Calcineurin homologous protein: a multifunctional Ca²⁺-binding protein family

Abstract

The calcineurin homologous protein (CHP) belongs to an evolutionarily conserved Ca²⁺-binding protein subfamily. The CHP subfamily is composed of CHP1, CHP2, and CHP3, which in vertebrates share significant homology at the protein level with each other and between other Ca²⁺-binding proteins. The CHP structure consists of two globular domains containing from one to four EF-hand structural motifs (calcium-binding regions composed of two helixes, E and F, joined by a loop), the myristoylation, and nuclear export signals. These structural features are essential for the function of the three members of the CHP subfamily. Indeed, CHP1–CHP3 have multiple and diverse essential functions, ranging from the regulation of the plasma membrane Na⁺/H⁺ exchanger protein function, to carrier vesicle trafficking and gene transcription. The diverse functions attributed to the CHP subfamily rendered an understanding of its action highly complex and often controversial. This review provides a comprehensive and organized examination of the properties and physiological roles of the CHP subfamily with a view to revealing a link between CHP diverse functions.

Calcineurin Homologous Protein: General Paradigm and Protein Structure

calcium functions as a secondary messenger regulating a great variety of cellular processes in a spatial and temporal manner (14). The Ca²⁺-signaling network is composed of many molecular components, including a large family of Ca²⁺-binding proteins characterized by the EF-hand structural motif (calcium-binding regions composed of two helixes, E and F, joined by a loop) such as the calcineurin homologous protein (CHP) subfamily (49, 59, 122, 127).

The first member of the CHP subfamily, CHP1 (also known as p22), was identified in 1996, by two laboratories in the search of proteins with completely different functions. Barroso et al. (10) were looking at novel proteins involved in membrane trafficking, while Lin and Barber (73) were examining new proteins binding to the regulatory domain of the Na⁺/H⁺ exchanger (NHE). The same year, a third group characterized an NHE-interacting protein (called p24), which had biochemical characteristics resembling those of proteins of the CHP subfamily. The molecular identity of p24 was not determined (42).

To date, the CHP subfamily is composed of three members, CHP1 (73), CHP2 (103), and CHP3 (also called tescalcin) (107), which in vertebrates share with each other significant homology at the protein level (CHP1 and CHP2 60 and 75%, CHP2 and CHP3 23 and 42%, CHP1 and CHP3 48 and 30% identity and similarity, respectively) (44, 96) (Fig. 1A). Interestingly, CHP1 and CHP2 form a group that is distinct from CHP3 (Fig. 1B). Furthermore, the CHP subfamily shares extensive amino acid sequence homology with the regulatory B subunit of phosphatase calcineurin (CnB) (Fig. 1B). Other members of EF-hand Ca²⁺-binding proteins such as recoverins, frequenin, calmodulin (CaM), and Ca²⁺- and integrin-binding protein show lower but significant similarity to CHPs (4, 10) (Fig. 1B). This suggests that CHPs most likely evolved from a single common ancestor, and CHPs and CnB probably are derived from a precursor EF-hand protein (73).

Fig. 1.

Secondary structure of calcineurin homologous proteins CHP1–CHP3 and their relationship with other EF-hand Ca²⁺-binding proteins. A: alignment of amino acid sequences of vertebrate CHP1–3 demonstrating conserved domains across vertebrates. Conservative secondary structural elements are emphasized, myristoylation site (black); EF1 and EF2 hands, which do not bind Ca²⁺ (yellow); EF3 and EF4 that bind Ca²⁺ (blue); nuclear export signal (NES; green); flexible linker (violet). The N- and C-lobes are also shown. B: phylogenetic tree of EF-hands Ca²⁺-binding protein family based on amino acid alignment. Phylogenetic trees were obtained using the Bayesian approach, which is implemented in MrBayes (125). Numbers in the nodes represent the posterior probability. Hs = Homo sapiens; Pt = Pan troglodytes; Mm = Mus musculus; Rn = Rattus norvegicus; Mmu = Macaca mulatta; Dr = Danio rerio; Bt = Bos taurus; Xl = Xenopus laevis; Gg = Gallus gallus; Md = Monodelphis domestica, Oc = Oryctolagus cuniculus; Xt = Xenopus tropicalis.

Structurally, CHP1 consists of two globular domains, the N-lobe and C-lobe, which are composed of α-helical structures containing four EF-hand motifs in pairs (96) (Fig. 2); the N-lobe is composed of EF1 and EF2 and the C-lobe EF3 and EF4. The CHP2 crystal structure is similar to that of CHP1 (4), while the CHP3 crystal structure still remains unknown. The N- and C-lobes of CHP1 and CHP2 are connected by a unique long flexible linker that is not found in other Ca²⁺-binding proteins, including CHP3 (4) (Figs. 1A and 2). The functional significance of this linker is not clear at the moment.

Fig. 2.

Overall structure of CHP1. Ribbon diagram is shown of the structure of CHP1 with conservative domains emphasized following the scheme as in the legend of Fig. 1: EF1 and EF2 hands, which are unable to bind Ca²⁺ (yellow); EF3 and EF4 that bind Ca²⁺ (blue); nuclear export signal (NES) (green); flexible linker (violet). The structure was obtained from the PDB file (MMDB ID: 34033) deposited by Naoe Y. (96, 156). For clarity, it was color-coded using PyMOL software, and Ca²⁺ ions bound with EF3 and EF4 were removed.

Protein alignment of vertebrate CHP1–3 reveals that they share similar features (Table 1 and Fig. 1A) such as 1) a myristoylation signal (MGxxxT/S); 2) EF-hand Ca²⁺-binding domains; and 3) a nuclear export signal (NES). The conservation of the myristoylation signal in CHPs suggests that CHPs have Ca²⁺-myristoyl switch, akin to recoverin (89). Indeed, the Ca²⁺-induced conformational change in CHP1 seems to require its myristoylation (102). CHP1 and CHP2 have four EF-hands, EF1–EF4 (4, 96). EF1 and EF2 display significant deviation from the canonical EF-hand sequence and fail to bind Ca²⁺, while EF3 and EF4 bind Ca²⁺ (4, 96). On the other hand, CHP3 has only one EF-hand that is capable of binding Ca²⁺ and aligns well with EF3 of CHP1 and CHP2 (44, 77) (Fig. 1A). The point mutation D123A in the EF-hand of CHP3 was shown to completely abolish Ca²⁺ binding (44).

Table 1.

Summary of structural domains, expression patterns, and basic biochemical properties of CHP subfamily of proteins

	Conserved Domain	Expression Pattern	Biochemical Property
CHP1	Myristoylation signal; 4 EF-hands (EF1*, EF2*, EF3, and EF4); NES-1 and NES-2; flexible linker	Brain, lung, testis, kidney, spleen, and heart (10, 73)	Apparent Kd (Ca2+) = ∼90 nM (101). In complex with NHE1, Kd (Ca2+) = ∼2 nM (101).
CHP2	Myristoylation signal; 4 EF-hands (EF1*, EF2*, EF3, and EF4); NES-1; flexible linker	Intestine, kidney, and stomach (54); various malignant cells (103)	Kd (Ca2+) = ∼1 nM (4)
CHP3	Myristoylation signal; 1 EF-hand	Developing testis (107); heart, brain, and stomach (44)	Kd (Ca2+) = 800 nM (44). In the presence of 1 mM Mg2+, Kd (Ca2+) = ∼3.5 μM (44).

