TRPM8 has Highly Conserved Amino Acid Sequences in Lipid-Water-Interface Regions

To study the importance of the LWI regions in TRPM8 evolution, we performed a conservation analysis using all vertebrate sequences. We have selected a stretch of five amino acids at the N- and C-terminal of each TM region which is approximately 6–10 Å in length (Fig. 1a). A total of 8 out of 12 LWI regions remain highly conserved, suggesting the importance of the LWI region in the overall channel function (Fig. 1b). Notably, the inner LWI regions are more conserved than the outer LWI region (stat value, p =  < 0.0001, unpaired non-parametric Mann–Whitney t-test). This indicates the overall selection pressure on the inner LWI region was more than the outer LWI region. This may also suggest that the factors that are present in the cytosolic region and the lipids that are present in the inner leaflet of the membrane have exerted more selection pressure on TRPM8 function. We also noted that overall, the LWI regions are more conserved than the TM regions or even full-length TRPM8 (stat value, p =  < 0.0001, unpaired non-parametric Mann–Whitney t-test), suggesting that LWI regions have more selection pressure than the TM or even full-length and are functionally more critical than the TM regions (Fig. 1b). The LWI representing the C-terminal peptide sequence is more conserved than the N-terminal sequence (stat value, p =  < 0.0001, unpaired non-parametric Mann–Whitney t-test) suggesting that the peptide directionality also plays a role in this selection.

Fig. 1

Amino acids present in the lipid-water-interface (LWI) regions of TRPM8 are highly conserved. a. Schematic representation of full-length human TRPM8 protein. The zoomed figure on the right side depicts the transmembrane regions of TRPM8, where green and red color represent N- and C-terminus LWI regions for each transmembrane sequence, respectively. The LWI regions located between TM2 and TM3 have overlapped with each other to some extent and are thus marked with yellow color. b. Conservation analysis of all the LWI regions separately, all inside, all outside, all N-terminal, and all C-terminal regions are shown. All LWI regions together remain more conserved than the full-length or even all TM regions. Conservation of Histone H4 has been considered as a positive control. c. Amino acids of different LWI regions are highly conserved throughout vertebrate evolution as shown by seqlogo analysis. Many of the conserved amino acids are also the hotspots for cancer-related mutations (shown by red star marks)

Next, we analyzed the seqlogo of all the LWI regions and compared that with mutations in TRPM8 that are listed in cancer patients which include lymphoid neoplasm, carcinoma, malignant melanoma, pancreatic intraepithelial neoplasia, hematopoietic neoplasm, carcinoid endocrine tumor, primitive neuroectodermal tumor- medulloblastoma, glioma etc. (Fig. 1c). We noted that several amino acids that are present in the LWI regions and are highly conserved have mutated in the case of cancer patients. Analysis of these mutations also suggests that the frequency of cancer is also observed to be higher in the LWI region. The frequency of the cancer mutation observed in the LWI region is ~ 61% (for 60 amino acids) and the frequency of the cancer mutation observed in the rest of the TRPM8 regions is 39% (1104–60 = 1044 amino acids). Which suggests that the LWI can actually serve as a potential “hotspot for cancer”. This strongly suggests the importance of TRPM8 in cancer and also suggests somatic mutations in TRPM8 at the LWI region as a possible causal factor for cancer progression (discussed later).

Analysis of Amino Acid Frequencies at the Lipid-Water-Interface Region of TRPM8

To gain the details of the molecular evolution of TRPM8 residues, we tested the frequency of all 20 amino acids in the LWI regions in different phylogenetic groups. We could not find any fish-specific TRPM8 sequence from the available databases. This limits us to perform our analysis in amphibian (A), Reptilian (R), Birds (B) and mammals (M) sub-groups. We noted that TRPM8 is missing in most of the fishes (Kastenhuber et al. 2012; Saito and Tominaga 2015; York and Zakon 2022). Systematic analysis of the residues present in all the LWI regions are summarized (Fig. 2, for details, see Supplementary Fig S1).

