isoform the exon was found to be

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

69

Fig. 4. The alternative sites A and C splice options of the PMCA pumps. Most splice variants have been extensively characterized in the four human and rat PMCA isoforms [70,71,100,102,103].

always inserted, at least in the analysis of the transcripts. Thus all the PMCA1 are only the x (or AII, using the alternative nomenclature [65]) splice variants [66]. In PMCA3 and 4 only splice variants x and z have been found, generated by the insertion or the exclusion, respectively, of a single exon of 42 nt in PMCA3 and of 36 nt in PMCA4. The pattern of alternative splicing is particularly complex for the PMCA2 transcript, where the insertion of exons generates four combinations. Three exons of 33, 60 and 42 nt, respectively, can be inserted in different ways: all inserted (splice variant w, or AIII), all excluded (splice variant z, or AI), the 42 nt exon included (splice variant x, or AII) or the 33 and 60 nt exons included (splice variant y, or AIV). The PMCA2y has been found only in rat, whereas the other isoforms have been detected in humans as well [67]. The inclusion of three exons in the PMCA2w variant has been shown to target the pump to apical membranes. In particular, the splice variant PMCA2w/a (also spliced at site C) is targeted to the stereocilia of the auditory system [68,69]. Splice site C involves the regulatory CaMBD and thus affects the affinity of the pump for CaM. All the isoforms undergo splicing also at C-site, albeit with different degrees of complexity [70–72]. C-site splicing modifies the regulatory domain as the inserts are not in frame, and lead to the premature truncation of the protein (a or CII variant instead of the longer unspliced variant b or

CI). The C-site splicing process is further complicated in each PMCA isoform by the presence of multiple internal donor sites within the spliced exons. Thus, five different variants could arise for PMCA1 (a, b, c, d and e, or CII, CI, CIII, CIV and CV), three for PMCA2 (a, b and c) and PMCA4 (a, b and d), and six for PMCA3 (a, b, c, d, e and f). However, as mentioned above, even if containing a shorter CaM-BD, the shorter variants of PMCA3 (a, e) and PMCA4 (a) are as effective in clearing Ca2+ from cells as non spliced variants [23]. The tissue specificity and localization of the PMCA pumps can also be influenced by a transcriptional regulation occurring in the non-coding regions of the genes, i.e., by the usage of different transcriptional start sites in the promoter, or by the expression of single 50 untranslated exons [73].

Molecular partners for PMCA PMCA pumps interact with numerous proteins. Some interacting partners are involved in the recruitment and maintenance of particular PMCA splice variants in specific membrane domains, others regulate the general PMCA ability to export Ca2+ from cells [74]. Interactions with molecules involved in signalling pathways have led to the

70

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

suggestion that PMCAs would also act as modulators of signal transduction [75]. Most partners for PMCA have been found to interact with a PDZ binding domain in the C-terminal region of the pump. Members of the MAGUK (Membrane-Associated Guanylate Kinase, SAP) family of kinases, a PDZdomain containing protein family, interact with PMCA isoforms 2b and 4b. MAGUKs are associated with the cortical actin cytoskeleton and are thought to act as scaffolding or clustering proteins for various membrane receptors and transporters. Thus, the binding of MAGUKs to the PMCAs could tether the Ca2+ pump to the membrane cytoskeletal network, maintaining them in specific membrane domains to control local Ca2+ concentrations. A study using the PMCA2b C–terminal region as a bait in a yeast two hybrid screening has shown that another PDZ-domain containing protein, NHERF2 (Na+/H+ Exchanger Regulatory Factor-2) is an interacting partner. This interaction is specific and isoform-selective, as pulldown and co-immunoprecipitation experiments using PMCA2b and PMCA4b, as well as NHERF1 and NHERF2 isoforms, only show interaction for PMCA2b and NHERF isoform 2 [76]. The study also demonstrated that PMCA2b co-localizes with NHERF2 at the apical membrane of polarized epithelial cells. Since apical targeting of the pump occurs even in the absence of exogenously expressed NHERF2, it could have a role in the stabilization of the pump in a particular microdomain of the cell. NHERFs could provide a scaffolding mechanism, bringing together transmembrane proteins, signalling molecules and the actin cytoskeleton in a regulated fashion, suggesting a function for NHERFs in the stabilization and membrane retention of interacting partners [76]. PMCA isoform 4b also interacts with a PDZ domain in Nitric Oxide Synthase I (NOS-I). Experiments in model cells have shown that the overexpression of PMCA4b reduces NOS-I activity. When cells are transfected with an inactive PMCA, the inhibition of NOS-I is greatly attenuated. Apparently, the down-regulation of NOS-I activity is caused by Ca2+ depletion in the immediate proximity of the enzyme [77]. Pull-down experiments have shown that another PDZ domain containing protein, PISP (PMCA-Interacting Single-PDZ domain), also interacts with all variants of the PMCAs. PISP is a ubiquitously expressed protein and its interaction with the PMCAs could influence their sorting to the plasma membrane [78]. Calcium/calmodulin-dependent Serine protein Kinase (CASK) is another MAGUK protein that interacts with PMCA4b. CASK is a co-activator of the transcription of T-element containing promoters. Overexpression of PMCA4b down-regulates the T-element-dependent reporter activity [79]. As was suggested for NOS-I, the negative regulation of CASK by PMCA could be the result of Ca2+ depletion in close proximity of the enzyme. Since CASK is located at brain synapses, PMCAs could be involved in the regulation of synaptic activity.