CHP, calcineurin homologous protein; NES, nuclear export signal; NHE, Na/H exchanger. *Denotes EF-hands unable to bind Ca²⁺.

At resting state, intracellular Ca²⁺ concentrations are generally maintained below 100 nM, although they can be rapidly and dynamically raised up to 1–10 μM upon specific stimuli (53). Indeed, one general way of classifying EF-hand Ca²⁺-binding proteins is based on their Ca²⁺ affinity and ability to respond to increasing Ca²⁺ signals. Proteins such as CaM, which are considered “Ca²⁺ sensors,” have affinities within the range of the stimulatory Ca²⁺ concentration (up to 10 μM) (28) and exhibit conformational changes that are Ca²⁺ induced, whereas proteins such as parvalbumin, which are considered “Ca²⁺ buffers,” constitutively bind Ca²⁺ with affinities of <100 nM and generally lack conformational changes upon Ca²⁺ binding (28).

Each of EF3 and EF4 in CHP1 can bind a Ca²⁺ ion [overall apparent constant of dissociation (K_d) is ∼90 nM; the Hill coefficient is ∼1.0]. This Ca²⁺-binding affinity is similar to that of the CnB (overall K_d is ∼70 nM) (139), although the K_d for other EF-hand Ca²⁺-binding proteins may vary broadly (23, 43, 105, 139). CHP2 binds Ca²⁺ with a higher affinity (K_d = 1 nM) (4) than CHP1. CHP3 is also capable of binding Ca²⁺ although the apparent K_d is ∼100 times higher (800 nM) compared with CHP1 and CHP2 (44, 77). CHP3 also binds magnesium (Mg²⁺), and its K_d for Ca²⁺ shifts up to 3.5 μM in the presence of 1 mM Mg²⁺. Since the physiological intracellular Mg²⁺ concentration is in the millimolar range, these results suggest that in living cells CHP3 constitutively binds Mg²⁺ and upon the momentary increase in intracellular Ca²⁺, Mg²⁺ may be exchanged for Ca²⁺ ions (44). Based on these observations, the CHP subfamily may serve as both Ca²⁺ sensors and Ca²⁺ buffers (Table 1).

CHP1 has two conserved nuclear export signals (NES-1 and NES-2) (94) (Figs. 1A and 2), while CHP2 has only NES-1, with CHP3 not having any canonical NES signal. Simultaneous mutations in both NES-1 and NES-2 lead to nuclear localization of CHP1 (94), indicating that one NES is sufficient to transport CHP1 from the nucleus to the cytosol (71).

CHP Localization Patterns

CHP1 is expressed in most tissues such as the brain, lung, testis, kidney, spleen, and heart (10, 73). Expression of CHP2 is largely restricted to normal intestinal epithelia (54), but it is induced in various malignant cells (103). Specifically, two transcripts (1.9 and 1.8 kb) of CHP2 mRNA were detected in both the small (jejunum and ileum) and large intestine. A small amount of the CHP2 transcript was found in the kidney, and traces were also observed in the stomach. The nature and function of these CHP2 transcript variants are unknown. CHP2 protein was detected only in the total lysate of the small and large intestine, but not in the kidney or stomach, which may reflect a low level of expression of CHP2 in those tissues (54). The functional significance of the specific localization of CHP2 in the intestine is unknown. Interestingly, the distribution of CHP1 and CHP2 transcripts in the epithelial layer of the small intestine was found to be different; CHP1 mRNA was detected in the connective tissue and epithelium, while CHP2 mRNA was found only in the epithelial layer on the small intestinal luminal side (54). The abundant localization of CHP2 in the epithelial layer of the small intestine lumen may be associated with its function in the absorptive epithelium of the intestine. The differential distribution of CHP1 and CHP2 may correlate with an undefined, although possibly different, role for the two isoforms in the intestinal cells. Nevertheless, CHP2 was originally identified in human cancer, being implicated in cancer progression; hence it may have an oncogenic potential. CHP3, which was originally isolated from the developing mouse testis and termed “tescalcin” (107), is detected predominantly in adult mouse heart, brain, and stomach (44), although in adult human tissues it appears to be restricted primarily to the heart (77) (Table 1). In summary, whether a specific localization reflects a precise function of the CHP subfamily members remains unclear.

Targets of CHP Encompass Diverse Functions

The three members of the CHP subfamily have multiple and diverse functions, ranging from the regulation of specific plasma membrane protein functions to carrier vesicle trafficking and gene transcription. Some of these functions are specific to each isoform of the subfamily, while others are common to all. However, some of these observations may reflect a lack of knowledge concerning each family isoform function rather than a true functional diversification between CHP isoforms. The discussion will be considered separately for CHP1, CHP2, and CHP3. The seemingly larger portion weighted toward CHP1 does not necessarily mean a greater importance but rather that there are more data available for CHP1.

CHP1.

There is a rather large database on CHP1 targets. To better organize the data, the targets of regulation will be broadly divided into plasma membrane transporters, vesicular trafficking, and gene transcription.

REGULATION OF PLASMA MEMBRANE PROTEIN FUNCTIONS.

NHE is targeted by the CHP subfamily. NHE is a family of ubiquitous proteins with a very wide array of physiological functions that catalyzes the electroneutral exchange of Na⁺ for H⁺. The human genome contains nine NHE isoforms (NHE1–9), two NHE-related genes (NHA1–2), and one sperm-specific NHE (20, 68, 99, 154). NHEs are ubiquitous and implicated in multiple physiological and pathophysiological processes. Structurally, NHEs have 2 major domains: an N terminus that harbors 12 putative membrane-spanning segments responsible primarily for cation permeation and exchange, and a C terminus that resides in the cytoplasm and functions mainly to regulate transport activity, membrane targeting, anchorage to and scaffold for the underlying actin cytoskeleton, and as a platform for the assembly of signaling complexes (16, 20, 37, 87, 98, 99).

Of the mammalian NHE isoforms, NHE1 was first identified in 1989 (130) and has received considerable attention because it is widely expressed and plays a vital role in many processes such as intracellular pH (pH_i) homeostasis and the maintenance of cell volume as well as cell shape, cell cycle, cell proliferation, cell migration, cell differentiation, cell adhesion, and resistance to apoptosis (24, 25, 47, 118). Lin and Barber (73) identified the interaction of CHP1 with NHE1 at the regulatory site of the NHE1 C terminus that was demonstrated to be critical for growth factor activation of the exchanger. Indeed, deletion of this NHE1 C-terminal domain completely abolishes growth factor-stimulated NHE1 exchange activity (151). Overexpression of CHP1 inhibited NHE1 activation by growth factor- and by mutation-activated GTPases. These findings indicate that overexpression of CHP1 in cells induces a constitutive CHP1-NHE1 binding, which probably is overriding the dissociation of endogenous CHP1 from NHE1 and blocking NHE1 activation upon specific stimuli. Lin and Barber (73) hypothesized that CHP1 may constitutively bind to NHE1 in the resting state and that CHP1 is released from the binding to the exchanger upon stimulation by growth factors. Release of CHP1 could either trigger NHE1 to take on an activating conformation or allow binding of an unspecified exchanger-positive regulator. The finding that CHP1 is a phosphoprotein and that its dephosphorylation is associated with increased NHE1 activity indicates that modifications in the CHP1 phosphorylation may regulate its ability to bind NHE1 and its action on NHE1 activity (73).