Fig. 2

Frequency analysis of residues at the LWI region indicates the trends of conservation, exclusion, positive and negative-selection during vertebrate evolution. a. Frequency analysis of Cys, Thr, Val, Arg are shown here (for all 20 amino acids, see supplementary Fig S1). Frequencies that remain conserved are shown by blue arrow, positive selection by a green background and negative selection with a red background. Certain residues show zero frequency throughout the vertebrate evolution suggesting complete exclusion of such amino acids (such as Cys) from the LWI region. The dotted line indicates the expected natural frequency of that amino acids. The green circle indicates the sudden increase in frequency from reptiles to birds (i.e. possible involvement of these residues for adjusting the channel function in altered body temperature). * = p < 0.01. b. Graphical representation of frequencies of amino acids in warm (red dots) and cold-blooded (blue dots) animals are shown. The dotted-line indicates the expected natural frequency of that amino acids. Mann–Whitney test was performed to calculate statistical significance. The values are: **** = p < 0.0001, ** = p < 0.001

Complete exclusion: We observed that several amino acids are completely excluded in the LWI region of TRPM8 sequences, suggesting that during the vertebrate evolution, these sequences never appeared in the LWI regions. For example, His, Ala, Met, Cys residues are excluded in the outer LWI region. Comparably, Glu, Gln, Pro, Asn, Ala, Cys, Ile, Phe, residues are excluded from the inner LWI region. Notably, Ala and Cys residues are totally excluded from both outer as well as inner LWI regions.

Positive selection: We noted that certain amino acids become positively selected, i.e. the trend of their frequency increased gradually during the vertebrate evolution. For example, Val in the outer LWI region has increased from lower vertebrates to mammals. In the same notion, His in the inner LWI region has increased from lower to higher vertebrates. Considering both outer and inner LWI regions, a positive selection of Val in total LWI is observed.

Negative selection: Opposite to positive selection, we observed that certain amino acids in specific regions have decreased during vertebrate evolution. For example, Ile has have declined trend infrequency in the outer LWI region. At the inner LWI region, Arg has declined trend in frequency during the vertebrate evolution. Considering both inner and outer LWI regions, Arg, Thr, and Ile residues have declined trend in frequency during the vertebrate evolution.

Amino acids with conserved frequencies: We noted that several amino acids are neither excluded nor selected positively or negatively. These amino acids remain at a conserved frequency during the vertebrate evolution. In other words, the presence of these amino acids in these frequencies is optimum, and critical for the channel function. We noted that at the outer LWI, there are amino acids such as Glu, Asp, Pro, Arg, Thr, Gly, Leu, Tyr, Phe, and Trp that remain at conserved frequencies. In a similar manner, at the inner LWI regions, a few amino acids, such as Lys, Val, Gly, and Tyr, remain at conserved frequencies. At the inner LWI region, Asp residue also remains at a conserved value though with some out layers. Considering both inner and outer LWI regions, Glu, Asp, Gln, Pro, His, Gly, Met, Tyr, and Phe remain mostly conserved during vertebrate evolution. Notably, in the case of TRPM8, overall, more amino acids are conserved at the outer LWI than the inner LWI region (discussed later).

Amino acids that remain more than the natural frequencies: While certain amino acids remain at conserved frequencies, there are cases when these residues appear more than the natural frequencies, suggesting more selection pressure at these amino acids. For example, at the outer LWI region, Pro, Arg, Phe, and Trp remain at higher values than the natural frequencies. Similarly, at the inner LWI region, Asp, Thr, Gly, Met, Leu and Tyr remain at higher values than their respective natural frequency. Arg also appears more than its natural frequency (with a trend of negative selection). Considering both inner LWI and outer LWI regions, Asp, Pro, Thr (With a trend of negative selection), Gly, Met, Tyr, and Phe appear more than their respective natural frequencies.

Amino acids that remain less than the natural frequencies: We noted that a few amino acids remain conserved throughout the vertebrate evolution, yet maintained at lower than the natural frequencies. This includes Glu, Asp, and Leu in the outer LWI regions. Similarly, Lys and Val remain at the less then natural frequencies at the inner LWI region. Considering both inner and outer LWI regions, Glu, Gln, and His are retained at lower than natural frequency.