A yeast two hybrid approach has shown that the C-terminal PDZ binding domain of all PMCAs also interacts with Ania-3, a member of the Homer family of scaffold proteins that couple NMDA (N-methyl-D-aspartate) receptors to metabotropic glutamate receptors and link extracellular signals to Ca2+ release from intracellular stores. The interaction could recruit the PMCAs to domains close to Ca2+ influx systems at specialized sites in neurons, and could represent a mechanism by which local calcium signalling and synaptic function can be modulated in neurons [80]. The last entry in the group of PDZ-domain-containing proteins is CLP36, a protein that interacts with the actin cytoskeleton in platelets. It has been found to interact with PMCA4b and appears to be involved in its translocation during platelet activation. A model has been proposed in which PMCA interacts with F-actin via CLP36. The presence of the pump in filopodia would regulate intracellular Ca2+ in these structures, and has a role in the late phase of platelet activation [81]. In addition to the PDZ binding domain, other PMCA regions also interact with protein partners. The main intracellular loop of PMCA4b (between TM4 and TM5) was found to bind the tumour suppressor Ras-Associated Factor 1, RASSF1. Co-expression of RASSF1 and PMCA4b inhibits the epidermal growth factor-dependent activation of RAS, compatible with a role for PMCA4b in the modulation of Ras-mediated signalling [75]. Other interactors that bind to the main intracellular loop of PMCAs are alpha-1 syntrophin and calcineurin (CaN). The overexpression of alpha-1 syntrophin in model cells increases the inhibitory effect of the PMCA on the NOS-1 activity. These proteins have been found to form a ternary complex in cardiac myocytes [82]. The main intracellular loop of PMCAs also binds the catalytic subunit of calcineurin A, suggesting that PMCA could be a negative modulator of calcineurinmediated signalling pathways [83]. The N-terminal region of PMCAs which has a very low degree of sequence homology among isoforms, has been used to identify possible isoform-specific binding partners. A recent study has shown that isoform e of 14-3-3 protein interacts with PMCA 1, 3 and 4, but not with isoform 2. The Ca2+ ejecting function of PMCAs is inhibited in model cells co-expressing the pumps and the 14-3-3 protein [84,85]. Fig. 5 offers a comprehensive summary of all interactors of PMCA pumps. Involvement of PMCAs in disease processes The Atp2b1 gene has been disrupted in mice by targeted deletion of the sequence encoding for the catalytic phosphorylation site [62]. Although at the preimplantation stage null mutants were present, by day 8.5 of pregnancy only heterozygous mutants were detectable [86]. The result suggests that PMCA1 has a housekeeping function during embryogenesis. Further information on the role of PMCA1

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

71

Fig. 5. A scheme of the topology of the PMCA pump with its proteins partners. Most partners interact with the C-terminal PDZ-binding domain of the pump. Other proteins interact with the N-terminal region and the main intracellular loop of the pump. The interacting proteins have a role in the recruitment and maintenance of PMCA to specific membrane domains and on the ability of PMCA to transport Ca2+. The sites of interaction with calmodulin (CaM), the PDZ-binding domain (PDZ BD), and the phospholipid-binding domain (PL BD) are also shown.

