Residues 248–252 and 300–304 of the cardiac Na⁺/Ca²⁺ exchanger are involved in its regulation by phospholemman

Abstract

Using split cardiac Na⁺/Ca²⁺ exchangers (NCX1), we previously demonstrated that phospholemman (PLM) regulates NCX1 by interacting with the proximal linker domain (residues 218–358) of the intracellular loop of NCX1. With the use of overlapping loop deletion mutants, interaction sites are localized to two regions spanning residues 238–270 and residues 300–328 of NCX1. In this study, we used alanine (Ala) linker scanning to pinpoint the residues in the proximal linker domain involved in regulation of NCX1 by PLM. Transfection of human embryonic kidney (HEK)293 cells with wild-type (WT) NCX1 or its Ala mutants but not empty vector resulted in NCX1 current (I_NaCa). Coexpression of PLM with WT NCX1 inhibited I_NaCa. Mutating residues 248–252 (PASKT) or 300–304 (QKHPD) in WT NCX1 to Ala resulted in loss of inhibition of I_NaCa by PLM. By contrast, inhibition of I_NaCa by PLM was preserved when residues 238–242, 243–247, 253–257, 258–262, 263–267, 305–309, 310–314, 315–319, 320–324, or 325–329 were mutated to Ala. While mutating residue 301 to alanine completely abolished PLM inhibition, mutation of any single residue 250–252, 300, or 302–304 resulted in partial reduction in inhibition. Mutating residues 248–252 to Ala resulted in significantly weaker association with PLM. The NCX1-G503P mutant that lacks Ca²⁺-dependent activation retained its sensitivity to PLM. We conclude that residues 248–252 and 300–304 in the proximal linker domain of NCX1 were involved in its inhibition by PLM.

phospholemman (PLM) or FXYD1 (37), a 72-amino acid sarcolemmal protein with a single transmembrane (TM) domain (28), regulates the activities of Na⁺-K⁺-ATPase (4, 6, 7, 10, 33) and Na⁺/Ca²⁺ exchanger (NCX1) (1, 5, 44) in the heart and L-type Ca²⁺ channels in transfected human embryonic kidney (HEK)293 cells (40). PLM coimmunoprecipitates with NCX1 in adult pig, rat, and guinea pig cardiac sarcolemma (1, 34, 40); L-type Ca²⁺ channels in guinea pig cardiac membranes (40); and with α-subunits of Na⁺-K⁺-ATPase in adult human, rabbit, bovine, rat, guinea pig, and mouse ventricular tissues (4, 6, 10, 33, 38). The signature FXYD motif in the extracellular NH₂-terminus of PLM stabilizes the interactions between α- and β-subunits of Na⁺-K⁺-ATPase (32) and regulates gating characteristics of L-type Ca²⁺ channels (12). The TM domain of PLM interacts with TM9 of Na⁺-K⁺-ATPase (32), whereas the cytoplasmic tail of PLM associates with the large intracellular loop of NCX1 (39). Phosphorylation of serine-68 of PLM by protein kinase (PK) A or PKC relieves its inhibition on Na⁺-K⁺-ATPase (7, 10, 13, 29, 33). By contrast, PLM phosphorylated at serine-68 inhibits NCX1 (34, 43).

NCX1 is an important sarcolemmal ion transporter since it mediates both Ca²⁺ efflux (3 Na⁺ in: 1 Ca²⁺ out; forward mode) and Ca²⁺ influx (3 Na⁺ out: 1 Ca²⁺ in; reverse mode) during an cardiac excitation-contraction cycle (3). NCX1 is a 938-amino acid (939 amino acid in the rat) protein consisting of an extracellular NH₂-terminal domain comprising the first 5 TM segments, a large intracellular loop (residues 218–764), and an intracellular COOH-terminal domain consisting of the last 4 TM segments (24, 30). The α-repeats in TM segments 2, 3, and 7 of NCX1 are important in ion transport activity (16, 23) while the large intracellular loop contains the regulatory domains of the exchanger (18–20). Specifically, the exchange inhibitory peptide (XIP) region (residues 219–238) (19), the proximal linker domain (residues 259–370), Ca²⁺ binding domains (CBD) 1 (residues 371–508) (18) and CBD2 (residues 501–689) (15), and the interaction site for endogenous XIP (residues 562–679) (20) all reside in the intracellular loop (Fig. 1). Using split NCX1 exchangers consisting of NH₂- or COOH-terminal domains with varying lengths of intracellular loop (25), we demonstrated that PLM interacts and associates with residues 218–358 of the proximal linker domain of NCX1 (39). Using overlapping NCX1 loop deletion mutants, we refined the sites of interaction between PLM and NCX1 to two regions spanning residues 238–270 and residues 300–328 (46). The present study utilizes alanine (Ala) linker scanning strategy to pinpoint the residues in NCX1 involved in its regulation by PLM.