This intriguing model of CHP1-NHE1 interaction was not confirmed in other studies. Pang et al. (102) produced a series of fusion proteins containing various regions of the carboxyl-terminal cytoplasmic domain (amino acids 503–815) of NHE1 and examined their interaction to CHP1 by Far-Western analysis. The first discrepancy between the studies by Pang et al. and Lin and Barber (73) was on the amino acid (aa) sequences of NHE1 C terminus required for CHP1 binding; aa 515–530 in Pang et al. (102) vs. aa 567–635 in Lin and Barber study (73). Although the reason for the discrepancy between these studies is not clear, recent structural study has confirmed the requirement of NHE1 aa sequences 515–530 for CHP1 binding (90).

The CHP1-binding domain of NHE1 was predicted to form a conserved α-helix (102). A helical wheel diagram for amino acids 518–535 of NHE1 revealed a cluster of hydrophobic residues. Mutants of NHE1 with substitution of hydrophobic for hydrophilic residues, which did not coimmunoprecipitate with CHP1, were expressed at the cell surface (102). NHE1 seems to be the principal target for CHP1 in the membrane because CHP1 per se was expressed uniformly in the cytosol but became partly localized at the cell surface upon NHE1 expression (102).

The activity of these mutants was dramatically reduced, despite the fact that nearly the same amount of NHE1 protein was expressed at the cell surface, implying that CHP1 binding to NHE1 is required for optimal NHE1 activity. CHP1 binding to NHE1 was found to affect the pH_i sensitivity of NHE1 by increasing its sensitivity at acidic pH_i (102). Indeed, cytoplasmic alkalization in response to several extracellular stimuli known to activate NHE1 (e.g., thrombin, hyperosmolarity) was not observed in cell-expressing mutants defective for the CHP1-binding region to NHE1. Interestingly, in the studies of Pang et al. (101, 102), overexpression of CHP1 did not reverse NHE1 activation by growth factors as proposed by Lin and Barber. (73). This discrepancy might reflect either differences in the experimental systems between the studies or that CHP1 action on NHE1 is more complex than described so far. CHP1 binding to NHE1 may inhibit or activate NHE1 activity, depending on both the cellular environment and the applied stimulus.

The complex action of CHP1 on NHE1 can be shown by the NHE1 sensitivity to intracellular ATP depletion being lost in cells expressing mutation of the CHP1-NHE1 binding site (102). Similar to a lack of CHP1-NHE1 interaction, ATP depletion has been associated with a drastic reduction in NHE1 activity and a loss of pH_i sensitivity, which may indicate a connection between NHE1 pH_i sensitivity, NHE1-CHP1 binding, and NHE1 sensitivity to intracellular ATP concentration. Interestingly, ATP depletion was associated with a reduction in the content of plasma membrane phosphatidylinositol 4,5-bisphosphate (PIP₂), and NHE1 has two putative binding sites for PIP₂ on the NHE1 C terminus; one of these is predicted to locate before and the other after the CHP1-binding region to NHE1 (1). It would be very informative to evaluate whether there is a functional connection between NHE1 binding to PIP₂ and CHP1.

As described above, CHP1 is myristoylated and contains four potential EF-hand Ca²⁺-binding motifs; two of these bind Ca²⁺ (EF3 and EF4). Myristoylation was not required for the interaction of CHP1 with NHE1, but it was required for Ca²⁺-induced conformational changes in CHP1 as for other Ca²⁺-binding proteins (89). Only double mutations of EF3 and EF4 impair the interaction of CHP1 with NHE1, a mutation of either EF3 or EF4 significantly reducing NHE1 activity and the NHE1 response to stimuli such as PKC activation, serum, and thrombin (101). These findings indicate that mutations at EF3 or EF4 impair NHE1 activity both at the baseline and upon stimuli.

Remarkably, the binding affinity of CHP1 for Ca²⁺ increased by ∼40-fold when CHP1 formed a complex with NHE1 (Table 1). The extraordinarily high affinity of CHP1-NHE1 for Ca²⁺ (∼2 nM) implies that the complex always contains two Ca²⁺ ions under physiological conditions. To demonstrate how the formation of the CHP1-NHE1 complex increases Ca²⁺-binding affinity, it was shown that most of the Ca²⁺ bound to CHP1 without the NHE1 fragment was released rapidly. In contrast, Ca²⁺ released from the CHP1 complex with NHE1 was much slower (101). These findings imply that the binding of CHP1 to Ca²⁺ in the CHP1-NHE1 complex may also play a structural (stabilization of the hydrophobic cleft) role and explain the findings of the impairment of the NHE1 baseline activity by a mutation at EF3 and EF4.

In summary, it is conceivable that the subdomain of the NHE1 C terminus that binds CHP1 is a key structure that permits putative pH_i sensing. The binding of Ca²⁺ may play an important role in maintaining CHP1 structure, thereby preserving the physiological pH_i sensitivity of NHE1. Mutation of arginine 440 in the fifth intracellular loop of NHE1, which connects transmembrane helices 10 and 11, markedly reduces the pH_i sensitivity of NHE1 (152). Specifically, substitution of arginine 440 with cysteine shifted the pH_i dependence of Na⁺ uptake to an acid side with a pK value of <6.2 while ATP depletion did not change pH_i sensitivity in cells expressing arginine 440 mutations. Although CHP1 binding to NHE1 and mutation of arginine 440 may alter the exchanger pH_i sensitivity independently, an attractive untested hypothesis could be that CHP1 and arginine 440 coordinately affect NHE1 pH_i sensitivity. For instance, CHP1 could bind NHE1 at two loci (aa sequence 515–530 and arginine 440). Thus substitution of arginine 440 with (an)other residue(s) would result in a partial disruption of CHP1-NHE1 interaction, thereby altering the pH_i sensitivity of the exchanger. Nevertheless, a second binding site of CHP1 to NHE1 in the fifth intracellular loop of NHE1 has not been studied thus far.

CaM also interacts with NHE1. Thus NHE1 appears to be regulated by two Ca²⁺-binding proteins. CHP1 could preserve the physiological pH_i sensitivity of NHE1, whereas CaM could play a role in sensing cytosolic Ca²⁺ (150).

Another unsolved contradiction in the field is that CHP1 was found to affect not only changes in NHE1 pH_i sensitivity, but lack of CHP1 expression reduced NHE1 protein expression at the plasma membrane and in whole cells without changes in NHE1 mRNA expression (82). These findings are in conflict with the previous data of Pang et al. (102) showing that NHE1 lacking CHP1 binding is trafficked to the plasma membrane, suggesting no requirement of CHP1 binding to NHE1 for the exchanger localization at the plasma membrane. To date, the reason for the discrepancy between the studies is not clear.