was provided by experiments where the presence of a null mutation in heterozygous PMCA1 mice favored apoptosis of cells of vascular smooth muscle [62]. Contractility measurements on these cells have documented the important role of PMCA1 in the extrusion of Ca2+ after carbachol stimulation or depolarization with KCl [87]. As mentioned, PMCA2 is highly expressed in the hair cells of the inner ear [88], in cerebellar Purkinje cells [89], and in the lactating mammary gland [56], and is also expressed at a lower level in the heart [58]. Both natural and generated mutants of PMCA2 are known, the characteristic phenotype being characterized by deafness and ataxia. In some cases, mice present progressive hearing loss, possibly due to the epistatical interaction of PMCA2 with other genetic loci (i.e., the ahl variant [90]). Two human families have been described in which mutations in the Atp2b2 gene are linked to the insurgence of deafness. In one, the autosomal recessive mutation augmented the hearing loss caused by a mutation in the cadherin 23 gene [91]). In the other family [64] (see Fig. 6), profound deafness was present only when a mutation in the cadherin 23 gene was present together with a mutation in the PMCA2 gene. If only the mutation in one of the two genes was present the subject was healthy, suggesting a new digenic mechanism for the development of hereditary hearing loss. The PMCA2 isoform in the stereocilia of the outer hair cells is splice variant w/a [68]. PMCA2 is characterized by the poor ability to rapidly increase activity in response to a sudden increase of Ca2+, but has higher basal Ca2+

ejection activity than other isoforms [64]. These properties are even more evident in splice variant w/a. The choice of this unusual variant in outer hair cells could thus reflect the special demands of Ca2+ homeostasis in the inner ear. In the inner ear the deflection of the stereocilia in response to sound stimulation causes the opening of the mechanoelectrical transduction channels (MET [92]), inducing the influx of Ca2+ into hair cells. PMCA2w/a pumps then the ion back to the endolymph in which stereocilia are embedded [93]. The impairment of the function of PMCA2w/a may lead to the lowering of Ca2+ concentration in the endolymph surrounding the stereocilia (as reported in the model mouse ‘‘deafwaddler”—dfw—[94]) and to the increase of the opening probability of MET channels, biasing them towards the fully open state [64]. As a result, outer hair cells would become insensitive to sound stimulation. The linkage between PMCA2 and cadherin 23 mutations reported in the two human cases is interesting as cadherin 23 is a component of the tip links that connect the top of a cilium in the hair bundle to the adjacent one. Even if at an endolymph concentration of Ca2+ as low as 6 lM, as reported in the dfw mouse, the integrity of tip links appears not to be compromised [95] the extrusion of calcium by PMCA2w/a may still have a role in maintaining it. The balance impairment induced by PMCA2 mutations is explained by defects in the vestibular system and in the cerebellar Purkinje cells, where PMCA2 is also highly expressed. The number of Purkinje cells was found to

72

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

Fig. 6. Digenic mechanism of inheritance deafness of the PMCA linked deafness (left panel). The subject II-2 is the only member of the family affected by hearing loss. He bears a mutation in the ATP2B2 gene and in CDH23 gene. Subject I-2 (the mother) only bears the ATP2B2 mutation and is healthy. Subject I-1 (the father) only bears the CDH23 mutation and he is also healthy. Impairment of function of the PMCA2 in handling Ca2+ homeostasis (right panel). The panel documents the decrease ability of PMCA2 to control Ca2+. The defect concerns the baseline calcium pumping activity of the pump. Ca was monitored using the Ca2+-sensitive photoprotein Aequorin. For details see [64,104].

