P2 Depletion Upon Receptor Activation

This Accepted Manuscript has not been copyedited and formatted. The final version may differ from this version.

Research Articles: Cellular/Molecular

Gαq sensitizes TRPM8 to inhibition by PI(4,5)P2 depletion upon receptor activation Luyu Liu, Yevgen Yudin, Janhavi Nagwekar, Chifei Kang, Natalia Shirokova and Tibor Rohacs Department of Pharmacology, Physiology and Neuroscience, Rutgers New Jersey Medical School, Newark NJ 07103 https://doi.org/10.1523/JNEUROSCI.2304-18.2019

Received: 6 September 2018

Revised: 26 April 2019

Accepted: 29 April 2019

Published: 24 May 2019

Author contributions: L.L., Y.Y., C.K., N.S., and T.R. designed research; L.L., Y.Y., J.N., and C.K. performed research; L.L., Y.Y., J.N., C.K., and T.R. analyzed data; L.L. wrote the first draft of the paper; L.L., Y.Y., J.N., C.K., N.S., and T.R. edited the paper; T.R. wrote the paper.

Conflict of Interest: The authors declare no competing financial interests.

T.R. was supported by NIH grants NS055159 and GM093290. The authors thank Dr. Joshua Berlin for his insightful comments, Dr. David Julius (UCSF) for providing the rat TRPM8 clone, Dr. David McKemy (University of Southern California) for providing the TRPM8 GFP reporter mouse line, Dr. Yasushi Okamura (Osaka University, Japan) for providing the ci-VSP and dr-VSP clones, Drs. Nikita Gamper (University of Leeds) and Andrew Tinker (University College London) for providing the tubby-R332H-YFP clone, Dr. Xuming Zhang (Aston University, Birmingham, UK) for providing the 3Gqiq clone. We also thank Linda Zabelka for maintaining the mouse colony and Luke Fritzky for his help using the Nikon A1R-A1 confocal system.

Corresponding author: Tibor Rohacs, [email protected], Department of Pharmacology, Physiology and Neuroscience, Rutgers New Jersey Medical School, 185 South Orange Ave, Newark NJ 07103

Cite as: J. Neurosci 2019; 10.1523/JNEUROSCI.2304-18.2019

Alerts: Sign up at www.jneurosci.org/alerts to receive customized email alerts when the fully formatted version of this article is published.

Accepted manuscripts are peer-reviewed but have not been through the copyediting, formatting, or proofreading process.

ͳ Gαq sensitizes TRPM8 to inhibition by PI(4,5)P2 depletion ʹ upon receptor activation ͵ Ͷ ͷ Short title: Mechanism of receptor-induced inhibition of TRPM8 ͸ ͹ ͺ Luyu Liu, Yevgen Yudin, Janhavi Nagwekar, Chifei Kang#, Natalia Shirokova, ͻ Tibor Rohacs ͳͲ ͳͳ Department of Pharmacology, Physiology and Neuroscience, Rutgers New ͳʹ Jersey Medical School, Newark NJ 07103 ͳ͵ ͳͶ Corresponding author: Tibor Rohacs, [email protected], Department of ͳͷ Pharmacology, Physiology and Neuroscience, Rutgers New Jersey Medical ͳ͸ School, 185 South Orange Ave, Newark NJ 07103 ͳ͹ ͳͺ #current address: Department of Medicine, Division of Endocrinology, Icahn ͳͻ School of Medicine at Mount Sinai, New York, NY 10029 ʹͲ ʹͳ ʹʹ ʹ͵ ʹͶ Number of pages: 25 (excluding figures) ʹͷ ʹ͸ Number of Figures: 10 ʹ͹ ʹͺ Number of words in: ʹͻ Abstract: 218 ͵Ͳ Introduction: 559 ͵ͳ Discussion: 1119 ͵ʹ ͵͵ ͵Ͷ The authors declare no financial conflict of interest ͵ͷ ͵͸ ACKNOWLEDGEMENTS: ͵͹ T.R. was supported by NIH grants NS055159 and GM093290. The authors thank ͵ͺ Dr. Joshua Berlin for his insightful comments, Dr. David Julius (UCSF) for ͵ͻ providing the rat TRPM8 clone, Dr. David McKemy (University of Southern ͶͲ California) for providing the TRPM8 GFP reporter mouse line, Dr. Yasushi Ͷͳ Okamura (Osaka University, Japan) for providing the ci-VSP and dr-VSP clones, Ͷʹ Drs. Nikita Gamper (University of Leeds) and Andrew Tinker (University College Ͷ͵ London) for providing the tubby-R332H-YFP clone, Dr. Xuming Zhang (Aston ͶͶ University, Birmingham, UK) for providing the 3Gqiq clone. We also thank Linda Ͷͷ Zabelka for maintaining the mouse colony and Luke Fritzky for his help using the Ͷ͸ Nikon A1R-A1 confocal system.

Ͷ͹ Ͷͺ Ͷͻ ABSTRACT ͷͲ ͷͳ The cold and menthol sensitive Transient Receptor Potential Melastatin 8 ͷʹ (TRPM8) channel is important for both physiological temperature detection and ͷ͵ cold allodynia. Activation of G-protein coupled receptors (GPCRs) by pro- ͷͶ inflammatory mediators inhibits these channels. It was proposed that this

ͷͷ inhibition proceeds via direct binding of Gαq to the channel. TRPM8 requires the

ͷ͸ plasma membrane phospholipid phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2

ͷ͹ or PIP2] for activity. It was claimed however that a decrease in cellular levels of ͷͺ this lipid upon receptor activation does not contribute to channel inhibition. Here

ͷͻ we show that supplementing the whole cell patch pipette with PI(4,5)P2 reduced

͸Ͳ inhibition of TRPM8 by activation of Gαq-coupled receptors in mouse dorsal root ͸ͳ ganglion (DRG) neurons isolated from both sexes. Stimulating the same ͸ʹ receptors activated Phospholipase C (PLC) and decreased plasma membrane

͸͵ PI(4,5)P2 levels in these neurons. PI(4,5)P2 also reduced inhibition of TRPM8 by