Amino acids that remain just equal to the natural frequencies: We noted that few amino acids remain conserved throughout the vertebrate evolution and their frequency match just equal to the natural frequency. These include Thr, Gly, and Tyr at the outer LWI region only.

All these analyses provide important information about the possible regulation of TRPM8 by Cys modification, glycosylation, and phosphorylation (discussed later).

Comparatives of LWI Residues in Human TRPM8 as Derived From Protter vs Cryo EM

Recently the Cryo-EM structure of human TRPM8 (PDB ID: 8BDC) become available (Palchevskyi et al. 2023). This allowed us to compare the TM and LWI residues as derived from Protter and in the Cryo-EM structure. We found that majority of the TM and LWI residues detected by these two different approaches match well (Supplementary Fig S2).

Ratio of Positive–Negative Residues and Hydrophilic-Hydrophobic Residues are Conserved in the Inner Lipid-Water-Interface Region of TRPM8

We calculated the frequency of total positively charged and total negatively charged residues, in the inner LWI region. In the inner LWI region, these values remain conserved (total positive charged residue =  ~ 17.8%, negatively charged residue =  ~ 10.7%) throughout the vertebrate evolution (Fig. 3a). We measured the ratio of positive to negative charged residues. Notably, the ratio remains conserved (positive to negative ratio at the inner LWI region =  ~ 1.6) in most vertebrate sequences tested. Similarly, we calculated the frequency of total hydrophobic (Trp, Phe, Tyr, Leu, Ile, Cys, Met, total value is 35.8%) and total hydrophilic residues (Ala, Arg, Asn, Asp, Gln, His, Pro, Ser, Thr, Lys, Gly, Val, total value is 64.2%) in inner LWI regions (Fig. 3b). This ratio also remains conserved (hydrophobic-hydrophilic at the inner LWI region =  ~ 1.8).

Fig. 3

The LWI region of TRPM8 has a unique pattern of amino acids. Shown are the ratio of frequency of hydrophobic, hydrophilic positive and negative amino acids remained conserved during vertebrate evolution. a. The graphs represent the total frequency of all positive, all negative and their ratio values in each species throughout the vertebrate evolution. The ratio remains conserved in the inner LWI region and semi-conserved in the outer LWI region. b. The frequency of total hydrophobic, total hydrophilic amino acids as well as they remain conserved throughout the vertebrate evolution. This conservation in ratio is observed in the inner LWI region, but not in the outer LWI region. Mann–Whitney test was performed to calculate statistical significance. The values are: **** = p < 0.0001, ** = p < 0.001

The frequency of all negatively charged residues at outer LWI remains conserved (~ 6.6). However, positively charged residues occur at higher frequencies (> 15) and remain conserved with some variations. Notably, the ratio of positive–negative at outer LWI is not conserved.

However, the total (inner and outer LWI regions together) frequency of hydrophobic residues remains conserved (~ 31). Similarly, the total frequency of the hydrophilic residues remains mostly conserved (~ 68.3). However, the ratio of hydrophilic to hydrophobic residues is less conserved as compared to the inner LWI. Similarly, the ratio of positive–negative charged residues is also not conserved when both inner and outer LWI are considered. These findings suggest that during vertebrate evolution, TRPM8 has a conserved pattern of amino acids, especially in its inner LWI region.

LWI Regions have Very Few Residues that Show Changes due to Body Temperature

During vertebrate evolution, warm-blooded animals evolved from cold-blooded animals and thus core body temperature increased several degrees. As TRPM8 is a temperature-gated on channel, changes in core body temperature is expected to induce changes in its channel opening properties. Thus, a change in core body temperature should be accompanied by changes in TRPM8 sequence so that it will readjust its function in different body temperature. So we hypothesized if a specific amino acid frequency is changed at the LWI region from cold-blooded animals to warm-blooded animals, or even a sudden change from reptilian to birds, then such changes may be triggered by the change in core body temperature. For that we have grouped all the cold-blooded animals (amphibians and reptiles) and warm-blooded animals (birds and mammals) in two distinct groups. We observed that only very few amino acids have significant and notable changes in terms of body temperature and most of the amino acids do not have any change (Supplementary Fig S3).

In that context, we observed that notable changes in Val in the outer LWI region between cold- blooded animals to warm-blooded animals. At the outer LWI region, Thr and Ile also have certain changes between these two groups. In the same context, Leu has notable changes in the inner LWI region. Asp has certain changes between these two groups. Considering both inner LWI and Outer LWI region, Val has notable changes. Other amino acids such as Asp, Thr, Ile, and Leu has certain changes. Therefore, these changes can be relevant for the altered thermogating behavior of TRPM8. Among all these, sudden change in Asn and Ser are observed between reptilians with birds.

Since the “cold response” of avian TRPM8 is very different from that of mammals, especially by its ability to get activated by low temperatures (Pertusa et al. 2018; Yang et al. 2020; Lu et al. 2022), we have also evaluated the amino acid differences in LWI regions between birds and mammals. We have found percentage values of Leucine and Tryptophan are significantly different in the inner LWI region. However, these two amino acids remained unchanged with respect to mammals in the outer LWI region. The total change also remained significantly different for Leu and Trp. On the other hand, a few amino acids, namely Lys, Asn, and Ser remained unchanged in the inner LWI region but were significantly changed at the outer LWI regions (Fig S1). These subtle differences can be important for the differences in the thermal activation of TRPM8 in these two groups.

TRPM8 has Several Cholesterol-Binding Conserved Motif Sequences

Cholesterol is a plasma membrane component and it impacts several transmembrane proteins. Cholesterol is also a vertebrate-specific molecule. Several membrane proteins show the presence of cholesterol binding. For that, we have used 67 vertebrate-specific TRPM8 sequences and analyzed the occurrence of possible cholesterol-binding sequences in TRPM8.

We observed a total of 40 CARC and 15 CRAC motifs that are present in the human-TRPM8 (Supplementary Table 2). Out of these, 14 CARC and 3 CRAC motifs are located at the TM and LWI regions suggesting that these regions have a high probability of interaction with cholesterol present in the membrane (Fig. 4a). We analyzed the conservation of these potential cholesterol-binding motifs. We observed that 1 CRAC motif (AA 678–688), 4 CARC motifs (7th, 8th, 9th, and 14th motifs, 8th and 9th motifs have overlapping regions) (AA 829–834, 842–853, 842–849, 901–909) are highly conserved (Fig. 4b). Except for the 1st, 3rd, 4th, and 6th CARC motifs, all these motifs have higher conservation than the full-length TRPM8, suggesting that cholesterol has a strong selection pressure on these motifs (discussed later).

Fig. 4

TRPM8 has several cholesterol-binding motif sequences in its LWI and TM regions which are conserved throughout the vertebrate evolution. a. Protter image demonstrating the presence of cholesterol recognizing sequences CARC and CRAC in the full-length TRPM8. The zoomed image in the right side shows the CRAC and CARC motifs in red and green respectively. b. The box plot shows divergence of different CARC (indicated in red) and CRAC (indicated in green) motifs. Several CARC- and CRAC-motifs present in TRPM8 remain conserved in all vertebrates. Histone H4 sequence has been used a control for highly conserved protein

Cholesterol Interaction with TRPM8 Differs with its Different Conformations

To analyze the possible interaction of cholesterol with TRPM8 and further regulation of TRPM8 by cholesterol, we performed docking of cholesterol on avian and human TRPM8. For this, Cryo-EM structure of TRPM8 with or without different ligands were used [PDB: 6BPQ (without any ligand), 6NR2 (with occupancy WS12), 6NR3 (with high occupancy Icilin with PIP2 and Ca2+), 6NR4 (with low occupancy Icilin with PIP2 and Ca2+), 6O6A (ligand-free), 6O6R (inhibitor AMTB-bound state), 6O72 (TC1-bound state), 6O77 (Ca2+-bound state) and 8BDC (human apo-TRPM8 in a closed-state) (Palchevskyi et al. 2023).

We observed that cholesterol interacts with avian TRPM8 (without any other ligands, PDB: 6BPQ) and good interaction has been observed at 8th CRAC-motif (one of the highly conserved motifs) (Fig. 5a). Cholesterol interaction is observed on the 8th CRAC-motif of TRPM8 in presence of activator also (with high occupancy Icilin, PDB ID: 6NR3) (Fig. 5b). Cholesterol interaction is also observed on the 14th CARC (another highly conserved motif) in the presence of another activator (WS12, PDB ID: 6NR2) (Fig. 5c). In all these cases H-bond formation between the OH-group of cholesterol with TRPM8 is observed. However, such H-bond formation is not observed in the case of TRPM8 with inhibitor-bound form (with AMTB, PDB ID: 6O6R) (data not shown).

Fig. 5

Cholesterol interacts with TRPM8 in apo- and in ligand-bound forms. a-d. Shown are the molecular docking of cholesterol (indicated in Red) with the Cryo-EM structures of Avian-TRPM8 (a-c) and Human TRPM8 (d) in different conformations. The respective binding energies (delta G value) are also indicated. e–f. Graphical representation of delta G values of cholesterol-binding with TRPM8 in different conformations of bird (d) and mouse (e). For each conformation, top 10 binding modes are plotted here

We have performed cholesterol docking with Human TRPM8 (in closed state, 8BDC) as well. Cholesterol docks well at the highly conserved region i.e. CARC8 (Fig. 5d). Cholesterol also docks well on the mouse TRPM8 (ligand-free) at CARC12 with a ∆G value of -8.9 (Fig S5). We have also found interaction of cholesterol at different CARC- and CRAC-motifs in ligand-bound form as well as ligand-free form in avian and mouse TRPM8 (data not shown).

Cholesterol also interacts well with mouse TRPM8 in ligand-free form, at CARC12 (Fig S4). In addition to this, cholesterol interaction was also observed with different conformations of TRPM8 in mouse. To analyze the changes in binding energy in different conformations, we calculated the top 10 binding modes of cholesterol with all available conformations of birds (Fig. 5e), and mice (Fig. 5f). Changes in binding energy are observed in the case of different confirmations, suggesting that cholesterol can interfere with the ligand bindings and vice versa.

The cholesterol docking in humans, mice, and avian TRPM8 indicates that CARC8 and CARC14 could be the "hot-spot" for possible cholesterol binding in all these three species. Notably, these 2 regions also remained highly conserved throughout the vertebrate evolution (Fig. 4b). From the data obtained from the Catalogue for Somatic Mutations in Cancer or COSMIC, we have observed that CARC8 and CARC14 regions have various somatic cancer mutations. We have found 4 cancer mutations (i.e., p.R842K, p.R851I, p.N852S, p.N852K) in the CARC8 region, while 6 somatic cancer mutations (i.e., p.R901C, p.S902*, p.S902L, p.I904M, p.Y905 = , p.L909M [Here, * symbol specifies nonsense mutation while ( =) is coding silent] are found in the CARC14 region. Recently interaction of cholesterol on the 3rd extracellular loop has been reported (Lee 2019). Therefore, we compared the cholesterol docking on the 3rd extracellular loop as well, especially on the CARC and CRAC motifs. We observed that in mouse TRPM8, cholesterol forms H-bond (with Glu953) on the 3rd extracellular-loop, especially in the ligand-free conformation (7WRA). We could not observe any possible H-bond formation in any other conformation. In other word, this may also suggest that the loss-of-interaction of cholesterol may prefer channel opening as observed in certain cases of TRPV1 and TRPV4 (Saha et al. 2022; Das and Goswami 2019). In our docking experiments (with Flycatcher TRPM8-cholesterol docking) we only focused on CARC-CRAC motifs, which we found to have several common sequences that matches with the Lee 2019 study (Eberhardt et al. 2021). Some notably residues are Leu842 from CARC-8, Val902 and Leu908 from CARC-14. Taken together the data suggests that CARC8 and CARC14 remain highly conserved and most prominent regions for cholesterol interaction. These two regions might have played a crucial role in TRPM8 evolution. These two regions may also have role in fine-tuning of TRPM8-specific function/s during the vertebrate evolution, and cholesterol interaction may have played a crucial role in such selection.

Alteration in the Cholesterol Level Affect the TRPM8 Localization on the Cell Surface

To study the TRPM8 correlation with cholesterol in vitro, we estimated the surface expression of TRPM8 in control and/or cholesterol reduced (by application of β-MCD) conditions. For this purpose, primary murine peritoneal macrophage has been chosen where TRPM8 expresses endogenously (Khalil et al. 2016). Cells were incubated with the TRPM8 modulators alone for 3:30 h, and subsequently cholesterol was reduced with the help of cholesterol-reducing agent β-MCD for 30 min wherever applicable (Fig. 6a). Subsequently the cells were fixed by PFA and labeling of TRPM8 at the surface was performed. At the same condition, surface expressed GM1 ganglioside was labeled with CTxB. We have observed major changes in TRPM8 surface expression due to β-MCD-treatment (Fig. 6b). In all these conditions, the cellular area remained mostly unchanged (Fig. 6c); however, the TRPM8 surface localization was significantly higher in cholesterol-reduced (β-MCD-treated) conditions, mostly irrespective of TRPM8 modulation (Fig. 6d). The surface staining for lipid-raft has also been significantly reduced in β-MCD treated condition (Fig. 6e) and so was the ratio of CTxB/TRPM8 (Fig. 6f). These data are suggestive of the major role of cholesterol in TRPM8 regulation.

Fig. 6

Cholesterol reduction promotes surface localization of TRPM8 in primary peritoneal macrophage. a. Schematic representation of the experiment and duration of treatments are depicted. b. Representative images show the TRPM8 intensity and cholesterol-enriched lipid raft intensity (stained with Cholera Toxin-B or CTxB). Fluorescence intensities are represented in rainbow RGB scale. The merge images represent TRPM8 in green and CTxB in red. c. Graph represents the cellular area after the treatment with TRPM8 modulators, with or without β-MCD treatment. d-e. Graphs represents % surface expression of TRPM8 (d) and Cholera Toxin-B (e). f. Graphical representation of ratio of CTxB and TRPM8. To calculate the statistical significance, non-parametric Kruskal–Wallis test was performed where #### = p < 0.0001

Several Mutations Causing Cancer are Located in the LWI and Cholesterol-Binding Regions of TRPM8

Intracellular Ca2+-signaling is intimately related to the signaling of cancer cells and/or tumor progression (Cui et al. 2017). We observed a large number of mutations that are present in the TRPM8 in tumor samples as somatic mutations. We have cataloged all these somatic mutations (Supplementary Table 3). A large number of mutations have been detected in the lipid-water-interface region of TRPM8. These mutations are also observed to be present in the LWI region, both in activator as well as inhibitor-bound forms (Fig. 7a, b). Similarly, a large number of mutations are also observed in the cholesterol-binding regions. In this regard, it is known that TRPM8 is involved in testosterone signaling relevant to prostate cancer (Grolez et al. 2019). Taken together, the data strongly suggest that TRPM8 cholesterol crosstalk is important for cellular functions and alteration of such crosstalk can be relevant in different cancer conditions.

Fig. 7

Series of somatic mutations in LWI and CARC-CRAC region of TRPM8 is found in cancer patients. The cancer mutation of LWI, and CARC-CRAC motifs are marked only in the TM regions and in LWI regions. a-b. A side view (left side) and transverse cut of side view (right side) of TRPM8 opened (a) and closed (b) structures with point mutations as found in cancer patients are shown. Mutations in LWI and CARC-CRAC regions only are shown in red and green respectively. The images in the middle show the top view of the structure at outer LWI (salmon color circle), the middle section of lipid bilayer and cytoplasmic region (cyan color circle). Highlighted amino acids represents the cancer mutation in LWI region (in red) and in the CARC-CRAC regions overlapping with LWI region (in yellow). These following positions, i.e. E782, P958 (for left panel, top), V784 (for right panel, top), P716, T732, S733, P734, L757, M758, V783, E832 (for left panel, below) and P734, I957 (for right panel, below) are marked in red color and these positions are located in CARC-CARC motifs where mutations in cancers are reported. Notably, V791, S827 (for left panel, top), Q785 (for right panel, top), R688, Q785, V791 (for left panel, below) and S827 (for right panel, below) are residues cancer mutations are reported and these residues indicate common positions for LWI as well as CRAC-CRAC motifs