Fig. 1.

Topology model for Na⁺/Ca²⁺ exchanger (NCX1). The mature NCX1 molecule is modeled to consist of 9 transmembrane (TM) segments with 2 reentrant loops (between TM2 and TM3 and between TM7 and TM8) as part of the conserved α-repeat motifs, which are important in ion transport activity (16, 23, 24). The NH₂-terminus is extracellular and COOH-terminus is intracellular. Between TM5 and TM6 is a large intracellular loop (residues 218–764), which contains the regulatory domains of the exchanger. Specifically, the proximal linker domain (residues 218–358), which interacts with phospholemman (39, 46), the exchange inhibitory peptide (XIP) region (residues 219–238) (19), the 2 calcium binding domains (CBD) 1 (residues 371–500) and 2 (residues 505–689) connected in tandem by a short linker (residues 501–504) (11, 15, 27), and the interaction site for endogenous XIP (residues 562–679) (20) all reside within the intracellular loop. The 2 specific segments [residues 248–252 (PASKT) and residues 300–304 (QKHPD)] in the proximal linker domain that mediates inhibition of NCX1 by phospholemman are shown.

METHODS

NCX1 mutants.

The full-length wild-type (WT) rat NCX1 in pAdTrack-CMV vector (45) was used to construct Ala mutants in discrete regions of the intracellular loop of NCX1, using Quik Change II XL in vitro mutagenesis system (Agilent Technologies). Initially NCX1 Ala mutants were constructed, five consecutive residues at a time, to span regions comprising residues 238–267 and residues 300–329. Subsequently, mutants with a single Ala substitution were constructed to span regions encompassing residues 248–252 and 300–304. To evaluate whether inhibition of NCX1 by PLM involved synergistic interactions between the two Ca²⁺-binding domains (14, 15, 18), NCX1-G503P mutant which lacks Ca²⁺-dependent activation (21) was constructed. After transformation into XL10-Gold Ultracompetent cells (Agilent), fidelity of mutations was authenticated by DNA sequencing. For simplicity, NCX1 Ala mutants are designated as x-yA for five consecutive Ala substitutions spanning residues x-y and as xA for single Ala substitution of residue x.

Transfection of HEK293 cells.

HEK293 cells were transfected with control pAdTrack-CMV vector alone (3 μg), or combination of vector (2 μg) and NCX1 (either WT or mutant, 1 μg), or combination of vector, PLM, and NCX1 (either WT or mutant, 1 μg each) as described in detail previously (1, 39, 43, 46). Cells were trypsinized at 24 h posttransfection using trypsin-EDTA, transferred to 35-mm dishes containing sterile glass coverslips, and incubated for an additional 24 h before electrophysiological measurements.

Na⁺/Ca²⁺ exchange current measurements.

