A New Cell-penetrating Peptide That Blocks the Autoinhibitory XIP Domain of NCX1 and Enhances Antiporter Activity

Abstract

The plasma membrane Na⁺/Ca²⁺ exchanger (NCX) is a high-capacity ionic transporter that exchanges 3Na⁺ ions for 1Ca²⁺ ion. The first 20 amino acids of the f-loop, named exchanger inhibitory peptide (XIP_NCX1), represent an autoinhibitory region involved in the Na⁺-dependent inactivation of the exchanger. Previous research has shown that an exogenous peptide having the same amino acid sequence as the XIP_NCX1 region exerts an inhibitory effect on NCX activity. In this study, we identified another regulatory peptide, named P1, which corresponds to the 562–688aa region of the exchanger. Patch-clamp analysis revealed that P1 increased the activity of the exchanger, whereas the XIP inhibited it. Furthermore, P1 colocalized with NCX1 thus suggesting a direct binding interaction. In addition, site-directed mutagenesis experiments revealed that the binding and the stimulatory effect of P1 requires a functional XIP_NCX1 domain on NCX1 thereby suggesting that P1 increases the exchanger activity by counteracting the action of this autoinhibitory sequence. Taken together, these results open a new strategy for developing peptidomimetic compounds that, by mimicking the functional pharmacophore of P1, might increase NCX1 activity and thus exert a therapeutic action in those diseases in which an increase in NCX1 activity might be helpful.

Introduction

The plasma membrane Na⁺/Ca²⁺ exchanger (NCX) is a high-capacity ionic transporter that exchanges three Na⁺ ions for one Ca²⁺ ion. Depending on electrochemical gradients and membrane potential, NCX functions in either the Ca²⁺-efflux (forward) or the Ca²⁺-influx (reverse) mode. NCX belongs to a multigene family comprising three isoforms, named NCX1 (ref. 1), NCX2 (ref 2), and NCX3 (ref. 3), all of which are widely expressed in the brain and in the skeletal muscle. However, only NCX1 is ubiquitously expressed in several tissues, including heart, smooth muscle, kidney, eye, secretory glands, and blood cells.⁴ In the heart, NCX1 represents the principal Ca²⁺ extrusion mechanism of cardiomyocytes that is important to maintain Ca²⁺ homeostasis after a heartbeat.⁵ Furthermore, recent data showed that NCX1 is also a key player in the initiation and maintenance of a stable heart rhythm,⁶ and more important, the class III antiarrhythmic drug dofetilide exerts its positive ionotropic action by enhancing NCX1 activity.⁷ For these reasons in the cardiovascular area, there is an interest in drugs that could enhance NCX1 activity. In the central nervous system, NCX1 is sprouted in neurons, astrocytes, oligodendrocytes, and microglia where it is regulated by several prosurvival pathways^8,9 and plays a relevant role in maintaining intracellular Na⁺ and Ca²⁺ homeostasis under different neurophysiological and neuropathological conditions. In particular, several studies indicate that an activator of NCX might be of potential value in neuroprotection against brain ischemia.^{10,11,12,13,14,15}

The development of new strategies capable of increasing NCX1 activity might represent a new potential pharmacological tool for those pathophysiological conditions, including stroke injury and heart arrhythmias, in which a selective stimulation of the exchanger activity might achieve beneficial effects.¹⁶ To date, although more than 16 chemical classes of NCX inhibitors have been identified,¹⁷ only 1 compound has been demonstrated to increase NCX activity.¹⁸ On the other hand, the first 20 amino acids (219–238aa) of the regulatory cytosolic f-loop of NCX1, named exchanger inhibitory peptide (XIP_NCX1), exert an autoinhibitory function on the activity of the exchanger since the substitution of some specific positive amino acids in this region increases peak and steady-state currents of NCX and removes its dependency from the Na⁺-dependent inactivation (Figure 1).^19,20 It has been hypothesized that the XIP_NCX1 domain might exert the autoinhibitory activity by binding to another unidentified sequence located on the f-loop of the exchanger.²¹ Conceivably, the target-binding region of the XIP_NCX1 domain should be the 562–688 amino acid sequence of the f-loop, since this region constitutes a molecular determinant for the inhibitory effect of an exogenous peptide having the same amino acid sequence as the XIP_NCX1 region.²² Altogether, these findings suggest that the binding between the N-terminal portion of the f-loop (P1 domain, 562–688aa) and the XIP_NCX1 domain (219-238aa) might inhibit NCX activity owing to conformational changes within the regulatory f-loop.

Figure 1.

Topological model of the Na⁺/Ca²⁺ Exchanger. (a) Transmembrane segments (TMSs) are represented by cylinders and are numbered. The α-repeat regions are boxed. Both XIP and P1 regions are shaded. (b) Complete amino acid sequence of the dog cardiac NCX1.1 isoform in which both XIP and P1 sequences are underlined and shaded.

