The plasma membrane Ca2+-ATPase-2 (PMCA2) is expressed in stereocilia of hair cells of the inner ear, whereas PMCA1 is expressed in the basolateral plasma membrane of hair cells. Both extrude excess Ca2+ from the cytosol. They are predicted to contain ten membrane-spanning segments, two large cytoplasmic loops as well as cytosolic N- and C-termini. Several isoform variants are generated for both PMCA1 and PMCA2 by alternative splicing, affecting their first cytosolic loop (A-site) and their C-terminal tail. To understand how these isoforms are differentially targeted in hair cells, we investigated their targeting regions and expression in hair cells. Our results show that a Leu-Ile motif in `b'-tail splice variants promotes PMCA1b and PMCA2b basolateral sorting in hair cells. Moreover, apical targeting of PMCA2 depends on the size of the A-site-spliced insert, suggesting that the conformation of the cytoplasmic loop plays a role in apical targeting.
The plasma membrane Ca2+-ATPases-1 and -2 (PMCA1 and PMCA2) are the major Ca2+ pumps of mammalian sensory hair cells (HCs). PMCA1 is located in the HC basolateral membrane, whereas PMCA2 is localized to the mechanosensory stereocilia at the apical pole of HCs (Dumont et al., 2001) (for a review, see Guerini et al., 2005). Both isoforms play major roles in Ca2+ extrusion from HCs to maintain a low cytosolic Ca2+ concentration. PMCA2 also contributes to the creation of a high extracellular Ca2+ concentration around the hair bundle and maintains the Ca2+ concentration of the endolymph (Wood et al., 2004). PMCA2 activity is required for normal hearing (Yamoah et al., 1998), and a mutation (G283S) in the mouse PMCA2 gene causes recessively inherited deafness (Street et al., 1998). A hypofunctional variant of PMCA2 was found in the heterozygous state associated with increased loss of hearing in human patients carrying homozygous mutation in CDH23 (Schultz et al., 2005). PMCA1 and PMCA2 display several splice variants due to the presence of two alternative splicing sites located in the first cytoplasmic loop (A-splice site) and in the C-terminal tail (C-splice region; Fig. 1) (for a review, see Strehler and Zacharias, 2001).
Hair cells of the inner ear are polarized epithelial cells with distinct apical and basolateral plasma membrane (PM) domains that contain different protein and lipid compositions. The apical pole contains a highly specialized stereociliary hair bundle that includes the mechanotransduction apparatus for auditory signals. The basolateral domain contains the synaptic contacts of both inner and outer HCs and, for outer HCs, it also contains the machinery for HC elongation and contraction (Kalinec et al., 1992). Since the apical and basolateral membranes are separated by junctions that do not allow protein and lipid flow, proteins found in only one domain, such as PMCA1 (basolateral) and PMCA2 (apical), require separate delivery systems. Although little is known about protein targeting in HCs, it has been studied extensively in other polarized epithelial cells.
Polarized protein distribution results from sorting mechanisms operating in the biosynthetic and recycling pathways that recognize specific sorting signals in proteins. Structural features, believed to operate as apical sorting signals include glycosylphophatidylinositol anchors (Lisanti et al., 1989; Ikonen, 2001), N-glycans (Scheiffele et al., 1995), O-glycans (Yeaman et al., 1997; Jacob et al., 2000), and protein sequences in the transmembrane domains (TMs) (Kundu et al., 1996; Lin et al., 1998), cytoplasmic domains (Chuang and Sung, 1998; Rodriguez-Boulan and Gonzalez, 1999; Nelson and Yeaman, 2001; Klapper et al., 2006) or extracytoplasmic loop (Qi et al., 2005). However, basolateral signals are typically formed by short, cytoplasmic peptide sequences (Le Gall et al., 1995; Yeaman et al., 1999; Mostov et al., 2000). Among the best-studied basolateral signals are Tyr-based motifs (consensus motif NPxY or YxxF) as well as Leu-Leu and di-hydrophobic residues (Matter et al., 1992; Aroeti et al., 1993; Hunziker and Fumey, 1994; Simonsen et al., 1997; Bonifacino and Dell'Angelica, 1999; Rodionov et al., 2000) or a single Leu residue (Deora et al., 2004).
The high sequence homology among the PMCA isoforms raises the interesting question of how they are selectively targeted to separate domains in polarized cells. A recent study showed that alternative splicing plays a major role in the targeting of PMCA2 in Madin-Darby canine kidney type-I (MDCK-1) epithelial cells. Chicka and Strehler showed that the A-splice-site insert regulates its targeting to the apical membrane (Chicka and Strehler, 2003). These authors showed that the size of the insertion at the A-splice site of PMCA2 controls its apical targeting, regardless of the C-terminal cassette; a 45-amino-acid (aa) insert at this site (the w-form) resulted in apical targeting, whereas smaller inserts resulted in basolateral targeting.
