Palmitoylation regulates cellular distribution of and transmembrane Ca flux through TrpM7

The bifunctional cation channel/kinase TrpM7 is ubiquitously expressed and regulates embryonic development and pathogenesis of several common diseases. The TrpM7 integral membrane ion channel domain regulates transmembrane movement of divalent cations, and its kinase domain controls gene expression via histone phosphorylation. Mechanisms regulating TrpM7 are elusive. It exists in two populations in the cell: at the cell surface where it controls divalent cation fluxes, and in intracellular vesicles where it controls zinc uptake and release. Here we report that TrpM7 is palmitoylated at a cluster of cysteines at the C terminal end of its Trp domain. Palmitoylation controls the exit of TrpM7 from the endoplasmic reticulum and the distribution of TrpM7 between cell surface and intracellular pools. Using the Retention Using Selective Hooks (RUSH) system, we demonstrate that palmitoylated TrpM7 traffics from the Golgi to the surface membrane whereas non-palmitoylated TrpM7 is sequestered in intracellular vesicles. We identify the Golgi-resident enzyme zDHHC17 and surface membrane-resident enzyme zDHHC5 as responsible for palmitoylating TrpM7 and find that TrpM7-mediated transmembrane calcium uptake is significantly reduced when TrpM7 is not palmitoylated. The closely related channel/kinase TrpM6 is also palmitoylated on the C terminal side of its Trp domain. Our findings demonstrate that palmitoylation controls ion channel activity of TrpM7 and that TrpM7 trafficking is dependant on its palmitoylation. We define a new mechanism for post translational modification and regulation of TrpM7 and other Trps.


Introduction
The bifunctional channel/kinase TrpM7 is a unique, dual-function protein that combines an N terminal ion channel with an intracellular C terminal protein kinase [1].The TrpM7 ion channel is a classic transient receptor potential (Trp) channel consisting of 6 transmembrane (TM) domains which tetramerise to form a channel, with a pore loop between TM domains 5 and 6 [2].Unlike the majority of Trp family members, the channel formed by TrpM7 is permeable to divalent (as well as monovalent) cations [3].The TrpM7 kinase domain is released from the ion channel by proteolysis and phosphorylates histones in the nucleus to regulate gene expression [4].
The superfamily of Trp channels is large and diverse, with multiple sensory roles [5].All members possess a characteristic Trp domain, an α-helix immediately after the sixth transmembrane domain, which lies parallel to the membrane beneath the transmembrane domains and controls channel gating [6].TrpM7 regulates numerous cellular processes, from receptor tyrosine kinase signalling, [7] to cell motility, [8] survival [3] and differentiation [9] and is essential for normal embryonic development [10,11] and organism divalent cation balance.[12,13] TrpM7 is also a key player in multiple pathologies.For example, variants that prevent assembly of TrpM6/TrpM7 hetero-oligomers cause hypomagnesemia and electrolyte disorders, [14] and TrpM7 variants are associated with neurodegenerative disorders [15] and stillbirth [16].Deletion of the TrpM7 kinase domain blunts vascular relaxation and exacerbates hypertension induced by Angiotensin II, [17] as well as promoting cardiac hypertrophy, fibrosis, and inflammation [18].Leptin-induced elevation of TrpM7 expression and ion channel activity in the carotid body is required for development of obesity-related hypertension [19].In neurons TrpM7-mediated Ca influx causes cellular Ca overload and cell death during anoxia [20].Understanding TrpM7 regulatory pathways is consequently a high priority to understand pathogenesis of and therapeutic opportunities for treating common, debilitating diseases.
TrpM7 exists in two distinct populations in cells.One population localises to the surface membrane.Here TrpM7 controls transmembrane ion fluxes by forming a channel permeable to magnesium, calcium and zinc, which is uniquely blocked by physiological concentrations of magnesium [3].The second population localises to intracellular vesicles which have been proposed to represent an intracellular zinc reservoir [21].TrpM7 releases zinc from these vesicles when reactive oxygen species are produced [21].The molecular basis by which TrpM7 is divided into these two populations has not been determined, but is clearly a high priority in order to understand control of both transmembrane ion fluxes and intracellular zinc mobilisation.
Since TrpM7 possesses multiple cysteine residues, we hypothesized that palmitoylation regulates TrpM7 activity and/or subcellular distribution.We mapped the TrpM7 palmitoylation sites to a cluster of cysteines at the C terminal end of the TrpM7 Trp domain.Mutagenesis of these cysteines to alanine blocked ER export of TrpM7, but replacement of this region with the analogous regions of TrpM2 (TrpM7-M2) or TrpM5 (TrpM7-M5) facilitated ER export of non-palmitoylated TrpM7.
To evaluate the role of palmitoylation in controlling TrpM7 distribution in post-ER compartments we used the retention using selective hooks (RUSH) system to arrest TrpM7 in the Golgi before inhibiting its palmitoylation.Delivery of non-palmitoylated TrpM7 to the cell surface was significantly reduced following release from the Golgi.In engineered cell lines delivery of TrpM7-M2 or TrpM7-M5 to the cell surface and TrpM7-mediated Ca fluxes were significantly reduced compared to wild type TrpM7.We conclude that by controlling delivery of TrpM7 to the plasma membrane palmitoylation controls TrpM7-mediated cellular Ca uptake.