On the basis of this assumption, a new pharmacological strategy could be to develop drugs capable of increasing NCX1 activity by blocking the XIP_NCX1 domain, thereby preventing its inhibitory binding to the 562–688aa region. To test this hypothesis, we used two different strategies: (i) a peptide having the same amino acid sequence as the 562–688aa region of the f-loop was endogenously expressed and (ii) a cell-penetrating proteic construct having the same amino acid sequence of the 562–688aa region of the f-loop and the HIV-Tat sequence was synthesized and purified. These proteins were used with the aim to: (i) bind the XIP_NCX1 domain on NCX1; (ii) cause a steric hindrance on this inhibitory region; (iii) reduce the inhibitory binding of XIP_NCX1 to the N-terminal region of the f-loop, thus ultimately increasing NCX1 activity. In particular, baby hamster kidney (BHK) cells stably expressing NCX1 were incubated with the cell-penetrating P1 protein provided with a Flag tag (P1^Flag) or transfected with several chimeric proteins comprising the enhanced fluorescent green protein (EGFP) marker fused with peptides having the same amino acid sequence as the 562–688aa (EGFP-P1) or 219-238aa (EGFP-XIP) regions of NCX1. The effect of the chimeric EGFP proteins on NCX currents (I_NCX) was then evaluated by means of patch-clamp technique in whole-cell configuration and single-cell fura-2 monitored Ca²⁺ concentrations. P1^Flag was firstly tested for time- and concentration-dependent cell uptake, and then, the effect of this protein on I_NCX was evaluated in BHK stably expressing NCX1. Our study indicated that P1 protein prevents NCX1 autoinhibition by binding to the XIP domain, thus leading to an increased activity of the exchanger. Furthermore, these results might lead to the development of peptidomimetic compounds that, by reproducing the stimulatory effect of the P1 peptide, might increase the activity of the exchanger.

Results

Stimulatory effect of the XIP-binding peptide P1 on NCX1 activity detected by patch-clamp technique and single cell fura-2-monitored Ca²⁺ concentrations

Since the plasma membrane is not permeable to peptides, two strategies were used to test the effects of P1 on NCX1 activity: (i) we generated eukaryotic plasmids expressing chimeric XIP and P1 proteins fused with the fluorescent EGFP probe, named EGFP-XIP and EGFP-P1, respectively and (ii) we expressed and purified a protein, that we named P1^Flag, consisting of five sequences: P1, the basic amino acid sequence from the transactivator of HIV-1 transcription (Tat) protein, and the three tags: Flag, hemoagglutinine (HA) and six histidines (6xHis). The expression of the chimeric peptides EGFP-XIP and EGFP-P1 in wild-type BHK or in BHK cells stably expressing the NCX1 isoform (BHK-NCX1) was verified by western blot with an antibody raised against the GFP epitope (Abcam, Cambridge, UK; Figure 2a). Positive transfected cells were identified by green fluorescence (Figure 2b,c). The penetration of P1^Flag in cytosol of BHK-NCX1 cells was verified by immunocitochemical analysis in a range of concentrations of 0.25–1.00 µmol/l (Figure 4a). Furthermore, immunocitochemical and immunoblot analysis showed that the uptake of P1^Flag in BHK cells was time dependent (Figure 4a,b), and no degradation of P1^Flag was observed.

Figure 2.

Effect of XIP and P1 peptides on NCX1 activity. (a) Immunoblot analysis of BHK cells transfected with EGFP, EGFP-XIP, or EGFP-P1 cDNA with GFP antibody. (b,c) Fluorescence microscopy images of BHK-NCX1 cells transfected with EGFP-XIP or EGFP-P1. (d–f) Representative superimposed traces of I_NCX1 in BHK-NCX1 cells transfected with EGFP-scrambled control (EGFP-Scr), EGFP-XIP or EGFP-P1 proteins at 0, 20, 40, and 60 seconds starting from the first recording as described in Materials and Methods section and (g) their quantification at +60 mV reported as mean ± SEM. Bars = 10 μm for both images.

Figure 4.

Interaction between EGFP-P1 and NCX1. (a) Top, representative scheme of P1^Flag structure. Bottom, immunocitochemical assay of the time- and concentration-dependent penetration of P1^Flag (red) in BHK-NCX1 cells. Hoecst dye (blue) was used to mark nuclei. (b) Western blot for P1^Flag, NCX1, and tubulin in BHK-NCX1 exposed for 0, 30, 60, or 90 minutes of incubation with 0.50 μmol/l P1^Flag in a serum-free medium. (c) Representative superimposed traces of I_NCX1 recorded in BHK-NCX1 cells exposed for 60 minutes to P1^Flag (black) or vehicle (gray), and their quantification in the forward and reverse modes of operation expressed as percentage mean ± SEM of three independent experimental sessions (n = 8 for each group). (d) Colocalization signal (red) of NCX1 and P1^Flag in BHK-NCX1 or BHK-NCX1K229Q cells exposed for 60 minutes to 0.50 μmol/l P1 or vehicle. Hoecst dye (blues) was used to mark nuclei. Red dotted puncta reveal proximity below 40 nm. (e) Western blot of NCX1 in cell lysates of wild-type BHK cells transfected with NCX1, EGFP-Scr, EGFP-P1, or EGFP-XIP proteins with NCX1 antibody. (f) Immunocitochemistry for NCX1 (green), Flag (red), and Hoecst (blue) signals in BHK-NCX1K229Q cells exposed to 0.50 μmol/l P1^Flag for 60 minutes. Bars = 5 μm for all images. *P < 0.05 versus vehicle-treated BHK-NCX1 cell group.