increase in the cerebella of PMCA2 KO mice, that of granule cells was instead reduced, and the molecular cell layer became thinner [96]. The structure of the vestibular system was normal except for the loss of otoconia (the Ca2+ carbonate crystals normally embedded in the otolithic membrane overlaying the sensory epithelium of saccule and utricle that sense linear acceleration and gravity [96]). Their absence could be explained by the inability of PMCA2 to efficiently eject Ca2+ to the endolymph: indeed, in animals bearing PMCA2 mutations the Ca2+ concentration in the endolymph of the cochlear duct decreased from 20– 25 lM to 6 lM, becoming insufficient for the formation of otoconia [97]. PMCA2 also has an important role in the regulation of Ca2+ concentration in the milk. Mammary epithelial cells (MECs) need to efficiently eject Ca2+ to supply it to milk produced in the gland during lactation. The PMCA2 in the apical membrane of MECs is the wb splice variant, which becomes highly expressed during lactation [56]. In PMCA2 KO mice the level of Ca2+ in the milk was found to be strongly reduced [98], in keeping with the key role of PMCA2 of MECs in the secretion of Ca2+. An up-regulation of SERCA2b and SPCA1 occurs under these conditions, but is evidently inadequate to compensate for the loss of PMCA2 [56]. A transcriptional down-regulation of PMCA2 has recently been found in several mouse models of Huntington disease (HD) and in the brains of HD patients. The down-regulation may contribute to the chronic cellular Ca2+ overload of HD neurons [99]. The PMCA4 is expressed ubiquitously [100] and could thus also represent a housekeeping gene. However, in contrast with those of PMCA1, PMCA4 null mutants survive, and it is thus possible to analyze their phenotype. KO mice were obtained by target disruption of the catalytic phosphorylation site of the gene in embryonic stem cells [62]. Since no differences were observed in the Mendelian ratio of the wild type, homozygous, and heterozygous offsprings,

the conclusion was reached that the disruption of the gene did not alter the embryonic development. PMCA4/mice did not show significant phenotype differences with respect to the wild type counterparts, except for male infertility. This was not surprising, as PMCA4 is expressed at particularly high levels in the testis (more than 90% of PMCA protein expressed there is isoform 4 [62]). The PMCA4 loss greatly affected sperm motility, but not the ability of the sperm to fertilize eggs [101]. The finding that PMCA4 null mutants showed no developmental defects, growth retardation, or reduced cell viability [62] suggests that PMCA4, even if widely expressed, has a specialized function. In addition to defects in sperm motility, PMCA4/ mice also had alterations of contraction, and propensity to apoptosis, in smooth muscles: PMCA4 is the most abundant isoform also in this cell type. However, when PMCA4 KO mice were crossed with the original strain, the disease phenotype was lost and was restored by the addition of a heterozygous PMCA1 null mutation. These data suggest that PMCA1 can act as a modifier in PMCA4 KO mice for the general control of Ca2+ homeostasis [62]. Conclusions The PMCA pump, discovered about forty years ago, has traditionally been a difficult object of study. Progress in the last few years has nevertheless been very significant, particularly in the areas that distinguish the PMCA pump from all other P-type pumps: from the properties of the unusually large number of isoforms, including the alternatively spliced variants, to the analysis of the multiplicity of interactors that may regulate the targeting and the function of the pump variants, to the discovery and molecular analysis of disease processes generated by mutations in the PMCA gene. Disease conditions affect predominantly nervous cells, which demand an absolutely precise control of Ca2+ homeostasis. The area concerning the specific functions of the PMCA pump in different cell types and even in dif-

