TRPMLs: in sickness and in health

Abstract

TRPML1, TRPML2 and TRPML3 belong to the mucolipin family of the TRP superfamily of ion channels. The founding member of this family, TRPML1, was cloned during the search for the genetic determinants of the lysosomal storage disease mucolipidosis type IV (MLIV). Mucolipins are predominantly expressed within the endocytic pathway, where they appear to regulate membrane traffic and/or degradation. The physiology of mucolipins raises some of the most interesting questions of modern cell biology. Their traffic and localization is a multistep process involving a system of adaptor proteins, while their ion channel activity possibly exemplifies the rare cases of regulation of endocytic traffic and hydrolysis by ion channels. Finally, dysregulation of mucolipins results in cell death leading to neurodegenerative phenotypes of MLIV and of the varitint-waddler mouse model of familial deafness. The present review discusses current knowledge and questions regarding this novel family of disease-relevant ion channels with a specific focus on mucolipin regulation and their role in membrane traffic and cell death. Since mucolipins are ubiquitously expressed, this review may be useful for a wide audience of basic biologists and clinicians.

trp is a diverse superfamily of ion channels whose founding members were identified as components of the receptor-activated signaling pathway driven by phospholipase C (44). Within a decade, the entire superfamily has been identified, and it is now clear that TRP channels are involved in a variety of sensory and transport functions including receptor-activated membrane permeability, temperature and chemical signaling, ion reabsorption and, likely, a number of other processes (47, 54, 68). Several TRP channels have been implicated in human diseases. Thus the dysregulation of polycystin 2 (TRPP2) causes autosomal dominant polycystic kidney disease (11, 76). Recent data identify mutations in genes coding for TRPC6 as the cause of focal segmental glomerulosclerosis (35, 56, 75). More examples of the TRP-related diseases are discussed in our recent review (34).

Conductance properties and regulation mechanisms have been defined for most of the TRP channels. Therefore, for most TRPs, the causal relations between channel downregulation and the resulting cellular dysfunction are clear. This is not the case for the subjects of this review—the members of the mucolipin family (or TRPMLs). These recently identified TRP channels are clearly involved in specific aspects of cellular activity, yet their properties and function have not been settled. The present review aims to summarize current knowledge and questions regarding their function. Mucolipins reside within the endocytic pathway—an intricate machine responsible for recycling of the cellular membranes and uptake of material from the extracellular space. In addition to harvesting the nutrients and repairing the plasma membrane, this pathway is directly involved in cellular signaling (by recycling membrane receptors and ion channels and taking up growth factors) and cell survival (autophagy and extracellular matrix modification). This sets mucolipins in a unique position as ion channels: not only, as discussed below, are they involved in metabolic and/or traffic functions but they appear to reside at the intersection of several vitally important cellular functions, including cell signaling and survival.

Furthermore, the proposed modes of mucolipins' action suggest completely new functions of ion channels. Several aspects of the endocytic function are regulated by inorganic ions. First, the membrane fusion process that drives the movement of endocytosed material and enzymes within the endocytic pathway depends on Ca²⁺ ions, presumably released from the endocytic organelles (51). If mucolipins are responsible for such Ca²⁺ release events, then the ion channels that drive membrane traffic in the endocytic pathway have been identified for the first time. On the other hand, the activity and the delivery of the digestive enzymes within the endocytic pathway depend on the acidity of the organelles in which these enzymes reside (73, 74). If mucolipins modulate endocytic function by regulating the ionic conditions within the endocytic organelles, then the first cation channels that directly regulate the metabolic function of the endocytic pathway have been identified. At present, it is difficult to choose between these two possibilities, and thus future studies promise very exciting discoveries.

This review begins with a description of current data on TRPML channels and concludes with a parallel analysis of their permeation properties, localization, and possible function that will specifically focus on the possible methodological issues regarding the present state of TRPML knowledge.

TRPMLs

The TRP superfamily consists of proteins with six transmembrane-spanning domains and the amino- and carboxy-terminal tails oriented toward the cytosol. TRPs are divided into two groups based on sequence and topological differences (68). Group 2 includes the TRPP and TRPML subfamilies, which share a high degree of sequence homology within the transmembrane domains and are distinguished from group 1 by the presence of a large extracellular loop between transmembrane segments 1 and 2. The TRPML subfamily includes three members (TRPML1–3) that share ∼75% amino acid similarity. One of the most distinctive characteristics of the TRPMLs is the short length of their cytosolic tails, which range from 61 to 72 amino acids. Other structural similarities shared by the members of the TRPML subfamily include the presence of several N-glycosylation sites in the first extracellular loop and a potential to be palmitoylated in a cystine-rich region located in the carboxy-terminal tail (Fig. 1). In addition, TRPMLs are characterized by their primary intracellular distribution into different subpopulations of membrane vesicles (see below). As in all TRP channels, pore sequences reside between the fifth and sixth transmembrane domains of TRPML. The presence of negatively charged glutamate and aspartate residues within the pore regions defines TRPML selectivity to cations and, perhaps, rectification properties. The pore sequences of TRPML1–3 are very similar (Fig. 1), and thus the differences in regulation and conductance characteristics of these channels discussed below are probably defined by the structural determinants located outside of the pore.