Patch-clamp recordings revealed that the transfection of EGFP-XIP cDNA in BHK-NCX1 cells caused a remarkable inhibition of NCX1 activity in both forward and reverse modes of operation as compared with EGFP-scrambled control (Figure 3a,b). In particular, this inhibition was time dependent, since it progressively increased from 0 to 20, 40, and 60 seconds (Figure 2e,g). By contrast, EGFP-P1 cDNA transfection (Figure 3d,e) or P1^Flag incubation (Figure 4c) significantly increased NCX1 activity in both forward and reverse modes of operation as compared to BHK-NCX1 cells transfected with EGFP-scrambled control or treated with vehicle of P1^Flag, respectively. Like the inhibitory action of EGFP-XIP, the stimulatory activity of EGFP-P1 was also time dependent, since it progressively increased from 0 to 20, 40, and 60 seconds (Figure 2f,g). The inhibitory and stimulatory effects of EGFP-XIP and EGFP-P1 peptides, respectively, detected by the patch-clamp technique, were confirmed when NCX activity was stimulated in the reverse mode of operation by Na⁺ removal and measured in single-cell fura-2 monitored Ca²⁺ increase (Figure 3c,f). The transfection of EGFP-P1 in BHK-NCX1 cells caused a remarkable downregulation of NCX1 protein level (Figure 3h), whereas it did not affect mRNA expression (Figure 3g). BHK-NCX1 cell lysates showed two additional proteolytic bands at 50 and 40 kDa (Figure 3i).

Figure 3.

Quantification of XIP and P1 effects on NCX1 activity. (a,d) Representative superimposed traces of I_NCX1 recorded at 40 seconds starting from the first stimulatory ramp in BHK-NCX1 cells transfected with EGFP-XIP (black), EGFP-scrambled control (gray), and EGFP-P1 (black) proteins; (b,e) their quantification in the forward and reverse modes of operation expressed as percentage mean ± SEM of three independent experimental sessions (n = 20 for each group). (c,f) Quantification of the effect of EGFP-XIP, EGFP-scrambled and EGFP-P1 proteins on NCX1 activity measured as Na⁺-dependent [Ca²⁺]_i increase by single-cell fura-2 AM microfluorimetry. Data were calculated as Δ% of plateau/basal [Ca²⁺]_i values after addition of Na⁺-free solution. (g,h) Quantification of NCX1 mRNA and protein levels in BHK-NCX1 cells stably expressing EGFP-Scr or EGFP-P1 proteins. NCX1 mRNA was normalized for endogenous hypoxanthine–guanine phosphoribosyltransferase mRNA, and NCX1 protein was normalized for the endogenous tubulin protein (n = 6). (i) Western blot of NCX1 in BHK-NCX1 cells stably expressing EGFP-Scr or EGFP-P1 proteins (n = 3). *P < 0.05 versus EGFP-Scr and untransfected BHK-NCX1 cell groups.

P1 peptide colocalizes with NCX1 protein

In situ proximity ligation assay revealed a colocalization of NCX1 and P1^Flag within a range of <40 nm in intact BHK-NCX1 cells by using specific NCX1 and Flag antibodies (Figure 4d), suggesting a direct interaction between the two proteins. On the other hand, no colocalization signal was observed in BHK cells expressing NCX1K229Q mutant in presence of P1^Flag (Figure 4d), although NCX1K229Q and P1^Flag proteins were recognized within the cells by the respective antibodies (Figure 4f).

The C-terminal sequence 108–125 of P1 peptide is the molecular determinant for the stimulatory effect on NCX1 currents

To identify the molecular determinants of the EGFP-P1 peptide responsible for the stimulatory effect on I_NCX1, its amino acid sequence was subdivided into four putative regions, named A (1–42), B (44–50), C (57–84), and D (108–125) (Figure 5a). The transfection of the first EGFP-P1Δ58–84 mutant into BHK-NCX1 cells showed an increase in NCX1 exchange activity in both forward and reverse modes of operation, as compared with the EGFP-scrambled control and nontransfected BHK-NCX1 cells (Figure 5b). The transfection of the second EGFP-P1Δ 44–84 mutant, however, was able to significantly increase NCX1 exchange activity only in the reverse mode of operation (Figure 5c). By contrast, the removal of the C-terminal amino acid region 108–125 from the P1 sequence totally reversed the stimulatory action, as evidenced by the inhibition of I_NCX1 in both forward and reverse modes of operation (Figure 5d).

Figure 5.

Effects of EGFP-P1 mutants on I_NCX1 measured by patch-clamp electrophysiology. (a) Amino acid sequence of P1 protein. Black boxes indicate the A, B, C, and D regions of the P1 protein. Left and right panels of b, c, and d represent superimposed traces of the currents recorded in BHK-NCX1 under control conditions, transfected with EGFP-Scr, EGFP-P1Δ57-84, EGFP-P1Δ44-84, or EGFP-P1Δ108-125, and their quantification as compared to untransfected BHK-NCX1 cells expressed as percentage mean ± SEM of three independent experimental sessions (n = 10 for each group). *P < 0.05 versus untransfected and EGFP-scrambled-transfected BHK-NCX1 cells groups.