In this study we investigated the targeting of PMCA1 and PMCA2 in HCs of the rat inner ear by transfecting GFP constructs of wild-type and mutant PMCA cDNAs into cultures of organs of Corti. We focused on the most divergent areas of the molecules corresponding to the first cytoplasmic loop and the alternatively spliced cassettes of the C-terminal. We find that the A-splice-site insert plays a key role in PMCA apical targeting. In addition, the C-splice-site insert influences their basolateral targeting.
Isolation of rat inner ear PMCA2 cDNA clones
A previous study on the sacculus of the bullfrog showed that the predominant PMCA2 splice form is 2va; however, this form is bullfrog specific and not found in mammals. Even though PMCA2 isoforms that contain the `w' and `z' inserts and the b-tail and c-tail were detected, they were only present at low abundance (Dumont et al., 2001). The major PMCA1 isoform in bullfrog is the 1xb variant (Dumont et al., 2001). To identify the forms of PMCA expressed in the mammalian organ of Corti, we carried out reverse transcriptase (RT)-PCR with primers that amplify all A and C splice forms on total RNA isolated from the organ of Corti from rat. Cloning the RT-PCR products relevant to the C-splice site, and sequencing more than 100 recombinant clones, showed only a 543-bp-sized insert corresponding to the a-tail variant of PMCA2 (Fig. 2). Similar experiments for the A-splice site showed two inserts of 627 and 762 bp corresponding to the z-loop- and w-loop-variants (Fig. 2). Gel electrophoresis of the PCR products confirmed the presence of the a-, w- and z-forms and showed the relative abundance of these forms, with the w-form predominating. These results suggest that 2wa and 2za are the major splice forms of PMCA2 present in the organ of Corti from rat. To verify that both of these inserts are present in endogenous PMCA2 from HCs, we used primers to amplify the entire molecule. Thirteen recombinant clones containing full-length PMCA2 cDNA were obtained; of these, 12 were the PMCA2wa form and one was the PMCA2za form (data not shown). Although not detected, we cannot rule out low levels of expression of the other splice forms, such as those reported by Dumont et al. for the sacculus of the bullfrog (Dumont et al., 2001).
We also carried out RT-PCR on total RNA with primers that would result in amplification of all A- and C-splice forms of PMCA1. For both splice-site RT-PCRs, two single bands of 519 bp and 409 bp, corresponding to the A-splice x-loop variant and to the C-splice b-tail variant, respectively, were isolated, demonstrating that PMCA1xb is the major variant transcribed in the inner ear.
To determine the expression patterns of PMCA1 and PMCA2 in HCs, we used antibodies directed specifically against their respective N-termini to label adult organs of Corti and vestibule explants from rat. Each antibody detects all splice variants of the corresponding PMCA. PMCA1 was exclusively detected in the basolateral PM of inner HCs (Fig. 2) and vestibular HCs (data not shown). It was also localized at low concentration to the basal PM of outer HCs of only the basal turn of the cochlea (data not shown). No traces of PMCA1 were detected in the stereocilia of any type of HCs. PMCA2, however, is consistently found exclusively at the apical PM of outer HCs (Fig. 2), the inner HCs of the apex of the cochlea (data not shown), and vestibular type-I and type-II HCs (Fig. 2). We were unable to detect PMCA2 in the basolateral PMs of vestibular, outer or inner HCs, which indicates that PMCA2 is expressed exclusively in the apical PM of HCs.
Different targeting profiles of PMCA2 isoforms in rat hair cells
In MDCK cells, GFP N-terminally-tagged PMCA2 isoforms were shown to be correctly folded and to behave identically to the non-tagged isoforms (Chicka and Strehler, 2003). We used GFP-tagged PMCA2 constructs in our studies to identify the sorting motifs in its splice variants. We began by examining the A-splice site forms wa and za, which we found to be present in the inner ear based on RT-PCR results.
PMCA2wa contains all three exons in the A-splice site and the a-tail aa sequence in its C-terminus (Fig. 1). It is targeted exclusively to the apical stereociliary hair bundle of HCs (Fig. 3A). Deletion of the a-tail from the PMCA2wa isoform does not alter its exclusive apical targeting (Fig. 3D). However, deletion of all three A-splice-site exons, as found in PMCA2za, leads to a non-polarized distribution to both basolateral and apical PMs (Fig. 3B). This same nonpolarized distribution was seen when we deleted the a-tail for PMCA2za (PMCA2za-del a) (Fig. 3C). This indicates that the presence of the three exons in the first cytosolic loop, as found in PMCA2wa, is crucial for the selective targeting of PMCA2 to the apical PM, whereas the a-tail does not influence targeting.
