Homer2 Protein Regulates Plasma Membrane Ca2+-ATPase-mediated Ca2+ Signaling in Mouse Parotid Gland Acinar Cells*

Background: Homer proteins bind multiple Ca2+-signaling proteins to shape the Ca2+ signal by poorly understood mechanisms. Results: Homer2 regulates PMCA expression and activity in parotid acinar cells. Conclusion: Homer2 acts as a regulator of PMCA-mediated Ca2+ clearance. Significance: Inhibition of Ca2+ clearance by Homer2 further clarifies its role in Ca2+ signaling. Homer proteins are scaffold molecules with a domain structure consisting of an N-terminal Ena/VASP homology 1 protein-binding domain and a C-terminal leucine zipper/coiled-coil domain. The Ena/VASP homology 1 domain recognizes proline-rich motifs and binds multiple Ca2+-signaling proteins, including G protein-coupled receptors, inositol 1,4,5-triphosphate receptors, ryanodine receptors, and transient receptor potential channels. However, their role in Ca2+ signaling in nonexcitable cells is not well understood. In this study, we investigated the role of Homer2 on Ca2+ signaling in parotid gland acinar cells using Homer2-deficient (Homer2−/−) mice. Homer2 is localized at the apical pole in acinar cells. Deletion of Homer2 did not affect inositol 1,4,5-triphosphate receptor localization or channel activity and did not affect the expression and activity of sarco/endoplasmic reticulum Ca2+-ATPase pumps. In contrast, Homer2 deletion markedly increased expression of plasma membrane Ca2+-ATPase (PMCA) pumps, in particular PMCA4, at the apical pole. Accordingly, Homer2 deficiency increased Ca2+ extrusion by acinar cells. These findings were supported by co-immunoprecipitation of Homer2 and PMCA in wild-type parotid cells and transfected human embryonic kidney 293 (HEK293) cells. We identified a Homer-binding PPXXF-like motif in the N terminus of PMCA that is required for interaction with Homer2. Mutation of the PPXXF-like motif did not affect the interaction of PMCA with Homer1 but inhibited its interaction with Homer2 and increased Ca2+ clearance by PMCA. These findings reveal an important regulation of PMCA by Homer2 that has a central role on PMCA-mediated Ca2+ signaling in parotid acinar cells.

neuronal transcription activity and thereby regulate dendritic spine morphogenesis, synapse remodeling, and synaptic clustering of CNS neurons (9,(17)(18)(19). Previous research found that Homer1 regulates Ca 2ϩ influx by associating IP 3 Rs with transient receptor potential channels (12). Homer2 tunes GPCR stimulus intensity by regulating the regulator of G protein signaling proteins and phospholipase C␤-promoting guanosine triphosphatase by G␣ in pancreatic acinar cells. Moreover, Homer2 and Homer3 bind nuclear factor of activated T cells by competing with calcineurin in T lymphocytes (14,20). However, the role of Homer proteins in Ca 2ϩ signaling in nonexcitable cells remains poorly characterized.
In this study, the role of Homer2 on Ca 2ϩ signaling in parotid gland acinar cells was investigated using Homer2 knock-out (Homer2 Ϫ/Ϫ ) mice. We report that Homer2 interacts with PMCA (particularly PMCA isoform 4) in model systems and in vivo and that this interaction regulates PMCA activity. The findings suggest a mechanism by which Homer proteins can regulate PMCA expression and PMCA-mediated Ca 2ϩ efflux in parotid acinar cells. We suggest that by inhibiting transient receptor potential channel-mediated Ca 2ϩ influx (12) and by differentially modulating Ca 2ϩ extrusion by PMCA (present data), the Homers serve to protect the cells from Ca 2ϩ toxicity by facilitating cytosolic Ca 2ϩ clearance to limit the Ca 2ϩ signal duration.
Animals and Preparation of Parotid Acinar Cells-Wild-type (WT) and Homer2 Ϫ/Ϫ mice have been described previously (14). The life span of Homer2 Ϫ/Ϫ mice is similar to that of WT littermates. All animal protocols were performed according to institution guidelines. Mice were sacrificed by cervical dislocation. The cells were prepared from the parotids of WT and Homer2 Ϫ/Ϫ mice by limited collagenase digestion as described previously (21). After isolation, the acinar cells were resuspended in an extracellular physiologic salt solution, the composition of which was as follows (in mM): 140 NaCl, 5 KCl, 1 MgCl 2 , 1 CaCl 2 , 10 HEPES, and 10 glucose, adjusted to pH 7.4 and 310 mosM.
Cell Culture and DNA Transfection-HEK293 cells (Korean Cell Line Bank, South Korea) were maintained in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum (Invitrogen) and 100 units/ml penicillin and streptomycin in a 5% CO 2 incubator. Approximately 1-5 ϫ 10 5 cells were seeded onto coverslips in 60-mm dishes and then incubated in antibiotic-free medium. The next day, DNA was mixed with Lipofectamine reagent (Invitrogen) and Opti-MEM, incubated for 20 min at room temperature, and then added to the cell culture media. The transfected cells were assayed at 38 -48 h after post-transfection.

Measurement of Intracellular Ca 2ϩ Concentration ([Ca 2ϩ ] i )-Parotid acinar cells from WT and Homer2
Ϫ/Ϫ mice were loaded with 5 M Fura-2/AM (Teflabs Inc., Austin, TX) and 0.05% pluronic F-127 for 60 min in physiologic salt solution. Fura-2 fluorescence was measured at the appropriate excitation wavelengths (340/380 nm) and emission at 510 nm wavelengths (ratio ϭ F 340 /F 380 ) using a Molecular Devices (Downingtown, PA) imaging system. The emitted fluorescence was monitored with a charge-coupled device camera (Photometrics, Tucson, AZ) attached to an inverted microscope. Fluorescence images were obtained at 2-s intervals. All data were analyzed using MetaFluor software (Molecular Devices).
Immunoprecipitation-The immunoprecipitation procedure was modified from Shin et al. (22) and Kim et al. (23). In brief, parotid and HEK293 microsomes were prepared by homogenizing a minced parotid and harvested HEK293 in a buffer containing 20 mM Mops (pH 6.7 with KOH), 250 mM sucrose, 1 mM EDTA, 1 mM MgCl 2 , 10 mM benzamidine, and 0.2 mM PMSF. The homogenized samples were centrifuged at 400 ϫ g for 10 min. The supernatants were collected and centrifuged at 900 ϫ g for 10 min at 4°C. To avoid protein degradation by digestive enzymes, immunoprecipitation was initiated immediately after completion of microsomal preparation. Microsomes were lysed in a buffer containing 50 mM Tris (pH 6.8 with HCl), 150 mM NaCl, 3 mM EDTA, 2 mM EGTA, and 0.5% Triton X-100 supplemented with protease inhibitors. The lysates were cleared by centrifugation at 14,000 ϫ g for 15 min. About 150 l of the extract (300 g of protein/sample) was incubated with 10 l of anti-PMCA (5F10) or 10 l of anti-FLAG antibodies for 2 h by rocking at 4°C. Protein A/G-agarose (Thermo Fisher Scientific Inc., Waltham, MA) was added to each mixture, and rocking was continued overnight at 4°C. Protein A/G-agarose was pelleted at 1,000 ϫ g for 10 s, and the beads were quickly washed with cold PBS. The immunoprecipitated proteins were separated by SDS-PAGE and probed with anti-FLAG (M2, 1:1,000) by overnight incubation at 4°C.

Measurement of [Ca 2ϩ
] Efflux-To measure directly the rate of Ca 2ϩ efflux by PMCA, we measured the appearance of Ca 2ϩ in the external medium using the procedure published by Zhao et al. (24) with slight modifications. Intact parotid acini from WT and Homer2 Ϫ/Ϫ mice were washed once and then suspended in 1 ml of medium containing 140 mM NaCl, 5 mM KCl, 10 mM glucose, and 10 mM HEPES, pH 7.4, with NaOH. 100 l of cell suspension were added to 1.5 ml of a similar medium containing 2 M free acid Fura-2 in a cuvette. After initiation of fluorescence recording, 10 M EGTA was added to reduce the extracellular Ca 2ϩ concentration ([Ca 2ϩ ] o ) to ϳ100 nM. After establishing a baseline leak for ϳ1 min, the cells were stimulated with 1 mM carbachol. At the end of the experiment, the signals were calibrated by adding 1 mM CaCl 2 and then 1 mM MnCl 2 to the medium as described previously (24). The design of experiments for PMCA stimulation while measuring cytoplasmic Ca 2ϩ was as follows. The cells were stimulated with 1 mM carbachol and 100 M CPA for about 15-20 s to release most of the Ca 2ϩ from internal stores and inhibit the SERCA Ca 2ϩ pumps. Efflux by PMCA was initiated by inhibition of the muscarinic receptors with the antagonist 10 M atropine and removal of external Ca 2ϩ to prevent Ca 2ϩ influx. CPA was maintained to inhibit the SERCA pumps.
Measurement of Ca 2ϩ Uptake and Release from Internal Stores-IP 3 -mediated Ca 2ϩ release from internal stores was measured in SLO-permeabilized cells as described previously (24). Cells were washed with a high K ϩ (120 mM KCl, 20 mM NaCl, 10 mM glucose, and 10 mM HEPES, pH 7.4, with NaOH), Chelex-treated medium and added to Chelex-treated medium containing an ATP regeneration system (composed of 3 mM ATP, 5 mM MgCl 2 , 10 mM creatine phosphate, and 5 units/ml creatine kinase), a mixture of mitochondrial inhibitors, 2 M Fluo-3, and 3 mg/ml SLO (Difco). After the addition of cells, the concentration of free Ca 2ϩ in this medium was ϳ350 -400 nM. In this medium, the cells were permeabilized almost instantaneously so that Ca 2ϩ uptake into the ER could be measured immediately. Uptake of Ca 2ϩ into the ER was allowed to continue until [Ca 2ϩ ] in the medium was stabilized. Then IP 3 was added in increasing concentrations to measure the extent of Ca 2ϩ release and the potency of IP 3 in mobilizing Ca 2ϩ from the ER. Subsequent addition of 1 mM carbachol was used to monitor the receptor-evoked Ca 2ϩ release.
Data Analysis and Statistics-All numeric values are represented as the mean Ϯ S.E. Statistical significance, determined to be p Ͻ 0.05, was calculated using the Student's unpaired t test.

Homer2 Deletion Does Not Affect the Polarized Expression of IP 3 Rs, IP 3 -mediated Ca 2ϩ
Release, or SERCA Activity-Comparison of the receptor-evoked Ca 2ϩ signaling in wild-type (WT) and Homer2 Ϫ/Ϫ acini revealed the altered signaling in the Homer2 Ϫ/Ϫ acini. To further examine this phenomenon, we analyzed the expression and activity of key Ca 2ϩ transporters. IP 3 Rs are established binding partners of Homer proteins (10,12). Therefore, we first examined the localization and expression of Homer2 and IP 3 Rs in parotid acinar cells from WT and Homer2 Ϫ/Ϫ mice. In WT parotid acinar cells, Homer2, and IP 3 Rs, positive staining was primarily observed in the apical pole. As expected, Homer2 staining was not detected in Homer2 Ϫ/Ϫ acini, and the expression and localization of all IP 3 R isoforms remained unchanged in Homer2 Ϫ/Ϫ acini (Fig. 1A).