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

ferent domains of the cell is now attracting increasing attention. It begins to be clear that the PMCA pump is not only critical in the maintenance of the appropriately overall cytosolic Ca2+ concentration: it also shapes the calcium signals spatially and temporally. One very important aspect in which no substantial progress has yet been made is the 3-D structure of the PMCA pump. A number of reasons suggest this will be a very difficult problem. It is nevertheless hoped that research developments will soon lead to substantial progress in this area as well. References [1] M.G. Palmgren, K.B. Axelsen, Biochim. Biophys. Acta 1365 (1998) 37–45. [2] J. Krebs, M. Vasak, A. Scarpa, E. Carafoli, Biochemistry 26 (1987) 3921–3926. [3] V. Niggli, E. Sigel, E. Carafoli, J. Biol. Chem. 257 (1982) 2350–2356. [4] C. Toyoshima, M. Nakasako, H. Nomura, H. Ogawa, Nature 405 (2000) 647–655. [5] C. Toyoshima, H. Nomura, Nature 418 (2002) 605–611. [6] L. Hao, J.L. Rigaud, G. Inesi, J. Biol. Chem. 269 (1994) 14268– 14275. [7] S. Luterbacher, H.J. Schatzmann, Experientia 39 (1983) 311–312. [8] E.E. Quist, B.D. Roufogalis, FEBS Lett. 50 (1975) 135–139. [9] I. Szasz, B. Sarkadi, A. Schubert, G. Gardos, Biochim. Biophys. Acta 512 (1978) 331–340. [10] S. Olson, M.G. Wang, E. Carafoli, E.E. Strehler, O.W. McBride, Genomics 9 (1991) 629–641. [11] P. Brandt, E. Ibrahim, G.A. Bruns, R.L. Neve, Genomics 14 (1992) 484–487. [12] F. Latif, F.M. Duh, J. Gnarra, K. Tory, I. Kuzmin, M. Yao, T. Stackhouse, W. Modi, L. Geil, L. Schmidt, et al., Cancer Res. 53 (1993) 861–867. [13] M.G. Wang, H. Yi, H. Hilfiker, E. Carafoli, E.E. Strehler, O.W. McBride, Cytogenet. Cell Genet. 67 (1994) 41–45. [14] R.M. Gopinath, F.F. Vincenzi, Biochem. Biophys. Res. Commun. 77 (1977) 1203–1209. [15] H.W. Jarrett, J.T. Penniston, Biochem. Biophys. Res. Commun. 77 (1977) 1210–1216. [16] A. Enyedi, M. Flura, B. Sarkadi, G. Gardos, E. Carafoli, J. Biol. Chem. 262 (1987) 6425–6430. [17] P. James, M. Maeda, R. Fischer, A.K. Verma, J. Krebs, J.T. Penniston, E. Carafoli, J. Biol. Chem. 263 (1988) 2905–2910. [18] R. Falchetto, T. Vorherr, J. Brunner, E. Carafoli, J. Biol. Chem. 266 (1991) 2930–2936. [19] R. Falchetto, T. Vorherr, E. Carafoli, Protein Sci. 1 (1992) 1613– 1621. [20] B. Elshorst, M. Hennig, H. Forsterling, A. Diener, M. Maurer, P. Schulte, H. Schwalbe, C. Griesinger, J. Krebs, H. Schmid, T. Vorherr, E. Carafoli, Biochemistry 38 (1999) 12320–12332. [21] M. Ikura, G.M. Clore, A.M. Gronenborn, G. Zhu, C.B. Klee, A. Bax, Science 256 (1992) 632–638. [22] D. Guerini, J. Krebs, E. Carafoli, J. Biol. Chem. 259 (1984) 15172– 15177. [23] M. Brini, L. Coletto, N. Pierobon, N. Kraev, D. Guerini, E. Carafoli, J. Biol. Chem. 278 (2003) 24500–24508. [24] B.L. Tempel, D. Shilling, The Plasma Membrane Calcium ATPase and Disease, ed., Springer, London, UK, 2007. [25] Y. Pang, H. Zhu, P. Wu, J. Chen, FEBS Lett. 579 (2005) 2397–2403. [26] D. Tang, W.L. Dean, D. Borchman, C.A. Paterson, Cell Calcium 39 (2006) 209–216. [27] P. Brodin, R. Falchetto, T. Vorherr, E. Carafoli, Eur. J. Biochem. 204 (1992) 939–946.