Fig. 1.

Structure/function map of TRPML family. Previously described naturally occurring mutations and empirically derived data on amino acid domains that define mucolipin traffic and regulation are shown.

TRPML1.

The best-characterized member of the TRPML family is TRPML1, also referred to as Mucolipin-1 or MCOLN1 because mutations in this protein have been associated with mucolipidosis type IV (MLIV) (5, 7, 60, 63). MLIV is an autosomal recessive disease characterized by severe mental and psychomotor retardation, diminished muscle tone or hypotonia, achlorhydria, and visual problems including corneal clouding, retinal degeneration, sensitivity to light, and strabismus (1–3). MLIV is relatively frequent among Ashkenazi Jews, with 1 in 100 of the population estimated to be a genetic carrier (4, 6). Two founder mutations that result in the absence of TRPML1 account for 95% of the cases in this group (3). TRPML1 is a 580-amino acid protein and has a molecular mass of 65 kDa. The predicted topology of the protein consists of six transmembrane-spanning domains with the amino- and carboxy-terminal tails oriented within the cytosol. The region of homology with the other members of the TRP family comprises transmembrane domains 3–6 (amino acids 349–507), which includes the pore located between transmembrane segments 5 and 6 (amino acids 461–475). Transmembrane domains 1 and 2 are connected by a long luminal loop that comprises four N-glycosylation sites and a cleavage site that may regulate TRPML1 channel activity (33, 41). TRPML1 is palmitoylated in three cysteine residues located at the carboxy-terminal tail (C⁵⁶⁵CC), and it has been suggested that palmitoylation modulates trafficking of the protein (72).

INTRACELLULAR DISTRIBUTION.

Studies using different epitope-tagged versions of TRPML1 expressed in several cell lines revealed that the protein primarily localizes to late endosomes-lysosomes (Fig. 2) (33, 41, 52, 64, 67, 72). The intracellular distribution of TRPML1 is regulated by two acidic dileucine motifs individually located near the ends of the amino- and carboxy-terminal tails (52, 67, 72). Like many other lysosomal proteins (e.g., CD63 and Lamp-1), TRPML1 can follow a direct or an indirect pathway to reach lysosomes. The direct pathway defines trafficking of the newly synthesized protein from the Golgi to late endosomes-lysosomes. This route is regulated by an acidic dileucine motif located at the amino-terminal tail (E¹¹TERLL) that mediates direct interaction with clathrin adaptors AP-1 and AP-3 (Ref. 72; S. Vergarajauregui and R. Puertollano, unpublished observations). In addition, TRPML1 can follow an indirect pathway, traveling from the Golgi to the plasma membrane from which it is internalized and sequentially delivered to early endosomes and late endosomes-lysosomes. Internalization from the cell surface is regulated by an acidic dileucine motif located at the carboxy-terminal tail (E⁵⁷³EHSLL) and is dependent on clathrin and clathrin adaptor AP-2 (72). As expected, simultaneous mutation of both dileucine motifs abolishes delivery of TRPML1 to late endosomes-lysosomes and causes mislocalization of the protein at the plasma membrane (52, 67, 72). Depletion of endogenous AP-1 by treatment with small interfering RNA (siRNA) also blocks transport of TRPML1 to late endosomes-lysosomes, indicating that the direct pathway is probably the principal route followed by TRPML1 (41).

Fig. 2.

Intracellular distribution of TRPMLs. Steady-state distribution of the 3 members of the TRPML family is shown. The proposed intracellular pathways followed by TRPML1 (orange arrows) and TRPML2 (brown arrows) are also represented. Organelles at the “classical” endocytic pathway are colored green. Organelles at the Arf6-regulated pathway are represented in dark brown. Note that the recycling endosomal compartment (REC) also receives cargo from the “classical” endocytic pathway and recycles it back to the plasma membrane through recycling tubules that differ from those containing TRPML2 and that for simplicity have not been included. EE, early endosome; LE, late endosome; LY, lysosome; RET, recycling endosomal tubules.