TrpM7 is palmitoylated
We investigated palmitoylation of TrpM7 in primary and immortalised cells using resin-assisted capture of acylated proteins (acyl-RAC) which uses cysteine-reactive beads to capture palmitoylated proteins in a hydroxylamine-dependent fashion after alkylation of nonpalmitoylated cysteines under strongly denaturing conditions.The success of each acyl-RAC assay was confirmed by probing for the palmitoylated lipid raft resident protein flotillin 2. We found endogenous TrpM7 was palmitoylated in vascular smooth muscle cells from humans and rats (Fig. 1A), cardiac ventricular tissue from mice, rats and humans (Fig. 1B) and human embryonic kidney cells (Fig. 1C).TrpM6-YFP and TrpM7-YFP fusion proteins were also palmitoylated when expressed in HEK cells (Fig 1C).We used two different TrpM7 antibodies for detection.One, raised against amino acids 1146-1165 of human TRPM7, reacted with all species investigated and detected full length and cleaved (1-1281, the ion channel domain) TrpM7.The second recognised only  full length human TrpM7.The degree of enrichment of cleaved TrpM7 in the acyl-RAC was identical to the degree of enrichment of full length TrpM7 (Fig. 1D), suggesting no relationship between palmitoylation and proteolytic processing.This finding also strongly suggests that all TrpM7 palmitoylation sites reside in the ion channel domain.

Palmitoylation site mapping in TrpM7
We set out to identify the palmitoylated cysteines in TrpM7 by first identifying the region(s) of the protein that are palmitoylated.We expressed the large cytosolic N and C terminal domains as YFP fusion proteins in HEK cells.Neither YFP-TrpM7 1-755 nor YFP-TrpM7 1096-1863  were robustly palmitoylated in this model system (Fig. 2A), suggesting the presence of the TM domains is important for TrpM7 palmitoylation.We therefore mutated candidate palmitoylation sites to alanine in full length TrpM7, principally focussing on cysteines that are solvent exposed and/or in close proximity to the membrane in the TrpM7 structure, [2] because the active site of the zDHHC-PAT enzymes lies at the membrane/cytosol interface [29].We concentrated first on a group of cysteines, C1143, C1144, C1146, which lie in a unstructured loop at the C terminal end of the TrpM7 Trp domain (Fig. 2B).We expressed mutants in HEK cells and monitored palmitoylation of TrpM7-YFP (by immunoblotting for GFP), endogenous TrpM7 (with an antibody specific for the human form of TrpM7) and the constitutively palmitoylated protein flotillin 2. Mutations C1143A/C1144A (2CA) or C1146A (1CA) did not change overall TrpM7-YFP palmitoylation assessed using acyl-RAC, but mutation of all three cysteines (3CA) together rendered TrpM7-YFP almost entirely non-palmitoylated (Fig. 2C).We therefore conclude that these three cysteines are the only palmitoylated residues in the protein.TrpM7 mutants 1CA and 2CA can therefore be regarded as 'partially palmitoylated': the acyl-RAC assay purifies proteins regardless of whether they are singly or multiply palmitoylated.