73

[28] T. Pinto Fde, H.P. Adamo, J. Biol. Chem. 277 (2002) 12784–12789. [29] V. Niggli, E.S. Adunyah, E. Carafoli, J. Biol. Chem. 256 (1981) 8588–8592. [30] E. Carafoli, M. Zurini, Biochim. Biophys. Acta 683 (1982) 279–301. [31] D. Choquette, G. Hakim, A.G. Filoteo, G.A. Plishker, J.R. Bostwick, J.T. Penniston, Biochem. Biophys. Res. Commun. 125 (1984) 908–915. [32] J. Verbist, T.W. Gadella Jr., L. Raeymaekers, F. Wuytack, K.W. Wirtz, R. Casteels, Biochim. Biophys. Acta 1063 (1991) 1–6. [33] E. Carafoli, Physiol. Rev. 71 (1991) 129–153. [34] P.H. James, M. Pruschy, T.E. Vorherr, J.T. Penniston, E. Carafoli, Biochemistry 28 (1989) 4253–4258. [35] W.L. Dean, D. Chen, P.C. Brandt, T.C. Vanaman, J. Biol. Chem. 272 (1997) 15113–15119. [36] T.C. Wan, M. Zabe, W.L. Dean, Thromb. Haemost. 89 (2003) 122– 131. [37] K.K. Wang, A. Villalobo, B.D. Roufogalis, Trends Cell Biol. 2 (1992) 46–52. [38] D. Guerini, B. Pan, E. Carafoli, J. Biol. Chem. 278 (2003) 38141– 38148. [39] F. Salamino, B. Sparatore, E. Melloni, M. Michetti, P.L. Viotti, S. Pontremoli, E. Carafoli, Cell Calcium 15 (1994) 28–35. [40] C.S. Brown, W.L. Dean, Platelets 18 (2007) 207–211. [41] J.D. Cavieres, Biochim. Biophys. Acta 771 (1984) 241–244. [42] A.M. Minocherhomjee, G. Beauregard, M. Potier, B.D. Roufogalis, Biochem. Biophys. Res. Commun. 116 (1983) 895–900. [43] T. Vorherr, T. Kessler, F. Hofmann, E. Carafoli, J. Biol. Chem. 266 (1991) 22–27. [44] V. Levi, J.P. Rossi, P.R. Castello, F.L. Gonzalez Flecha, FEBS Lett. 483 (2000) 99–103. [45] D. Kosk-Kosicka, T. Bzdega, J. Biol. Chem. 263 (1988) 18184– 18189. [46] D.L. Sackett, D. Kosk-Kosicka, J. Biol. Chem. 271 (1996) 9987– 9991. [47] V. Levi, J.P. Rossi, P.R. Castello, F.L. Gonzalez Flecha, Biophys. J. 82 (2002) 437–446. [48] T. Fujimoto, J. Cell Biol. 120 (1993) 1147–1157. [49] J.E. Schnitzer, P. Oh, B.S. Jacobson, A.M. Dvorak, Proc. Natl. Acad. Sci. USA 92 (1995) 1759–1763. [50] H.R. Hope, L.J. Pike, Mol. Biol. Cell 7 (1996) 843–851. [51] P.W. Shaul, E.J. Smart, L.J. Robinson, Z. German, I.S. Yuhanna, Y. Ying, R.G. Anderson, T. Michel, J. Biol. Chem. 271 (1996) 6518– 6522. [52] G.E. Shull, J. Greeb, J. Biol. Chem. 263 (1988) 8646–8657. [53] A.K. Verma, A.G. Filoteo, D.R. Stanford, E.D. Wieben, J.T. Penniston, E.E. Strehler, R. Fischer, R. Heim, G. Vogel, S. Mathews, et al., J. Biol. Chem. 263 (1988) 14152–14159. [54] E.E. Strehler, D.A. Zacharias, Physiol. Rev. 81 (2001) 21–50. [55] T.P. Stauffer, D. Guerini, E. Carafoli, J. Biol. Chem. 270 (1995) 12184–12190. [56] T.A. Reinhardt, A.G. Filoteo, J.T. Penniston, R.L. Horst, Am. J. Physiol. Cell Physiol. 279 (2000) C1595–C1602. [57] B.J. Brown, H. Hilfiker, S.J. DeMarco, D.A. Zacharias, T.M. Greenwood, D. Guerini, E.E. Strehler, Biochim. Biophys. Acta 1283 (1996) 10–13. [58] J. Greeb, G.E. Shull, J. Biol. Chem. 264 (1989) 18569–18576. [59] S. De Jaegere, F. Wuytack, J.A. Eggermont, H. Verboomen, R. Casteels, Biochem. J. 271 (1990) 655–660. [60] A. Howard, S. Legon, J.R. Walters, Am. J. Physiol. 265 (1993) G917–G925. [61] P. Brandt, R.L. Neve, A. Kammesheidt, R.E. Rhoads, T.C. Vanaman, J. Biol. Chem. 267 (1992) 4376–4385. [62] G.W. Okunade, M.L. Miller, G.J. Pyne, R.L. Sutliff, K.T. O’Connor, J.C. Neumann, A. Andringa, D.A. Miller, V. Prasad, T. Doetschman, R.J. Paul, G.E. Shull, J. Biol. Chem. 279 (2004) 33742–33750. [63] T. Domi, F. Di Leva, L. Fedrizzi, A. Rimessi, M. Brini, Ann. NY Acad. Sci. 1099 (2007) 237–246.