CHANNEL ACTIVITY AND REGULATION.

The conductance properties and regulation of TRPML1 have not been settled. At present, there are almost as many versions of TRPML1 permeability and selectivity as there were attempts to define it. These are summarized in Table 1. It is likely that the difficulties in establishing the TRPML1 conductance properties stem are mainly caused by difficulties in delivery of this intracellular ion channel to the plasma membrane, which may affect channel recordings using the standard whole cell patch clamp.

Table 1.

Present state of knowledge regarding TRPML current properties

TRPML	Current Properties	System	Method of Recording	Reference
TRPML1	Linear IV, K+, Na+, and Ca2+ permeable, inhibited by Ca2+. Single-channel conductance of 80–90 pS in 100 mM K+ or Na+ and 33 pS in 100 mM Ca2+.	Xenopus oocyte, HEK-293 overexpression	Whole cell voltage clamp and single-channel patch clamp	36, 37
TRPML1	Outwardly rectifying, K+ permeable, inhibited by H+ and Ca2+. 130–150 pS conductance in 150 mM K+, multiple conductance substates, most frequent of which is a 35-pS substate. Inhibited by amiloride. Various conductance substates for TRPML1 mutants.	In vitro translated	Bilayer patch clamp	12, 55
TRPML1	Outwardly rectifying, Na+ and H+ permeable, inhibited by Ca2+	HEK-293 overexpression	Whole cell voltage clamp	33, 62
TRPML1	No currents under control conditions, inward current activated by the Va mutations, Na+ and Ca2+ permeable, activated by H+	HEK-293 overexpression	Whole cell voltage clamp, fluorescent Ca2+ measurements	77
TRPML1	Linear conductance, K+ permeable, blocked by amiloride and low pH. 130-pS wild-type channel conductance in 150 mM K+ and ∼30-pS mutant channel conductance.	In vitro translated	Bilayer patch clamp	23
TRPML1	Inward current in lysosomes permeable to Na+, Ca2+, and Fe2+	HEK-293 overexpression	Whole lysosomal voltage clamp and single channel	19
TRPML2	No currents under control conditions, no current induced by the Va mutations	HEK-293 overexpression	Whole cell voltage clamp	77
TRPML2	Ca2+ influx activated by the Va mutations	HEK-293 overexpression	Fluorescent Ca2+ measurements	25
TRPML3	No currents under control conditions, inward current activated by the Va mutations, Na+ and Ca2+ permeable, inhibited by H+	HEK-293 overexpression	Whole cell voltage clamp	25, 77
TRPML3	Inward current activated by extracellular Na+ removal and by the Va mutations, Na+ and Ca2+ permeable, inhibited by H+	HEK-293 overexpression	Whole cell voltage clamp	31, 32
TRPML3	Outward currents under control conditions, inward current activated by the Va mutations, Na+ and Ca2+ permeable, inhibited by H+	HEK-293 overexpression	Whole cell voltage clamp	45

Va, varitint-waddler.

The first reports on TRPML1 selectivity were obtained in Xenopus oocytes and HEK-293 cells (36, 37). The channels were measured with the cell-attached patch clamp and reconstitution into planar lipid bilayers. These experimental systems yielded monovalent cation channels that are somewhat more selective for Ca²⁺ over Na⁺ and have linear conductance (36, 37). The TRPML1 Ca²⁺ permeability reported in these studies gave rise to the idea that TRPML1 is a Ca²⁺ channel that releases lysosomal Ca²⁺ and drives the fusion or fission of organelles within the endocytic pathway (37, 49).

The next series of reports on TRPML1 reconstituted into planar lipid bilayers by another group yielded outwardly rectifying monovalent cation-permeable channels that are blocked by Ca²⁺ and H⁺ (12, 55). The outward rectification means that these channels prefer to move cations into the lysosomal lumen.

Whole cell patch-clamp recordings of HEK-293 cells overexpressing TRPML1 were reported to yield similar outwardly rectifying monovalent cation channels that were blocked by Ca²⁺ (62). In contrast to the previously mentioned studies, TRPML1 was reported to conduct H⁺, which was taken as an indication that TRPML1 is a H⁺ leak channel whose function is regulation of lysosomal pH (62). That cathepsin B, the protease that likely cleaves TRPML1, inhibited this conductance was interpreted as inhibition of TRPML1 by cathepsin B-mediated cleavage (33). Later, this conductance was suggested to be negatively regulated by protein kinase A-driven TRPML1 phosphorylation (71).