We next examined the role of the b-tail in PMCA2 targeting by appending the b splice form to PMCA2z, which lacks an apical A-splice site insert. This PMCA2zb isoform is targeted exclusively basolaterally in HCs (Fig. 3E). This basolateral signal is not dominant over the A-splice-site signal however, because a construct containing both the cytosolic loop exons and the b-tail (PMCA2wb) is delivered to both the apical and basolateral PMs (Fig. 3F). These results suggest that the b-tail contains a signal that results in a vectorial basolateral targeting in the absence of A-splice site inserts. In the presence of A-splice site inserts, targeting is to both poles suggesting a competition between the apical and basolateral signals present in A-splice site inserts and the b-tail, respectively.
We were also interested in studying the targeting properties of PMCA1 in HCs; PMCA1 has over 85% aa sequence homology with PMCA2. The wild-type PMCA1xb, which is the only isoform detected by RT-PCR in the inner ear, is targeted exclusively basolaterally in HCs (Fig. 4B). The b-tail-deleted PMCA1xb variant is targeted to both apical and basolateral PM of HCs (Fig. 4C), suggesting a crucial role of the b-tail for the vectorial basolateral targeting of this pump. This demonstrates that the b-tails of both PMCA1 and PMCA2, which show a high degree of aa homologies (Fig. 1D), encompass the molecular determinant for their targeting. Given the high degree of homology between the two b-tails, we have focused our mutagenesis study on the PMCA2xb-encoding construct to identify the motif(s) necessary for basolateral targeting.
A Leu-Ile motif is required for vectorial basolateral targeting of PMCA1 and PMCA2 in HCs
To identify the crucial aa-residues for the basolateral targeting of PMCA2 present in the b-tail (103 aa), we used the isoform PMCA2xb, which traffics identically to PMCA2zb and to PMCA1xb, and is targeted exclusively to the basolateral PM of HCs (Fig. 6B) and MDCK-I cells (supplementary material Movie 1), to perform site directed mutagenesis experiments on the b-tail amino-acid sequence. We studied the role of the potential AP-adaptor-binding tyrosine-based motif YEGL (at positions 1121-1124) and the C-terminal PDZ-domain-interacting motif ETSL (at positions 1209-12), in the basolateral targeting of PMCA2xb. We generated single mutants 2xb-Y1121A and 2xb-del ETSL[1209-12], and the double mutant 2xb-Y1121A-del ETSL and imaged their location in HCs. All three variants were exclusively basolaterally targeted (Fig. 5A-C), showing that these two motifs do not influence the basolateral targeting of PMCA2. We then performed serial deletions in the b-tail of the PMCA2xb isoform. Deletion of the 46 C-terminal aa of the b-tail in PMCA2xb wild type did not alter the targeting of the resulting variant 2xb-del46[1167-ter] in HCs (Fig. 6C). However, the deletion of 65 aa of the C-terminal (1148-ter) of PMCA2xb – variant 2xb-del65[1148-ter] – resulted in the delivery to both apical and basolateral PMs (Fig. 6D). The variant PMCA2za-del a gave similar results (Fig. 3). The deletion of 14 aa at positions 1153-1166 also altered the targeting of PMCA2xb – the variant 2xb-del14[1153-66] (Fig. 6E) – and the variant 2xb-Y1121A-del14-delETSL (Fig. 5D). Thus, these 14 aa are crucial for targeting PMCA2xb exclusively basolaterally. Partial deletion of this region showed that the seven-aa sequence Leu-Ile-Asp-Asp-Thr-Asp-Leu (at positions 1153-1159) was necessary for the exclusively basolateral localization (Fig. 6F). The adjacent downstream sequence Glu-Glu-Asp-Ala-Ala-Leu-Lys did not appear to be important for the targeting of the pump, because the deletion of these residues did not change the exclusively basolateral delivery of PMCA2xb (Fig. 6G). A closer look at aa-sequence alignment of the b-tail of rat and human PMCA1, PMCA2, PMCA3 and PMCA4, showed that three hydrophobic residues, Pro-Leu-Ile at position 1152-1154 in PMCA2xb, are fully conserved among PMCA1, PMCA2 and PMCA3, and are Pro-Leu-Leu in human PMCA4 (Fig. 7C). Leu-Leu or Ile-Leu motifs have been found to act as basolateral targeting motifs in several proteins including E-Cadherin, CD4, GLUT4 and TRP-1 (for a review, see Bonifacino and Traub, 2003). The conserved residues Pro-Leu-Ile (1183-1185) might play identical roles in all PMCAs. Pro1152Ala substitution in wild-type PMCA2xb did not alter the basolateral targeting of the protein (Fig. 7D). However, Leu1153Ala and Ile1154Ala substitutions in wild-type PMCA2xb altered the basolateral targeting of the protein in HCs (Fig. 7E,F). However, the substitution of the downstream acidic aa residues into Ala did not alter the localization of PMCA2xb (Fig. 7H). Moreover, the single substitution of Leu1165Ala, which is similar to the CD147 basolateral targeting motif (see aa alignment in Fig. 7B), did not alter the exclusively basolateral targeting of the pump (Fig. 7G). Combining both substitutions also did not alter the targeting (Fig. 7I). Together, these observations allowed us to conclude that residues Leu1153-Ile1154 are necessary for the targeting of PMCA2xb. These hydrophobic residues are conserved in all mammalian PMCA b-tails (see aa alignment in Fig. 7C) and are likely to be necessary for the localization of these members of the PMCA family.