74

F. Di Leva et al. / Archives of Biochemistry and Biophysics 476 (2008) 65–74

[64] R. Ficarella, F. Di Leva, M. Bortolozzi, S. Ortolano, F. Donaudy, M. Petrillo, S. Melchionda, A. Lelli, T. Domi, L. Fedrizzi, D. Lim, G.E. Shull, P. Gasparini, M. Brini, F. Mammano, E. Carafoli, Proc. Natl. Acad. Sci. USA 104 (2007) 1516–1521. [65] E. Carafoli, FASEB J. 8 (1994) 993–1002. [66] R. Bobe, R. Bredoux, E. Corvazier, C. Lacabaratz-Porret, V. Martin, T. Kovacs, J. Enouf, Platelets 16 (2005) 133–150. [67] H.P. Adamo, A.K. Verma, M.A. Sanders, R. Heim, J.L. Salisbury, E.D. Wieben, J.T. Penniston, Biochem. J. 285 (Pt 3) (1992) 791–797. [68] J.K. Hill, D.E. Williams, M. LeMasurier, R.A. Dumont, E.E. Strehler, P.G. Gillespie, J. Neurosci. 26 (2006) 6172–6180. [69] M. Grati, N. Aggarwal, E.E. Strehler, R.J. Wenthold, J. Cell Sci. 119 (2006) 2995–3007. [70] T.P. Stauffer, H. Hilfiker, E. Carafoli, E.E. Strehler, J. Biol. Chem. 268 (1993) 25993–26003. [71] T.P. Keeton, S.E. Burk, G.E. Shull, J. Biol. Chem. 268 (1993) 2740– 2748. [72] J. Santiago-Garcia, J. Mas-Oliva, D. Saavedra, A. Zarain-Herzberg, Mol. Cell Biochem. 155 (1996) 173–182. [73] R.S. Silverstein, B.L. Tempel, Neuroscience 141 (2006) 245–257. [74] S.J. DeMarco, E.E. Strehler, J. Biol. Chem. 276 (2001) 21594–21600. [75] A.L. Armesilla, J.C. Williams, M.H. Buch, A. Pickard, M. Emerson, E.J. Cartwright, D. Oceandy, M.D. Vos, S. Gillies, G.J. Clark, L. Neyses, J. Biol. Chem. 279 (2004) 31318–31328. [76] S.J. DeMarco, M.C. Chicka, E.E. Strehler, J. Biol. Chem. 277 (2002) 10506–10511. [77] K. Schuh, S. Uldrijan, M. Telkamp, N. Rothlein, L. Neyses, J. Cell Biol. 155 (2001) 201–205. [78] G.M. Goellner, S.J. DeMarco, E.E. Strehler, Ann. NY Acad. Sci. 986 (2003) 461–471. [79] K. Schuh, S. Uldrijan, S. Gambaryan, N. Roethlein, L. Neyses, J. Biol. Chem. 278 (2003) 9778–9783. [80] V. Sgambato-Faure, Y. Xiong, J.D. Berke, S.E. Hyman, E.E. Strehler, Biochem. Biophys. Res. Commun 343 (2006) 630–637. [81] L.D. Bozulic, M.T. Malik, D.W. Powell, A. Nanez, A.J. Link, K.S. Ramos, W.L. Dean, Thromb. Haemost. 97 (2007) 587–597. [82] J.C. Williams, A.L. Armesilla, T.M. Mohamed, C.L. Hagarty, F.H. McIntyre, S. Schomburg, A.O. Zaki, D. Oceandy, E.J. Cartwright, M.H. Buch, M. Emerson, L. Neyses, J. Biol. Chem. 281 (2006) 23341–23348. [83] M.H. Buch, A. Pickard, A. Rodriguez, S. Gillies, A.H. Maass, M. Emerson, E.J. Cartwright, J.C. Williams, D. Oceandy, J.M. Redondo, L. Neyses, A.L. Armesilla, J. Biol. Chem. 280 (2005) 29479–29487.