The XIP-binding domain is a molecular determinant for the stimulatory effect of the P1 peptide on NCX1 currents

Deletion mutagenesis of the amino acid region 241–680 of the f-loop (NCX1Δ241–680), which leaves an intact XIP domain and the C-terminal region of P1, did not prevent the stimulatory effect of EGFP-P1 transfection on I_NCX1, since a significant increase in NCX1 exchange activity was still observed (Figure 6a). On the contrary, when lysine229 was substituted with glutamine, the ensuing new mutant NCX1K229Q displayed no changes in I_NCX upon exposure to the chimeric EGFP-P1 protein (Figure 6b).

Figure 6.

Effects of EGFP-P1 on the currents of NCX1 mutants. (a,b) The left panels represent superimposed traces of the currents recorded in BHK stably expressing NCX1Δ241-680 or NCX1K229Q transfected with EGFP-Scr or EGFP-P1; the right panels show the quantification of the effects as compared to untransfected BHK-NCX1 cells expressed as percentage mean ± SEM of three independent experimental sessions (n = 10 for each group). *P < 0.05 versus untransfected BHK-NCX1 and EGFP-scrambled control groups.

EGFP-P1 peptide stimulates NCX2 isoform and inhibits NCX3 isoform

To determine whether EGFP-P1 was selective only for NCX1 or also for the other two isoforms, BHK-NCX2 and BHK-NCX3 cells were transfected with EGFP-P1. In BHK-NCX2, EGFP-P1 increased the antiporter activity in both forward and reverse modes of operation, as measured by patch-clamp technique (Figure 7a). By contrast, in BHK-NCX3, EGFP-P1 decreased I_NCX3 in both forward and reverse modes of operation (Figure 7b).

Figure 7.

Effects of EGFP-P1 on NCX2 and NCX3 currents. (a,b) The left panels represent superimposed traces of the currents recorded in BHK-NCX2 or BHK-NCX3 transfected with EGFP-Scr or EGFP-P1; the right panels show the quantification of the respective effects as compared to untransfected BHK-NCX2 or BHK-NCX3 cells expressed as percentage mean ± SEM of three independent experimental sessions (n = 15 for each group). *P < 0.05 versus EGFP-scrambled control group. (c) Amino acid alignment of XIP (219–238aa) and P1_108-128 (668–688aa) sequences of NCX1, NCX1K229Q, NCX2, and NCX3. Conserved amino acids among exchangers are indicated with dots.

Discussion

In this study, we report a new strategy to increase NCX1 activity by producing a steric hindrance on the XIP_NCX1 domain (219–238aa), a region of the cytoplasmic f-loop involved in the endogenous autoinhibition of the exchanger activity. This steric hindrance was obtained with two strategies: (i) by transfecting BHK-NCX1 cells with a chimeric protein comprising a peptide corresponding to the amino acid region 562–688 of the f-loop, named P1, and the EGFP; (ii) by incubating BHK-NCX1 cells with a cell-penetrating protein, named P1^Flag, comprising HIV-Tat, P1, and Flag sequences. The stimulatory effect of P1 peptide on I_NCX1, working either in the forward or in the reverse mode of operation, relied on the ability of this protein to bind the XIP_NCX1 domain on NCX1.

Consistent with previous results,^19,20 the expression of the EGFP-XIP chimeric protein in the cytosol remarkably inhibited I_NCX1, measured by the patch-clamp technique and by single-cell fura-2 monitored Ca²⁺ increase. By contrast, the expression of the EGFP-P1 protein in BHK-NCX1 cells increased NCX1 activity in both forward and reverse modes of operation. Since the P1-induced increase of the antiporter activity was not caused by an increase in NCX1 mRNA or protein levels, we hypothesized that the opposite effects of P1 and XIP peptides might be ascribed to the steric hindrance exerted on the target sequence present on NCX1 f-loop. Indeed, the presence of the eicosapeptide XIP inhibits NCX1 activity by binding its P1_NCX1 domain (aa 562–688), a region of the f-loop almost entailed in the calcium-binding domain 2 that is involved in the Ca²⁺-dependent activation/inactivation of the antiporter.²³ By opposite, P1 protein binds the 219-238aa region (XIP_NCX1 domain) of the NCX1 f-loop, preventing the inhibitory binding between the XIP_NCX1 domain (aa 219–238) and the P1_NCX1 region (aa 562–688; Figure 8), thus resulting in a stimulatory effect. In line with this hypothesis of competition, the inhibitory and the stimulatory effects of both EGFP-XIP and EGFP-P1 proteins, respectively, became even more evident when repetitive stimulations were delivered at 20, 40, and 60 seconds from the first ramp. This effect could be explained by f-loop conformational changes, induced by ramp stimulations, that render both the XIP_NCX1 and P1_NCX1 domains ever more accessible to the competitive binding exerted by EGFP-P1 or EGFP-XIP, respectively.

Figure 8.

Cartoon illustrating the proposed mechanism of P1 for NCX1 autoinhibition. (a,b) During resting conditions NCX1 is autoinhibited by a conformational change in the f-loop due to the binding interaction between XIP and P1 domains. (c) The transfection of the P1 peptide causes a steric hindrance on the XIP domain when it is accessible. P1, by binding to the XIP domain, stabilizes NCX1 in the active state. (d) The removal of the region 241–680 in NCX1 puts XIP and the region 680–685 of the P1 domain in such a close proximity that the exchanger cannot be further inhibited by the XIP peptide.