The b-tail basolateral targeting information is transposable to other transmembrane proteins and is recognized by LLC-PK1 epithelial cells
Different types of epithelial cells vary widely in the final distribution of PM proteins or in the pathways that these proteins follow to the cell surface (Keller and Simons, 1997). In the pig-kidney epithelial (LLC-PK1) cell line, some basolateral membrane proteins exhibit a defect in targeting, resulting in their appearance at the apical surface (Roush et al., 1998; Folsh et al., 1999; Gan et al., 2002). This defect is due to the absence of the μ1B subunit of the adaptin AP1B in these cells (Ohno et al., 1999; Roush et al., 1998; Folsh et al., 1999). We verified the absence of the μ1B subunit by visualizing the apical endocytosis of labelled-transferrin in polarized LLC-PK1 cells (data not shown). Although AP1B has a preference for a consensus Tyr motif through the μ1 subunit (Ohno et al., 1998), it has been shown that the β subunit (Rapoport et al., 1998), and/or γ-σ1 subunit combination (Janvier et al., 2003) of AP1 can bind to Leu-Leu motifs. Alternatively, an accessory protein like PACS-1 might be acting as a piggy-back linker (Rapoport et al., 1998; Wan et al., 1998). Moreover, AP1B has been implicated in targeting cargo in the biosynthetic pathway as well as in the recycling route (Folsh et al., 1999; Gan et al., 2002). In the case of PMCA, transfected LLC-PK1 cells exhibited exclusively basolateral polarity for PMCA2xb (Fig. 8A) and PMCA1xb (data not shown). Thus, the sorting machinery responsible for basolateral sorting of PMCA1xb and PMCA2xb is present in both HCs and LLC-PK1 cells and recognizes LI(1153-54).
To test the transposability of the basolateral targeting information in PMCA1 and PMCA2 b-tails, we transfected LLC-PK1 cells with constructs encoding the interleukin-2 receptor α-chain (Tac) – which is 200 aa long and has a single C-terminal TM domain (GenBank CAH73597) – and Tac-2b-tail chimera (encoding Tac fused C-terminally to the PMCA2 b-tail aa sequence). Tac was predominantly targeted to the apical PM of LLC-PK1 (Fig. 8B), MDCKI and HCs (data not shown). However, Tac-2b-tail chimera exhibited exclusively basolateral polarity in LLC-PK1 (Fig. 8C). However, the Tac-2b-tail chimera with Leu1153Ala and Ile1154Ala substitutions was predominantly apically targeted in LLC-PK1 (Fig. 8D), similar to the PM localization of Tac alone. These observations support the transposability of the basolateral targeting information encompassed in PMCA1 and PMCA2 b-tails.
Apical targeting of b-tail lacking PMCA2 isoforms is determined by the size of their first cytoplasmic loop
We showed by RT-PCR that PMCA2wa, lacking the b-tail, is the major PMCA2 isoform detected in the inner ear (Fig. 2). This isoform has the A-splice site cassettes that play a dominant role in its exclusive apical targeting in HCs (Fig. 3) and MDCKI cells (supplementary material Movie 2). Thus, we used the cDNA construct encoding this isoform fused to EGFP and generated a series of mutations in the alternatively spliced cassettes of the first cytoplasmic loop to identify the crucial determinants for its exclusively apical targeting. Mutations included constructs encoding EGFP-tagged PMCA2ya by deleting exon 3, PMCA2xa by deleting exon 1 and exon 2, PMCA2wa-del(ex1) by deleting exon 1, PMCA2wa-del(ex2) by deleting exon 2 and PMCA2wa-del(ex1-ex3) by deleting exon 1 and exon 3. In HCs, only PMCA2wa, PMCA2ya and PMCA2wa-del(ex1) were targeted exclusively to the apical PM (Fig. 9). The A-splice-insert sizes are 45, 31 and 34 aa for 2wa, 2ya and 2wa-del(ex1), respectively. Isoforms 2za, 2wa-del(ex2) and 2wa-del(ex1-ex3) with insert sizes of 0, 25 and 20 aa, respectively, were targeted to both apical and basolateral poles (Figs 3, 9). All isoforms targeted exclusively to the apical pole contained exon 2, suggesting that exon 2 contains sufficient information for apical targeting. However, the presence of exon 2 alone was insufficient to exclusively target the molecule to the apical pole, because 2wa-del(ex1-ex3) was targeted to the apical and the basolateral pole (Fig. 9). To further investigate the role of exon 2 in targeting, we generated constructs encoding PMCA2wa lacking nine aa in exon 2 [2wa-del(9-ex2)] and PMCA2wa lacking 11 aa in exon 2 [2wa-del(11-ex2)]. We also substituted the sequence of exon 2 in the PMCA2wa isoform with a completely random aa sequence to generate the isoform 2wa-(ex2-sub). All of the generated isoforms were targeted exclusively to the apical PM (Fig. 9), indicating that no specific apical targeting motif is present in exon 2. Rather, our results suggest that the size of the A-splice-site insert is an important factor for apical targeting. None of the constructs eliminated apical targeting completely, indicating that the absence of an A-splice-site insert results in the delivery to both poles – as seen for 2za (Fig. 3) – which might occur by default.