Palmitoylation and intracellular trafficking of TrpM7
We visualised the subcellular location of wild type, 1CA, 2CA and unpalmitoylatable 3CA TrpM7-YFP in transiently transfected HEK cells.Wild type TrpM7 predominantly localised to punctate intracellular vesicles (Fig. 3A), as has been described previously [21].3CA TrpM7-YFP was localised only to a perinuclear compartment that we identified as the endoplasmic reticulum (ER) using the marker protein dsRed-ER.Partially palmitoylated TrpM7 mutants 1CA and 2CA displayed intermediate localisation, with some located in punctate intracellular vesicles and some in the ER (Fig. 3B).We concluded from these experiments that TrpM7 is palmitoylated by an ER resident zDHHC-PAT, and that palmitoylation is required for TrpM7 to exit the ER.
We sought to investigate the role of palmitoylation in the life cycle of TrpM7 after it leaves the ER.The ER retention of 3CA TrpM7 precluded us working with this non-palmitoylated mutant.We therefore aimed to manipulate palmitoylation of TrpM7 in a post-ER compartment, and used the retention using selective hooks (RUSH) system to first arrest TrpM7 in the secretory pathway [30].RUSH relies on the interaction of a streptavidin-fused 'hook', confined to a subcellular compartment of choice, with streptavidin binding protein (SBP) fused to the protein of interest.Treatment with biotin releases the protein of interest because it competes with SBP for the streptavidin fused to the hook.
We engineered HEK-derived 293 T-REx cells stably expressing the ER hook streptavidin-CD74 or the Golgi hook streptavidin-Golgin-84, [31] and fused SBP to the N terminus of TrpM7-YFP.SBP-TrpM7-YFP localised normally when expressed in HEK cells but was retained in the ER or  1-3).SBP-TrpM7-3CA remained confined to the ER in both cell types.Biotin treatment induced the release of WT SBP-TrpM7 from either the ER or the Golgi (Supplementary Figures 2, 3).

Golgi when expressed in 293 T-REx cells expressing the corresponding hooks (Supplementary Figures
TrpM7 unequivocally operates at the cell surface, [3] but the cell surface pool (a low concentration in a large volume) is difficult to detect using confocal imaging compared to the vesicular pool (a high concentration in a small volume) [21].We therefore used membrane-impermeable biotinylation reagents to monitor abundance of SBP-TrpM7 at the surface membrane.We first manipulated palmitoylation of SBP-TrpM7 confined in the Golgi by treating cells with the broad-spectrum zDHHC-PAT inhibitor 2-bromopalmitate and the broad-spectrum thioesterase inhibitor palmostatin B (Fig 4A).2-BP reduced TrpM7 palmitoylation but palmostatin B was without effect, suggesting TrpM7 is depalmitoylated by a thioesterase which is insensitive to this agent.After treating with drugs for 4 h we released SBP-TrpM7 using biotin and measured its delivery to the cell surface membrane.All experiments monitored cell surface abundance of the housekeeping, surface membrane resident Na pump, whose surface abundance is not impacted by 2-BP treatment [32].We confirmed that no intracellular proteins were purified in the fraction designated surface membrane by blotting for the glycolytic enzyme GAPDH.Less SBP-TrpM7 was delivered to the surface membrane in cells pre-treated with 2-bromopalmitate before TrpM7 was released from the Golgi (Fig. 4B).We conclude from these experiments that as well as controlling its ER exit, TrpM7 palmitoylation promotes its trafficking from the Golgi to the surface membrane.Interestingly biotin release of TrpM7 reduced its palmitoylation compared to when it was held in the Golgi (Fig. 4A), suggesting TrpM7 is depalmitoylated in post-Golgi compartments.