A direct interaction between the NCX1 and P1 is suggested by the results obtained by the in situ proximity ligation assay that showed a colocalization of wild-type NCX1 with the newly synthesized penetrating form of P1 in intact BHK cells. However, an indirect interaction between NCX1 and P1 might not be completely excluded. The putative binding might be most likely facilitated by the prevalence of negative charged amino acids on the P1 sequence and by the presence of positive charged amino acids on the XIP domain. Moreover, the assumption that EGFP-P1 exerts its effect by binding to the XIP_NCX1 domain on the f-loop of NCX1 is further validated by the results obtained with the NCX1 mutant, NCX1K229Q, which is devoid of a functional XIP domain.¹⁹ Indeed, in BHK cells expressing NCX1K229Q mutant, there was no colocalization signal with the cell-penetrating protein of P1. Moreover, the transfection of EGFP-P1 cDNA in these cells failed to modulate I_NCX1 in both modes of operation. In accordance with these results, some negative charged metabolic molecules such as the phosphatidylinositides PI, PIP, and PIP₂ (refs. 24,25,26) and GM1 ganglioside^27,28 were identified to directly bind the XIP domain on NCX1 and exert a stimulatory effect on NCX1 activity. On the other hand, nowadays, no evidence has been provided that there are proteins binding P1 domain on NCX1.

To deeper investigate the involvement of the different regions of the P1 peptide in NCX1 activity, the sequence of this small protein, made up of 127 amino acids, was divided into four main regions: region A (1–42aa) provided with a prevalence of negative charged amino acids; region B (44–50aa) provided with a prevalence of neutral amino acids; region C (57–84aa) provided with both positive and negative charged amino acids; and region D (108–125aa) a highly conserved region between the three exchanger isoforms. Deletion mutagenesis showed that the removal of the C region (EGFP-P1Δ57-84) reduced the stimulatory effect on NCX1 but did not hamper the ability of the mutant peptide to increase NCX1 activity. Deletion of region 44–84, comprising both the B and C regions, further decreased the stimulatory effect of the peptide on I_NCX1. Indeed, the increase in I_NCX in the forward mode of operation was not significant. These data suggested that although the central portion (B and C regions) of the P1 peptide influences the binding of the XIP domain, it is not crucial for the stimulatory effect. By contrast, the removal of the D region (108–125aa), i.e., mutant EGFP-P1Δ108–125, converted the stimulatory effect of the EGFP-P1 protein on I_NCX1 into an inhibitory effect, thus demonstrating the determinant role of this amino acid sequence. Accordingly, Maxwell et al.²⁹ showed that the deletion of six amino acids in the corresponding amino acid region D of NCX1, i.e., mutant NCX1Δ680-685, renders this exchanger insensitive to XIP-dependent inactivation by intracellular Na⁺, presumably for its inability to form or sustain inactive states. These results further support the idea that the last five amino acids of the D region in the P1 domain (680–685 on NCX1) are fundamental for regulating the activity of the exchanger. Consistently, the mutant NCX1Δ241–680, in which the XIP region and the active site of P1 on the f-loop (680-685aa on NCX1) are very close and can therefore more easily interact, possesses a very low activity comparable to that observed in BHK-NCX1 cells when they are inhibited by the EGFP-XIP protein. In accordance with these results, no XIP-dependent reduction of NCX activity has so far been reported in this exchanger mutant.²⁰ On the other hand, the EGFP-P1 protein increased I_NCX of NCX1Δ241-680 mutant by preventing the physiological inhibitory interaction between the two regions of the f-loop.

Another aspect of this study that deserves discussion is the effect of the P1 peptide on the currents of the other two isoforms, NCX2 and NCX3. Indeed, although P1 increased I_NCX2, the stimulation was lower than that observed for I_NCX1. The different magnitude of the stimulatory effect of P1 on NCX1 and NCX2 currents might be explained by the differences in the sequences of the XIP domains of NCX1 (XIP_NCX1) and NCX2 (XIP_NCX2) proteins. Indeed, XIP_NCX1 is provided with a neutral residue at position 16 (glycine234) and two positive charged basic residues at positions 17 (lysin235) and 19 (arginine237). Instead, at these same positions, XIP_NCX2 displays the negatively charged residue aspartate and two neutral residues, proline and serine. Thus, the different charges characterizing these amino acid residues might explain the reduced effect of P1 on the XIP_NCX2 domain of NCX2 observed in the present study. Analogously, Linck et al.³⁰ demonstrated that the XIP_NCX2 peptide exerts a weak inhibitory effect on NCX1 activity, thus supporting the idea that the binding between the XIP_NCX2 and the P1_NCX1 domains is weaker than that observed between XIP_NCX1 and P1_NCX1.

Unlike the stimulatory effect observed on I_NCX1 and I_NCX2, P1 exerted a mild inhibitory action on I_NCX3. Although there are no definitive explanations for this contrasting phenomenon, some evidence might help to explain this difference. First, NCX3 is the only exchanger isoform whose XIP-mediated Na⁺-dependent autoinhibition is not reverted by the presence of high intracellular ATP levels,³⁰ suggesting the presence of a peculiar mechanism of autoregulation. Second, although we observed that the XIP_NCX3 sequence is similar to that of XIP_NCX2, the further substitution of the two amino acids, valin227 and tyrosin228 with methionine and histamine, respectively, might have decreased the affinity for the binding of P1. Finally, we also evidenced that the P1 protein contains a region that inhibits the exchanger activity. Indeed, when the D region was removed from P1 (mutant P1Δ108-125), NCX1 activity was inhibited. On the basis of this evidence, it is possible to hypothesize that although P1 is unable to bind to the XIP_NCX3 domain and thus to stimulate NCX3 activity, an amino acid region of the peptide can still directly, or indirectly, inhibit the activity of the exchanger protein, as observed in the P1Δ108-125 mutant.

Altogether, these findings reveal a potentially effective new strategy for increasing the activity of NCX1 and NCX2 at the level of the cytoplasmic f-loop.

The effect of the endogenous expression of P1 protein, by cDNA transfection, or, more importantly, the exogenous cell-penetrating Tat-P1 construct on the activation of NCX1 might have a relevant clinical perspective in stroke. In fact, it has been largely demonstrated that the knocking down of NCX1 or the genetic ablation of NCX2 isoforms worsens ischemic brain damage,^10,31 thus suggesting that a strategy aimed to increase the activity of NCX1 and NCX2 might have a therapeutic relevance for reducing the damage induced by cerebral ischemia. In support of this promising therapeutic strategy, a compound that increases the activity of NCX1 and NCX2 in both forward and reverse modes of operations displayed a neuroprotective effect in an animal model of stroke.¹⁸

Furthermore, the results of the present study might also have a clinical interest in some cardiovascular diseases. Indeed, the overexpression of NCX1 may rescue postinfarction rat myocytes from contractile abnormalities³² and the pharmacological inhibition of NCX1 expression and activity aggravates postinfarction myocardial dysfunction and the diastolic function.^33,34 In addition, recent reports highlighted the key role of NCX1 in the initiation and maintenance of a stable heart rhythm,^6,35 whereas the overexpression of the NCX1 inhibitor phospholemman S68E has a detrimental effect on cardiac function both in terms of depressed cardiac function and increased arrhythmogenesis in transgenic mice.³⁶

Altogether, these findings reveal a potentially effective new strategy for increasing the activity of NCX1 at the level of the cytoplasmic f-loop. More specifically, by transfecting cells or by using the cell-penetrating form of Tat-P1, a small protein that by binding to the autoinhibitory XIP domain of the NCX1 f-loop removes its autoinhibitory effect, we were able to stimulate the exchange activity of NCX1. This new strategy might allow to rationally develop cell-penetrating proteins or peptidomimetic compounds that, by reproducing the functional pharmacophore of P1, required for the binding to the XIP domain on NCX1, might increase its activity and possibly exert a positive effect in those pathologies such as stroke, or some cardiovascular diseases, in which an increase in the activity of the antiporter might be therapeutically helpful.

Materials and Methods

Cell cultures. Stably transfected BHK cells with canine cardiac NCX1.1, rat brain NCX2.1, or NCX3.3 were grown on plastic dishes in a mix of Dulbecco's Modified Eagle's medium and Ham's F12 media (1:1; Gibco; Invitrogen, Milan, Italy) supplemented with 5% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin (Sigma, St Louis, MI) (complete medium). Cells were cultured in a humidified 5% CO₂ atmosphere; the culture medium was changed every 2 days. For microfluorimetric studies, cells were plated on glass coverslips (Fisher, Springfield, NJ) coated with poly-l-lysine (30 µg/ml; Sigma), and used at least 12 hours after seeding.

Generation and expression of mutant NCXs. EGFP-XIP and EGFP-P1 were generated by cloning in frame the corresponding cDNA of NCX1 in EGFP-C2 expressing vector (Clontech Laboratories, Mountain View, CA). NCX1 and P1 mutants were generated by means of the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Successful construction of the modified cDNAs was verified by sequencing both strands (Primm, Milan, Italy). EGFP-XIP, EGFP-Scr, EGFP-P1, and mutant exchangers were transiently or stably transfected in BHK cells according to the Lipofectamine 2000 (Invitrogen, Carlsbad, CA) protocol. Stable cell lines expressing mutant exchangers were selected by G418 resistance and by a Ca²⁺-killing procedure.³⁷ Scrambled control, i.e., EGFP-Scr, was generated by cloning the reverse cDNA of P1 into the EGFP-C1 expressing vector.