In addition, NHE1 degradation by the ubiquitin-proteasome system participates to the loss of NHE1 in CHP1-deficient cells (132). Binding of CHP1 to NHE1 may stabilize the NHE1 structure, allowing it to stay at the plasma membrane and/or reach the plasma membrane (83). Indeed, CHP1 expression in CHP1-deficient cells rescued low NHE1 protein expression (82). Finally, inhibition of the proteasomal pathway did not completely inhibit the reduction of NHE1 total protein in CHP1-deficient cells (82), thus suggesting that other mechanisms (e.g., decrease in NHE1 synthesis) may be involved in the loss of NHE1.

The CHP1-binding domain is well conserved among mammalian NHE isoforms from 1 to 5 (102). Indeed, CHP1 binds to NHE1–5 (54). Analogously to NHE1, mutants of other NHE isoforms (e.g., NHE2 and NHE3) that disrupt binding exhibited a markedly reduced exchange activity. NHE isoforms exhibit very different modes of regulation by many physiological factors. This functional diversity suggests that different NHE isoforms have fundamental differences in the regulatory mechanism. For instance, NHE3, which is mainly localized in the brush-border membrane of the renal proximal tubule and small intestine (2, 15, 159), mediates the absorption of the bulk of filtered sodium and fluid (115, 116, 159). NHE3 is highly regulated by a large number of factors (17, 159). One such agonist is the endogenous purine nucleoside adenosine that modulates many physiological processes, and in stress or injury, such as during ischemia, it offers cytoprotection, thus preventing tissue damage (33).

NHE3 binds CHP1 (35, 102), and both the constitutive function of NHE3 and its acute regulation by adenosine involve CHP1. Activation of the adenosine A₁ receptors (A1R) by selective agonists inhibits the exchanger activity and surface protein expression via phospholipase C/Ca²⁺/PKC signaling pathways (36). Amino acid residues within the CHP1-binding region in NHE3 are essential for A1R control of NHE3 activity. Indeed, A1R activation increases the binding between CHP1 and NHE3, and interference with this binding reverses the A1R inhibition of NHE3 activity. These findings imply that CHP1 binding to NHE3 may be a downstream signal to PKC activation (35), which was also proposed for NHE1 regulation by PKC (101). However, in a system with either CHP1 overexpression or knockdown, CHP1's effect on NHE3 seems more complex. CHP1 increases NHE3 constitutive transport function and protein abundance, showing that CHP1 is not only necessary for exchanger regulation but also for its constitutive transport function. Once more, the function of CHP1 binding to NHE isoforms seems contradictory. On one hand, CHP1 binding to NHE3 mediated by acute activation of A1R inhibits NHE3 and, on the other hand, overexpression of CHP1, which presumably increases CHP1-NHE3 binding, activates NHE3. One possible explanation for this apparent discrepancy between the studies is that the first study investigated the effect of acute CHP1 action on NHE3 while the second studied the effect of chronic CHP1. As highlighted above, it is conceivable that these functional diversities induced by CHP1-NHE3 binding may reflect specific settings. One possible model is that ezrin phosphorylation is necessary for the CHP1-dependent control of NHE3 baseline transport activity and surface protein expression (34). Ezrin is a member of the ezrin/radixin/moesin family that is highly enriched in apical microvilli of epithelial cells and is responsible for the membrane targeting of membrane proteins by directly binding them to the actin cytoskeleton (19, 39, 51, 75). NHE3 binds directly and indirectly to ezrin (26, 157, 158). Direct binding of NHE3 to ezrin is in close proximity to the NHE3-CHP1-binding region and affects several aspects of basal NHE3 trafficking (26), thus lending support to a model of CHP1 and ezrin existing in a regulatory complex of NHE3 function (34).

REGULATION OF CARRIER VESICLE TRAFFIC FROM ORGANELLES TO THE PLASMA MEMBRANE.

Ca²⁺ triggers many forms of exocytosis in different types of eukaryotic cells. To fulfill this function, Ca²⁺ signals should be transduced to other protein(s) involved in vesicle exocytosis. A highly conserved family of proteins, such as synaptotagmins, serves as a sensor for Ca²⁺-induced exocytosis (104). EF-hand Ca²⁺-binding proteins are proposed to function as Ca²⁺ sensors (76, 104). Indeed, frequenin, a Drosophila member of the recoverin family of EF-hand Ca²⁺-binding proteins, was shown to facilitate neurotransmitter release in the neuromuscular junctions (113).

CHP1 was identified in a rat liver cDNA library screen using a specific antibody generated against a homogeneous protein population, present in specialized exocytic vesicles of polarized epithelial cells (10). In a cell-free assay, CHP1 was found to be required for targeting and fusion of transcytotic vesicles (TCV) with the apical plasma membrane. The addition of CHP1 antibody to the fusion reaction dramatically reduced fusion efficiency, while specific mutations of the EF-hands, which compromise Ca²⁺-binding to CHP1, render CHP1 unable to function in transcytotic targeting/fusion. N-myristoylation was found to be essential for CHP1 function in exocytic traffic, as the addition of non-N-myristoylated recombinant CHP1 to the fusion reaction between TCV and the plasma membrane did not reverse the inhibitory effect induced by CHP1 antibody (10).

In summary, CHP1 might act as a Ca²⁺ sensor for constitutive exocytosis (Fig. 3B) within the concentration range of those shown for other constitutive fusion events (0.1–1 μM). This is because 1) it is widely expressed and conserved; 2) it is required for vesicle targeting/fusion; and 3) it undergoes conformational change upon the binding of Ca²⁺ (13, 84, 92, 140). The molecular mechanisms of the action of CHP1 on constitutive exocytosis are unknown. However, like other members of the EF-hand superfamily, it most likely uses conformational changes to transduce Ca²⁺ signals to other protein(s) involved in vesicle exocytosis.

Fig. 3.

Proposed model of the diverse signal pathways regulated by CHP1. A: Na⁺/H⁺ exchanger 1 (NHE1) in lamellipodia and membrane ruffles of polarized fibroblast like-cells (12). This is one of NHE1-specialized plasma membrane localizations that correlates with the formation of the CHP-NHE complex at the cell surface and regulation of NHE1 function by CHP1. B: fusion of transcytotic vesicles with the plasma membrane and possible molecular model of microtubule and actin interaction with CHP1 functional control of vesicle transport at the plasma membrane of epithelial cells. C: Golgi transport and similar function of CHP1 as for ER transport. D and E: ER transport and function of CHP1 to facilitate vesicle movements along microtubule cytoskeleton. F and G: CHP1 translocation to the nucleus mediated by binding to death-associated protein kinase (DAP-kinase)-related apoptosis-inducing kinase (DRAK2).