A recently published series of reports employed TRPML1 that was mutated to resemble the TRPML3 form shown to induce the varitint-waddler (Va) mouse phenotype. The emerging consensus on the Va form of TRPML3 (discussed in more detail below) is that these mutations cause pathological activation of TRPML3 leading to cell death. The Va mutations were proposed to severely disrupt the pore (or near pore) architecture and, similar to the voltage-sensitive K⁺ channels, lock the channel in the open state (see Ref. 16 for theoretical analysis of these substitutions). It was reported that the recombinant wild-type TRPML1 expressed in HEK-293 cells was inactive insofar it could be ascertained with the standard whole cell patch clamp (77) or fluorescent Ca²⁺ measurements (25) unless the mutation corresponding to the Va A419P substitution in TRPML3 was introduced in TRPML1 (25, 77) (Fig. 1). The resulting channel activity was inwardly rectifying, Ca²⁺ selective, and impermeable to H⁺ but activated by low pH (77). Since native TRPML1 has not been recorded, it is difficult to judge whether and to what extent this mutation changes the TRPML1 regulation. Nonetheless, this interesting development lends support to the proposed TRPML1 role in Ca²⁺ release that drives membrane fusion within the endocytic pathway. Inward rectification (preferred direction of current from lysosomal lumen into cytoplasm), Ca²⁺ permeability, and positive regulation of current by low pH characteristic for lysosomal lumen are certainly better suited for Ca²⁺ release function than the activity described in the previous two paragraphs. It is unclear whether the Va mutations “activated” TRPML1 or promoted its delivery to the plasma membrane. This raises an important question of whether TRPML1 is a spontaneously active leak channel or whether it is activated by a yet unidentified cellular messenger. Ca²⁺ release from lysosomes has been shown to promote the fusion of organelles within the lower portion of the endocytic pathway (51), but it is unclear what triggers such release. Nicotinic acid adenine dinucleotide phosphate (NAADP) was proposed to be such a messenger; the channel activity reported to be activated by NAADP in liver preparations (78) resembles the initial reports on TRPML1 activity.

CELLULAR FUNCTION.

MLIV is a lysosomal storage disorder (LSD) characterized by accumulation of enlarged vacuolar structures containing phospholipids, sphingolipids, mucopolysaccharides, and gangliosides (8, 57, 60). However, and in contrast to most LSDs, accumulation of undegraded substrates in lysosomes is not due to the lack of specific lysosomal hydrolases. Instead, the present models of TRPML1 function focus on its role in organelle biogenesis or regulation of lysosomal hydrolysis. Analysis of patient cells revealed that the rate of lactosylceramide (LacCer) traffic through the late endocytic pathway is severely impaired in MLIV (15, 62), thus suggesting that TRPML1 might regulate sorting of certain lipids. The role of TRPML1 in trafficking was further supported by functional analysis of the Caenorhabditis elegans ortholog of TRPML1, CUP-5 (22). CUP-5-null cells accumulate enlarged vacuolar structures that contain both late-endosomal and lysosomal markers (22), leading to the suggestion that reformation of lysosomes from late endosome-lysosome hybrid structures is blocked in the absence of CUP-5 (39). In mammal cells, stable small hairpin RNA (shRNA) clones of mouse macrophages with reduced expression of TRPML1 show delayed transport of fluid-phased markers to lysosomes, impaired exit of LacCer from lysosomes, and reduced transport of major histocompatibility complex (MHC) II to the plasma membrane (64). In addition, degradation of platelet-derived growth factor receptor (PDGFR) is delayed in MLIV fibroblasts (70). Therefore, TRPML1 function is required for efficient trafficking of lipids and proteins along the late endocytic pathway.

However, a fundamental question remains unsolved: are the trafficking defects observed in TRPML1-deficient cells a secondary consequence of lipid accumulation in lysosomes or do they reflect a direct role of TRPML1 in the regulation of trafficking events? Recently, Kiselyov's and Weisz's laboratories addressed this question by depleting TRPML1 in HeLa cells with siRNA. Interestingly, they found that acute TRPML1 deficiency led to a reduction in lysosomal pH, while delivery to lysosomes and degradation of both lipid and protein components of the low-density lipoprotein (LDL) complexes was not affected. LacCer and EGF traffic and metabolism were unaffected as well (40). These results suggest that the primary function of TRPML1 might be the regulation of lysosomal pH and that trafficking defects observed in MLIV result from the accumulation of undigested material in lysosomes. In addition, the role of TRPML1 in lysosomal pH regulation is in agreement with the reported permeability of the protein to H⁺ and the overacidification of lysosomes observed in MLIV fibroblasts (62).