Measurement of intracellular Ca²⁺ concentrations. [Ca²⁺]_i was measured by single-cell computer-assisted video imaging. In brief, BHK cells were loaded with 10 µmol/l fura-2 acetoxymethyl ester (Fura-2/AM) for 30 minutes at 37 °C in normal Krebs' solution containing the following: 5.5 mmol/l KCl, 160 mmol/l NaCl, 1.2 mmol/l MgCl₂, 1.5 mmol/l CaCl₂, 10 mmol/l glucose, and 10 mmol/l HEPES-NaOH, pH 7.4. At the end of the fura-2/AM loading period, the coverslips were placed into a perfusion chamber (Medical System Greenvale, NY) mounted onto the stage of an inverted Zeiss Axiovert 200 microscope (Carl Zeiss, Jena, Germany) equipped with a FLUAR 40 oil objective lens. The experiments were carried out with a digital imaging system consisting of a MicroMax 512BFT cooled charge-coupled device camera (Princeton Instruments, Trenton, NJ), a LAMBDA 10-2 filter wheeler (Sutter Instrument, Novato, CA), and a Meta-Morph/MetaFluor Imaging System software (Molecular Devices, Sunnyvale, CA). After loading, cells were alternatively illuminated at wavelengths of 340 and 380 nm by a Xenon lamp. The emitted light was passed through a 512-nm barrier filter. Fura-2 fluorescence intensity was measured every 3 seconds. Forty to 65 individual cells were selected and monitored simultaneously from each coverslip. All of the results were presented as cytosolic Ca²⁺ concentrations. Since K_D for fura-2 was assumed to be 224 nmol/l, Grynkiewicz's equation³⁸ was used for calibration. NCX activity was evaluated as Ca²⁺ uptake through the reverse mode by switching normal Krebs' medium to Na⁺-deficient NMDG⁺ medium (Na⁺-free): 5.5 mmol/l KCl, 147 mmol/l N-methyl glucamine, 1.2 mmol/l MgCl₂, 1.5 mmol/l CaCl₂, 10 mmol/l glucose, and 10 mmol/l HEPES, pH 7.4. These experiments were performed with the irreversible and selective inhibitor of the sarco-(endo)plasmic reticulum Ca²⁺ ATPase thapsigargin (1 µmol/l), as previously described.³⁹

Electrophysiology. I_NCX were recorded by the patch-clamp technique in whole-cell configuration, as previously reported.^40,41,42 Currents were filtered at 5 kHz and digitized using a Digidata 1322A interface (Molecular Devices). Data were acquired and analyzed using the pClamp software (version 9.0; Molecular Devices). The I_NCX was recorded starting from a holding potential of −60 mV to a short-step depolarization at +60 mV (60 ms).⁴³ Then, a descending voltage ramp from +60 to −120 mV was applied. The current recorded in the descending portion of the ramp (from +60 to −120 mV) was used to plot the current–voltage relation curve. The magnitudes of I_NCX were measured at the end of +60 mV (reverse mode) and at the end of −120 mV (forward mode). Because Ni²⁺ blocks I_NCX, NiCl₂ (5 mmol/l) was routinely added to measure the NCX-independent currents. The Ni²⁺-insensitive components were subtracted from total currents to isolate I_NCX. Cells were perfused with external Ringer solution containing 126 mmol/l NaCl, 1.2 mmol/l NaHPO₄, 2.4 mmol/l KCl, 2.4 mmol/l CaCl₂, 1.2 mmol/l MgCl₂, 10 mmol/l glucose, and 18 mmol/l NaHCO₃, pH 7.4. Tetraethylammonium (TEA; 20 mmol/l), 50 nmol/l tetrodotoxin, and 10 µmol/l nimodipine were added to the Ringer's solution to block TEA-sensitive K⁺, tetrodotoxin-sensitive Na⁺, and L-type Ca²⁺ currents. The dialyzing pipette solution contained 100 mmol/l potassium gluconate, 10 mmol/l TEA, 20 mmol/l NaCl, 1 mmol/l magnesium ATP, 0.1 mmol/l CaCl₂, 2 mmol/l MgCl₂, 0.75 mmol/l EGTA, and 10 mmol/l HEPES, adjusted to pH 7.2 with Cs(OH)₂. Possible changes in cell size occurring upon specific treatments were calculated by monitoring the capacitance of each cell membrane, which is directly related to membrane surface area, and by expressing the current amplitude data as current densities (measured as picoamperes per picofarad). Capacitive currents were estimated from the decay of capacitive transients induced by 5-mV depolarizing pulses from a holding potential of −80 mV and acquired at a sampling rate of 50 kHz. The membrane capacitance was calculated according to the following equation: C_m = τ_c·I_o/ΔE_m(1 − I_∞/I_o), where C_m is membrane capacitance, τ_c is the time constant of the membrane capacitance, I_o is the maximum capacitive current value, ΔE_m is the amplitude of the voltage step, and I_∞ is the amplitude of the steady-state current.

Evaluation of mRNA and protein levels. BHK cells were firstly stably transfected with NCX1 or NCX1K229Q mutant, and then transfected transiently with EGFP-P1 or EGFP-Scr cDNA. For reverse transcription, 2.0 µg of each RNA extract from each cell group was reverse transcribed by SuperScript III (Invitrogen), according to Invitrogen protocol by using random hexamers. Total cDNA was amplified by Real-Time PCR (7500fast; Applied Biosystems, Monza, Italy) for canine NCX1 and normalized for the endogenous hamster hypoxanthine–guanine phosphoribosyltransferase signal. For protein expression analysis, whole-cell protein extracts from each BHK cell group were obtained and processed as previously reported.⁴⁰ Nitrocellulose membranes were incubated with anti-NCX1 antibody (rabbit polyclonal; Swant, Bellinzona, Switzerland; 1:1,000 dilution) or with mouse monoclonal antitubulin (1:1,000; Santa Cruz Biotechnology, Heidelberg, Germany). These nitrocellulose membranes were first washed with 0.1% Tween 20 and then incubated with the corresponding secondary antibodies for 1 hour (GE Healthcare, Little Chalfont, UK). Immunoreactive bands were detected with the enhanced chemiluminescence (ECL) (GE Healthcare). The optical density of the bands, normalized with tubulin (Sigma), was determined by Chemi-Doc Imaging System (Bio-Rad, Segrate, Italy).