Non-palmitoylated TrpM7 chimaeras show reduced abundance at the cell surface
We investigated palmitoylation site conservation amongst the TrpM family members.There are two major branches of the TrpM family, the α branch contains TrpMs 1, 3, 6 and 7, and the β branch TrpMs 2, 4, 5 and 8. [33,34] The TrpM7 palmitoylated cysteines are conserved in all α family members as well as TrpMs 4 and 8 in the β branch (Fig. 5A).Mutation of the analogous cysteines in TrpM6 rendered it non-palmitoylated (Fig. 5A), suggesting that palmitoylation at the C terminal end of the Trp domain is a common feature of this family.
Notably, TrpM2 and TrpM5 (the only TrpM family members to lack potential palmitoylation sites) are established to operate at the cell surface membrane, suggesting that some TrpM family members can traverse the secretory pathway without being palmitoylated.An alignment of only murine TrpMs 2, 5 and 7 reveals that the palmitoylated cysteines in TrpM7 are replaced with positively charged and hydrophobic amino acids in TrpM2 and TrpM5 (Fig. 5B).We reasoned that these replacements might facilitate membrane engagement of this region of the protein, which might substitute for palmitoylation and facilitate ER exit.We therefore replaced the 'CCVC' palmitoylated region of TrpM7 with 'KRIV' (TrpM2) or 'KQVF' (TrpM5) to generate two chimaeras TrpM7-M2-YFP and TrpM7-M5-YFP.Both chimaeras were non-palmitoylated (Fig. 5C) and their gross subcellular distribution assessed using confocal imaging was not different from WT TrpM7-YFP (Fig. 5D).We assessed the delivery of these proteins to the cell surface membrane using membrane-impermeable biotinylation reagents.Significantly less non-palmitoylated chimaeric TrpM7 was delivered to the cell surface than wild type (Fig. 5E).We conclude from these experiments that palmitoylation is one of the principal determinants of the fate of TrpM7 in the cell.Palmitoylated TrpM7 is delivered to the surface membrane, whereas non-palmitoylated TrpM7 is localised to intracellular vesicles.

zDHHC17 and zDHHC5 palmitoylate TrpM7
A recent high resolution proteomic characterisation of TrpM7 identified the Golgi-localised palmitoylating enzyme zDHHC17 as a putative interaction partner [35].zDHHC17 recruits substrates via its intracellular N terminal ankyrin repeats, which interact with the zDHHC ankyrin repeat-binding motif (zDABM) sequence consensus (VIAP)(VIT) XXQP in these substrates [36].The final proline in this motif is invariant [36] and central to the interaction with zDHHC17 ankyrin repeats [37].Lemonidis et al. created position-specific scoring matrices to identify zDABM sequences in human proteins using Scansite 3 [38].The six highest scoring matches to the zDABM in TrpM7 are around P586 (z score 4.8), P26 (z score 3.1), P1131 (z score 3.0), P21 (z score 2.7), P1068 (z score 2.5), and P769 (z score 2.4).P1068 is extracellular and we ruled out P26 and P21 because peptide arrays demonstrated that D (in position 25) and I (in position 20) are strongly disfavoured immediately N terminal to the zDABM proline [38].We therefore evaluated whether P586, P769 or P1131 are important for TrpM7 palmitoylation.These 3 residues reside in regions of TrpM7 that are either disordered or were not resolved in the TrpM7 cryoEM structure and which are surface exposed and may therefore engage in protein interactions.Tryptophan is strongly disfavoured in most positions in the zDABM [38] so we generated the mutants P586W, P769W and P1131W.Palmitoylation of P769W and P1131W but not P586W TrpM7 was significantly reduced in comparison to wild type TrpM7 (Fig. 6A, B).We conclude that zDHHC17 mediates palmitoylation and therefore sorting of TrpM7 in the Golgi.
We investigated the post-Golgi control of TrpM7 palmitoylation by the cell surface localised enzyme zDHHC5 by evaluating the palmitoylation of endogenous TrpM7 in an engineered zDHHC5 knockout cell line.[23,39] Re-expression of zDHHC5 in these cells elevated TrpM7 palmitoylation (Fig. 6C, D).We conclude that TrpM7 palmitoylation remains dynamic following its delivery to the cell surface, and is controlled by zDHHC5.