Whole cell patch clamp recordings were performed at 30°C as detailed previously (1, 39, 43, 46). Our conditions were carefully designed to block K⁺ (Cs⁺ substitution), Ca²⁺ (1 μM verapamil), Na⁺-K⁺-ATPase (1 mM ouabain), and Cl⁻ (30 μM niflumic acid) currents. In addition, the ionic compositions of the pipette (205 nM Ca²⁺ and 12.25 mM Na⁺) and extracellular solutions (5 mM Ca²⁺ and 140 mM Na⁺) were biased toward measurement of reverse Na⁺/Ca²⁺ exchange (outward Na⁺/Ca²⁺ exchange current, I_NaCa). Cells were held at the calculated equilibrium potential for Na⁺/Ca²⁺ exchanger (−73 mV under prevailing ionic conditions) for 5 min before stimulation. This precaution minimized fluxes through NCX1 before the voltage ramp and thus allow intracellular Na⁺ ([Na⁺]_i) and Ca²⁺ concentrations ([Ca²⁺]_i) to equilibrate with those present in the pipette solution. Only cells that fluoresced green (excitation 380, emission 510 nm), indicating successful transfection, were selected for current measurements. Cells were stimulated with a descending-ascending voltage ramp (from +100 to −120 back to +100 mV, 500 mV/s), in the absence and presence of 1 mM CdCl₂. The difference current measured during the descending voltage ramp was taken to be I_NaCa. I_NaCa of each cell was divided by its whole cell capacitance (C_m) to account for variation in cell sizes.

Coimmunoprecipitation of PLM and NCX1 and PLM and NCX1 Ala mutants.

Crude membranes were prepared from transfected HEK293 cells expressing WT NCX1 and PLM, or PLM and 248–252A mutant, or PLM and 300–304A mutant, as described previously (1, 39). They were resuspended in a minimal volume of buffer III [in mM: 140 NaCl, 25 imidazole, 1 EDTA, and a combination of complete protease inhibitor and phosphatase inhibitor cocktails (pH 7.4)], adjusted to 2 mg in 300 μl of buffer III, and combined with C₁₂E₈ detergent in 100 μl of buffer III (detergent:protein ratio of 2:1) at room temperature for 20 min. After addition of 400 μl of buffer III, samples were centrifuged at 37,000 g at 4°C. The supernatant was transferred to a fresh tube, and protein content was determined. Solubilized crude membrane preparations (400 μg) were preincubated with 40 μl of protein-A-agarose for 3 h at 4°C before relevant antibodies were added. Precleared supernatants were incubated with either 5 μg of preimmune rabbit IgG (control), or 5 μg of polyclonal PLM antibody (C2), or 5 μg of NCX1 antibody (R3F1) overnight at 4°C. The next day, 40 μl (50% slurry) of washed suspended protein-A-agarose beads was added to each sample, and the mixture was incubated for a further 2 h at 4°C. Beads were pelleted, washed 4 to 6 times with 1.5 ml of buffer III containing 0.05% C₁₂E₈, and resuspended in 40 μl of sample buffer (0.2 M dithiothreitol). Beads were boiled for 5 min at 95°C, and immunoblotting was performed.

GST pull-down assays.

GST-NCX1 fusion proteins were synthesized, purified, and used in GST pull-down assays as described in detail previously (39, 46). Briefly, PCR fragments of the intracellular loop of NCX1 (218–270, 218–270/248–252A, 300–373, 300–373/300–304A, and 250–300) were cloned in pGEX-6P-1 (Amersham). Fidelity of cDNA was confirmed by sequencing. After transformation into BL21 Eschericia coli, constructs were screened under the control of an isopropyl 1-thio-β-d-galactopyranoside-inducible promoter for the expression of GST-NCX1 fusion proteins at the predicted molecular weights. Purified GST-NCX1 fusion protein products (8, 39, 46) were coupled to GSH-Sepharose (Amersham). For pull-down assays, His-tagged PLM obtained from transformed BL21(DE3)pLysS bacteria (35) was used. C2 antibody (1:10,000) was used to detect PLM in these experiments. We have previously demonstrated that WT PLM harvested from transfected HEK293 cells and His-tagged PLM gave similar results in GST pull-down assays (39).

Statistical analysis.

All results are expressed as means ± SE. For the analysis of I_NaCa as a function of group (e.g., NCX1 vs. NCX1 + PLM) and voltage, two-way ANOVA was used to determine statistical significance. For comparison of His-PLM pulled down by GST-WT NCX1 and GST-NCX1 Ala mutants, paired Student's t-test was used. A commercial software package (JMP version 7, SAS Institute; Cary, NC) was used. P ≤ 0.05 was taken to be statistically significant.

RESULTS

Expression of NCX1 and Ala mutants results in functional I_NaCa.

We have previously demonstrated that under the conditions used in our patch-clamp experiments, the Cd²⁺-sensitive current is I_NaCa (1, 39, 43, 46). HEK293 cells transfected with pAdTrack-CMV vector did not exhibit I_NaCa (Fig. 2), in agreement with our previous results (1, 46). By contrast, HEK293 cells transfected with NCX1 or its Ala mutants all displayed the characteristic I_NaCa (Figs. 2–4).

Fig. 2.

Mutating residues 248–252 to alanine (Ala) results in loss of inhibition of Na⁺/Ca²⁺ current (I_NaCa) by phospholemman (PLM) in transfected human embryonic kidney (HEK)293 cells. I_NaCa measured in HEK293 cells transfected with NCX1 Ala mutants (5 consecutive Ala substitutions at a time) spanning the region encompassing residues 238–267 with (diamonds; n = 6, 7, 8, 6, 2, and 6 for 238–242A, 243–247A, 248–252A, 253–257A, 258–262A, and 263–267A mutants, respectively) or without PLM (triangles; n = 4, 6, 8, 6, 6, and 8 for 238–242A, 243–247A, 248–252A, 253–257A, 258–262A, and 263–267A mutants, respectively) are shown. HEK293 cells transfected with pAdTrack-CMV vector have very low endogenous currents under our measurement conditions (open circles, top left panel; n = 8). Symbols represent means ± SE. Error bars are not shown if they fall inside the boundaries of symbols. Except for 248–252A mutant (group, P < 0.54; voltage, P < 0.0001; and group × voltage interaction effects, P < 0.52), two-way ANOVA indicated significant (P < 0.0001 for group, voltage, and group × voltage interaction effects) inhibition of I_NaCa by PLM for all other NCX1 Ala mutants are shown.

Fig. 4.

Effects of single residue mutation on inhibition of I_NaCa by PLM. I_NaCa measured in HEK293 cells transfected with single NCX1 Ala mutants, with (diamonds; n = 5, 6, 8, 10, 3, 10, 6, 7, and 7 for 248A, 250A, 251A, 252A, 300A, 301A, 302A, 303A, and 304A mutants, respectively) and without PLM (triangles; n = 4, 11, 9, 8, 5, 4, 7, 5, and 10 for 248A, 250A, 251A, 252A, 300A, 301A, 302A, 303A, and 304A mutants, respectively) are shown. Residue 249 of wild-type rat NCX1 is Ala. I_NaCa in cells transfected with NCX1-G503P mutant, which lacks Ca²⁺-dependent activation (21), with (diamonds; n = 4) and without PLM (triangles; n = 6) is also shown. Symbols represent means ± SE. Error bars are not shown if they fall inside the boundaries of symbols. Except for 301A mutant (group, P < 0.91; voltage, P < 0.0001; and group × voltage interaction effects, P < 0.56), two-way ANOVA indicated significant (group, P < 0.037; voltage, P < 0.0001; and group × voltage interaction effects, P < 0.0001) inhibition of I_NaCa by PLM for all other single NCX1 Ala mutants are shown. For NCX1-G503P mutant, two-way ANOVA indicated significant (P < 0.0001 for group, voltage and group × voltage interaction effects) inhibition of I_NaCa by PLM.

PLM inhibits I_NaCa by interacting with regions spanning residues 248–252 (PASKT) and 300–304 (QKHPD) of the intracellular loop of NCX1.

Coexpression of PLM with WT NCX1 resulted in significant (P < 0.0001) suppression of I_NaCa in HEK293 cells (∼52% inhibition; data not shown), in agreement with our previous observations (1, 39, 43, 46). Using overlapping NCX1 deletion mutants, we previously demonstrated that two regions spanning residues 238–270 and 300–328 are involved in regulation of NCX1 by PLM (46). Therefore, we performed Ala linker scanning of these two regions. For the region spanning residues 238–270, mutating residues 248–252 (PASKT) to Ala resulted in loss of regulation of NCX1 by PLM (Fig. 2). For the region spanning residues 300–328, only residues 300–304 (QKHPD), when mutated to Ala, resulted in loss of inhibition of I_NaCa by PLM in transfected HEK293 cells (Fig. 3).

Fig. 3.

Mutating residues 300–304 to Ala results in loss of inhibition of I_NaCa by PLM in transfected HEK293 cells. I_NaCa measured in HEK293 cells transfected with NCX1 Ala mutants (5 consecutive Ala substitutions at a time) spanning the region encompassing residues 300–329, with (diamonds; n = 4, 4, 5, 6, 9, and 6 for 300–304A, 305–309A, 310–314A, 315–319A, 320–324A, and 325–329A mutants, respectively) or without PLM (triangles; n = 9, 14, 7, 6, 7, and 9 for 300–304A, 305–309A, 310–314A, 315–319A, 320–324A, and 325–329A mutants, respectively) are shown. Symbols represent means ± SE. Error bars are not shown if they fall inside the boundaries of symbols. Except for 300–304A mutant (group, P < 0.16; voltage, P < 0.0001; and group × voltage interaction effects, P < 0.07), two-way ANOVA indicated significant (P < 0.0001 for group, voltage, and group × voltage interaction effects) inhibition of I_NaCa by PLM for all other NCX1 Ala mutants are shown.

The effects of single Ala mutations in the two regions encompassing residues 248–252 and residues 300–304 on I_NaCa inhibition by PLM were next examined (Fig. 4). We did not mutate residue 249 since it is Ala in WT rat NCX1. Whereas mutating residue 301 to Ala completely abolished PLM inhibition, mutation of any single residue 250–252, 300, or 302–304 resulted in partial reduction of inhibition of I_NaCa by PLM (Fig. 4).

Mutating residues 248–252 to Ala weakens association of NCX1 with PLM.

We have previously shown by coimmunoprecipitation studies that PLM physically associates with NCX1 in transfected HEK293 cells (1, 39). In addition, GST pull-down assays establish that PLM associates with the intracellular loop (residues 218–764) of NCX1, specifically residues 218–371 and 508–674, but not CBD 1 (residues 371–508) (39). In the present study, we first confirmed that PLM coimmunoprecipitated WT NCX1 in transfected HEK293 cells (Fig. 5). Mutating residues 248–252 or residues 300–304 to Ala did not abolish association with PLM (Fig. 5).

Fig. 5.

Reciprocal copurification of PLM with wild-type (WT) NCX1, 248–252A and 300–304A mutants in transfected HEK293 crude membrane extracts. Top: immunoblots of WT NCX1, 248–252A, and 300–304A immunoprecipitates from 400 μg of solubilized membrane preparations using 5 μg of anti-NCX1 antibody (R3F1), or 5 μg of anti-PLM antibody (C2), or control rabbit serum (IgG) and probed with R3F1. Crude membranes (input) were prepared as previously described (1, 39). Molecular markers are indicated on the left.

To evaluate whether association of PLM with the proximal linker domain of WT NCX1 was significantly weakened by mutating residues 248–252 and residues 300–304 to Ala, we performed GST pull-down assay. We first confirmed our previous observations (46) that PLM associated with GST-NCX1/218–270 and GST-NCX1/300–373 but not GST-NCX1/250–300 (Fig. 6). From three experiments using new sets of GST fusion proteins prepared from independent bacterial cultures, in terms of pulling down His-PLM, GST-NCX1/218–270/248–252A was 58.9 ± 15.2% as efficient as GST-NCX1/218–270 (P < 0.04), and GST-NCX1/300–373/300–304A was 56.8 ± 16.8% as efficient as GST-NCX1/300–373 (P < 0.13) (Fig. 6).

Fig. 6.

Physical association of PLM with residues 218–270 was weakened by mutating residues 248–252 to Ala. Purified GST or GST-NCX1 fusion proteins linked to GSH-sepharose beads were incubated with His-tagged PLM (1 μg), and GST pull-down assay performed as described in methods. Top: GST and GST-NCX1 fusion proteins detected by Coomassie blue staining. MW, molecular weight marker. 218–270 Mut, NCX1 fragment spanning residues 218–270 in which residues 248–252 were mutated to Ala; 300–373 Mut, NCX1 fragment spanning residues 300–373 in which residues 300–304 were mutated to Ala. Signal intensities (normalized to signal intensity of GST) of 218–270, 218–270 Mut, 300–373 and 300–373 Mut are 0.31, 0.47, 0.45, and 0.70, respectively. Bottom: His-tagged PLM detected by C2 antibody. Signal intensities (normalized to signal intensity of input His-PLM) of His-PLM pulled down by 218–270, 218–270 Mut, 300–373 and 300–373 Mut are 0.49, 0.21, 0.52, and 0.22, respectively. For 218–270 Mut, the His-PLM signal corrected for the difference in GST-fusion products is (0.21 × 0.31)/0.47 = 0.14. Therefore, compared with 218–270, 218–270 Mut is 0.14/0.49 or 28.5% as efficient in terms of pulling down His-PLM. For 300–373 Mut, the efficiency in pulling down His-PLM is 26.9% compared with 300–373. Composite results from three separate GST pulldown experiments are given in results.

Inhibition of NCX1 by PLM is preserved despite lack of Ca²⁺-dependent activation.

Recent evidence suggests specific synergistic interactions between CBD1 and CBD2 of the exchanger (11, 14, 17, 27). To evaluate the possibility that PLM alters the synergistic interactions between CBD1 and CBD2, thereby resulting in NCX1 inhibition, we utilized NCX1-G503P mutant, which lacks Ca²⁺-dependent activation (21). As shown in Fig. 4, I_NaCa in HEK293 cells expressing NCX1-G503P mutant was smaller compared with WT NCX1 or its Ala mutants. This may reflect lack of Ca²⁺-dependent activation. More germane to the present study is the observation that I_NaCa measured in HEK293 cells expressing NCX1-G503P mutant was strongly inhibited by cotransfection with PLM.

DISCUSSION

Cumulative evidence supports a direct regulatory role of PLM on NCX1 activity in the heart (1, 5, 43, 44). Since PLM modulates the activities of not only NCX1 but also Na⁺-K⁺-ATPase (4, 6, 7, 10, 33) and possibly L-type Ca²⁺ channels (40) in the heart, detailed knowledge of the molecular interaction sites is desirable not only for the understanding of mechanisms of modulation but also for design of drugs that interfere with regulation of one ion transporter but not the other by PLM.

One mechanism by which PLM regulates NCX1 function is by affecting the synergistic interactions between CBD1 and CBD2 (11, 14, 17, 27), thereby modulating Ca²⁺-dependent activation of NCX1 (21). This is unlikely for the following reasons. First, I_NaCa measured in HEK293 cells expressing split NCX1 exchangers in which residues 359–763 (encompassing both CBD1 and CBD2) are absent is inhibited by PLM (39). Second, intracellular NCX1 loop fragment encompassing residues 371–508 (CBD1) has no physical association with PLM as indicated by GST pulldown assays (39). Third, NCX1-G503P mutant, which lacks Ca²⁺-dependent activation of NCX1 (21), is still inhibited by PLM (Fig. 4). Finally, the K_0.5 for Ca²⁺-dependent activation of NCX1 is 20–200 nM in intact cardiac myocytes (9, 22). Given that bulk [Ca²⁺]_i in cardiac cells fluctuates between 100 and 1,000 nM during an excitation-contraction cycle (3), and that the submembranous [Ca²⁺] sensed by NCX1 is likely higher than that measured in bulk cytoplasm (42), it is reasonable to expect that NCX1 is fully activated in the heart under physiological conditions. Indeed, in the intact mouse (41) and guinea pig (22) [but not ferret (41)] cardiac myocytes, Ca²⁺-dependent activation of NCX1 is not demonstrable at physiological [Ca²⁺]_i. Despite full activation of NCX1 by prevailing [Ca²⁺]_i in mouse myocytes, inhibition of NCX1 by PLM is clearly present (43). Collectively, these observations suggest that regulation of NCX1 by PLM is unlikely to involve Ca²⁺-dependent activation of the exchanger.

Recent evidence obtained with chemical cross-linking (31) and fluorescence resonance energy transfer (FRET) (17) techniques suggests that NCX1 exists as dimers in insect and Xenopus cell membranes. Dimerization of NCX1 may be important for the proper targeting of the protein to the plasma membrane (26) or may affect NCX1 function via Ca²⁺-dependent structural rearrangements within NCX1 dimers (17). PLM conceivably may regulate NCX1 function by altering oligomerization of NCX1 molecules in the membrane. While the existence of NCX1 oligomers has been elegantly demonstrated in model expression systems, it is not known whether NCX1 exists as monomer or oligomer in native cardiac membranes. Moreover, the stoichiometry of interaction between PLM and NCX1 in cardiac cells is unknown. Much additional work is required before this hypothesis can be accepted or rejected.

The major finding of the present study is that using Ala linker scanning and electrophysiological techniques, we identified two small regions spanning residues 248–252 and 300–304 in the intracellular loop of NCX1 as important in its inhibition by PLM. With the exception of 301A mutant, which completely abolished NCX1 inhibition by PLM, single Ala substitution mutants resulted in reduction in inhibition of NCX1 by PLM, suggesting several residues spanning 248–252 or 300–304 act in concert to effect inhibition of NCX1 by PLM.

Although the structure of the proximal linker domain is unknown, one hypothetical model of the region encompassing residues 259–370 suggests that Val²⁶¹ is in close proximity with Ala³¹⁴ (15). In addition, our results with both NCX1 overlapping loop deletion mutants (46) and Ala substitution mutants indicate that two regions spanning residues 248–252 and residues 300–304 of intracellular loop of NCX1 were required for inhibition of I_NaCa by PLM. Therefore, the two regions encompassing residues 248–252 and residues 300–304 are likely to be held in close proximity with each other and suggest some “structure” in the proximal linker domain. An alternative interpretation is that one PLM molecule interacts with the region encompassing residues 248–252 of one NCX1 molecule while a second PLM molecule interacts with the region encompassing residues 300–304 of the other NCX1 molecule in the NCX1 dimer. This hypothesis is made plausible by the recent findings that both NCX1 (17, 31) and PLM (2, 36) exist as oligomers in model lipid membranes and heterologous expression systems.

There are caveats to the present study. The first is that mutating five consecutive residues in the intracellular loop of NCX1 to Ala may alter its secondary and tertiary structures such that the interaction with PLM is affected. This is likely the case for 248–252A mutant since its association with PLM was weaker. However, internally consistent results were obtained using fundamentally different experimental techniques (electrophysiology and GST pull-down) and different molecular approaches [split NCX1 exchangers (39), overlapping NCX1 loop deletion mutants (46), and Ala linker scanning], suggesting robustness of our experimental data and conclusions. Another concern is that in the present study, we did not measure the percentage of PLM that was phosphorylated at serine-68, which is the active specie that inhibits NCX1. Approximately 30–45% of PLM in unstimulated HEK293 cells coexpressing PLM and WT NCX1 is phosphorylated at serine-68 (43, 46). In addition, the fraction of PLM phosphorylated at serine-68 was similar in HEK293 cells expressing various NCX1 loop deletion mutants compared with WT NCX1 (46). Therefore, we felt it unlikely that the Ala substitution mutants would grossly distort the fraction of PLM phosphorylated at serine-68.

In summary, by expressing Na⁺/Ca²⁺ exchanger and its Ala substitution mutants in HEK293 cells, we identified two small regions spanning residues 248–252 and residues 300–304 of the Na⁺/Ca²⁺ exchanger as important in its regulation by phospholemman. In addition, mutating residues 248–252 to Ala weakened the association of Na⁺/Ca²⁺ exchanger with phospholemman.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Grants HL-58672 and HL-74854 and by American Heart Association Scientist Development Grant F64702 (T. O. Chan).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).