[84] C. Linde, F. Di Leva, T. Domi, S.C.E. Tosatto, M. Brini, E. Carafoli, Cell Calcium, in press. [85] A. Rimessi, L. Coletto, P. Pinton, R. Rizzuto, M. Brini, E. Carafoli, J. Biol. Chem. 280 (2005) 37195–37203. [86] V. Prasad, G.W. Okunade, M.L. Miller, G.E. Shull, Biochem. Biophys. Res. Commun. 322 (2004) 1192–1203. [87] L. Liu, Y. Ishida, G. Okunade, G.E. Shull, R.J. Paul, Am. J. Physiol. Cell Physiol. 290 (2006) C1239–C1247. [88] R.A. Dumont, U. Lins, A.G. Filoteo, J.T. Penniston, B. Kachar, P.G. Gillespie, J. Neurosci. 21 (2001) 5066–5078. [89] W.L. Stahl, T.J. Eakin, J.W. Owens Jr., J.F. Breininger, P.E. Filuk, W.R. Anderson, Brain Res. Mol. Brain Res. 16 (1992) 223–231. [90] Q.Y. Zheng, K.R. Johnson, Hear Res. 154 (2001) 45–53. [91] J.M. Schultz, Y. Yang, A.J. Caride, A.G. Filoteo, A.R. Penheiter, A. Lagziel, R.J. Morell, S.A. Mohiddin, L. Fananapazir, A.C. Madeo, J.T. Penniston, A.J. Griffith, N. Engl. J. Med. 352 (2005) 1557–1564. [92] D.P. Corey, M. Sotomayor, Nature 428 (2004) 901–903. [93] E.N. Yamoah, E.A. Lumpkin, R.A. Dumont, P.J. Smith, A.J. Hudspeth, P.G. Gillespie, J. Neurosci. 18 (1998) 610–624. [94] J.D. Wood, S.J. Muchinsky, A.G. Filoteo, J.T. Penniston, B.L. Tempel, J. Assoc. Res. Otolaryngol. 5 (2004) 99–110. [95] J.A. Assad, G.M. Shepherd, D.P. Corey, Neuron 7 (1991) 985–994. [96] P.J. Kozel, R.A. Friedman, L.C. Erway, E.N. Yamoah, L.H. Liu, T. Riddle, J.J. Duffy, T. Doetschman, M.L. Miller, E.L. Cardell, G.E. Shull, J. Biol. Chem. 273 (1998) 18693–18696. [97] V. Prasad, G. Okunade, L. Liu, R.J. Paul, G.E. Shull, Ann. NY Acad. Sci. 1099 (2007) 276–286. [98] T.A. Reinhardt, J.D. Lippolis, G.E. Shull, R.L. Horst, J. Biol. Chem. 279 (2004) 42369–42373. [99] A. Kuhn, D.R. Goldstein, A. Hodges, A.D. Strand, T. Sengstag, C. Kooperberg, K. Becanovic, M.A. Pouladi, K. Sathasivam, J.H. Cha, A.J. Hannan, M.R. Hayden, B.R. Leavitt, S.B. Dunnett, R.J. Ferrante, R. Albin, P. Shelbourne, M. Delorenzi, S.J. Augood, R.L. Faull, J.M. Olson, G.P. Bates, L. Jones, R. Luthi-Carter, Hum. Mol. Genet. 16 (2007) 1845–1861. [100] T.P. Keeton, G.E. Shull, Biochem. J. 306 (Pt 3) (1995) 779–785. [101] K. Schuh, E.J. Cartwright, E. Jankevics, K. Bundschu, J. Liebermann, J.C. Williams, A.L. Armesilla, M. Emerson, D. Oceandy, K.P. Knobeloch, L. Neyses, J. Biol. Chem. 279 (2004) 28220–28226. [102] D.A. Zacharias, S.J. Dalrymple, E.E. Strehler, Brain Res. Mol. Brain Res. 28 (1995) 263–272. [103] S.E. Burk, G.E. Shull, J. Biol. Chem. 267 (1992) 19683–19690. [104] M. Brini, R. Marsault, C. Bastianutto, J. Alvarez, T. Pozzan, R. Rizzuto, J. Biol. Chem. 270 (1995) 9896–9903.