We next examined whether Homer2 deletion affects Ca 2ϩ uptake into the ER, as well as the activity of IP 3 Rs and their response to IP 3 . Parotid acinar cells from WT and Homer2 Ϫ/Ϫ mice were permeabilized with SLO within 10 -15 s, and the [Ca 2ϩ ] in the incubation media was reduced to 50 -80 nM within 2 min at 37°C by SERCA-mediated Ca 2ϩ uptake into the IP 3 -mobilizable pool, similar to the method used with pancre- atic acinar cells (14). The rate and extent of Ca 2ϩ uptake into the ER was similar in WT and Homer2 Ϫ/Ϫ cells, providing the first indication that, unlike in the pancreas (14), Homer2 dele-tion did not affect parotid acinar cell SERCA activity. The addition of increasing concentrations of IP 3 and the muscarinic agonist carbachol resulted in a similar increase in [Ca 2ϩ ] due to  Ca 2ϩ release from stores of SLO-permeabilized WT and Homer2 Ϫ/Ϫ cells (Fig. 1, B and C). Thus, no compensatory effects in expression, localization, and IP 3 R activity were observed in Homer2 Ϫ/Ϫ mice.
PMCA Expression Is Selectively Increased in Specific Tissues from Homer2 Ϫ/Ϫ Mice-Next, we examined the expression of the SERCA and PMCA pumps. Because both PMCA1 and PMCA4 are expressed in salivary gland cells, we used a pan anti-PMCA antibody, 5F10, which detects both isoforms, to determine PMCA expression. To verify this PMCA expression, we used the anti-PMCA antibody JA9, which is specific for PMCA4 (14,24). Notably, immunostaining experiments suggested increased PMCA expression in the apical region of Homer2 Ϫ/Ϫ parotid acinar cells, whereas the SERCA2 expression remained unaffected in these mice (Fig. 1D). To further analyze how Ca 2ϩ pump expression was affected by Homer2 deletion, we examined protein expression using Western blotting analyses and parotid membranes prepared from WT and Homer2 Ϫ/Ϫ parotid acinar cells. As shown in Fig. 2, A and B, PMCA expression in parotid acinar cells from Homer2 Ϫ/Ϫ mice increased significantly to 2.5 Ϯ 0.1-fold greater than parotid acinar cells from WT mice (n ϭ 4, p Ͻ 0.001). However, the expression of SERCA2b (1.2 Ϯ 0.2-fold of WT, n ϭ 4) remained unchanged in the parotid membranes. Similar results were obtained with submandibular gland membranes. Interestingly, however, opposite results were observed in pancreas membranes, suggesting a tissue-specific adaptive response to Homer2 deletion. Furthermore, the protein levels of the PMCA4 isoform were higher in Homer2 Ϫ/Ϫ parotid mem-branes compared with WT (3.7 Ϯ 0.6-fold of WT, n ϭ 4, p Ͻ 0.05, Fig. 2C).

Rate of [Ca 2ϩ ] efflux Is Increased in Homer2
Ϫ/Ϫ Parotid Cells-The major routes for Ca 2ϩ clearance in nonexcitable cells, such as parotid acinar cells, are Ca 2ϩ uptake into the ER by the SERCA pumps and Ca 2ϩ efflux across the plasma membrane by PMCA (25)(26)(27). The results shown in Figs. 1B and 2 suggest that Homer2 deletion does not affect SERCA expression and activity but increases the PMCA expression. To determine whether increased PMCA protein expression translates to increased PMCA activity in intact cells, we examined PMCAmediated Ca 2ϩ clearance in WT and Homer2 Ϫ/Ϫ parotid acini. In the first protocol, cells were stimulated with a high concentration of carbachol and treated with the SERCA inhibitor CPA to release ER Ca 2ϩ and maximally activate Ca 2ϩ influx to cause a large increase in cytoplasmic Ca 2ϩ . Ca 2ϩ clearance was then initiated by terminating cell stimulation with atropine while simultaneously inhibiting SERCA activity. Fig. 3A shows that the addition of 10 M atropine in a Ca 2ϩ -free solution resulted in an immediate clearance of Ca 2ϩ primarily by PMCA (Fig.  3A). Comparing the slope of Ca 2ϩ clearance revealed that Ca 2ϩ clearance in Homer2 Ϫ/Ϫ cells is 1.5-fold faster compared with WT cells (1.5 Ϯ 0.3-fold, n ϭ 4, p Ͻ 0.05, Fig. 3B). In a second protocol, we assayed PMCA activity by measuring the change in [Ca 2ϩ ] o in cells incubated in media with low external Ca 2ϩ concentration and stimulated with 1 mM carbachol. As shown in Fig. 3C, [Ca 2ϩ ] o increased significantly in response to carbachol stimulation in both cell types. Importantly, the change in Only Ab denotes control IP using antibodies without extract in the IP assay; ϪAb denotes control IP using extract without antibodies in the IP assay; and ϩAb denotes IP using extract and antibodies in the IP assay. C, co-IP of Homer proteins (FLAG-tagged) and PMCA (wild-type, PPXXF-like motif mutants (Mt_PMCA),or PMCA4 with deleted PDZ-binding motif (Del_PMCA4) co-transfected in HEK293 cells. Extracts prepared from the cells co-transfected with the respective FLAG-Homers and the indicated PMCA constructs were used to IP either PMCA with the 5F10 antibodies or the Homers with anti-FLAG antibodies, and the precipitates were probed with anti-FLAG antibodies to detect the Homers. Note that Mt_PMCAs reduced the interaction of PMCA with Homer2 but not the interaction of PMCA with Homer1a and Homer1c. However, the interaction of PMCA with Homer1a and Homer1c was decreased when the Del_PMCA4 mutant was used. [Ca 2ϩ ] o was ϳ1.5-fold higher in Homer2 Ϫ/Ϫ cells compared with WT cells (n ϭ 4, p Ͻ 0.01, Fig. 3D).

Regulation of PMCA-mediated Ca 2؉ Signaling by Homer2
Homer2 Interacts with and Regulates PMCA Expression-To identify potential interacting sites between PMCA and the Homer proteins, we searched for a PPXXF-like motif in PMCA that may interact with the Ena/VASP homology 1 domain of Homer proteins (16). The only such potential motif is present in the N-terminal 91-98 residues of several PMCA subtypes; however, proline and phenylalanine are separated by more than two residues (Fig. 4A). To determine whether this region functions as a Homer-binding motif to mediate the PMCA-Homer2 interaction, we first examined whether Homer2 selectively binds PMCA in parotid acinar cells from WT mice using coimmunoprecipitation assays. As shown in Fig. 4B, PMCA coimmunoprecipitates with endogenous Homer2 suggesting that endogenous Homer2 associates with PMCA.
To examine whether Homer proteins bind PMCA1 and/or PMCA4 in cell lysates and whether mutations in the potential Homer-binding motif of PMCA affect the interaction, we performed co-immunoprecipitation assays using HEK293 cells that transiently express Homer2 and PMCA or PMCA mutants. As shown in Fig. 4C, Homer2 co-immunoprecipitated with PMCA1b and PMCA4b but did not co-immunoprecipitate with the PMCA mutants (Mt_PMCA), which changes the PP/LL and F/R of the PPXXF-like motif. These results indicate that the PPXXF-containing region functions as a Homer-binding motif and mediates Homer2 interaction with PMCAs. To determine whether these mutants interact with other Homer proteins, we examined the association between Homer1 and PMCAs in transfected HEK293 cells. Mutation of PMCA did not affect the interaction with short Homer1a, which lacks the coiled-coil domain, or long Homer1c. This result suggests significant specificity of the Homer-binding motif on PMCA toward Homer2.
PMCA isoforms 2b and 4b have PDZ domain-binding ligands in their C terminus that interact with various scaffolding proteins such as membrane-associated guanylate kinase, NHERF2 (Na ϩ /H ϩ exchanger regulatory factor-2), NOS-I (nitric-oxide synthase I), and Homer1 (Ania-3) (28 -31), some of which interact with the Homers. Therefore, we examined whether mutation of the PDZ ligand of PMCA4 (marked Del_PMCA4) affected the interaction of Homers with PMCA. Co-expression of Homers and Del_PMCA4 showed reduced interaction of Homer1a and Homer1c with Del_PMCA4 compared with their interaction with WT PMCA4. However, interaction of Homer2 was unaffected (Fig. 4C). These results indicate that the PDZ ligand of PMCA4b may function to mediate Homer1 variant interaction with PMCA.