Another question that remains largely unresolved concerns the mechanism by which defects in TRPML1 function result in mental and psychomotor retardation. Recent evidence suggests that defective lysosomal function in MLIV due to lipid buildup has important repercussions for the autophagic pathway (Fig. 3). Autophagy is a crucial clearance mechanism that protects against the accumulation of toxic protein aggregates and damaged organelles. In MLIV, defective lysosomal function impairs autophagosome degradation, resulting in the accumulation of p62-enriched protein inclusions and ubiquitinated aggregates in the cytosol (70). Dysfunctional autophagy also leads to the accumulation of aberrant mitochondria with reduced Ca²⁺ buffering capacity and renders cells more susceptible to proapoptotic stimuli (28). Neurons are particularly sensitive to defects in autophagy because of their incapability to divide and their high sensitivity to the accumulation of toxic aggregates. Therefore, autophagic dysfunction may contribute to the neurodegeneration observed in MLIV patients. Defective autophagy has also been reported in other LSDs characterized by severe neurodegeneration (48, 59), revealing a very interesting connection between lysosomal function, autophagy, and neuronal death. Defective lysosomal function might also affect other cellular processes. For example, CUP-5-null embryos show increased lethality due to reduced degradation of yolk proteins. This causes a reduction in nutrient availability that promotes autophagy activation and, because autophagosomes cannot be efficiently degraded, apoptosis occurs, leading to developmental defects (58). Loss of CUP-5 also causes accumulation of MRP-4, an endosomal/lysosomal member of the ATP-binding cassette (ABC) transporter superfamily that further contributes to lysosomal malfunction (58). Finally, gene expression profiling of MLIV fibroblasts revealed variation in the levels of expression of multiple genes implicated in processes as diverse as lysosome biogenesis, cholesterol and lipid metabolism, signal transduction, and organelle acidification, among others (9), indicating that MLIV is a multifaceted disease that affects multiple cellular systems. The recent generation of a murine model for MLIV that accurately reproduces the phenotype of the disease (69) will hopefully provide additional understanding of the function of TRPML1 and will allow the design of therapies for the treatment of this devastating disorder.

Fig. 3.

Autophagy in mucolipidosis type IV (MLIV). Proposed mechanism of autophagic abnormalities in MLIV, and perhaps other lysosomal storage disorders (LSDs), is shown. In normal cells, old defective organelles and ubiquitinated proteins destined to degradation are routed into autophagosome for degradation. Autophagic defect in cells affected by MLIV preclude or impede such degradation, leading to accumulation of cellular waste and cell death.

TRPML2.

TRPML2 is a 566-amino acid protein with a predicted mass of 65 kDa and shows ∼60% amino acid homology with TRPML1. TRPML2 is the least characterized member of the family, likely because mutations in this protein have not been associated with any disorder in either humans or animal models to date.

INTRACELLULAR DISTRIBUTION AND CHANNEL ACTIVITY.

Heterologous expression of epitope-tagged TRPML2 in cells showed that the protein localizes to plasma membrane, long tubular structures, and vesicles (Fig. 2; Ref. 30). TRPML2-positive vesicles colocalize with CD63 and Lamp-1 and accumulate LysoTracker (30, 61, 67), indicating that a fraction of TRPML2 is present in late endosomes-lysosomes. The long tubular structures correspond to recycling endosomes from the Arf6 pathway. The Arf6 pathway is regulated by the small GTPase Arf6 and includes internalization of proteins from the plasma membrane by a clathrin-independent process, transport to a recycling compartment, and delivery back to the cell surface through long, tubular endosomes (18). Numerous proteins, including MHC I, interleukin-2 receptor, β₁-integrin, and many glycosylphosphatidylinositol-anchored proteins (GPI-APs), travel along the Arf6-regulated pathway (10, 46, 53). In HeLa cells, TRPML2 colocalizes with MHC I and GPI-APs in tubular structures. In addition, expression of a constitutively active mutant of Arf6 causes sequestration of TRPML2, MHC I, and GPI-APs into the same enlarged vacuolar structures (30). These data indicate that TRPML2 uses the Arf6 pathway to cycle between the plasma membrane and recycling endosomes.

Expression of mouse TRPML2 in HEK-293 cells did not produce any measurable conductance, leading to the suggestion that TRPML2 might be a nonfunctional channel (77). However, a shorter mouse TRPML2 isoform lacking the first 28 amino acids has been reported (NM_001005846) that shows channel activity when overexpressed in cells (16). These data suggest that differential expression of the shorter and longer isoforms may regulate TRPML2 activity in cells and raise the interesting question of how the first 28 residues of the protein exert their inhibitory function. In addition, substitution of alanine 424 by proline, a mutation equivalent to the one that renders TRPML3 constitutive active (see below), dramatically increases the activity of TRPML2 (16), suggesting that TRPML2 and TRPML3 may share similar structural regulatory mechanisms. Further work will be required to determine whether TRPML2 is a channel on its own or acts as a regulator of other channels.

CELLULAR FUNCTION.

TRPML2 is expressed in B-lymphocytes at the pre-B-cell, mature B-cell, and plasma membrane cell stages, and its level of expression is modulated by Bruton's tyrosine kinase, a protein that plays a crucial role in B-lymphocyte development (38). In addition, overexpression of carboxy-terminal green fluorescent protein (GFP) version TRPML2 in DT40 B-lymphocytes induced accumulation of enlarged lysosomal structures that contain cholesterol and are accessible to endocytosed B-cell antigen receptors (61). Therefore, it has been suggested that TRPML2 might participate in the regulation of the specialized lysosomal compartment of B-lymphocytes and thus be critical for a normal immune response.

It has also been reported that TRPML2 exerts a regulatory role in the trafficking of certain proteins along the Arf6-regulated pathway. Overexpression of TRPML2 induces a strong activation of Arf6, while transfection of HeLa cells with an inactive version of TRPML2 (D⁴⁶³D/KK) delayed the recycling of internalized GPI-APs back to the plasma membrane (30).

TRPML3.

TRPML3 has been cloned as a genetic determinant of the Va mouse phenotype, a model of hearing loss and pigmentation defects. In contrast to most such models, Va is apparently caused not by the loss but by the gain of function of the molecule of interest (discussed below). Human TRPML3 is an intracellularly expressed protein made of 533 amino acids; its predicted mass is ∼64 kDa. Unlike TRPML1, TRPML3 does not appear to be cleaved (31).

INTRACELLULAR DISTRIBUTION.

Consistent with the hearing and vestibular defects observed in the Va mouse, TRPML3 is highly expressed in the inner ear of wild-type mice, including the stria vascularis of the cochlea and the organ of Corti sensory hair cells (17, 45, 77). In hair cells, endogenous TRPML3 mostly localizes to intracellular vesicles in the subcuticular region and the pericuticular necklace. Lower levels of TRPML3 are also observed at the plasma membrane of the stereocilia (17, 65).

Heterologously expressed TRPML3 localizes to the plasma membrane and intracellular vesicles (45, 77). Most of the TRPML3-labeled vesicles contain early endosomal markers such as Hrs, while a smaller fraction of vesicles are positive for CD63 (J. A. Martina and R. Puertollano, unpublished observations). Therefore, TRPML3 seems to be distributed in the early endocytic pathway (early endosomes-late endosomes), in contrast to TRPML1, which localizes to the late endosomal pathway (late endosomes-lysosomes) (Fig. 2). This distribution is in agreement with electrophysiology results showing that the activity of TRPML3 is maximal at a pH between 6 and 6.5 (i.e., the pH usually found in early endosomes) (see below) while TRPML1 is more active at a lower pH characteristic of lysosomes (pH 4.5–5). Finally, recombinant TRPML3 accumulates at espin-elongated microvilli that resemble stereocilia when expressed in LLC-PK₁-CL4 epithelia cells, a culture model for hair cells (45).

Recently, Venkatachalam et al. (67) reported that when expressed alone TRPML3 remains trapped at the endoplasmic reticulum (ER) and that coexpression with TRPML1 or TRPML2 promotes formation of heterooligomers (TRPML1-TRPML3 or TRPML2-TRPML3) and allows transport of TRPML3 to lysosomes. However, these results remain controversial, as expression of TRPML3 in MLIV fibroblasts shows that TRPML3 can efficiently reach the endocytic pathway in the absence of TRPML1 (Puertollano, unpublished observation). Although we cannot discard the possibility that endogenous TRPML2 is sufficient to promote exit of TRPML3 from the ER in MLIV cells, it seems improbable because of the predicted excess of recombinant TRPML3. Additional experiments will be needed to resolve this disagreement.

CHANNEL ACTIVITY AND REGULATION.

In contrast to TRPML1, there is, by and large, a consensus regarding TRPML3 properties. Several reports have shown that TRPML3 is an inwardly (from lumen into cytoplasm) rectifying monovalent cation channel that is permeable to Ca²⁺ and suppressed by H⁺ (16, 31, 32, 45, 77). The inactivation of TRPML3 by acidity is congruent with its primary localization within the upper, less acidic, portions of the endocytic pathway. The structural determinants of TRPML3 regulation by acidity have been traced down to a chain of histidines within the large amino acid loop connecting the first and second transmembrane domains of TRPML3 (32).

As with TRPML1, the reports on TRPML3 regulation are fragmentary. One series of reports that utilized whole cell patch clamp and fluorescent Ca²⁺ measurements suggested that unless the Va substitutions are introduced TRPML3 activity in the plasma membrane is undetectable under overexpression conditions (25, 45, 77). Another group demonstrated TRPML3 currents under control conditions. These currents were also augmented by the Va substitutions and, strikingly, by the temporary omission of Na⁺ from the extracellular (luminal in the case of organellar TRPML3) solution (31). The TRPML3 regulation by Na⁺ appears to involve some of the sites mediating the H⁺ dependence of TRPML3 (32). The same group reported that the latter is significantly affected by the Va mutations, raising the question of the extent to which TRPML3 regulation is affected by the Va mutations (32). A detailed analysis of the possible structure/function consequences of the Va substitutions can be found in the recent excellent review by Cuajungco and Samie (16).

CELLULAR FUNCTION.

The reports on TRPML3 role in cells revolve around two questions: what is its function within the endocytic pathway, and why do the Va mutations cause cell death? TRPML3 overexpression was reported to affect delivery of cargo from the plasma membrane to lysosomes; on the basis of this fact TRPML3 was suggested to regulate trafficking along the endocytic pathway (H. J. Kim and S. Muallem, personal communication; J. A. Martina and R. Puertollano, personal communication). Upregulation of TRPML3 activity by mutations was proposed to cause cell depolarization and/or overload with Ca²⁺ leading to cell death (16, 25, 45, 77). As with TRPML1, pinpointing TRPML3 function depends on identification of its activation mechanism, which would allow testing of the critical junctures translating ion permeation through these channels to the exact aspects of endocytic function regulated by TRPML.

TRPML heteromultimerization.

The widely discrepant reports on TRPML1 permeability characteristics are perhaps the most serious issue obscuring its function in the endocytic pathway. It is becoming apparent that in addition to the methodological and expression differences discussed below, heteromultimerization, i.e., formation of conductive complexes comprising different TRPML subunits, can be a decisive factor in defining TRPML conductance properties in the given system.

Many, if not all, TRP channels heteromultimerize; the most abundant evidence for the ensuing changes in the channel's permeability comes from the TRPC family. Several groups reported drastic changes in ion permeation properties when various TRPC subunits were coexpressed (reviewed in Ref. 26). Despite clear differences in localization, TRPMLs do coexist in a subset of endocytic compartments (30, 67), and thus it is entirely possible that they heteromultimerize and share the pore, resulting in new conductance and regulatory inputs. This issue was raised repeatedly during the recent symposium on mucolipin function in disease, and the future research into TRPML function will almost certainly take TRPML heteromultimerization into consideration, thus promising important new developments in this exciting field.

Discussion and Future Directions

Although the importance of the TRPMLs seems clear, there are still many fundamental problems that remain to be elucidated. These problems include characterization of the intracellular distribution of endogenous TRPMLs, measurement of channel activity in specific intracellular organelles, analysis of the oligomeric state and conformation of endogenous TRPMLs, and identification of the posttranslational modifications and/or regulatory proteins that turn these channels on and off.

Endogenous TRPMLs have been notoriously difficult to detect in immunohistological experiments because of the lack of reliable antibodies. For this reason, most of the studies on TRPMLs have made use of recombinant proteins that might potentially be mistargeted because of overexpression. In addition, it is unclear whether the addition of epitope tags may affect the properties or distribution of these proteins. Therefore, production of antibodies that recognize endogenous TRPMLs by electron microscopy and immunocytochemistry will be of great importance. Recently, Bach's laboratory reported the generation of antibodies that recognize the three members of the TRPML family in immunofluorescence experiments. Preliminary data with these antibodies appear to confirm some of the results obtained by expressing recombinant proteins, including the distribution of endogenous TRPML2 and TRPML3 at the plasma membrane and endosomal structures other than lysosomes (D. A. Zeevi and G. Bach, personal communication). Future experiments will allow a more careful characterization of TRPML distribution. In the meantime, some of the antibodies obtained against TRPMLs by several groups work efficiently for immunoblotting and immunoprecipitation, allowing the option of addressing TRPML distribution through biochemical methods such as cell fractionation and cell surface biotinylation. Proteomic analysis might also be a suitable approach to investigate the presence of TRPMLs in different subpopulations of membrane vesicles.

An additional controversy that will have to be resolved is the selectivity of TRPML channels. To date, most studies have analyzed plasma membrane currents of TRPMLs on heterologous expression. One obvious possibility is that the channels behave in a different way in intracellular membranes than when they are expressed on the cell surface. This may be due to the different luminal environment and lipid composition of the membranes but also to the modulation of channel function by association with other proteins or by posttranslational modifications such as phosphorylation (71). Localization of regulatory subunits or kinases in specific cellular compartments could then provide a particular adjustment of the channel activity.

One way to overcome this problem is to characterize TRPML permeability properties in native lysosome membranes. Recently, Dong and colleagues measured TRPML1 activity by directly patch-clamping late endosomal and lysosomal membranes. These experiments led to the unexpected conclusion that TRPML1 behaves as an iron release channel (19), thus indicating that the selectivity of the channel in late endosomes-lysosomes might indeed differ from that observed at the plasma membrane. However, whole lysosome current recordings required the artificial enlargement of lysosomes by treatment with vacuolin-1, a drug that disrupts the segregation of inner and limiting membranes characteristic of endosomes and lysosomes (13). It will be important to corroborate that treatment with vacuolin-1 does not affect the normal properties of these organelles.

Another alternative is the use of biochemical methods to isolate specific cellular organelles. For example, lysosomes from control and TRPML1-knockout mice can be easily isolated by using differential sedimentation through a Percoll gradient. Comparison of the electrochemical properties of the two lysosomal populations might provide interesting clues on the primary function and specificity of the TRPML1 channel. This approach has recently allowed the identification of the Cl⁻/H⁺ antiporter ClC7 as the primary chloride permeation channel in lysosomes (24). Furthermore, depletion of the different members of the TRPML family by siRNA (alone or in combination) followed by purification and characterization of the biochemical properties of specific organelles (e.g., membrane potential, ion permeability, luminal pH) would also offer critical information.

The elucidation of the three-dimensional protein structure of ion channels can also be extremely revealing. Although this is not an easy task because of the well-known difficulty of overproducing and crystallizing protein membranes, successes in crystallographic analyses of some ion channels such as the KcsA potassium channel (20), the chloride channel (21), and the voltage-gated potassium channel (29) have been reported. In addition, the structure of full-length TRPV1 channel has recently been solved by single-particle electron cryomicroscopy (cryo-EM) (42). Although the resolution obtained with cryo-EM is not ideal, the use of computer modeling allows the generation of protein models that can be used for the design and interpretation of testable experimental hypothesis. A similar approach will provide exciting molecular insights into the function and structure of TRPML channels.

Finally, the generation of animal models for MLIV has provided essential information about the cellular function of TRPML1 and should facilitate the development of eventual therapies for the treatment of this disease. It is expected that the generation of TRPML2- and TRPML3-knockout animals will be equally useful for determining the physiological roles of these proteins.

As with many other complex biological problems, the complete understanding of TRPML function will require the use of a broad spectrum of physiological, biophysical, and cell biological techniques.

Concluding Remarks

In the last few years our understanding of the TRPML family has improved significantly. It now seems established that TRPMLs mainly distribute along the endocytic pathway and participate in the sorting of lipids and proteins. The characterization of TRPMLs as ion channels reveals that, in addition to protein and lipid organization, the luminal ionic composition of an organelle also plays a key role in its function. This notion is supported by the fact that different drugs that affect the ionic composition of endocytic organelles, such as monensin, bafilomycin, and chloroquine, cause alterations in the morphology and function of the endocytic pathway (14, 43, 50, 66). Alterations in the function of TRPMLs are predicted to have a profound effect on the ionic composition of specific endocytic organelles, leading to defective vesicle-mediated mechanisms that sort lipids and proteins to different locations in the cell. For example, overacidification of late endosomes and lysosomes in TRPML1-deficient cells (40, 62) might alter the enzymatic activity of luminal hydrolases as well as the oligomeric state of transmembrane proteins, thus affecting organelle function. In addition, changes in pH might affect the recruitment of cytosolic proteins to the endosomal/lysosomal membranes. Thus recruitment to endosomal membranes of the small GTPase Arf6, a protein implicated in transport vesicle formation and protein sorting, is regulated by pH (27). Alteration in the luminal constitution of specific organelles might also have important implications in signaling, as attenuation of signaling cascades requires dissociation of ligands from their respective receptors and delivery of receptors to lysosomes for degradation. In conclusion, defects in TRPML function are predicted to have important repercussions in organelle acidification, vesicle fusion, endosome maturation, and signaling, indicating the crucial role played by this family of proteins under normal and pathological conditions.

GRANTS

R. Puertollano is supported by the Intramural Research Program of the National Heart, Lung, and Blood Institute, National Institutes of Health and K. Kiselyov by funding from the Mucolipidosis 4 Foundation.