Our results suggest that A-splice-site inserts are not directly responsible for targeting of PMCA2 in HCs but can influence the targeting. Therefore, the targeting signals might reside elsewhere on the first cytoplasmic loop or the bordering TM2 and TM3. Thus, we investigated the role in apical trafficking of two known domains in the first cytoplasmic loop of PMCA2. We generated a 12 aa deletion in the conserved region of the phospholipid-binding site (G-domain), which is located downstream of the A-splice site (Filoteo et al., 1992). The encoded isoform, 2wa-del12(380-91) was targeted exclusively apically (Fig. 10), showing that the targeting of PMCA2 is independent of the G-domain. We also reproduced the Gly283Ser mutation that was detected in the deafwaddler (dfw) mutant mouse and is located upstream of the A-splice site. The 2wa-dfw(G283S)-encoded protein was targeted exclusively apically (Fig. 10), showing that in HCs, this spontaneous mutation does not alter the apical targeting properties of PMCA2.
To better understand the mechanisms by which proteins are selectively targeted in hair cells of the mammalian inner ear, we investigated the trafficking of two relatively abundant proteins, PMCA1 and PMCA2. These are highly homologous proteins but have very different distribution patterns in hair cells, with PMCA1 confined to the basolateral membrane and PMCA2 to the apical membrane. We focused on the areas of the molecules that are most divergent, the alternatively spliced cassettes of the first intracellular loop and the C-terminus. Using GFP-tagged constructs of PMCA1 and PMCA2 and transfection of hair cells in explant cultures, we show that both sites play roles in determining apical or basolateral destinations.
PMCA1 and PMCA2 isoforms are expressed in rat HCs
A previous study on the sacculus of the bullfrog showed that the predominant PMCA2 splice form is 2va, a form not found in mammals, and the prominent PMCA1 is 1xb (Dumont et al., 2001). Using RT-PCR we found that two isoforms of PMCA2, PMCA2wa and PMCA2za, are the predominant forms expressed in the inner ear of the rat. Our semi-quantitative analysis showed that PMCA2wa is the more abundant isoform. We detected a single form of PMCA1, PMCA1xb. Expression of GFP-tagged constructs of these isoforms reproduced the patterns of distribution obtained with antibodies specifically recognizing PMCA1 and PMCA2. Whereas GFP-PMCA2wa was found only in the apical membrane, GFP-PMCA2za was present in both apical and basolateral membranes. The absence of immunochemical staining in the basolateral membrane with anti-PMCA2 antibodies reflects the very low abundance of the PMCA2za isoform, supporting our RT-PCR results. Alternatively, PMCA2za-detected mRNA could be either silenced in HCs or expressed by cochlear and vestibular ganglion neurons that are a potential contamination of the sensory epithelia preparation from which we extracted total mRNA.
A Leu-Ile motif in the b-tail is a basolateral signal for PMCA1 and PMCA2
The presence of the b-tail influences the trafficking of both PMCA1 and PMCA2. Deletion of the b-tail of PMCA1 leads to its targeting to both apical and basolateral membranes, replacement of the a-tail with the b-tail on PMCA2wa also results in its targeting to the apical and the basolateral membrane. One potential motif in the b-tail that could influence trafficking is the PDZ-binding domain at its C-terminus. PDZ interactions are known to influence the trafficking of many proteins (Cohen et al., 1998; Simske et al., 1996; Rongo et al., 1998). However, deletion of the last four aa of the b-tail of PMCA2xb did not alter its vectorial basolateral targeting. However, deletion of the entire b-tail of PMCA1xb, which is targeted exclusively to the basolateral membrane, resulted in targeting to both poles, showing that signal(s) in the b-tail play a role in basolateral targeting. Presumably, elimination of this signal results in a default mode of trafficking to both membranes. A second motif we investigated is a tyrosine motif (YEGL), a consensus binding-site for the μ subunits of AP adaptors (for a review, see Bonifacino and Traub, 2003). For the basolateral delivery, YxxΦ basolateral signals contained in PM proteins interact with adaptors of the clathrin type (AP1B) or non-clathrin type (AP4) at the trans Golgi network or the recycling endosomes. Mutation of this site did not influence basolateral trafficking. A third site that might influence trafficking is Leu1151 located downstream of an acidic cluster. The Leu-acidic cluster is homologous to the motif responsible for basolateral targeting of CD147 (Fig. 7) (Deora et al., 2004). This motif is functional in various epithelia including retinal pigment epithelia (RPE) of newborn rats (Marmorstein et al., 1996).
Through a deletion analysis, we identified a di-hydrophobic (Leu-Ile) motif that is conserved as a basolateral targeting motif in all rat and human PMCA1, PMCA 2 and PMCA3 b-splice variants. In PMCA4b, this di-hydrophobic motif is also well-conserved as Phe-Leu in rat, and as Leu-Leu sequence in human (see Fig. 7C). This di-hydrophobic motif has also been described as a basolateral targeting signal in many other proteins, including furin (for a review, see Bonifacino and Traub, 2003). The motif in PMCA is similar to the one identified for the basolateral targeting of E-cadherin (Fig. 7A) and conserved in many other members of the cadherin family (Miranda et al., 2001). Almost all known Leu-Leu-type motifs fit the consensus motifs (DE)xxxL(LI) and DxxLL (for a review, see Bonifacino and Traub, 2003). However, neither the sequences in the PMCAs, nor those in cadherins fit these consensus motifs. Leu-Leu motifs are generally found in an acidic context and are bordered by an acidic cluster, either N-terminally like in furin (Simmen et al., 1999), or C-terminally like in PMCAs and cadherins. In some cases, such as in furin, the acidic clusters (EEDE) located N-terminally to the Phe-Ile motif have been shown to be important for the function of the motif in basolateral targeting. However, we show that, in PMCA the acidic cluster located downstream of the dihydrophobic residues does not influence the basolateral targeting. One possible mechanism for the function of the Leu-Ile motif in basolateral trafficking involves Rab11 and the exocyst complex. Rab11 has been shown to regulate the basolateral delivery of wild-type Drosophila E-cadherin (dE-cadherin), which depends on a Leu-Leu motif (Lock and Stow, 2005). Rab11 and Armadillo interact with the exocyst components Sec15 and Sec10, respectively. Also, loss-of-function of Sec5, Sec6 and Sec15, components of the exocyst complex, results in the accumulation of dE-cadherin in a large, Rab11-recycling, endosomal compartment and inhibits the delivery of dE-cadherin to the PM (Langevin et al., 2005). These observations support a model whereby the exocyst regulates dE-Cadherin trafficking from recycling endosomes to sites on the epithelial cell membrane where Armadillo is located. A similar mechanism could regulate the trafficking of PMCA1 in HCs. A Leu-Leu motif has also been shown to regulate the trafficking of Shal K+ channels to dendrites (Rivera et al., 2003).
Apical targeting of PMCA2 depends on the alternatively spliced cassettes at the A-splice site
Targeting studies of GFP-constructs of the two PMCA2 isoforms PMCA2wa and PMCA2za detected in cochlea, suggested that the presence of the three alternatively spliced cassettes at the A-splice site plays a role in the apical targeting of PMCA2wa. PMCA2wa was targeted exclusively to the apical membrane whereas PMCA2za was delivered to both poles. Sequential deletion of the three alternatively spliced exons showed that PMCA2wa, PMCA2ya and PMCA2wa-del(ex-1) were all delivered exclusively to the apical membrane (Fig. 3A, Fig. 9D,B). However, larger deletions, such as in PMCA2wa-del(ex-2), PMCA2wa-del(ex1-ex3) and PMCA2za, were all delivered to both apical and basolateral membranes, suggesting that the size of the insert influenced the trafficking. This was confirmed by substituting an unrelated aa sequence with exon 2, in which case the product was – like PMCA2wa, which has the same-sized insert of 45 amino acids – delivered exclusively to the apical membrane.
The maximum number of aa for PMCA1, PMCA3 and PMCA4 A-splice-site cassettes is 13, 14 and 12 aa, respectively (for a review, see Strehler and Zacharias, 2001). We showed that in PMCA2, A-splice inserts of 31 aa or more are required to target the protein exclusively to the apical membrane (Fig. 9). Therefore, whatever the A-splice-site variant is, the b-tail containing isoforms of PMCA3 and (human) PMCA4 should be targeted exclusively basolaterally, as do PMCA1xb and PMCA2zb and PMCA2xb (Fig. 4B, Fig. 3E, Fig. 6B). The nature of the PMCA3 isoform, expressed in rat vestibular HCs and immunolocalized basolaterally (Dumont et al., 2001), has not been reported. RT-PCR data combined with our targeting results, showing the requirement of the b-tail for basolateral targeting, suggest that PMCA3b is the major splice form expressed in HCs (our unpublished observation).
Our finding that the size of the A-splice site, but not its composition, is important for apical targeting suggests that a particular motif, which controls targeting, is not present in the alternatively spliced cassettes. Rather, these results suggest that the size of the insert alters the conformation of the cytoplasmic loop and may affect the availability of motifs elsewhere in the loop. We investigated two possible sites in the loop for their implication in the apical targeting of PMCA2, the mutation site that causes deafness in the deafwaddler mouse and the phospholipid-binding domain. Mutation of either of these sites did not alter trafficking of PMCA2. Further investigations on PMCA2 apical targeting elements would include the remaining sequences in the loop as well as the bordering TMs.