Protein synthesis. According to previous studies,^44,45,46 a protein construct consisting of 6xHis (HHHHHH), HIV-Tat (GRKKRRQRRRQ), hemagglutinin (YPYDVPDYA), Flag octapeptide (DYKDDDDK) sequences, and the 562–688aa (P1) region of dog NCX1.1 named P1^Flag (Figure 4a), was rationally designed by our research group to: (i) cross biological membranes; (ii) bind the XIP domain on NCX1 on the citosolic side; and (iii) be recognized by specific antibodies. P1^Flag protein was expressed in Escherichia coli BL21 and purified to >85% from inclusion bodies by GenScript, Hong Kong, China. In particular, a single colony of E. coli BL21, harboring a plasmid containing Trx-His tag to facilitate the purification, was cultured and induced with 0.5 mmol/l isopropil-β-D-1-tiogalattopiranoside at 15 °C overnight. Cell pellets were lysed by sonication, precipitated by centrifugation, and then were dissolved in urea. Denatured protein was purified from endotoxin by Nickel column to 0.1–1 EU/µg as detected by limulus amebocyte lysate method. Target protein was eluted with a stepwise gradient of imidazole. Fractions were pooled, refolded, and sterilized by 0.22 μm filter and packaged aseptically. Proteins were analyzed by dodium dodecyl sulfate–polyacrylamide gel electrophoresis and western blot by using standard protocols for molecular weight and purity measurements. The primary antibody for western blot was mouse-anti-his mAb (GenScript; Cat.No.A00186). The concentration was determined by Bradford protein assay with bovine serum albumin (BSA) as a standard. P1^Flag was diluted in a storage buffer containing 50 mmol/l Tris–HCl, 150 mmol/l NaCl, 10% glycerol, pH 8.0. The storage buffer was used as vehicle.

P1^Flag protein uptake assay. Cellular uptake assay of P1^Flag was performed on BHK cells stably expressing NCX1 and plated on glass coverslips 24 hours before the experiment. After phosphate-buffered saline (PBS) washing, cells were incubated with serum-free complete medium containing vehicle or 0.25 to 1 µmol/l of P1^Flag for 90 minutes or 0.50 µmol/l of P1^Flag for 30 to 90 minutes. Afterward, BHK-NCX1 cells were washed several times with PBS and incubated with complete medium until immunocitochemical assay. No cell death was observed. An incubation of 0.50 μmol/l of P1^Flag for 60 minutes was used in electrophysiological studies, in situ proximity ligation assay, and immunocytochemistry experiments.

In situ proximity ligation assay. For in situ proximity ligation assay, BHK cells stably transfected with NCX1 or NCX1K229Q were plated on glass coverslips for 24 hours and then incubated with a serum-free complete medium containing 0.50 µmol/l P1^Flag, or vehicle, for 60 minutes. Afterward, BHK cells were washed several times with PBS and incubated with complete medium for 30 minutes. Then, cells were rinsed twice in cold 0.01 mol/l PBS at pH 7.4 and fixed at room temperature in 4% (w/v) paraformaldehyde for 20 minutes. Following three washes in PBS, cells were blocked for 45 minutes in PBS containing 3% BSA. They were then incubated overnight at 4 °C with a solution of PBS+BSA 3% containing the following antibodies: anti-NCX1 (rabbit polyclonal antibody; Swant; dilution 1:1,500) and anti-FLAG (mouse; Sigma; dilution 1:1,500). Next, cells were treated by using the Red Duolink kit (Sigma-Aldrich, Milan, Italy) according to the manufacturer's protocol. An immunocitochemical analysis for NCX1 and Flag immunosignals was performed on the BHK cell group expressing NCX1K229Q and incubated with P1^Flag as a control.

Immunocitochemistry and image acquisition. For immunocitochemistry experiments, BHK cells were rinsed twice in cold 0.01 mol/l PBS at pH 7.4 and fixed at room temperature in 4% (w/v) paraformaldehyde for 20 minutes. Following three washes in PBS, cells were blocked for 45 minutes in PBS containing 3% BSA. They were then incubated overnight at 4 °C with a solution of PBS+BSA 3% containing the following antibodies: anti-NCX1 (rabbit polyclonal antibody; Swant; dilution 1:1,500) and anti-FLAG (mouse; Sigma; dilution 1:1,500). Next, cells were washed in PBS, incubated with antirabbit Cy2-conjugated antibody (Jackson Immuno Research Laboratories, West Grove, PA; dilution 1:200) and antimouse Cy3-conjugated antibody (Jackson Immuno Research Laboratories; dilution 1:200) for 1 hour at room temperature under dark conditions, and washed again with PBS. Finally, they were mounted with a SlowFadeTM Antifade Kit (Molecular Probes-Invitrogen) and analyzed by confocal microscope. Immunofluorescece images were obtained using a Zeiss inverted 700 confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany) and a 63× oil immersion objective. The illumination intensity of the 543 Xenon and 488 Argon lasers, used to excite cy3 and cy2 fluorescence, was kept to a minimum of laser output to avoid phototoxicity.

Statistical analysis. All data were expressed as mean ± SEM. Statistical comparisons between experimental groups were performed using the t-test or one-way analysis of variance when required. P < 0.05 was considered statistically significant.