Influence of palmitoylation on TrpM7-mediated cellular Ca uptake
We assessed ion channel activity of TrpM7 using 293 T-REx cells stably expressing either tetracycline inducible TrpM7-YFP or TrpM7-M5-YFP.In these stably transfected cells TrpM7-M5-YFP was not palmitoylated and was less abundant at the surface membrane compared to wild type (Supplementary Figure 4).Ca uptake in the presence of 1 mM or 5 mM extracellular Ca was measured in cells loaded with the fluorescent Ca sensor Fluo-4 [40].Cellular Ca uptake was indistinguishable in cells not treated with tetracycline.TrpM7-mediated Ca uptake was significantly reduced in cells expressing TrpM7-M5-YFP compared to TrpM7-YFP (Fig. 7A, 7B).We conclude that by controlling TrpM7

Discussion
This investigation set out to determine whether TrpM7 is palmitoylated and if so how palmitoylation influences the behaviour of the dual channel/kinase TrpM7.We report that palmitoylation regulates multiple steps in the TrpM7 life cycle.Ultimately by controlling delivery of TrpM7 to the surface membrane palmitoylation controls TrpM7mediated cation fluxes at the plasma membrane.

Palmitoylation and TrpM7 channel activity
The substantial reduction in TrpM7-mediated cellular Ca uptake when TrpM7 is not palmitoylated is largely due to the reduced delivery of non-palmitoylated TrpM7-M5 to the surface membrane.Interestingly we note that the very substantial reduced delivery of TrpM7-M5 to the cell surface compared to wild type TrpM7 (~70% less) is not fully matched by the decline in cellular Ca influx in the same cells (~50% in the presence of 5 mM extracellular Ca).This raises the possibility that the single channel activity of palmitoylated TrpM7 is different from nonpalmitoylated TrpM7.In addition, the fact that non-palmitoylated TrpM7 is diverted to the vesicular pool means that channel activity and/or regulatory pathways of this pool may be different from cell surface TrpM7.Although in this investigation we focussed on transmembrane Ca uptake as an index of TrpM7 ion channel activity, we suggest that TrpM7-mediated Mg and Zn fluxes (which make important contributions to multiple pathologies [41][42][43]) will be identically regulated by TrpM7 palmitoylation.
The mechanisms controlling channel activation of Trp superfamily members are diverse, but a common feature of Trp channel gating is its regulation by the anionic lipid phosphatidylinositol 4,5 bisphosphate (PIP2) which binds to positively charged amino acids in the Trp box within the channel's Trp domain.[44][45][46][47] Multiple Trp channels are inhibited when PIP2 is hydrolysed, including TrpM7 [48].The mobility and/or flexibility of the Trp domain, likely regulated by PIP2 binding, are key determinants of channel gating.The palmitoylation sites in TrpM7 lie at the C terminal end of the Trp domain, in a region where the protein structure is poorly resolved [2].It is conceivable that palmitoylation here may influence the ability of the Trp domain, a rigid α-helix, to move relative to the membrane, which would consequently alter channel behaviour.

TrpM7 palmitoylation by zDHHC17
Like other polytopic membrane proteins, TrpM7 is palmitoylated throughout the secretory pathway by zDHHC-PATs confined to individual cellular compartments [31].We did not investigate which zDHHC-PAT palmitoylates TrpM7 in the ER, but identified zDHHC17 in the Golgi as one of the principal determinants of TrpM7 palmitoylation, and consequently of its cellular fate.We identified two different prolines within consensus zDABMs whose presence was required for TrpM7 palmitoylation.Notably, a large proportion (12/17, 70%) of proteins containing multiple potential zDABMs were found to interact with zDHHC17 through more than one of these motifs [38].Since removal of one zDABM clearly impairs the ability of the other to direct TrpM7 palmitoylation, our data suggest some co-operativity to substrate recruitment and/or palmitoylation by zDHHC17.
Enhanced cation influx through TrpM7 has been implicated in numerous pathologies, from anoxia-induced Ca overload in the brain, [20] to leptin induced hypertension in the carotid body [19].There are therefore multiple potential therapeutic applications for a TrpM7 channel inhibitor.Our investigation signposts a hitherto unexplored route for inhibiting TrpM7-mediated cation flux for therapeutic gain, by manipulating its palmitoylation status to reduce delivery to the cell surface.Substrate recognition by zDHHC17 is the best characterised of the zDHHC-PAT family [37,38].Targeting substrate recruitment by other zDHHC-PATs effectively blocks palmitoylation of their substrates [39].We suggest the same will apply to the relationship between TrpM7 and zDHHC17.

Palmitoylation and TrpM7 passage through the secretory pathway
The retention of the non-palmitoylated TrpM7 mutant 3CA in the ER indicates that TrpM7 palmitoylation first occurs early in its lifetime, is mediated by ER-resident zDHHC-PATs, and is required for ER export.Palmitoylation can control the distribution of proteins in membranes based on their curvature, [49,50] meaning that palmitoylated proteins tend to cluster at sites of vesicular budding.However, since the non-palmitoylated chimaeras TrpM7-M5 and TrpM7-M2 can both leave the ER, we suggest that palmitoylation is required to stabilise the TrpM7 structure by anchoring the C terminal end of the Trp domain to the membrane, rather than for clustering TrpM7 into ER export vesicles.In other words, the ER retention of 3CA TrpM7 occurs because of 'quality control' rather than 'failure to export'.We speculate that the mixed basic / hydrophobic nature of 'KRIV' (TrpM2) or 'KQVF' (TrpM5) is sufficient to engage the membrane and substitute for palmitoylation in stabilising nascent TrpM7 in the ER.
The role of palmitoylation in sorting proteins for vesicular export from the Golgi is well established [50].Our data is consistent with TrpM7 destined for the surface membrane being palmitoylated by zDHHC17 in the Golgi (leading to its segregation into plasma membrane-bound vesicles), with TrpM7 destined for the intracellular pool remaining non-palmitoylated.Since TrpM7 leaving the ER must be palmitoylated this implies the existence of a TrpM7 depalmitoylating enzyme that remains to be identified, but which in our experiments is insensitive to the broad-spectrum thioesterase inhibitor palmostatin B. Notably, no zDHHC-PATs were found in a proteomic characterisation of TrpM7 intracellular vesicles, [21] which implies that the vesicular pool of TrpM7 remains non-palmitoylated.Our finding that the cell surface resident enzyme zDHHC5 palmitoylates TrpM7 implies that its palmitoylation status (and therefore possibly channel activity) is dynamically regulated once it has reached the surface membrane.
The  domain to the membrane may therefore be a common feature of the Trp family.

Ethics statement
Primary ventricular myocytes from neonatal rats and adult rabbits and vascular smooth muscle cells (VSMCs) from adult rats were utilized in this investigation.All protocols were approved by the University of Glasgow Animal Welfare and Ethics Review Board.Hearts and vessels from neonatal and adult rats were collected post-mortem after sacrificing animals using a method designated Schedule 1 by the Animals (Scientific Procedures) Act 1986.Rabbit hearts were excised from terminally anaesthetized, heparin-treated animals under the authority of a Project License granted by the UK Home Office.
Human endocardium samples from organ donors were obtained from the Gill Cardiovascular Biorepository at the University of Kentucky, to which patients and families of organ donors provided written consent.
Human VSMC were isolated from gluteal biopsies of subcutaneous fat from normotensive and hypertensive volunteers at the Ottawa Hospital, Ontario, Canada, as we previously described [51,52].Ethical approval was obtained from the research ethics board of the Ottawa Hospital Research Institute (OHRI), Canada (#997,392,132).Written informed consent was obtained for all study participants in accordance with the Declaration of Helsinki.

Plasmids and mutagenesis
A plasmid expressing murine TrpM7 with a C terminal yellow fluorescent protein fusion was kindly provided by Dr Vladimir Chubanov, Walther-Straub-Institut für Pharmakologie und Toxikologie.
All mutagenesis utilised either Quikchange II site directed mutagenesis kit (Agilent) or InFusion cloning (Takara).TrpM7 fragments were subcloned into pEYFP-C1 (Clontech) and TrpM7-YFP was subcloned into pcDNA5 FRT/TO (Invitrogen) using InFusion cloning.All mutagenesis and cloning used oligonucleotide primers designed according to the kit manufacturer's instructions.

Primary cells
Primary VSMCs were isolated from small arteries obtained from gluteal biopsies subcutaneous fat and mesenteric arteries from WKY rats, by enzymatic digestion, as previously described [52].Briefly, excess fat, connective tissue and adventitia from isolated arteries were removed.Vessels were incubated in Ham's F-12 culture medium containing 1% gentamicin, collagenase type I, elastase, soybean trypsin inhibitor, and bovine serum albumin for 60-90 min at 37 • C under constant agitation.The digested tissue was further dissociated by repeated aspiration through a syringe with a 20-gauge needle and filtered through a 100 μm nylon filter (BD Biosciences, Oxford, UK) to remove debris.Cell suspension was centrifuged at 800 g for 3 min, re-suspended in Ham's F-12 culture medium containing 10% FBS and seeded onto 25 cm 2 flasks for the first 48 h.Thereafter, human VSMCs were maintained in DMEM supplemented with Smooth Muscle Growth Supplement (SMGS; Thermo Fisher Scientific) and rat VSMC were grown in DMEM supplemented with 10% FBS and 5% antibiotics (100 IU/ml penicillin, 100 μg/ml streptomycin)).
Adult rabbit ventricular cardiomyocytes were isolated from New Zealand white rabbits (3-4 kg, male, 12 weeks of age), as previous described [53].

Cell culture
HEK-293 cells and HEK-derived 293 T-REx cells were cultured using standard conditions.Transfection of plasmid constructs was achieved using Lipofectamine 2000 (Invitrogen) according to manufacturer's instructions.Cells were transfected with TrpM7-YFP 48 h before experimentation.
Cells stably expressing ER and Golgi hooks [31] and zDHHC5 knockout cells are described elsewhere [23,39].Cells stably expressing TrpM7 and its mutants were generated by co-transfecting 293 T-REx cells with pcDNA5 FRT/TO TrpM7 and pOG44 (Invitrogen) then selecting using Hygromycin and Blasticidin.Tetracycline induction of mRNA expression was 24-48 h before experimentation.

Palmitoylation assays
Palmitoylated proteins were purified using resin-assisted capture of acylated proteins (acyl-RAC), as described previously [54].Briefly, free cysteines were alkylated with MMTS, and palmitoylated proteins captured using thiopropyl Sepharose in the presence of 250 mM neutral hydroxylamine to cleave thioester bonds.

Confocal imaging
Cells were seeded on the poly-L-lysine coated glass coverslips (Ø16mm) and incubated at 37 • C for 48 h after transfection or tetracycline induction of expression.Cells were briefly washed with PBS followed by fixation with 4% paraformaldehyde (PFA) diluted in PBS for 20 min at room temperature.10 mM Glycine/PBS was applied to quench any unreacted aldehyde twice for 5 min each time, then cells were washed, and coverslips mounted with Duolink® mounting medium (Fisherbrand™), containing 4 ′ ,6-diamidino-2-phenylindole (DAPI).
For immunofluorescence, cells were washed with PBS 3 times after the quenching step, then incubated with 0.1% Triton X-100/PBS at room temperature for 10 min to permeabilise membranes.Coverslips were incubated in anti-HA primary antibody (1:200) followed by anti-rat secondary (1:400 Alexa Fluor 546 or Alexa Fluor 633 conjugated, Thermo Fisher Scientific) in 0.1% BSA/PBS then washed and mounted.
Confocal images were acquired using a Zeiss LSM 520 microscope or Zeiss LSM 880 microscope with a 63x oil immersion objective (Carl Zeiss, Cambridge, UK).