Materials and Methods
Whole mount preparation of adult rat organs of Corti
Three-week to 6-week-old rats were anesthetized with CO2 and decapitated according to National Institutes of Health (NIH) guidelines. Cochleae were freshly dissected and fixed with 4% paraformaldehyde (PFA), injected into the membranous labyrinth through the round and oval windows and incubated for 30 minutes at room temperature. After removing the bone, the tectorial membrane was removed and organ of Corti pieces were isolated from different turns of the cochlea and processed for immunofluorescence.
Cloning of rat organ of Corti PMCA2 isoforms and cDNA constructs
Full-length cDNAs encoding PMCA2 isoforms PMCA2wa (1199 aa) and PMCA2za (1154 aa) expressed in rat inner ear sensory epithelia were obtained from total mRNA prepared from P2 to P15 rat vestibular neuroepithelia and organ of Corti sensory patches by using Triazol (Life Technologies, MD). Gene-specific reverse transcription (RT) was performed on DNase-treated mRNA with the primer 5′-GGGCCTCAGCTAAAGCG-3′ and MuMLV reverse transcriptase (GibcoBRL, CA), followed by a PCR reaction with the forward primer: 5′-atgtgaagcttg-GGTGATATGACCAACAGCGACTTTTAC-3′ and the reverse primer: 5′-gcaaggtacc-GCTAAAGCGACGTCTCCAGGCTGTGG-3′ using High-Fidelity Taq Polymerase (Invitrogen, CA) (lowercase sequences correspond to inserted flanking HindIII and KpnI restriction sites used for directional cloning in pEGFP-C2). The RT-PCR products were cloned into the pCR2.1 TA cloning vector (Invitrogen, CA), and their A-splice region and their 3′-termini sequences were sequenced to identify the PMCA2wa and PMCA2za isoforms.
To characterize the splicing events that occur in rat cochlea and vestibular hair cells, we randomly reverse-transcribed 5 μg of RNase-free DNase-I-treated total mRNA and PCR-amplified the cDNA regions encoding for the C-terminal tails with primers C-PMCA2-F 5′-CCGGGGCCTGAATCGGATCC-3′ and C-PMCA2-R 5′-GGATGGGGCTCCCTGGACTTG-3′ or C-PMCA1-F 5′-GATGTTGAAGAGATTGACCATGCCG and C-PMCA1-R 5′-CTTACAATCAGAGTGATGTTTCCAAACTATG, and part of the second intracellular loop, including the A-splice-site cDNA sequence with primers A-PMCA2-F 5′-GAAAGCTCGCTCACAGGGGAG-3′ and A-PMCA2-R 5′-ACCCATGGTCTCACAGGCG-3′ or A-PMCA1-F 5′-GACGAGAGTTCGTTGACTGGCGAG and A-PMCA1-R: GATTGGCGTGCACTCAGCAAGCCACGG. After 25 cycles of amplification of the PMCA1 and PMCA2 C-splice-site regions, and PMCA1 A-splice-site region, samples of the RT-PCR product were loaded on a 2%-agarose gel containing ethidium bromide. For cDNA amplicons of the PMCA2 A-splice-site region, 4 μl of each sample was tested after 15, 18, 21 and 25 cycles of amplification, and emitted fluorescence under UV-light excitation was collected on digital images (NIH). cDNA amplicons were randomly cloned into the pCR2.1 TA cloning vector (Invitrogen, CA), and more than a hundred clones for each RT-PCR product were sequenced for their content using an ABI-3100 Avant (Applied Biosystems, CA).
The full-length human PMCA1xb coding sequence was assembled from PCR fragments following a strategy described earlier, and cloned into the expression vector pMM2 (Adamo et al., 1992). The entire cDNA coding from residue Gly2 to the C-terminal residue Leu1220 was amplified from this construct with primers SacI-hPMCA1-F 5′-ggcggcgcggagctcGGGCGACATGGCAAACAACTCAG and KpnI-hPMCA1-R 5′-ggccggtaccTCAGAGTGATGTTTCC, and directionally cloned into the pEGFP-C2 vector (Clontech, CA). Sequencing of the full-length cDNA revealed two aa substitutions in the first cytoplasmic loop of the encoded proteins (M267R and N281R) and two aa substitutions in the second cytoplasmic loop (V588I and K747R) (supplementary material Fig. S1). The imaged HCs (Fig. 4), which expressed EGFP-tagged PMCA1xb and PMCA1xb-del b were transfected with constructs encoding proteins containing the substitutions. The reversion of both substitutions in the first cytoplasmic loop did not change the targeting properties of the encoded proteins (data not shown). EGFP-tagged PMCA2wa and PMCA2za were generated by directional cloning of inserts from clones described above into HindIII and KpnI cloning sites of the pEGFP-C2 vector (Clontech, CA). The construct encoding EGFP-tagged human PMCA2xb (1212 aa) has been previously described (Chicka and Strehler, 2003). Chimeric fusions between cDNAs encoding either the a or b C-terminal splice variants (59aa or 103 aa, respectively) and the cDNA encoding for TAC were generated in a pCDNA3 expression vector (Invitrogen, CA) using standard PCR methods. A QuickChange site-directed mutagenesis kit (Stratagene, CA) was used to introduce stop codons, deletions, and substitutions in the EGFP-tagged PMCA1xb, PMCA2xb, PMCA2za and PMCA2wa, as well as TAC chimera. All constructs were verified by DNA sequencing.
Organotypic cultures of organ of Corti and vestibular sensory patches from rat, and ballistic transfection of hair cells
Organotypic cultures of organ of Corti and vestibular sensory epithelia from rat were prepared according to Sobkowicz et al. (Sobkowicz et al., 1993), and modified as described by Rzadzinska et al. (Rzadzinska et al., 2004). Postnatal day 0-4 rat pups were anesthetized with CO2 and decapitated according to NIH guidelines; their temporal bones were isolated and placed into Leibovitz L-15 media (GIBCO, CA). The dissected organ of Corti was split into two to four pieces for culturing, and the otoconia were removed from above the vestibular sensory epithelia. Tissues were attached to dishes with coverslips in the bottom that had been coated with CellTak (BD Biosciences, MD) and placed in culture dishes (MatTek, MA). Cultures were incubated at 37°C and 5% CO2 in DMEM F-12 media supplemented with 5% fetal bovine serum (FBS) containing 10 μg/ml ampicillin (GIBCO, CA). Explants were maintained in culture for 1-7 days. The inner ear explants were transfected with 1 μm gold particles coated with plasmid DNA, by using a Helios Gene Gun (BioRad Laboratories, CA). Explants were fixed with 4% paraformaldehyde (PFA) in PBS 24 hours after transfection, permeabilized with 0.5% Triton X-100 in PBS and counterstained for actin with Rhodamine-coupled phalloidin for 30 minutes. The specimens were finally mounted with Antifade Mounting medium (Molecular Probes, CA).
Cell culture and transient transfections of MDCK-I and LLC-PK1 cells
LLC-PK1 (CL-101) and MDCKI (NBL-2; CCL-34) cells were obtained from American Type Culture Collection (ATCC; VA). We used a standard transfection technique with Lipofectamine-2000 as described by the manufacturer (Invitrogen, CA). Briefly, MDCK cells were grown on glass coverslips and LLC-PK1 were polarized on 0.4 μm TransWells in DMEM medium (Life Technologies, MD) supplemented with 10% serum. Cells were washed with serum-free DMEM 3 hours before transfection. DNA-Lipofectamine complex prepared in serum-free medium was added to the apical side. Cells were incubated with DNA-Lipofectamine for 6 hours at 37°C, 5% CO2. After 6 hours, the medium was replaced by regular DMEM serum-containing medium. Monolayers were analyzed 24 hours after transfection by immunofluorescence.
Immunofluorescence and confocal microscopy
The following primary antibodies were used: rat anti-ZO-1 (1:300 dilution; Chemicon, CA), rabbit anti-PMCA2 (1:300 dilution; SWANT, Switzerland), rabbit anti PMCA1 (1:300 dilution; Abcam, MA), mouse anti-TAC (ATCC, VA), secondary Alexa Fluor-555 goat anti-rat, Alexa Fluor-468 rat anti-goat, or Alexa Fluor-468 goat anti-mouse (Molecular Probes, CA). Cell monolayers and cochlear cultures were fixed for 20 minutes with 4% PFA prepared in PBS and quenched with 50 mM NH4Cl in PBS (containing 1 mM CaCl2 and 1 m MMgCl2). Cells were then permeabilized with 0.5% Triton X-100 for 15 minutes and blocked with 10% FBS/PBS for 30 minutes. Primary antibodies were applied for 1 hour at room temperature. The specimens were washed three times for 15 minutes with PBS and incubated with fluorescent secondary antibodies. High-resolution confocal images (2048×2048 pixels) were taken at room temperature with an LSM510 Zeiss microscope with a 63× NA=1 oil objective (Carl Zeiss, NY). 3D-images were built from stacks of serial (0.8 μm) x-y (en face) confocal images (512×512 pixels) using Volocity software (Improvision, MA).
We thank L. A. Dunbar and R. S. Petralia (NIDCD/NIH) for their helpful discussions while assembling and writing the manuscript, and B. Kachar (NIDCD/NIH) and R. Weigert (NHLBI/NIH) for their critical reading of the manuscript. This work was supported by the Intramural Research Program of NIDCD. E.E.S. was supported by grant GM28835 from the NIH.