͸Ͷ activation of heterologously expressed Gαq-coupled muscarinic M1 receptors. Co-

͸ͷ expression of a constitutively active Gαq protein that does not couple to PLC ͸͸ inhibited TRPM8 activity, and in cells expressing this protein decreasing

͸͹ PI(4,5)P2 levels using a voltage sensitive 5’-phosphatase induced a stronger ͸ͺ inhibition of TRPM8 activity than in control cells. Our data indicate that upon

͸ͻ GPCR activation, Gαq binding reduces the apparent affinity of TRPM8 for

͹Ͳ PI(4,5)P2 and thus sensitizes the channel to inhibition induced by decreasing

͹ͳ PI(4,5)P2 levels. ͹ʹ ͹͵

͹Ͷ ͹ͷ SIGNIFICANCE STATEMENT ͹͸ ͹͹ Increased sensitivity to heat in inflammation is partially mediated by inhibition of ͹ͺ the cold- and menthol sensitive TRPM8 ion channels. Most inflammatory ͹ͻ mediators act via GPCR-s that activate the Phospholipase C pathway leading to

ͺͲ the hydrolysis of PI(4,5)P2. How receptor activation by inflammatory mediators ͺͳ leads to TRPM8 inhibition is not well understood. Here we propose that direct

ͺʹ binding of Gαq both reduces TRPM8 activity, and sensitizes the channel to

ͺ͵ inhibition by decreased levels of its cofactor PI(4,5)P2. Our data demonstrate the

ͺͶ convergence of two downstream effectors of receptor activation Gαq and

ͺͷ PI(4,5)P2 hydrolysis in regulation of TRPM8. ͺ͸ ͺ͹

ͺͺ INTRODUCTION ͺͻ ͻͲ TRPM8 is a non-selective Ca2+ permeable cation channel activated by cold ͻͳ temperatures, as well as chemical agonists such as menthol, icilin, and WS12 ͻʹ (Almaraz et al., 2014). It is expressed in the primary sensory neurons of dorsal ͻ͵ root ganglia (DRG) and trigeminal ganglia (TG), and mice lacking these channels ͻͶ display deficiencies in sensing moderately cold (Bautista et al., 2007; Colburn et ͻͷ al., 2007; Dhaka et al., 2007) as well as noxious cold temperatures (Knowlton et ͻ͸ al., 2010). TRPM8 may also play a role in cold allodynia induced by nerve injury ͻ͹ (Xing et al., 2007). ͻͺ ͻͻ DRG neurons express receptors for various inflammatory mediators, such as ͳͲͲ bradykinin, ATP and prostaglandins. Most of those receptors couple to

ͳͲͳ heterotrimeric G-proteins in the Gαq family, and lead to activation of

ͳͲʹ phospholipase C (PLC) (Rohacs, 2016). Activation of these Gαq-coupled ͳͲ͵ receptors induces thermal hyperalgesia (Caterina et al., 2000), via mechanisms ͳͲͶ including sensitizing the heat- and capsaicin-activated Transient Receptor ͳͲͷ Potential Vanilloid (TRPV1) channels (Cesare and McNaughton, 1996; Tominaga ͳͲ͸ et al., 2001) and inhibiting the cold-activated TRPM8 (Premkumar et al., 2005; ͳͲ͹ Zhang et al., 2012). ͳͲͺ

ͳͲͻ The mechanism of the inhibition of TRPM8 by Gαq-coupled receptors is not fully ͳͳͲ understood. Protein Kinase C (PKC) has been demonstrated to be involved in ͳͳͳ sensitization of TRPV1 (Numazaki et al., 2002; Bhave et al., 2003; Zhang et al., ͳͳʹ 2008), but its involvement in TRPM8 inhibition is controversial (Premkumar et al.,

ͳͳ͵ 2005; Zhang et al., 2012). It was proposed that binding of Gαq to TRPM8 ͳͳͶ mediates inhibition of the channel (Zhang et al., 2012; Li and Zhang, 2013). ͳͳͷ Similar to most TRP channels (Rohacs, 2014), TRPM8 requires the membrane

ͳͳ͸ phospholipid phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2 or PIP2] for activity ͳͳ͹ (Liu and Qin, 2005; Rohacs et al., 2005; Daniels et al., 2009; Zakharian et al., ͳͳͺ 2010). Ca2+ influx through the channel was shown to activate a Ca2+ sensitive

ͳͳͻ PLC isoform, and the resulting decrease in PI(4,5)P2 levels limits channel activity, ͳʹͲ leading to desensitization (Rohacs et al., 2005; Daniels et al., 2009; Yudin et al.,

ͳʹͳ 2011; Yudin et al., 2016). Stimulation of Gαq-coupled receptors leads to the

ͳʹʹ activation of PLCβ isoforms, which hydrolyze PI(4,5)P2, but it was questioned

ͳʹ͵ whether this leads to a decrease in PI(4,5)P2 levels in DRG neurons (Liu et al.,

ͳʹͶ 2010), and it was argued that a decrease in PI(4,5)P2 levels does not play a role ͳʹͷ in TRPM8 channel inhibition (Zhang et al., 2012). ͳʹ͸

ͳʹ͹ Here we revisited this question, and demonstrate that the decrease in PI(4,5)P2

ͳʹͺ levels plays an important role in inhibition of TRPM8 by Gαq-coupled receptor ͳʹͻ activation in DRG neurons. First we show that application of a cocktail of

ͳ͵Ͳ inflammatory mediators that activate Gαq-coupled receptors decreased PI(4,5)P2 ͳ͵ͳ levels and inhibited TRPM8 activity in DRG neurons. Supplementing the whole

ͳ͵ʹ cell patch pipette with PI(4,5)P2 alleviated the inhibition of TRPM8 by these

ͳ͵͵ receptor agonists. Excess PI(4,5)P2 also decreased the inhibition of ͳ͵Ͷ heterologously expressed TRPM8 by activation of muscarinic M1 receptors. Co-

ͳ͵ͷ expression of a constitutively active Gαq that is deficient in PLC activation ͳ͵͸ inhibited TRPM8 activity, which is consistent with earlier results (Zhang et al.,

ͳ͵͹ 2012). Co-expression of the PLC-deficient Gαq also increased the sensitivity of

ͳ͵ͺ TRPM8 to inhibition by decreasing PI(4,5)P2 levels using a specific voltage ͳ͵ͻ sensitive phosphoinositide 5-phosphatase, providing a mechanistic framework of

ͳͶͲ how Gαq inhibits TRPM8. Overall, our data indicate that a decrease in PI(4,5)P2

ͳͶͳ levels and direct binding of Gαq converge on inhibiting TRPM8 activity upon cell ͳͶʹ surface receptor activation. ͳͶ͵ ͳͶͶ MATERIALS AND METHODS ͳͶͷ ͳͶ͸ DRG neuron isolation and preparation ͳͶ͹ ͳͶͺ All animal procedures were approved by the Institutional Animal Care and Use ͳͶͻ Committee at Rutgers New Jersey Medical School. Dorsal root ganglion (DRG)

ͳͷͲ neurons were isolated from 2- to 6-month-old wild-type C57BL6 ͳͷͳ (RRID:IMSR_JAX:000664) or TRPM8-GFP mice (Takashima et al., 2007) ͳͷʹ (RRID:IMSR_JAX:020650) as described previously (Yudin et al., 2016) with ͳͷ͵ some modifications. We have extensively characterized electrophysiological ͳͷͶ properties of GFP-positive DRG neurons in this reporter mouse in our earlier ͳͷͷ work (Yudin et al., 2016). DRG neurons were isolated from mice of either sex, ͳͷ͸ anesthetized, and perfused via the left ventricle with ice-cold Hank's buffered salt ͳͷ͹ solution (HBSS; Invitrogen). DRGs were harvested from all spinal segments after ͳͷͺ laminectomy and removal of the spinal column and maintained in ice-cold HBSS ͳͷͻ for the duration of the isolation. Isolated ganglia were cleaned from excess dorsal ͳ͸Ͳ root nerve tissue and incubated in an HBSS-based enzyme solution containing 3 ͳ͸ͳ mg/ml type I collagenase (Worthington) and 5 mg/ml Dispase (Sigma) at 37°C for ͳ͸ʹ 35 min, followed by mechanical trituration by repetitive pipetting through an uncut ͳ͸͵ 1000 μl pipette tip. Digestive enzymes were then removed after centrifugation of ͳ͸Ͷ the cells at 80 × g for 10 min. Cells were then either directly re-suspended in ͳ͸ͷ growth medium and seeded onto glass coverslips coated with a mixture of poly- ͳ͸͸ D-lysine (Invitrogen) and laminin (Sigma), or were first transfected using the ͳ͸͹ Amaxa nucleoporator according to manufacturer's instructions (Lonza ͳ͸ͺ Walkersville). Briefly, 60,000–100,000 cells obtained from each animal were re- ͳ͸ͻ suspended in 100 μl nucleofector solution after complete removal of digestive ͳ͹Ͳ enzymes. The cDNA used for transfecting neurons was prepared using the ͳ͹ͳ Endo-Free Plasmid Maxi Kit from QIAGEN. Vectors and reagent for transduction

ͳ͹ʹ with the Green PI(4,5)P2 Sensor (CAG-Promoter) #D0405G and Red PI(4,5)P2 ͳ͹͵ sensor (#D0405R) were obtained from Montana Molecular. Before seeding onto ͳ͹Ͷ glass coverslips, approximately 60,000–100,000 cells were re-suspended in 500 ͳ͹ͷ μl transduction mix after complete removal of digestive enzymes. Neurons were ͳ͹͸ maintained in culture for at least 24 h before measurements in DMEM-F12 ͳ͹͹ supplemented with 10% FBS (Thermo Scientific), 100 IU/ml penicillin and 100 ͳ͹ͺ μg/ml streptomycin. ͳ͹ͻ ͳͺͲ HEK cell culture and preparation

ͳͺͳ ͳͺʹ Human embryonic kidney 293 (HEK293) cells were obtained from the American ͳͺ͵ Type Culture Collection (ATCC) (catalogue number CRL-1573, ͳͺͶ RRID:CVCL_0045) and were cultured in minimal essential medium (Invitrogen) ͳͺͷ containing supplements of 10% (v/v) Hyclone-characterized FBS (Thermo ͳͺ͸ Scientific), 100 IU/ml penicillin, and 100 μg/ml streptomycin. Transient

ͳͺ͹ transfection was performed at ∼70% cell confluence with the Effectene reagent ͳͺͺ (QIAGEN) according to the manufacturer's protocol. Cells were incubated with ͳͺͻ the lipid–DNA complexes overnight (16-20 h). The cDNA used for transfecting ͳͻͲ HEK cells was prepared using the Endo-Free Plasmid Maxi Kit from QIAGEN.

ͳͻͳ The cDNA for the 3Gqiq construct was obtained from Dr. Xuming Zhang’s

ͳͻʹ laboratory (Zhang et al., 2012). The Q209L mutant of 3Gqiq was generated using ͳͻ͵ the QuickChange Mutagenesis Kit (Agilent). Transduction was performed at

ͳͻͶ ∼70% cell confluence with BacMam green PI(4,5)P2 sensor with CAG promoter ͳͻͷ and BacMam M1 receptor (Montana Molecular) according to manufacturer’s ͳͻ͸ protocol. Cells were incubated with transduction mix overnight (16-24 h). We ͳͻ͹ then trypsinized and re-plated the cells onto poly-D-lysine-coated glass ͳͻͺ coverslips and incubated them for an additional at least 2 h (in the absence of the ͳͻͻ transfection reagent) before measurements. All mammalian cells were kept in a

ʹͲͲ humidity-controlled tissue-culture incubator maintaining 5% CO2 at 37°C. ʹͲͳ ʹͲʹ Electrophysiology ʹͲ͵ ʹͲͶ For whole-cell patch-clamp recordings the room where the experiments were ʹͲͷ performed was heated up to 27–30°C, as described previously (Yudin et al., ʹͲ͸ 2016). This slightly higher than conventional room temperature was used ʹͲ͹ because TRPM8 may show some basal activity at 22-24 °C. Patch pipettes were ʹͲͺ pulled from borosilicate glass capillaries (1.75 mm outer diameter, Sutter ʹͲͻ Instruments) on a P-97 pipette puller (Sutter Instrument) to a resistance of 4–6 ʹͳͲ MΩ. After formation of seals, the whole-cell configuration was established and

ʹͳͳ currents were measured at a holding potential of −60 mV (DRG neurons) or with ʹͳʹ a voltage ramp protocol from -100 mV to +100 mV (HEK293 cells) using an ʹͳ͵ Axopatch 200B amplifier (Molecular Devices). Currents were filtered at 2 kHz ʹͳͶ using the low-pass Bessel filter of the amplifier and digitized using a Digidata ʹͳͷ 1440 unit (Molecular Devices). In some experiments, membrane potential pulses ʹͳ͸ of indicated lengths to 100 mV were applied to activate the voltage-sensitive ʹͳ͹ phosphatase (Ci-VSP), as described earlier (Velisetty et al., 2016). ʹͳͺ Measurements were conducted in solutions based on either Ca2+-free (NCF)

ʹͳͻ medium containing the following (in mM): 137 NaCl, 4 KCl, 1 MgCl2, 5 HEPES, 5 ʹʹͲ MES, 10 glucose, 5 EGTA, pH adjusted to 7.4 with NaOH, or Ca2+ containing

ʹʹͳ (NCF) medium consisting of (in mM): 137 NaCl, 5 KCl, 1 MgCl2, 10 HEPES, 10

ʹʹʹ glucose, 2 CaCl2, pH adjusted to 7.4 with NaOH (Lukacs et al., 2013a). ʹʹ͵ Intracellular solutions for DRG measurements (NIC-DRG) consisted of the

ʹʹͶ following (in mM): 130 K-Gluconate, 10 KCl, 2 MgCl2, 2 Na2ATP, 0.2 Na2GTP,

ʹʹͷ 1.5 CaCl2, 2.5 EGTA, 10 HEPES, pH adjusted to 7.25 with KOH. Intracellular ʹʹ͸ solutions for HEK293 whole-cell measurements (NIC-HEK) consisted of the

ʹʹ͹ following (in mM): 130 KCl, 10 KOH, 3 MgCl2, 2 Na2ATP, 0.2 Na2GTP, 0.2 CaCl2,

ʹʹͺ 2.5 EGTA, 10 HEPES, pH adjusted to 7.25 with KOH (Lukacs et al., 2013b). DiC8

ʹʹͻ PI(4,5)P2 (Cayman Chemical, Ann Arbor, MI) was dissolved in NIC. Carbachol ʹ͵Ͳ and all chemicals in the inflammatory cocktail were purchased from Sigma ʹ͵ͳ Aldrich, BIM IV was from Santa Cruz Biotechnology (Dallas, TX), and WS12 was ʹ͵ʹ from Alomone Labs, (Minneapolis, MN). ʹ͵͵ ʹ͵Ͷ Ca2+ imaging ʹ͵ͷ ʹ͵͸ Ca2+ imaging measurements were performed with an Olympus IX-51 inverted ʹ͵͹ microscope equipped with a DeltaRAM excitation light source (Photon ʹ͵ͺ Technology International). DRG neurons or HEK cells were loaded with 1 μM ʹ͵ͻ fura-2 AM (Invitrogen) for 50 min in extracellular solution supplemented with ʹͶͲ 0.1mg/ml Bovine Serum Albumin before the measurement at room temperature ʹͶͳ (27–30°C). Images at 340 nm and 380 nm excitation wavelengths were recorded

ʹͶʹ with a Roper Cool-Snap digital CCD camera; emission wavelength was 510 nm.

ʹͶ͵ Measurements were conducted in NCF solution supplemented with 2 mM CaCl2. ʹͶͶ Image analysis was performed using the Image Master software (PTI). ʹͶͷ ʹͶ͸ Confocal fluorescence imaging ʹͶ͹ ʹͶͺ Confocal measurements for figures 3,4 and 7 were conducted with an Olympus ʹͶͻ FluoView-1000 confocal microscope in the frame scan mode (Olympus) using a ʹͷͲ 60X water immersion objective at room temperature (~25°C). Green fluorescence ʹͷͳ and YFP fluorescence was measured using excitation wavelength of 473 nm; ʹͷʹ emission was detected through a 515/50 nm band-pass filter. Wild type DRG ʹͷ͵ neurons or HEK cells were transfected with YFP-Tubby R332H or YFP-Tubby ʹͷͶ WT cDNA at least 24 hours before confocal measurements. Before each ʹͷͷ experiment, neurons or HEK cells were serum-deprived in Ca2+ free or 2 mM ʹͷ͸ Ca2+-containing NCF solution for at least 20 min. For transduction experiments, ʹͷ͹ Wild type DRG neurons or HEK cells were transduced at least 24 hours before

ʹͷͺ confocal measurements by BacMam green PI(4,5)P2 sensor with CAG promoter ʹͷͻ and BacMam M1 receptor. Before each experiment, neurons or HEK cells were ʹ͸Ͳ serum-deprived in DPBS (Sigma-Aldrich) for 30 min in the dark. Image analysis ʹ͸ͳ was performed using Olympus FluoView-1000 and ImageJ. Confocal ʹ͸ʹ measurements in TRPM8-GFP reporter mouse DRG neurons with the red

ʹ͸͵ PI(4,5)P2 sensor (Figure 5) were conducted with a Nikon A1R-A1 confocal laser ʹ͸Ͷ scanning microscope system using a 20X objective at room temperature (~25°C). ʹ͸ͷ GFP fluorescence was measured at excitation wavelength 488 nm; emission was ʹ͸͸ detected through a 525/50 nm band-pass filter. Red fluorescence was measured ʹ͸͹ at excitation wavelength 561 nm; emission was detected through a 595/50 band- ʹ͸ͺ pass filter. Confocal images were collected using the NIS-Elements imaging ʹ͸ͻ software. Image analysis was performed using NIS-Elements, Metamorph and ʹ͹Ͳ ImageJ softwares. ʹ͹ͳ ʹ͹ʹ Total Internal Reflection Fluorescence (TIRF) Imaging

ʹ͹͵ ʹ͹Ͷ TIRF measurements were performed on an Olympus IX-81 inverted microscope, ʹ͹ͷ equipped with an ORCA-FLASH 4.0 camera (Hamamatsu), and a single line ʹ͹͸ TIRF illumination module. Excitation light was provided by a 488 nm argon laser ʹ͹͹ through a 60x NA 1.49 TIRF objective. For the experiments in Figure 8 HEK293 ʹ͹ͺ cells were transfected with the human M1 receptor and either TRPM8 tagged ʹ͹ͻ with GFP on its N-terminus (GFP-TRPM8), or YFP-tubby. The GFP-TRPM8 ʹͺͲ clone was generated by subcloning the TRPM8 coding region into the pEGFP-C1 ʹͺͳ vector between the Xho1 and EcoR1 sites using standard molecular biological ʹͺʹ techniques. Cells were plated on 25 mm round glass cover-slips, and solutions ʹͺ͵ were manually applied to the chamber with gentle mixing. For the experiments in ʹͺͶ Figure 9E,F, HEK293 cells were transfected with GFP-TRPM8, or GFP-TRPM8

ʹͺͷ plus 3GqiqQ209L. Data were analyzed using the Metamorph and Image J ʹͺ͸ software. ʹͺ͹ ʹͺͺ Experimental Design and Statistical analysis ʹͺͻ ʹͻͲ Data analysis was performed in Excel and Microcal Origin. Data collection was ʹͻͳ randomized. Data were analyzed with t-test, or Analysis of variance with Fisher ʹͻʹ post hoc test, p values are reported in the figures, results of the analysis of ʹͻ͵ variance are reported in the figure legends. Data are plotted as mean +/- ʹͻͶ standard error of the mean (SEM) for most experiments. ʹͻͷ ʹͻ͸ RESULTS ʹͻ͹

ʹͻͺ PI(4,5)P2 alleviates inhibition of TRPM8 by activation of Gαq-coupled ʹͻͻ receptors in DRG neurons ͵ͲͲ

͵Ͳͳ Mouse DRG neurons express a number of different Gαq-protein coupled ͵Ͳʹ receptors (Thakur et al., 2014). To identify receptors co-expressed with TRPM8, ͵Ͳ͵ first we performed Ca2+ imaging experiments in DRG and TG neurons isolated

ͳͲ

͵ͲͶ from TRPM8-GFP reporter mice. We consecutively applied 500 μM menthol, 2 ͵Ͳͷ μM bradykinin, 1 μM prostaglandin E2 (PGE2), and 500 nM endothelin (ET) (Fig. ͵Ͳ͸ 1A, B). Most DRG neurons (62%) and TG neurons (84%) that were activated by ͵Ͳ͹ menthol responded to either bradykinin, PGE2 or ET, but no individual agonist ͵Ͳͺ stimulated all TRPM8 positive neurons. ͵Ͳͻ ͵ͳͲ Next we measured Ca2+ signals in DRG neurons in response to the more specific

͵ͳͳ TRPM8 agonist WS12 and to activate various Gαq-coupled receptors we applied ͵ͳʹ an inflammatory cocktail containing 100 μM ADP, 100 μM UTP, 500 nM ͵ͳ͵ bradykinin, 100 μM histamine, 100 μM serotonin, 10 μM PGE2 and 10 μM ͵ͳͶ prostaglandin I2 (PGI2) (Fig. 1C, D), similar to that used by (Zhang et al., 2012). ͵ͳͷ We found that 11% of all DRG neurons were GFP positive and 57% of GFP- ͵ͳ͸ positive neurons responded to WS12 (Fig. 1E,F). This response rate is ͵ͳ͹ comparable to earlier results, indicating the incomplete overlap of endogenous ͵ͳͺ TRPM8 expression and GFP expression from the TRPM8 promoter-GFP BAC ͵ͳͻ clone (Takashima et al., 2007; Yudin et al., 2016). Within these GFP positive ͵ʹͲ neurons, almost all neurons activated by WS12 responded to the inflammatory ͵ʹͳ cocktail. Only one neuron (out of 23 GFP positive neurons) was stimulated by ͵ʹʹ WS12 but not by the inflammatory cocktail (Fig. 1F). ͵ʹ͵ ͵ʹͶ Next we preformed whole-cell patch clamp experiments on DRG neurons ͵ʹͷ isolated from TRPM8-GFP reporter mice and tested the effect of supplementing

͵ʹ͸ the patch pipette solution with the water-soluble diC8 PI(4,5)P2. We used a ͵ʹ͹ protocol of consecutive one minute applications of 10 μM WS12 in Ca2+-free NCF ͵ʹͺ solution to avoid desensitization. We used WS12 instead of menthol, as it is ͵ʹͻ more specific for TRPM8 and the latter may also activate other channels such as ͵͵Ͳ TRPA1 (Yudin et al., 2016). The inflammatory cocktail was applied one min

͵͵ͳ before the third application of WS12 (Fig. 2 A, B). Without diC8 PI(4,5)P2, the ͵͵ʹ inflammatory cocktail reduced WS12-induced current amplitudes by 53% (Fig. 2

͵͵͵ A,C,D). When we included diC8 PI(4,5)P2 in the patch pipette solution, WS12- ͵͵Ͷ induced currents decreased only by 9% on average after the application of the

ͳͳ

͵͵ͷ inflammatory cocktail (Fig. 2 B,C,D). Baseline current amplitudes induced by ͵͵͸ WS12 before the application of the inflammatory cocktail were not different

͵͵͹ between control and PI(4,5)P2 dialyzed neurons (Fig. 2C). These findings show

͵͵ͺ that TRPM8 channel activity is inhibited by Gαq-protein coupled receptors in DRG

͵͵ͻ neurons, and this inhibition is alleviated by excess PI(4,5)P2. ͵ͶͲ

͵Ͷͳ Receptor-induced decrease of PI(4,5)P2 in the plasma membrane ͵Ͷʹ

͵Ͷ͵ There are conflicting results on whether plasma membrane PI(4,5)P2 levels

͵ͶͶ decrease in DRG neurons in response to stimulating endogenous Gαq-coupled ͵Ͷͷ bradykinin receptors (Liu et al., 2010; Lukacs et al., 2013b), and there is no

͵Ͷ͸ information on the effects of other Gαq-coupled receptors (Rohacs, 2016).

͵Ͷ͹ Therefore we tested if there is a decrease in PI(4,5)P2 levels in response to the ͵Ͷͺ inflammatory cocktail we used in our electrophysiology experiments. We

͵Ͷͻ transfected DRG neurons with the YFP-tagged R322H Tubby PI(4,5)P2 sensor ͵ͷͲ using the Amaxa nucleoporator system. This YFP-tagged phosphoinositide ͵ͷͳ binding domain of the Tubby protein contains the R322H mutation that decreases

͵ͷʹ its affinity for PI(4,5)P2, thus increasing its sensitivity to small, physiological

͵ͷ͵ decreases in PI(4,5)P2 levels (Quinn et al., 2008; Lukacs et al., 2013b). The

͵ͷͶ Tubby-based fluorescent sensor binds to PI(4,5)P2 specifically; it is located in the

͵ͷͷ plasma membrane in resting conditions; when PI(4,5)P2 levels decrease, this ͵ͷ͸ indicator is translocated to the cytoplasm, which we monitored using confocal ͵ͷ͹ microscopy. Application of the inflammatory cocktail induced rapid decrease of ͵ͷͺ fluorescence in the plasma membrane and a less pronounced, variable increase ͵ͷͻ in the cytoplasm in the neuron cell bodies (Fig. 3 A, B, middle panels) as well as ͵͸Ͳ in the neuronal processes (Fig. 3 A, B, bottom panels), indicating decreased

͵͸ͳ PI(4,5)P2 levels. Fig. 3 C-E summarizes these results. ͵͸ʹ

͵͸͵ We also monitored PI(4,5)P2 levels using a different fluorescence based sensor

͵͸Ͷ the BacMam PI(4,5)P2 sensor kit from Montana Molecular. This green PI(4,5)P2

͵͸ͷ sensor is based on a dimerization-dependent fluorescent PI(4,5)P2 binding

ͳʹ

͵͸͸ protein (Tewson et al., 2013). The kit also contains cDNA for the M1 muscarinic

͵͸͹ receptor, a Gαq-protein coupled receptor, to serve as a positive control. Isolated

͵͸ͺ neurons transduced with green BacMam PI(4,5)P2 sensor showed prominent ͵͸ͻ plasma membrane labeling. Administration of inflammatory cocktail induced ͵͹Ͳ robust decrease of fluorescence levels in the plasma membrane, indicating

͵͹ͳ decreased PI(4,5)P2 levels. Carbachol was applied at the end of the experiment ͵͹ʹ to activate M1 receptors, which induced a further decrease of fluorescence in ͵͹͵ most DRG neurons (Fig. 4 A-F). ͵͹Ͷ

͵͹ͷ We also tested receptor-induced depletion of PI(4,5)P2 in HEK293 cells

͵͹͸ transduced with the green BacMam PI(4,5)P2 sensor and BacMam M1 receptor. ͵͹͹ After 24-hour culture, HEK cells showed prominent plasma membrane labeling. ͵͹ͺ Administration of carbachol in 2 mM Ca2+ NCF solution induced robust decrease ͵͹ͻ of fluorescence levels in the plasma membrane, no changes were detected in ͵ͺͲ “bleaching” control experiments (Fig. 4 G-L). ͵ͺͳ

͵ͺʹ To test if PI(4,5)P2 levels decreased in TRPM8 positive cells, we transduced

͵ͺ͵ DRG neurons isolated from TRPM8-GFP mice with a red BacMam PI(4,5)P2 ͵ͺͶ sensor (Fig. 5). We found that application of the inflammatory cocktail decreased

͵ͺͷ PI(4,5)P2 levels in both GFP positive (Fig. 5A) and GFP negative neurons (Fig. ͵ͺ͸ 5B). While all 9 GFP positive neurons we tested responded to the cocktail, only 8 ͵ͺ͹ out of 12 GFP negative neurons showed a clear response. The decrease in

͵ͺͺ PI(4,5)P2 was significantly larger in GFP positive neurons compared to ͵ͺͻ responding GFP negative neurons (Fig. 5D,F). Overall we conclude that

͵ͻͲ activation of Gαq-coupled receptors decreases PI(4,5)P2 levels in the plasma ͵ͻͳ membrane of TRPM8 positive DRG neurons. ͵ͻʹ

͵ͻ͵ PI(4,5)P2 alleviates inhibition of TRPM8 activity by heterologously

͵ͻͶ expressed Gαq-coupled receptors ͵ͻͷ

ͳ͵

͵ͻ͸ We showed that PI(4,5)P2 alleviated inhibition of TRPM8 activity by Gαq-coupled ͵ͻ͹ receptors in DRG neurons. Next we tested if this was reproducible in HEK293 ͵ͻͺ cells co-expressing TRPM8 and M1 receptors, which were reported to inhibit ͵ͻͻ TRPM8 (Li and Zhang, 2013). We performed whole-cell patch clamp experiments ͶͲͲ with, or without supplementing the patch pipette solution with the water-soluble

ͶͲͳ diC8 PI(4,5)P2 (Fig. 6 A, B, D). First we used a protocol of consecutive 20-second ͶͲʹ applications of 200 μM menthol in Ca2+-free NCF solution. Carbachol was ͶͲ͵ applied 80 seconds before the fourth application of menthol. Current amplitudes ͶͲͶ induced by menthol were partially reduced by carbachol to 58% (at +60 mV) and ͶͲͷ 60% (at -60 mV) of currents induced by menthol only. When the patch pipette

ͶͲ͸ contained diC8 PI(4,5)P2, there was no decrease in current amplitudes after ͶͲ͹ carbachol application (Fig. 6B,D). We also tested whether protein kinase C ͶͲͺ (PKC) played a role in this inhibition (Fig. 6C,D). Previous studies reported ͶͲͻ contradictory results on the effects of PKC inhibitors on TRPM8 inhibition

ͶͳͲ induced by Gαq-coupled receptors (Premkumar et al., 2005; Zhang et al., 2012). Ͷͳͳ We co-applied the PKC inhibitor BIM IV (1 μM) with 100 μM carbachol for 80 Ͷͳʹ seconds before the fourth application of menthol. We found no significant effect Ͷͳ͵ of BIM IV on carbachol-induced inhibition of TRPM8 (Fig. 6C,D), similar to the ͶͳͶ findings of (Zhang et al., 2012). As a positive control for BIM, we show that it Ͷͳͷ inhibited sensitization of capsaicin-induced TRPV1 currents in HEK293 cells by Ͷͳ͸ stimulating endogenous purinergic receptors with 100 μM ATP (Fig. 6E-H). Ͷͳ͹ Ͷͳͺ Next we performed similar experiments in NCF solution containing 2 mM Ca2+ Ͷͳͻ (Fig. 6 I-K). Upon application of carbachol, TRPM8 currents induced by menthol ͶʹͲ were inhibited by 82% at +60 mV and 71% at -60 mV. The effect was partially

Ͷʹͳ alleviated by including diC8 PI(4,5)P2 in patch pipette solution; current inhibition Ͷʹʹ decreased to 63% (at +60 mV) and 54% (at -60 mV) under these conditions (Fig. Ͷʹ͵ 6 K). ͶʹͶ Ͷʹͷ Next we examined whether the enhanced inhibition in the presence of

2+ Ͷʹ͸ extracellular Ca was because of different levels of PI(4,5)P2 depletion. We

ͳͶ

Ͷʹ͹ transfected HEK293 cells with the YFP-tagged wild-type Tubby and M1 receptor, Ͷʹͺ and monitored translocation with confocal microscopy. The YFP-tagged wild type

Ͷʹͻ Tubby has higher affinity for PI(4,5)P2 than the YFP-tagged R322H Tubby (Quinn Ͷ͵Ͳ et al., 2008), therefore it may be more sensitive to changes at high levels of

Ͷ͵ͳ PI(4,5)P2 depletion. HEK293 cells transfected with YFP-tagged WT tubby Ͷ͵ʹ showed prominent plasma membrane labeling and application of 100 μM Ͷ͵͵ carbachol both in 2 mM Ca2+ NCF solution and in Ca2+-free NCF solution induced Ͷ͵Ͷ translocation of fluorescent sensors from the plasma membrane to the cytoplasm Ͷ͵ͷ (Fig. 7 A-E). Fluorescence levels in 2 mM Ca2+ NCF solution decreased more in Ͷ͵͸ plasma membrane and increased more in cytosol than in Ca2+-free NCF solution Ͷ͵͹ (Fig. 7C) and this difference was statistically significant (Fig. 7E). This indicates

Ͷ͵ͺ that M1 receptor activation induced a larger decrease in PI(4,5)P2 levels in the Ͷ͵ͻ presence of extracellular Ca2+, compared to that in the absence of extracellular ͶͶͲ Ca2+. This correlates well with the stronger inhibition of TRPM8 in the presence ͶͶͳ of extracellular Ca2+. We also used the lower affinity YFP-tagged R322H Tubby ͶͶʹ in 2 mM Ca2+ NCF solution and in Ca2+ -free NCF solution and found that there ͶͶ͵ was no significant difference between the carbachol-induced fluorescence

ͶͶͶ changes (Fig. 7F-H). This is likely due to the large decrease of PI(4,5)P2 levels in ͶͶͷ both conditions, differentiating between which is outside the dynamic range of ͶͶ͸ this probe. ͶͶ͹

ͶͶͺ Activation of Gαq coupled receptors does not decrease cell surface ͶͶͻ expression of TRPM8 ͶͷͲ

Ͷͷͳ To test if activation of Gαq coupled receptors decreases surface expression of Ͷͷʹ TRPM8 channels, we transfected HEK293 cells with GFP-tagged TRPM8 and Ͷͷ͵ M1 muscarinic receptors and performed Total Internal Reflection Fluorescence ͶͷͶ (TIRF) imaging experiments. TIRF selectively illuminates a narrow region above Ͷͷͷ the glass cover slide, which consists mainly of the plasma membrane at the Ͷͷ͸ bottom of the cells attached to the glass. This technique has been used to study Ͷͷ͹ channel trafficking as well as to study phosphoinositide dynamics (Yao and Qin,

ͳͷ

Ͷͷͺ 2009; Stratiievska et al., 2018). Application of 100 μM carbachol did not induce Ͷͷͻ any detectable decrease of the TIRF signal in HEK cells transfected with GFP- Ͷ͸Ͳ tagged TRPM8 and M1 muscarinic receptors (Fig. 8A,B,D,E). Fluorescence Ͷ͸ͳ showed a slow decrease due to photobleaching, which was indistinguishable Ͷ͸ʹ from that in negative control cells without the application of carbachol (Fig. 8D). Ͷ͸͵ As a positive control to show that carbachol diffuses to the narrow space Ͷ͸Ͷ between the plasma membrane and the cover glass to an extent that is sufficient

Ͷ͸ͷ to activate M1 receptors, we also transfected cells with the tubby-YFP PI(4,5)P2 Ͷ͸͸ sensor and M1 receptors. Application of carbachol in these cells induced a rapid Ͷ͸͹ and robust decrease of TIRF signal (Fig. 8C,D,E), indicating translocation of the

Ͷ͸ͺ PI(4,5)P2 probe from the plasma membrane to the cytoplasm, signifying receptor

Ͷ͸ͻ activation and depletion of PI(4,5)P2. Ͷ͹Ͳ

Ͷ͹ͳ Gαq sensitizes TRPM8 to inhibition by PI(4,5)P2 depletion Ͷ͹ʹ

Ͷ͹͵ Our data so far shows that PI(4,5)P2 depletion plays an important role in TRPM8

Ͷ͹Ͷ inhibition. To test the relationship between PI(4,5)P2 and Gαq regulation of

Ͷ͹ͷ TRPM8, we used a chimera between Gαq and Gαi that does not couple to PLC,

Ͷ͹͸ but inhibits TRPM8 activity (Zhang et al., 2012). The 3Gqiq chimera and its

Ͷ͹͹ constitutively active variant 3GqiqQ209L were shown not to decrease PI(4,5)P2 Ͷ͹ͺ levels when expressed in HEK cells using the same tubby-YFP sensor we used Ͷ͹ͻ for our experiments (Zhang et al., 2012). First we performed Ca2+ imaging ͶͺͲ experiments, and compared responses to menthol application in HEK293 cells

Ͷͺͳ co-transfected with TRPM8 and either 3Gqiq or 3GqiqQ209L (Fig. 9 A, B). Ͷͺʹ Menthol-induced Ca2+ signals were significantly smaller in cells expressing

Ͷͺ͵ 3GqiqQ209L than in cells expressing wild 3Gqiq or TRPM8 only (Fig. 9 A-D). This ͶͺͶ is consistent with the whole-cell patch-clamp results of Zhang et al. 2012, who

Ͷͺͷ found that 3GqiqQ209L inhibited TRPM8 activity and 3Gqiq did not. Ͷͺ͸ Ͷͺ͹ Next we tested weather the decreased menthol-induced Ca2+ signals in cells

Ͷͺͺ expressing the 3GqiqQ209L construct was due to decreased surface expression

ͳ͸

Ͷͺͻ of TRPM8. We co-expressed 3GqiqQ209L with the GFP-tagged TRPM8 in ͶͻͲ HEK293 cells, measured plasma membrane fluorescence using TIRF Ͷͻͳ microscopy and compared it to cells expressing GFP-TRPM8 alone. Figure 9E,F

Ͷͻʹ shows that in cells co-expressing the 3GqiqQ209L the TIRF signal of GFP- Ͷͻ͵ TRPM8 did not decrease, but showed a slight, statistically non-significant ͶͻͶ increase. These data indicate that the decrease in menthol induced Ca2+ signal

Ͷͻͷ in cells expressing 3GqiqQ209L was not caused by decreased surface expression Ͷͻ͸ of TRPM8. Ͷͻ͹

Ͷͻͺ Next we tested if the presence of Gαq has any effect on the level of inhibition

Ͷͻͻ evoked by decreasing PI(4,5)P2 levels. To bypass the PLC pathway, we used a

ͷͲͲ voltage-sensitive 5’ phosphatase (Ci-VSP) to decrease PI(4,5)P2 levels by ͷͲͳ converting this lipid to PI(4)P (Iwasaki et al., 2008). Whole-cell patch-clamp ͷͲʹ experiments were performed in HEK293 cells co-transfected with TRPM8 and Ci- 2+ ͷͲ͵ VSP, with or without 3GqiqQ209L in Ca -free NCF solution (Fig. 10 A-D). ͷͲͶ Consistent with the Ca2+ imaging experiments, menthol-induced currents before

ͷͲͷ the activation of ciVSP were smaller in cells expressing the 3GqiqQ209L (Fig. ͷͲ͸ 10C). Ci-VSP was activated by depolarizing pulses to 100 mV for 0.1 s, 0.2 s, 0.3 ͷͲ͹ s, and 1 s; menthol-induced currents were gradually reduced by 15%, 34%, 49%,

ͷͲͺ and 68% respectively in cells without 3GqiqQ209L. In cells expressing

ͷͲͻ 3GqiqQ209L, the same depolarizing pulses induced higher levels of inhibition of ͷͳͲ menthol-evoked currents, 68%, 80%, 72%, and 85% respectively (Fig. 10 D).

ͷͳͳ These results indicate that TRPM8 is inhibited more by PI(4,5)P2 depletion in the

ͷͳʹ presence of Gαq. ͷͳ͵

ͷͳͶ We conclude that PI(4,5)P2 alleviates inhibition of TRPM8 activity by Gαq-coupled

ͷͳͷ receptors in DRG neurons. Gαq not only directly inhibits TRPM8 as proposed by ͷͳ͸ Zhang et al 2012, but also sensitizes the channel to inhibition by decreasing

ͷͳ͹ PI(4,5)P2 levels, therefore the two pathways converge to inhibit TRPM8 activity ͷͳͺ upon GPCR activation. ͷͳͻ

ͳ͹

ͷʹͲ DISCUSSION ͷʹͳ ͷʹʹ Converging evidence demonstrate that TRPM8 channels require the plasma

ͷʹ͵ membrane phospholipid PI(4,5)P2 for activity. TRPM8 currents show a ͷʹͶ characteristic decrease (rundown) in excised inside out patches, which can be

ͷʹͷ restored by applying PI(4,5)P2 to the cytoplasmic surface of the membrane patch

ͷʹ͸ (Liu and Qin, 2005; Rohacs et al., 2005). Increasing endogenous PI(4,5)P2 levels ͷʹ͹ with MgATP also restored TRPM8 activity in excised patches after current

ͷʹͺ rundown (Yudin et al., 2011). PI(4,5)P2 is also required for the cold- or menthol- ͷʹͻ induced activity of the purified TRPM8 protein in planar lipid bilayers, indicating a ͷ͵Ͳ direct effect on the channel (Zakharian et al., 2009; Zakharian et al., 2010).

ͷ͵ͳ PI(4,5)P2 was also shown to modulate the cold threshold of the channel (Fujita et

ͷ͵ʹ al., 2013). In accordance with dependence of the activity of TRPM8 on PI(4,5)P2,

ͷ͵͵ it was shown that selectively decreasing PI(4,5)P2 levels in intact cells is ͷ͵Ͷ sufficient to inhibit channel activity either by using specific chemically inducible ͷ͵ͷ phosphoinositide phosphatases (Varnai et al., 2006; Daniels et al., 2009), or ