Introduction

The transient receptor potential (TRP) superfamily is a multifunctional set of membrane proteins that function as ion channels and extend communication lines between the cell and its environment, what may be called sensory physiology or somatosensation. TRP channels in cellular membranes are modulated by a vast array of physical or chemical stimuli, including radiation (in the form of temperature, infrared radiation or light), pressure (osmotic or mechanical) and natural compounds. Excellent reviews about TRP channels are available [1-7], and thus an overview of the TRP superfamily is beyond the scope of the present review. The subject of discussion here is the transient receptor potential vanilloid 2 (TRPV2) channel, one of the most mysterious and intriguing members of this superfamily, with a paucity of data concerning the endogenous function of the channel. The intention of the review is to shed some light on the controversies concerning TRPV2 by integrating the latest reports on TRPV2 with historical data on this ion channel towards understanding the TRPV molecular mechanism from a more structural perspective.

The TRP channel superfamily

TRP channels were originally identified within the rhabdomeres of the photoreceptor in Drosophila [8] and the first member was cloned in the late 1980s [9]. TRP channels constitute an extensive ion channel superfamily represented across the phylogenetic tree from yeast to human. They are indeed the second largest family of voltage-gated-like ion channels after the potassium channel family [10]. Many TRP channels are polymodal signal integrators that trigger cellular signalling pathways via non-selective cation flux (e.g. Ca²⁺, Mg²⁺ and Na⁺). Different TRP channels show different permeability and selectivity to cations, both monovalent and divalent, whereas most trivalent cations act as channel blockers. The specifics of cation permeability are physiologically important since cations play roles in cell function regulation, such as fertilization, muscle contraction or exocytosis.

As a relevant example of TRP channel function in mammals, several TRP channels located in the dorsal root ganglion (DRG) nociceptive sensory neurons respond to nociceptive stimuli to cause pain. From a pharmaceutical and biomedical point of view, there are vast and diverse implications including some still to be determined. For instance, TRPs are lead targets in the treatment of acute and chronic pain [11, 12]. Nevertheless the role of TRP channels is not limited to the pain field. TRP channels have been implicated in several physiological and pathological processes such as cancer, genetic disorders and other rare diseases, and not only restricted to somatosensation.

A set of common sequence features among TRPs translates to shared structural features. Mainly, these channels share a membrane topology of six transmembrane segments (S1–S6), with a pore-forming loop between S5 and S6. Similarly to the tridimensional structure of potassium channels, TRPs are arranged in the membrane as tetramers as the basic functional unit. The S1–S4 segments and the N- and C-terminal cytoplasmic domains are modulating domains important to the gating of the channel, with the S5 andS6 segments defining the pore and selectivity filter (Fig. 1A).

TRPs are not classified by their functional role but based on amino acid sequence identity and similarity. The mammalian TRP channel superfamily is classified into six subfamilies with significant sequence similarity within the transmembrane domains (TMDs) but very low similarity in their N- and C-terminal cytoplasmic regions (Fig. 1B). The six subfamilies are named based on founding members: TRPA (ANKTM1); TRPC (canonical); TRPM (melastatin); TRPML (mucolipin); TRPP (polycystin); and TRPV (vanilloid). A seventh subfamily, TRPN (NompC), is present in invertebrates and some vertebrates, although it is absent in mammals. Here we focus on the TRPV subfamily and how TRPV2 distinguishes itself from its closest TRPV homologs.

TRPV2 within the TRPV subfamily

The first identified TRP channel of the TRPV family was OSM-9 (osmotic avoidance abnormal family member 9) from Caenorhabditis elegans [13]. OSM-9 is required for olfaction, mechanosensation and olfactory adaptation. OCR-1 to OCR-4 (osm-9/capsaicin receptor related) are the other TRPV channels in C. elegans, for a total of five [14]. In Drosophila melanogaster the TRPV members are Nanchung and Inactive, which are involved in sensory perception [15]. The TRPV mammal subfamily is divided into two groups based on sequence homology, TRPV1–4 and TRPV5–6 (Fig. 1B). The group formed by TRPV5 and TRPV6 shares high sequence identity (~ 75%) but low identity with the TRPV1–4 group (~ 20%) [6, 16]. TRPV5 and TRPV6 are calcium-selective channels important for general Ca²⁺ homeostasis and heterotetramers between them have been described [17]. All four channels within the TRPV1–4 group show temperature-invoked currents when expressed in heterologous cell systems, ranging from activation at ~ 25 °C for TRPV4 to activation at ~ 52 °C for TRPV2. However, a physiological role in temperature sensing has not been demonstrated for TRPV2, and the thermal responses of TRPV2 knockout mice have been shown to be similar to those of wild-type mice [18]. In contrast, TRPV1 has been convincingly shown to play an important role in thermosensation using knockout mice [19]. The original characterization of the TRPV3 knockout mice [20] and TRPV4 knockout mice [21] showed some impairment in thermosensation. However, more recent work showed that any thermosensory defects in the TRPV3 and TRPV4 knockout mice are minor and strain-dependent [22]. Therefore, a physiological thermosensory role is the exception (TRPV1) rather than the rule for TRPV channels.

As detailed below, there is increasing support for a role of TRPV2 in various osmosensory or mechanosensory mechanisms. The mechanosensory function of select TRPV channel family members seems to have been maintained through evolution. For instance, the C. elegans TRPV channels OSM-9 and OCR-2 are essential for osmosensation and mechanosensation [23]. Similarly, Nanchung and Inactive, the TRPV-related proteins in D. melanogaster, located in the mechanosensitive Johnston's organ in the fly antennae, are responsible for the sense of hearing [15, 24].

Two laboratories nearly simultaneously cloned TRPV2 orthologs. Human and rat TRPV2 were cloned based on their homology to TRPV1, and accordingly named vanilloid-receptor-like protein 1(VRL-1) [25]. Mouse TRPV2 was cloned and named growth-factor -regulated channel (GRC) based on the fact that it induces calcium currents upon stimulation of cells by insulin growth factor 1 (IGF-1) [26]. Of note, TRPV2 has also infrequently been referred to as Osm-9-like TRP channel 2 (OTRPC2) [27]. TRPV2 mediates cationic currents with higher divalent permeability (P) (Ca²⁺ > Mg²⁺ > Na⁺ ~ Cs⁺ ~ K⁺; P_Ca²⁺/P_Na⁺ = 2.94; P_Mg²⁺/P_Na⁺ = 2.40) [25]. The sequence identity among the best characterized orthologs (human, rat and mouse) ranges from 75% to 90% (Fig. 2). TRPV2 shares 50% sequence identity with TRPV1, the best characterized member of this subfamily. TRPV1 is a polymodal channel activated by anandamide, capsaicin, resiniferatoxin, heat, low pH and eicosanoids [28]. TRPV3 is a thermosensor for non-noxious heat and responds to natural compounds found in plants such as oregano, camphor and thyme [29]. TRPV4 is a constitutively active Ca²⁺-permeable channel modulated by non-noxious heat [21] and also by pressure and hypotonicity [30, 31]. TRPV4 has also been associated with hearing loss [32].

Heteromerization of TRPV1–4 channels has been documented, although mostly in heterologous expression systems. Heterotetramerization of TRPV2 with TRPV1 has been described in vitro in HEK cells [17, 33, 34]. TRPV1 and TRPV2 have also been shown to tightly colocalize in vivo in rat DRG [17, 33, 35]. However, physiological roles of any TRPV2 heterotetramerization have yet to be identified.

TRPV2 structure and evolution

TRPV2 presents a high conservation level of identity within mammalian orthologs (see Fig. 2 for human, rat and mouse sequence alignment). Specific TRPV2 orthologs are only found in tetrapod vertebrates, but not in fish where there is a TRPV1/2 isoform. One evolutionary hypothesis is that a common ancestor between TRPV1 and TRPV2, referred to as TRPV1/2, duplicated generating two genes [36]. This gene duplication is ambiguous in time. One hypothesis is that it could have occurred in the common ancestor between fishes and tetrapods and then the copy for TRPV2 was lost in fishes. The second possibility argues for the gene duplication occurring after the divergence between tetrapods and vertebrates. Saito and Shingai argue that based on genomic organization of TRPVs the first possibility is more likely [36].

TRPV2 sequence features are best understood through comparisons with the other TRPV subfamily members. Excellent reviews on the structural organization of TRPV channels are available [7, 37, 38]. Briefly, TRPV2 consists of a large N-terminal cytoplasmic domain (~ 390 residues), followed by six transmembrane segments (~ 250 residues) containing a pore-forming loop, and a C-terminal cytoplasmic domain (~ 100 residues). Specific TRPV2 sequence features are summarized in Fig. 2.

The N-terminal cytoplasmic region of TRPV2 can be further subdivided into a distal N-terminal region, an ankyrin repeat domain (ARD) and a membrane-proximal linker region. The distal N-terminal cytoplasmic region comprises the first ~ 60–70 residues, which were not ordered and therefore not visible in the first crystal structure obtained for a human TRPV2 N-terminal region [39]. This distal N-terminal region is rather divergent in length, sequence and function for different TRPV channels. For instance, the TRPV4 N-terminus is longer, ~ 130 residues, and contains a proline-rich segment that interacts with the SH3 domain of PACSIN [40]. In rat TRPV2, deletion of up to 65 N-terminal residues did not significantly alter channel function as assessed in vitro, whereas deletion of 83 or more residues greatly reduced plasma membrane expression in HEK293 cells [41], supporting the assignment of a domain boundary around 60–70 residues from the N-terminus.

Residues ~ 70–320 of TRPV2 form the ARD, which contains six ankyrin repeats. This soluble domain has yielded the only piece of atomistic tridimensional information currently available for TRPV2: the crystal structures of the human and rat TRPV2-ARD were published nearly simultaneously [39, 42]. ARDs are common protein–protein recognition domains, and although by no means unique to TRP channels they are clearly important for the modulation of TRP channels from the TRPA, TRPC, TRPN and TRPV subfamilies [43]. Each TRPV2 ankyrin repeat (ANK1–6) is defined by two antiparallel helices (inner and outer) followed by a loop region linking to a β-hairpin structure (named Fingers 1–5). These motifs are arranged linearly, with repeats packing side-by-side and the β-hairpin structures projecting out ~ 90° relative to the helical axes. Each ankyrin repeat is typically rotated a few degrees counterclockwise relative to the previous repeat, leading to an overall helical arrangement of the repeats [44]. The concave surface thus generated is often used in protein–ligand interactions to modulate the function of ARD-containing proteins. This is an important observation when comparing the sequence conservation profiles of the ARDs of TRPV1 and TRPV2 (Fig. 3). When the sequence conservation within a set of orthologous sequences is mapped onto the molecular surface of the ARD, TRPV2 shows a lower degree of conservation than TRPV1. The most conserved region on the TRPV2-ARD surface does map to the concave surface, partially overlapping with the larger conserved region on the TRPV1-ARD. As detailed below, the TRPV1-ARD binds to ATP, but the ATP binding region, highly conserved in TRPV1, is not present in TRPV2 (Fig. 3). Overall, the different conserved surface profiles of TRPV1 and TRPV2 suggest that their ARDs bind to a different set of regulatory ligands.

Connecting the TRPV2-ARD to the first transmembrane segment (S1) is a highly conserved ~ 65-residue segment, which has been referred to as the membrane proximal domain. Functional analyses of modular chimeras between TRPV1 and TRPV2, and their functional analyses expressed in HEK cells, suggest that the membrane proximal domain is an important structural module for the temperature sensitivity of TRPV channels [45].

As mentioned above, the TMD (residues ~ 390 and ~ 640) of TRPV2 is predicted to contain six transmembrane segments (S1–S6) and a pore-forming loop between S5 and S6 (predictions by tmhmm [46] and ∆gpred [47]). Accordingly, we have experimentally refined the position of the TMD for rat TRPV2 to be between residues 389 and 638 (unpublished results) as shown in Fig. 2. Heat, pH, voltage sensitivity and agonist binding sites for TRPV1 have been related to residues in the TMD (both at transmembrane segments and loops) but none of the results obtained for TRPV1 could be translated for TRPV2 [48]. For example, residues in the S4–S5 linker region may be important for voltage sensitivity of TRPV1; however, no effect has been observed when mutating analogous residues in the TRPV2 S4–S5 linker [49].

An interesting feature in the TRPV2 TMD is the presence of an N-glycosylation consensus motif (NXT/S) in the S5–S6 loop (Fig. 2). While the first asparagine is only found in rat and hamster TRPV2, the second one is conserved in nearly all TRPV2 orthologs. This glycosylation site is analogous to the N604 glycosylation site for TRPV1 [50]. An interesting study showed that, in rat F11 DRG cells, heterologously expressed rat and mouse TRPV2 are mostly expressed in the plasma membrane in a glycosylated form, whereas the endogenous TRPV2 isoform (rat) remains in the cytoplasm in a non-glycosylated form [51]. These results suggest that glycosylation plays a role in TRPV2 trafficking towards the plasma membrane, as has been shown for other TRPs [52, 53]. Recent reports indicate that the glycosylation of TRPV2 may play an anchoring role under the modulation of the Klotho factor [54] (further discussed below).

The TRPV2 pore-forming domain and selectivity filter, located in the S5–S6 region, is highly conserved (Fig. 2). Although the selectivity filter does not show a clear consensus sequence across the TRP channel superfamily [55, 56], the TRPV1–4 group shows high overall conservation (Fig. 2, inset). TRPV2 does differ in divalent/monovalent cation permeability from the rest of the non-selective TRPV subfamily members (TRPV1, TRPV3 and TRPV4), with the lowest P_Ca²⁺/P_Na⁺ at 2.94 (TRPV1 ~ 10, TRPV3 ~ 10, TRPV4 ~ 6) [3]. This difference in selectivity could be due to the conserved aspartate-to-glutamate substitution in TRPV2 (highlighted in Fig. 2, inset). TRPV4 shows a similar aspartate-to-glutamate substitution at a different location in the selectivity filter, which may similarly account for somewhat lower P_Ca²⁺/P_Na⁺ ~ 6 selectivity compared with P_Ca²⁺/P_Na⁺ ~ 10 for TRPV1 and TRPV3. Mutagenesis studies in the pore-forming region have indicated that mouse TRPV2 heterologously expressed in HEK cells may be constitutively active, inducing high cytotoxicity, since a change of charge mutant (E594K) reduced cytotoxicity [57].

The C-terminal cytoplasmic domain spanning residues 639–761 in rat TRPV2 contains relevant features for the assembly of the channel, such as oligomerization (TRP domain), phosphatidylinositol 4,5-bisphosphate (PIP₂) binding and calmodulin binding domains [58-60]. Although the sequence determinants driving TRPV2 oligomerization have not been proven directly, studies on TRPV4 [61] and TRPV1 [60] provide strong arguments for the role of a C-terminal coil-coiled domain in the homotetramerization of the channels. The predicted coil-coiled domain in TRPV1 overlaps with the TRP domain, just after the S6 segment. The TRP domain consists of ~ 20 amino acids, and for TRPV1 it has been shown that it participates in channel oligomerization and gating transduction [60, 62, 63].

TRPV2 regulatory interactions

Several regulatory interactions between TRPV2 cytoplasmic domains and various signalling molecules have been investigated. A two-hybrid screening using the first ~ 390 residues of human TRPV2, corresponding to the whole N-terminal cytoplasmic region, identified an interaction of TRPV2 with recombinase gene activator (RGA) [64] in the RBL2H3 mast cell line. Although it was later demonstrated that a physiological complex containing both TRPV2 and RGA is crucial for the cAMP-driven trafficking of TRPV2 to the plasma membrane, it remains unclear whether RGA and TRPV2 interact directly [65].

The TRPV2 N-terminal region has also been implicated in another protein–protein interaction. Mouse TRPV2 residues 1–167 were used as bait to identify the human GSRP-56 (Golgi-localized spectrin-repeat containing protein 56), which is an alternative splicing isoform of Syne-1 (synaptic nuclear envelope protein 1) [66]. A physiological role for this interaction has yet to be elucidated.

The N-terminal ARD plays important roles in the regulation of the TRPV channels. For instance, mutations to its concave surface affect the activity of TRPV1, TRPV3 and TRPV4 [67-70]. In the case of TRPV1, the channel activity can be modulated by ATP or calmodulin, and these two ligands were found to bind competitively to the TRPV1-ARD concave surface [70]. This finding was extended to TRPV3 and TRPV4, although the details of the modulatory effects differ [68]. Intriguingly, TRPV2 is the only member of the TRPV1–4 group that shows no binding of ATP or calmodulin to its ARD and is insensitive to changes in intracellular ATP concentration [67-70].

The TRPV2 Ca²⁺-dependent desensitization is dependent on PIP₂ depletion from the membrane [58]. The C-terminal TRP domain contains the TRP box (I₆₉₅WKLQR₇₀₁ in rat TRPV1, I₆₅₉WKLQK₆₆₄ in rat TRPV2), which has been implicated in PIP₂ and calmodulin binding. Alanine mutagenesis of TRPV1 residues I696, W697 and R701 (K in TRPV2) severely affected channel gating, raising the free energy of channel activation [62]. Mercado et al. defined a PIP₂ binding domain in TRPV2 within the C-terminal proximal residues 647–715 [58] (Fig. 2). It has been proposed that the phospholipase C hydrolysis of PIP₂ is the main regulatory process for TRPV2 and other TRP channels [71]. Recently, Julius and his team have shown compelling evidence that TRPV1 is intrinsically temperature sensitive by using an extremely elegant TRPV1 reconstitution model [72]. Besides being the first published report for TRPV1 activity in a biochemically defined reconstituted membrane system, they describe the inhibition of TRPV1 by C-terminal binding of phosphoinositides. A number of studies had previously provided evidence favoring a sensitizing role of phosphoinositides [58, 71]. In contrast, the Julius laboratory had already previously described PIP₂-mediated inhibition of TRPV1 in cells [73], and they now provided strong evidence of a similar inhibitory role in a reconstituted system. Julius and colleagues present a comprehensive and testable model of how the physiology of phosphoinositide levels can regulate TRPV1 activity, which may be a conserved mechanism for TRPV2.

A C-terminal TRPV2 calmodulin binding domain was initially predicted and later characterized in vitro by the Gordon laboratory [58, 74]. This calmodulin binding domain was further characterized and restricted to the 654–683 region of human TRPV2, almost fully overlapping with the TRP domain [59]. Despite showing binding of calmodulin to a C-terminal fragment of TRPV2 (residues 684–753, Fig. 2) in a Ca²⁺-dependent manner, the Gordon laboratory could not show that calmodulin is a major player in TRPV2 desensitization [58].

A few additional predicted and/or characterized TRPV2 interacting proteins have been cataloged in the TRP channel interacting protein (TRIP) database [75, 76], some of which are further detailed in later sections.

TRPV2 physiology and signalling pathways

TRPV2 is a homotetrameric N-glycosylated protein that translocates from intracellular membrane compartments to the plasma membrane after stimulation of the phosphatidylinositol 3-kinase (PI3K) and other kinase signalling pathways. TRPV2 proteins expressed heterologously in HEK293 cells show a species-dependent variability in thermal sensing and ligand activation (further described below in TRPV2 modulation and pharmacology). From a physiological perspective, TRPV2 has been associated with several functions depending on the considered tissue. Although originally described as a noxious heat thermosensor, several reports point towards TRPV2 functioning instead as a mechanoreceptor and/or osmosensor. Strikingly, TRPV2 knockout mice display normal thermal and mechanical nociception responses [18]. Analysis of sensory ganglia development (DRG and trigeminal ganglia) resulted in no significant differences comparing wild-type and knockout tissues. The acute nociception and hyperalgesia responses to thermal and mechanical stimuli in the knockout mice were normal. To discard compensation effects of other heat-gated channels on the TRPV2 knockout mice, the authors studied TRPV1/TRPV2 double knockout mice and the thermal phenotype was statistically indistinguishable from the single TRPV1 knockout. From a chemical ligand perspective, in contrast to most studies where TRPV2 was expressed heterologously, studies of TRPV2-null cultured neurons showed no differential calcium responses dependent on cannabinoid, probenecid or 2-aminoethoxydiphenyl borate (2-APB) compared with wild-type, although the wild-type responses were highly inconsistent [18]. Therefore, the TRPV2 knockout mouse analyses indicate that TRPV2 is not sensitive to these compounds under physiological conditions.

The most distinctive features of the knockout mice are reduced perinatal viability and embryo and adult body weight. Although the study of the knockout mice has thus far shed little light on the physiological roles of TRPV2, studies derived from culturing knockout mice cells indicate that macrophage function is clearly affected [77], which may correlate with perinatal lethality relating TRPV2 to immune activity (discussed below and excellently reviewed recently in [78]). In summary, the physiological role of TRPV2 is probably one of the most unsettled and controversial among TRP channels. Below we discuss the available data on TRPV2 physiology and signalling pathways, but first we introduce the pharmacological agents that have been shown to modulate TRPV2. The sparse pharmacology toolkit available to study TRPV2 is important for evaluating our knowledge of TRPV2 physiology.

Conclusions and future perspectives

We have reviewed the current information about the sequence determinants and physiology of the TRPV2 ion channel. The TRPV2 tissue distribution is quite broad and the cellular signalling pathways that regulate its function seem to show some tissue-specific variations. In vivo, the knockout TRPV2 mouse model argues against the role of TRPV2 as a mechanosensor and thermosensor. The fact that TRPV1/TRPV2 double knockout mice show a thermal phenotype that is indistinguishable from the single TRPV1 knockout should discourage future classification of TRPV2 as a thermosensor, at least in mice. Most in vitro cellular studies have shown TRPV2 to be activated downstream of physical stimuli such as mechanical stretch, heat, osmotic swelling, and endogenous and exogenous chemical modulators such as hormones, growth factors, chemotactic peptides (fMetLeu, neuropeptide head activator etc.), lysophospholipids and cannabinoids. Depending on the cell type, TRPV2 is regulated by general or specific signalling pathways involving phospholipase C, the PI3K pathway or other kinases. However, an important theme is emerging that TRPV2 is regulated through translocation and trafficking from internal cytosolic pools towards the plasma membrane, where the levels of PIP₂ in the surroundings of TRPV2 seem to play a crucial role in TRPV2-mediated Ca²⁺ signalling. Concerning the translocation mechanism, heterologous expression studies of recombinant TRPV2 in HEK cells have shown that PI3K activation of TRPV2 does not drive TRPV2 trafficking towards the cell surface [57] in contrast to other studies [26, 77, 114, 116, 122]. The set of controversies related to TRPV2, such as thermosensation, translocation, LPS dependence etc., should call for extra attention on drawing definitive conclusions, since experimental conditions (e.g. expression systems, antibodies, detection tags/epitopes) seem to be a major source of conflicting observations for TRPV2 activation. Another major limitation in TRPV2 biochemistry and physiology studies is the scarcity of specific endogenous ligands and/or pharmacological drugs, making TRPV2 essentially an orphan TRP channel. The most reliable pharmacological modulators for TRPV2 are probenecid [87] and cannabidiol [135], as agonists, which are also known to activate/inhibit other TRP channels (e.g. TRPV1, TRPA1, TRPM8 [81, 87]), which can lead to complications in interpreting experiments in physiological systems coexpressing several TRP channels. Regarding inhibition, tranilast has been identified, so far, as a TRPV2-specific antagonist [89].

To deorphanize TRPV2 and to understand TRPV2's molecular mechanisms, a combination of several approaches can be taken. First, integrating the current physiological knowledge of TRPV2 with structural biology approaches will open new perspectives on how this polymodal tetrameric channel is modulated. New structural knowledge on TRPV2 will also provide more general knowledge on the TRPV subfamily, including its closest homolog TRPV1, and the TRP superfamily. For example, the structural characterization of the N-terminal ARD has already yielded information about differential TRPV2 regulation compared with the rest of the TRPV subfamily. In line with the structural biology aspect, biochemical and/or structural evidence for TRPV2 modulator binding would be ideal for a better understanding of TRPV2 molecular mechanisms. Second, obtaining information on the TRPV2 interactome has already provided insightful information such as protein–protein interactions implicated in regulation, translocation and trafficking. Further studies on the TRPV2 interactome may open new physiological perspectives on TRPV2's elusive role(s). Finally, the development of methods to identify modulators of TRPV2 in a high-throughput screening fashion should open new therapeutic perspectives for a channel with potential implications in several types of cancer and other disorders such as muscular dystrophies, diabetes and central nervous system disorders. Interestingly, several pharmaceutical companies have started to patent compounds and methods to screen properties of TRPV2 (reviewed in [28]). We can therefore anticipate some synergy in which pharmacological advances will enable experiments to deepen our understanding of TRPV2 physiology.