The microtubule cytoskeleton is believed to participate in most of the intracellular membrane-trafficking events, including exocytotic vesicle traffic to the plasma membrane (7, 38). Dependence on microtubules was demonstrated for the intracellular organization of the endoplasmic reticulum (ER) network (11, 67), the Golgi apparatus (144), mitochondria (120), lysosomes (91), endosomes, and plasma membrane (7, 46). Ca²⁺ controls cell polarity and shape by modulating membrane trafficking and microtubule dynamics in polarized secretory epithelia (9). Specifically, the early secretory pathway is regulated by the Ca²⁺ efflux from the intracellular stores (27, 55, 114), and the ER and Golgi apparatus are well-studied intracellular stores that can release Ca²⁺ upon stimulation (97, 110, 112). Regulatory EF-hand proteins, such as CaM, involved in membrane trafficking also associate with microtubules (56, 109, 114, 117).

CHP1 distributes itself along microtubules during interphase and mitosis in various cell lines (145), and it associates not only with the microtubules but also with membrane-bound organelles (6). This is compatible with the hypothesis that CHP1 is a signal protein mediating organelle assembly with the microtubules. Indeed, in a membrane-microtubule-binding assay that allowed separation of the binding of CHP1 to microtubules from their association with membranes, CHP1 increased the amount of ER, and to a lesser extent Golgi, membranes associated with microtubules (6). These findings indicate that CHP1 plays a role in the interactions of ER and Golgi membranes with microtubules.

CHP1-dependent microtubule-membrane association is Ca²⁺ dependent. However, CHP1 was found to bind microtubules independently of Ca²⁺ (6), thus indicating that Ca²⁺ is only responsible for modulating the ability of CHP1 to associate with microsomal membranes. Indeed, CHP1's association with microsomal membranes is modulated by a Ca²⁺-myristoyl switch, indicating that in the Ca²⁺-free state, the myristoyl moiety in CHP1 is inaccessible to membranes, as has been demonstrated for recoverin (3, 142). The Ca²⁺-mediated conformational change in CHP1, which exposes the myristoyl group, promotes the association of CHP1 with the membrane. Although this Ca²⁺-myristoyl switch model does not apply to all EF-hand proteins (69, 135–138), it explains the changes undergone by recoverin's three-dimensional structure upon binding Ca²⁺ (3, 142).

Experiments with CHP1 antibody delivery by digitonin-based bulk microinjection supported the model that CHP1 also regulates microtubule dynamics and stability by affecting the organization of the ER network in two distinct manners. First, dependent on CHP1's ability to undergo Ca²⁺-induced conformational changes, CHP1 affects ER network formation. Second, independent of its ability to undergo Ca²⁺-induced conformational changes, CHP1 mediates the retraction of the ER apparatus toward microtubule bundles. CHP1 associates with membranes of the ER network that act as Ca²⁺ stores and was found to release large amounts of Ca²⁺ upon stimulation. Thus CHP1 may be exposed to rising local concentrations of Ca²⁺ that would allow CHP1 to undergo Ca²⁺-induced conformational changes and thus transduce intracellular Ca²⁺ signals to downstream effectors.

In summary, it is important to identify molecules that link membrane traffic to the microtubule network to control and stimulate cellular processes, such as exocytosis. CHP1 is an excellent candidate for such a protein because it mediates the interactions between microtubules and membrane-bound organelles and the targeting/fusion of exocytotic vesicles to the plasma membrane. The CHP1-microtubule association might indicate a function of CHP1 in facilitating vesicle movement along microtubules, or alternatively this association might regulate the dynamics and organization of microtubules that are also involved, although indirectly, in vesicle trafficking.

This prompts the question as to how CHP1 may facilitate vesicle movement along the microtubule. This question triggers even broader queries such as how the cytoskeleton contributes to the delivery of carrier vesicles between organelles and then from the organelles to the plasma membrane. In addition, how do the morphology and location of organelles as well as the targeting and directionality of vesicle traffic depend on the microtubule cytoskeleton organization?

These questions are far from being addressed; however, some work has established the mechanisms by which CHP1 may affect the traffic of carrier vesicles along the microtubule cytoskeleton. First, CHP1 does not possess any characteristic microtubule-binding domain; therefore, it probably associates with microtubules indirectly via cytosolic microtubule-binding factor(s). Indeed, CHP1 interacts with several microtubule-associated proteins within the 30- to 100-kDa range. One such CHP1-binding partner, a 35- to 40-kDa microtubule-binding protein, has been identified in liver and kidney microtubule pellets as GAPDH (5). Several lines of evidence indicate a functional relationship between CHP1 and GAPDH: 1) CHP1 coimmunoprecipitates with GAPDH; the N-myristoyl group and the last six C-terminal amino acid residues are involved in the interaction of CHP1 with GAPDH, hence in CHP1-microtubule association; 2) CHP1 and GAPDH line up along the microtubules in punctate structures in cells; and 3) GAPDH facilitates the CHP1-dependent interactions between microsomal membranes and microtubules by enhancing CHP1 ability to bind microtubules (5).

GAPDH is a classic glycolytic enzyme that also participates in microtubule bundling, membrane fusion, phosphotransfer activity, nuclear RNA export, DNA replication, and DNA repair (133). Moreover, GAPDH associates with both actin and the microtubule cytoskeleton (131, 146, 147). GAPDH seems responsible for ∼70% of the CHP1-microtubule-binding activity, with other yet unknown proteins accounting for the remaining ∼30%. CHP1 may bind GAPDH directly through a nonlinear binding site that should include at least the N-myristoyl group and the extreme C-terminal region. CHP1 interacts with microtubules through cytosolic protein(s) other than GAPDH in an N-myristoylation- and C-terminal-independent manner (5).

Taken together, these findings suggest that GAPDH facilitates CHP1-dependent microtubule-membrane interactions by affecting the ability of CHP1 to associate with microtubules but not with microsomal membranes (Fig. 3, C–E).

Molecular motors bound on the surface of cargo organelles drive intracellular vesicle transport (Fig. 3B). They interact with and ride on cytoskeletal structures such us microtubules or actin filaments that serve as conduits for the cargo movement (Fig. 3B). In the current model, it is believed that long-distance movement of organelles is supported by microtubules whereas local transport is via actin filaments. The dual system collaborates to deliver organelles to specific cellular destinations (21, 40, 111). Indeed, the same organelles can bear both microtubule- and actin filament-associated motors (Fig. 3B). Molecular motors such as members of the kinesin, dynein, and myosin superfamilies are implicated in trafficking of vesicle and organelle. Specifically, kinesin and dynein motors are involved in long-range vesicle transport along microtubules, whereas myosin motors mediate short-range transport on actin filaments (21, 41) (Fig. 3, B–E).

Kinesin and most kinesin-related proteins drive vesicles along microtubules toward the cell periphery (50). In a yeast two-hybrid screen of a rat brain cDNA library, CHP1 was found to interact with a new member of the kinesin superfamily, KIF1Bβ2 (Fig. 3, B–E). The CHP1-binding region in KIF1Bβ2 is located at the carboxyl-terminal end of the motor domain and is Ca²⁺ dependent and specific for KIF1Bβ2 (95). Subcellular fractionation showed colocalization of KIF1Bβ2 with a synaptic transport vesicle marker protein, synaptophysin (95). This is in agreement with the findings that deletion of the KIF1Bs gene leads to impaired transport of synaptic vesicle proteins to the distal region of the axon (162). This implies some transporting function of KIF1Bs to the nerve terminal and that KIF1Bβ2-CHP1 binding may modulate the transport of synaptic vesicles in neuronal cells. Although the functional significance of KIF1Bβ2-CHP1 binding is still unclear, the observation that an EF-hand Ca²⁺-binding protein is involved in the kinesin-motor protein function is interesting.

Considering the fact that CaM is known to interact with myosin and modulate short-range transport on actin filaments (86, 121), CHP1 may facilitate the Ca²⁺-mediated mobility of kinesin-motor proteins on microtubule filaments. However, the coordination of these types of transport and the mechanisms that control the switching between microtubule- and actin filament-based organelle transport are extremely complex. Switching of vesicles between the two major cytoskeletal systems is tightly controlled by signal cascade events (124).

In summary, CHP1 that is required for targeting/fusion of exocytotic vesicles with the plasma membrane (10) and is involved in microtubule-membrane interaction (6, 145) may easily serve as a key signal molecule in Ca²⁺-dependent intracellular transport of organelles along cytoskeletal structures (Fig. 3B).

REGULATION OF GENE TRANSCRIPTION AND NUCLEAR FUNCTION.

CHP1 is named after calcineurin because of its high homology with this phosphatase. However, CHP1 also controls calcineurin function. Calcineurin is a heterodimeric enzyme composed of the catalytic subunit CnA and the Ca²⁺-binding regulatory subunit CnB (63, 93). This phosphatase requires the Ca²⁺-dependent binding of CaM to the CnA-CnB complex at a domain distinct from the CnB-binding site for maximal activity (108). Activation of calcineurin has been involved in dephosphorylation and nuclear translocation of the nuclear factor of activated T cells (NFAT), which in turn drives the transcription of response genes such as the cytokines IL-2 and IL-4. These mechanisms are involved in the control of T cell proliferation and the immune response (8, 31, 148).

CHP1 overexpression in HeLa and Jurkat cells impaired the nuclear translocation and the transcriptional activity of NFAT (74). Indeed, CHP1 inhibits calcineurin phosphatase activity in vivo by binding to the CnA subunit and probably impairing the assembly of the heterotrimeric calcineurin holoenzyme (74). CHP1 association with CnA did not reconstitute the calcineurin activity in the presence of either CnB or CaM, thus implying that CHP1's association with CnA may affect CnB and CnA binding, and preventing the conformational change in CnA required for its activation. Hence, CHP1 could function in a manner analogous to the action of the calcineurin activity inhibitors cyclosporin A and FK506. Activation of T cell signaling also induces a decrease in CHP1 abundance that may involve a positive feedback mechanism for releasing an inhibitory action of CHP1 during T cell receptor signaling. Therefore, CHP1 may represent an emerging class of endogenous calcineurin regulators.

Furthermore, considering the extensive homology of CHP1 with CnB, CHP1 might serve as a regulatory protein in response to Ca²⁺ with another putative CnA-like catalytic subunit protein. Death-associated protein kinase (DAP-kinase)-related apoptosis-inducing kinase (DRAK2) was isolated from a rat brain cDNA library. Transient expression of DRAK2 induces morphological changes typically found in apoptotic cells (129). CHP1 binds DRAK2 in vitro, and at least 67 residues of the C-terminal end region of the death-associated protein kinase domain are responsible for this binding to CHP1. Interestingly, heterologous expression of both proteins in COS cells caused a drastic change in the localization of CHP1 within the same cell (81). A significant fraction of CHP1 moved to the nucleus without a change in the localization of DRAK2, which was mainly in the nucleus (Fig. 3, F and G). A small but significant fraction of CHP1 and DRAK2 colocalized in proximity to the membrane surface of COS cells, suggesting that CHP1 and DRAK2 also may interact at the cell surface.

DRAK2 belongs to an emerging DAP kinase family of serine/threonine protein kinases (61). Other members of the DAP kinase family have been proposed to play a key role in the transduction of apoptotic signal cascade. DAP kinase is believed to act as a positive intermediary of the proapoptotic signals and to cause cell death. Moreover, DRAK2 is believed to be involved in the execution of the apoptosis. DAP and DRAK2 kinases were both shown to bind CaM (29, 64), and the peptides of these proteins without the CaM-binding domain enhanced the apoptosis (29, 64). These findings raised the possibility that DRAK2 might be regulated by CHP1.

Indeed, CHP1 specifically and significantly inhibits the kinase activity of DRAK2 (∼85% inhibition, IC₅₀ for CHP1 1.5 μM) for both autophosphorylation and phosphorylation of the exogenous substrate (myosin light chain) in vitro. The CHP1-induced inhibitory effect was Ca²⁺ dependent whereas the binding between CHP1 and DRAK2 was not (65). This shows that the inhibition of the kinase activity by CHP1 is not due to simple CHP1-DRAK2 association and that the CHP1-DRAK2 complex association and dissociation with Ca²⁺ may cause a conformational switch between an inactive and active state. Nevertheless, the physiological function of CHP1 in DRAK2 regulation is still unknown, as it is not clear why CHP1 moves to the nucleus in response to DRAK2 binding. Two possibilities are hypothesized: 1) the translocation of the CHP1-DRAK2 complex to the nucleus mediated by a nuclear localizing signal on DRAK2; and 2) the recycling of CHP1 between the nucleus and the Golgi apparatus and sequestration of CHP1 in the nucleus by DRAK2 (Fig. 3, F and G). This second hypothesis would fit with the CHP1 function in the membrane traffic as the Golgi apparatus plays a central role in this process. Additionally, some CHP1 and DRAK2 were found at the plasma membrane where NHE1, another CHP1-binding partner, is also expressed. Hence, CHP1 might facilitate translocation of NHE1 from the Golgi to the plasma membrane and DRAK2 might contribute in this event.

In the nucleus, CHP1 also associates with the upstream binding factor (UBF) and inhibits ribosomal RNA synthesis, thus indicating that nuclear-localized CHP1 is regulated and has a distinct function compared with CHP1 in other cellular compartments (57). UBFs are transcription factors that bind ribosomal DNA and form a preinitiation complex with RNA polymerase I to increase transcribed ribosomal DNA.

Cytoplasmic CHP1 translocates to the nucleus in serum-deprived fibroblast-like cells. This nuclear localization of CHP1 is dependent on the NES motifs on CHP1 C terminus. The nuclear transcription factors UBF1 and UBF2 were identified as binding partners of CHP1 by immunoprecipitation experiments by matrix-assisted laser desorption ionizing-mass spectrometry. Interestingly, CHP1 and UBF immune complexes were formed mainly in the nucleus of serum-deprived cells and in the absence of Ca²⁺. Whether CHP1 binds directly to UBF, however, remains uncertain. Functionally, CHP1 nuclear retention, probably via restriction of UBF translocation to the nucleolus, inhibits serum-stimulated ribosomal DNA transcription.

In summary, CHP1 translocates in the cytoplasm in a serum- and Ca²⁺-induced mode. CHP1 located in the nucleus functions as a blocker of ribosomal RNA synthesis.

CHP2.

A large majority of studies of CHP2 involve its interaction with NHE1. CHP2 is another member of the CHP subfamily and was first identified in human cancer patients with hepatocellular carcinoma tumors. As with CHP1, CHP2 binds to the juxtamembrane region of the cytoplasmic domain of NHE1 to NHE3, and binding of CHP2 to the exchangers determines its localization at the plasma membrane. However, CHP2 interacts more strongly with NHE1 than does CHP1 and increases NHE1 activity. The function of CHP2 binding to other NHE isoforms is unknown.

Specifically, structural data on CHP2 interaction to NHE1 has revealed that CHP2 may have a dual role in NHE1 activation. CHP2 binding to NHE1 results in an increase in both maximal exchanger activity (V_max) and the H⁺ affinity of NHE1 (4). On the other hand, deletion of the CHP2 flexible linker, which connects the N- and C-lobes of CHP2 (Fig. 1A), results in a reduction of the exchanger activity by acidic shift of the pH_i dependency of Na⁺ uptake but not by changes in V_max (4).

These findings indicate that CHP2 may be dynamically bound to NHE1 and has a dual mode of action on NHE1: 1) CHP2 binding to NHE1 increases both V_max and H⁺ affinity of NHE1 and acts as an activator of NHE1; and 2) CHP2, via its flexible linker, mediates the fine tuning of changes in NHE1 activity in response to changes in pH_i by shifting its H⁺ affinity but not its V_max. There is no similar structural evidence for the CHP1-NHE1 interaction, although the functional studies reported above support of a dual mode of action of CHP1 on NHE1.

CHP2 seems to protect cells from serum deprivation-induced death by increasing pH_i (103). Increased pH_i caused by NHE1 activation provides a permissive or an obligatory condition for proliferation and differentiation of cells (30, 100, 119, 153, 155). Conversely, decreased pH_i, attributed to reduced NHE1 activity, resulted in growth arrest or cell death (79, 123, 128, 143). Furthermore, activation of NHE1 is also associated with high pH_i in the absence of serum and oncogenic transformation (24). CHP2 may be involved in the maintenance of the abnormally high pH_i in malignantly transformed cells. An implication of CHP2 in cancer progression is compatible with overexpression of CHP2 in ovarian cancer cells (58).

Although it is not clear whether changes in the phosphorylation of CHP1 is a key event in the serum-induced activation of NHE1 in normal cells, the difference in phosphorylation status could be one possible mechanism to explain the observed difference in the mode of serum-dependent regulation of NHE1 by CHP1 and CHP2. Indeed, there are several potential phosphorylation sites in CHP1, such as putative phosphorylation sites for PKC (serine 112) or calmodulin-dependent protein kinase II (threonine 7, serine 33, and serine 37), that are not conserved in CHP2 (103).

In view of the multiple biological functions of the CHPs, CHP2 may be involved in the regulation of cellular functions other than NHE1 activity, which may contribute to CHP2's oncogenic potential. Indeed, CHP2 expression accelerates proliferation, and CHP2 knockdown suppresses proliferation in several cell lines (70). CHP2 overexpression also altered the tumorigenic capacity of HEK293 cells in vivo. Additionally, CHP2 transfectants appeared to be more aggressive, invading the spleen, liver, and kidney, while no control cells invaded these organs in control animals (70). Deregulation of the calcineurin-signaling pathways is implicated in tumor development and progression. Aberrant expression or activation of calcineurin and the well-known substrate NFAT has been documented in a number of tumors (22, 66, 78, 85). As there is a functional distinction between CHP1 and CHP2 and their regulation of NHE1, CHP1 and CHP2 display significant divergence in the regulation of the calcineurin/NFAT-signaling pathway. CHP1 negatively regulates this pathway (74), while CHP2 enhances calcineurin activity and induces activation of NFAT by its translocation to the nucleus (70). Furthermore, CHP1 and CHP2 exert opposite effects on cell proliferation. Proliferation of CHP1 transfectants was suppressed (74) while proliferation of CHP2 transfectants was enhanced (70, 71). The mechanisms of CHP2 function on calcineurin activity are not clear, but CHP2 may substitute CnB and induce changes in molecular conformation that either increases the affinity of the catalytic domain substrate or weakens the interaction between the catalytic domain with the autoinhibitory domain. CHP1 may act similarly on the calcineurin but with opposite effects.

CHP3.

Similarly to CHP1 and CHP2, CHP3 is located in the cytoplasm, particularly in and around the nucleus of cells. It is also found in the plasma membrane region corresponding to the lamellipodia and membrane ruffles where NHE1 is enriched (32, 44, 102) (Fig. 3A). Furthermore, the localization of CHP3 in lamellipodia and membrane ruffles is serum dependent (72). As is the case with other members of the CHP subfamily, CHP3 binds to NHE1 (72). CHP3 shares a similar functional activity with CHP1 (44), although the functional regulation of CHP3 binding to NHE1 is controversial. CHP3 was found first to bind NHE1 at two different aa sequences (either 505–571 or 633–815) (72, 160). The first sequence is homologous with the binding site of both CHP1 and CHP2 to NHE1 (102, 103). In one study, it was stated that CHP3 interacted with and inhibited NHE1, an effect enhanced by the presence of serum. Interestingly, the C-terminal domain of NHE undergoes changes in conformation that are pH_i and, possibly, protein-protein interaction dependent (52, 72). Indeed, the NHE1 C terminus is responsible for the sensitivity of NHE1 to pH_i (52). In this study, the authors proposed that the binding of CHP3 to the NHE1 C terminus, which inhibits NHE1, may induce changes in the conformation of the NHE1 tail and hence in its sensitivity to pH_i (72). In the same work, it was found that Ca²⁺ induces an increase in the CHP3-NHE1 complex but reverses the inhibitory effect of CHP3 on NHE1. This implies that Ca²⁺-induced enhancement of CHP3 binding to NHE1 could be an important mechanism of NHE regulation (72).

Contrary to these observations, CHP3 upregulates rather than downregulates the NHE1 cell surface activity, and this effect is not due to an alteration in pH_i sensitivity but to promotion of NHE1 maturation and half-life at the cell surface, thereby increasing the abundance of functional NHE1 (160). Mechanistically, the CHP3-mediated increase in surface NHE1 protein levels could arise from accelerated processing of newly translated NHE1 along the biosynthetic pathways, but could also reflect increased stability of the mature protein at the cell surface, although these two processes are not necessarily mutually exclusive. Hence, CHP3, like CHP1, seems to regulate both pH_i sensitivity and protein expression of NHE1.

Origin of the CHP Subfamily: Evolutionary History and Clues to Its Function

A series of homology-based searches using BLAST tools against the NCBI databases reveal that CHP1 is predominantly present in the Opisthokonts group, specifically in the Filasterea, Choanoflagellatea, and Animalia subgroups of the Eukaryote superkingdom (60). Our phylogenetic analysis indicates that CHP1 most likely appeared from an EF-hand protein precursor in an ancestor of Capsaspora (Filasterea), a sister group to Choanoflagellatea and Animalia, dating to >850 million years ago (Fig. 4A) (48, 62). It is striking that CHP1 shows a high degree of conservation with the maintenance of critical structural domains such as the myristoylation site (excluding zebra finch, Taeniopygia guttata, and roundworm, Ascaris suum), four EF-hands, and two NES across diverse organisms. Interestingly, the presence of the unique long flexible linker between the N- and C-lobes of CHP1 is restricted mainly to birds and mammals, although a shorter version appears to be present in amphibians. This sequence may reflect some specific functions of CHP1 acquired later on rather than a simple hinge sequence as was shown for CaM (4, 161). This observation raises the question as to whether the long flexible linker is a result of CHP1 structural modification intended to target NHEs or other binding partners (106).

Fig. 4.

Phylogenetic relationship of CHP1 homology and distribution of CHP1–CHP3 and NHE1–NHE5 in eukaryotes. A: phylogenetic tree of CHP1 homology based on an amino acid alignment. Phylogenetic trees were obtained using the Bayesian approach, which is implemented in MrBayes (125). Numbers in the nodes represent the posterior probability. B: metazoan phylogeny of CHP1–CHP3 in relationship to members of NHE family (NHE1–NHE5 and NHE9), which show high homology to NHE in the sequence of the CHP-binding region (4). The presence of various proteins is denoted by rectangles, not placed relative to timescale. Black frames around rectangles indicate proteins, which were not annotated as NHE in the NCBI databases, but demonstrate high similarity to NHE1–5. The number in parenthesis denotes the number of species represented in that branch, the genome sequences of which are available in the NCBI databases. The phylogeny was drawn according to the tree of life website (http://tolweb.org/tree/) (48).

Contrary to CHP1, CHP2 and CHP3 were found only in vertebrates (Fig. 4B). However, it is difficult to assess whether CHP2 and CHP3 are, indeed, exclusive to vertebrates because, it must be noted that there is a strong bias introduced with only a handful of complete draft genome sequences available at present. Nonetheless, the disparate distribution of CHP1, CHP2, and CHP3 raises several points: 1) since CHP2 and CHP3 appear in vertebrates, it is tempting to speculate that CHP1 might have given rise to CHP2 and CHP3 via a gene duplication event; and 2) another possibility is that the precursors of CHP2 and CHP3 were substantially lost in other lineages except for vertebrates (Figs. 1B and 4B).

CHPs interact with a plethora of proteins, such as members of the NHE family. NHE1–NHE5 have a conserved putative site for binding CHPs (4, 54). The presence of various binding partners is one of the many evolutionary forces that influence protein evolution (106). NCBI databases were extensively surveyed to detect the presence of CHP1–3 and NHE1–5 in various organisms. In summary, we found a patchy distribution of NHE1–5 relative to CHP1–3 across the metazoan phylogeny (Fig. 4B). It is likely that this pattern might be a reflection of lineage-specific gain/loss events. On the other hand, this observation could be due in part to the fact that there are limited numbers of complete draft genome sequences available in the surveyed NCBI databases.

It is also intriguing as to whether the presence of CHP2 and CHP3 in vertebrate correlates with a more efficient and complex Na⁺ absorption function and the necessity to differentially regulate apical NHE3 and basolateral NHE1 of intestinal and possibly renal epithelia. To maximize the efficiency of the vectorial Na⁺ transport from the lumen to the interstitial space, a differential regulation of NHE1 and NHE3 may be necessary. This would ensure that in specific conditions NHE1 and NHE3 would not be simultaneously activated (18). Although these are interesting conjectures, what remains unknown is whether CHPs and NHEs are coevolving. Further studies on phylogenetic analysis will shed more light on this inquiry.

CHP Prospective: Are Its Multifunctions Connected?

There is no doubt that CHPs are multifunctional. CHP facilitates membrane movement along the microtubule cytoskeleton, participating in the delivery of carrier vesicles between organelles and from organelles to the plasma membrane (5, 10, 145). Does CHP support only long-distance movement of organelles on microtubules or also local transport on actin filaments? What about the switch of carrier vesicles between microtubule and actin filaments (Fig. 3B)? Of note, CHP was found to be part of the NHE3-ezrin complex (34), with ezrin being responsible for targeting membrane proteins to the cell surface by binding to the actin cytoskeleton (19, 39, 51, 75). This would support a CHP function also at the actin cytoskeleton level.

CHP regulates the activity of both NHE1 and NHE3, possibly by changing their pH_i sensitivity as well as the level of protein expression. Are these two functions of CHP on NHEs connected? For example, CnB-deficient yeast accumulates sodium and demonstrates a decrease in adaptation to salt stress due to a decrease in expression of the yeast Na⁺-ATPase (88). Additionally, is the action of CHP on NHEs connected to the participation of CHP in membrane movement along the cytoskeleton? Is CHP supporting the movement of NHEs along the exo- and/or endocytotic pathway? CHP accelerates the processing of newly translated NHE1 along the biosynthetic pathways (160) and reduces its degradation by the proteasome system (160). This would support the involvement of CHP in trafficking of NHEs to and from the cell surface. Additionally, microtubule-dependent vesicle transport is integral in the maintenance and regulation of transporters and channels on the apical plasma membrane of organs such as the kidney (45), supporting a functional regulation by CHP of trafficking of NHEs.

CHP affects the function of phosphatases and kinases. How does this relate to the overall picture of the function of CHP? What about the nuclear vs. cytosolic vs. plasma membrane localization of CHP? CaM has a function at the plasma membrane, cytosol, and also the nucleus (76, 126, 134); indeed, it regulates transcription either directly via binding to transcription factors (141) or indirectly via modulating protein kinases or phosphatases that control the activity of transcription factors (80).

In summary, CHP is probably a signal molecule, as are other EF-hand Ca²⁺-binding proteins (49, 76, 122, 127), that couples local Ca²⁺ concentration changes to a variety of specific signal-response cascades. Nevertheless, we are left with too many unanswered questions on the function of this family of EF-hand Ca²⁺-binding protein. Generation of animal models will ultimately be required to study the overall physiological impact (and associations with human diseases and diagnostics) of this family of proteins. Indeed, it was recently found that loss of Caenorhabditis elegans (C. elegans) chp1 orthologous, pbo-1 (posterior body contraction mutant) function results in a weakened defecation muscle contraction and a caloric-restriction phenotype of C. elegans. Pbo-1 affects defecation in C. elegans by regulating differentially the activity of intestinal NHE exchangers located at the apical and basolateral intestinal membranes of C. elegans (149). Further studies using this and other animal models may aid identifying the global effects of losing CHP function.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: F.D.S. and V.B. provided conception and design of research; F.D.S., K.V., and V.B. prepared figures; F.D.S. and V.B. drafted manuscript; F.D.S., K.V., O.M., and V.B. edited and revised manuscript; F.D.S., K.V., O.M., and V.B. approved final version of manuscript.