Preparation of cell surface proteins
Surface membrane proteins were biotinylated on extracellular primary amines using 1 mg/ml sulfo-NHS-SS-biotin in Dulbecco's PBS, then purified using streptavidin Sepharose as described previously [55].To confirm the integrity of the assay cell surface fractions were routinely immunoblotted for the cell surface protein Na/K ATPase α1 subunit and the intracellular protein GAPDH.

Calcium uptake
Cells were seeded into 96-well poly-L-lysine coated microplates and incubated at 37 • C for 48 h in the presence or absence of tetracycline to induce TrpM7 expression.Cells were loaded with Fluo-4 for 1 hour at 37 • C using the Fluo-4 Direct Calcium Assay Kit (Thermo Fisher Scientific) in an extracellular buffer composed of 140 mM NaCl, 6 mM KCl, 10 mM HEPES, 5 mM Glucose (pH 7.4) according to the manufacturer's instructions.Fluorescence intensity (Ex = 494 nm and Em = 516 nm) was measured at 37 • C using a POLARstar OPTIMA Multidetection Microplate Reader (BMG LABTECH) before and after the addition of 1

Fig. 1 .
Fig. 1.TrpM7 is palmitoylated in cells and tissues A -Palmitoylated proteins were prepared from vascular smooth muscle from normotensive (NT) and hypertensive (HT) patients or Wistar Kyoto (WKY) rats, and immunoblotted as indicated.TrpM7FL: antibody reacts only with full length human TrpM7; TrpM7IC: antibody reacts with cleaved (150 kDa) and full length (250 kDa) TrpM7 from all species investigated; UF: unfractionated cell / tissue lysate; Palm: purified palmitoylated proteins.B -Palmitoylated proteins were prepared from ventricular muscle from neonatal rats, adults rats, rabbits, humans and immunoblotted as indicated.C -YFP tagged TrpM7 and TrpM6 transiently transfected in HEK cells and endogenous TrpM7 in HEK cells are palmitoylated.Neg: palmitoylation assay negative control in which hydroxylamine is excluded from the purification.D -Relative enrichment of full length (250 kDa) and cleaved (150 kDa) TrpM7 from human ventricular muscle in the palmitoylation assay (n = 13).

X
.Gao et al.

Fig. 4 .
Fig. 4. Palmitoylation and TrpM7 intracellular trafficking (2) A -Palmitoylation of SBP-TrpM7-YFP arrested in the Golgi in cells treated with the zDHHC-PAT inhibitor 2-bromopalmitate (2-BP), the thioesterase inhibitor palmostatin B (PalmB) or vehicle (Ctl).Also shown is the impact of releasing TrpM7 from the Golgi hook by treating with biotin.The graph shows the relative enrichment of TrpM7 and flotillin 2 (Flot2) in the palmitoylation assay.**: P<0.01, *: P<0.05 compared to control, one-way ANOVA followed by Dunnett's multiple comparisons test (n = 3).B -Cell surface abundance of SBP-TrpM7-YFP after release from the Golgi following treatment with 2-bromopalmitate, palmostatin B or vehicle.Samples were immunoblotted for the cell surface protein Na/K ATPase α1 subunit (Na pump) and the intracellular enzyme GAPDH.The graph shows the relative enrichment of TrpM7 and the Na pump in the cell surface fraction.*: P<0.05 compared to control, one-way ANOVA followed by Dunnett's multiple comparisons test (n = 3).
TrpM7 palmitoylation sites are retained in most members of the TrpM family.Clustal alignment of members of the wider Trp superfamily also identifies analogous cysteines in most members of the TrpV and TrpC families (Fig S5).Notably TrpC6, TrpV3 and TrpV4, the only channels lacking cysteine residues analogous to TrpM7, also possess clusters of basic and hydrophobic amino acids like those found in TrpM2 and TrpM5.A requirement to anchor the C terminal end of the Trp