A previous study suggests that exogenous expression of Homer1a, Homer1c, or Homer2a had no effect on endogenous PMCA function, whereas knockdown of Homer1 slowed PMCA-mediated Ca 2ϩ clearance in neuronal cells (44). We re-examined these effects in nonexcitable cells by expressing the Homers in HEK293 cells and measured native PMCA activity. Transfection of Homer1a enhanced native PMCA activity, whereas transfection of the long Homers, Homer1c or Homer2, had no effect (Fig. 5A). These findings suggest that the level of endogenous long Homers are saturating with respect to PMCA in HEK293 cells, and thus further expression had no inhibitory effect. Accordingly, Homer1a likely relieved the tonic inhibition by the long Homers to activate the native PMCA activity. Therefore, to examine the relationship between the Homerbinding motif and PMCA activity, we measured the effect of Homer proteins on Ca 2ϩ extrusion by expressed WT and mutant PMCA pumps. Overexpression of WT PMCAs increased Ca 2ϩ clearance, and expression of the PPXXF-like mutants additionally accelerated Ca 2ϩ clearance (Fig. 5, B and E). Co-transfection of Homer2 and WT PMCA pumps significantly slowed Ca 2ϩ clearance. This inhibitory effect was abolished when Homer2 was co-expressed with the PMCA mutants (Fig. 5, C and E). However, Homer1a did not increase PMCA activity when cells were co-transfected with WT and mutant PMCA pumps compared with the overexpression of Homer1a or WT and mutant PMCA pumps (Fig. 5, D and E). These results indicate that Homer2 interaction with the PPXXF-like motif on PMCA is essential to regulate Ca 2ϩ clearance and Ca 2ϩ signaling in nonexcitable cells.

DISCUSSION
In this study, we demonstrate a novel interaction between Homer2 and PMCA in native parotid acinar cells. Furthermore, we provide evidence in support of a critical role for Homer2 in modulating PMCA activity and thus Ca 2ϩ signaling. Previous reports that focused on the molecular structure of Homer proteins in the CNS indicated that these proteins bind GPCRs, such as mGluR1/5, and IP 3 Rs, as well as act as scaffolding proteins for assembling Ca 2ϩ -signaling complexes in cellular microdomains (8 -10, 13, 32). The apical region of exocrine cells is equivalent to the CNS synaptic region, where signaling complexes are clustered to form a "trigger zone" from which all forms of Ca 2ϩ signals, including Ca 2ϩ waves, are initiated (33,34). Consistently, immunocytochemical studies have demonstrated that all IP 3 R types are highly enriched in the apical region (14,35,36). Accordingly, we found that Homer2 and IP 3 Rs are expressed in the apical pole of parotid acinar cells. However, Homer2 deletion had no effect on the expression or function of any of the IP 3 R isoforms, suggesting that Homer2 may have another role on Ca 2ϩ signaling in parotid acinar cells.
Further analyses revealed that Homer2 affected PMCA expression in the apical membranes of parotid acinar cells. Hence, the most interesting finding from our study is the adaptive increase in PMCA protein expression in Homer2-deficient parotid and submandibular gland acinar cells. Like SERCA, the PMCAs have a crucial role in maintaining Ca 2ϩ homeostasis   37). Previous studies suggested that PMCAs are expressed in both the lateral and apical regions of pancreatic, submandibular gland, and parotid cells, as well as in the brain (31, 38 -41). In addition, PMCAs co-localize with mGluR1, IP 3 R1, and Homer proteins, including Homer1a, Homer1b/c, and Homer3, in neurons (31,40,41). Here, we discovered that Homer2 interacts with PMCA and decreases the rate of [Ca 2ϩ ] i clearance in native parotid acinar cells, which is absent in Homer2-deficient mice. These results are similar to previous findings showing up-regulation of specific PMCA isoforms due to adaptations in Ca 2ϩ signaling and Ca 2ϩ -dependent cellular functions in pancreatic and submandibular gland acinar cells from serca2-deficient mice (24). Therefore, it appears that PMCA expression and function is particularly sensitive to perturbations in Ca 2ϩ signaling, and cells modify PMCA activity to adjusting the Ca 2ϩ signal.
Adaptations of PMCA expression may be regulated by Ca 2ϩ itself. Several studies have shown that Ca 2ϩ can alter the expression levels of Ca 2ϩ -signaling components, such as pumps and channels, thereby maintaining flexibility in Ca 2ϩsignaling remodeling systems (3). In primary cultured cerebellar granule cells, PMCA2, PMCA3, and IP 3 R1 expression is upregulated during the activation of Ca 2ϩ -dependent events. In contrast, type 2 Na ϩ /Ca 2ϩ exchanger and PMCA4 are rapidly down-regulated (1,42,43). An interesting aspect of our findings is that Homer2-deficient parotid cells exhibit increased expression and activity of PMCA but not of SERCA2. However, Homer2-deficient pancreatic acinar cells increase SERCA2 but not PMCA expression, suggesting that PMCA adaptation occurs in a tissue-specific manner, perhaps reflecting specific cellular functions and localization.
PMCAs interact with partner molecules through their PDZbinding motif in the C terminus (28 -30). In this study, we characterized an additional novel binding region (PPXXF-like motif) of PMCA isoforms that participate in the interaction with, and inhibition of, PMCA by Homer2. Thus, mutation of the PPXXF-like motif reduced the interaction of PMCA1 and PMCA4 with Homer2 (Fig. 4C) and prevented inhibition of ectopically expressed PMCA by Homer2 (Fig. 5E). In contrast, the PMCA4 PDZ motif does not appear to be essential for the interaction with Homer2 because its deletion did not prevent the interaction of PMCA4 with Homer2 (Fig. 4C).
Interestingly, the role of the PPXXF-like motif appears to be specific for the interaction of PMCAs with Homer2. Thus, mutation of the PPXXF-like motif does not affect the interaction of PMCAs with a short or long Homer1 (Fig. 4C) or activation of PMCA activity by Homer1a in co-transfected HEK293 cells (Fig. 5, D and E). These results were unexpected in view of the reports that Homer1a (Ania-3) retains binding to the PDZbinding motif in the C terminus of PMCA isoforms and activates [Ca 2ϩ ] i clearance by PMCA when expressed with long Homer1 or the N terminus of Homer2 as an analog of Homer1a in neuronal cells (31,44). Using expression in HEK293 cells, we confirmed interaction of Homer1 with the PDZ motif of PMCA. We found that this interaction is eliminated by deletion of the PDZ motif on PMCA4, as was reported previously (31). However, unlike a previous report, the Ca 2ϩ clearance rate by native PMCA was increased by overexpression of short Homer1a but not by overexpression of long Homers (Fig. 5A) (44). These experiments suggest that the level of long Homers in HEK293 cells (and perhaps other nonexcitable cells) is saturating with respect to PMCA, and thus additional expression of the long Homers caused no further inhibition. A prediction of this interpretation is the deletion of the long Homers, and their inhibition by the short Homer1a should accelerate PMCA activity and Ca 2ϩ clearance. This was indeed the case, in which deletion of Homer2 in mice or expression of Homer1a in HEK293 cells increased the native PMCA activity.
In summary, the key findings of the our study are as follows: 1) Homer2 binding to the PPXXF-like motif of PMCAs inhibited PMCA activity, and 2) expression and binding of Homer1 to the PDZ domain of PMCAs increased their activity. Thus, PMCAs undergo dual regulation by Homer proteins, i.e. inhibition by the Homer2 and stimulation by Homer1. In this manner, Homer proteins can regulate the duration of the Ca 2ϩ signal in parotid acinar cells to either extend or shorten the signal by inhibition or stimulation of PMCA activity, respectively.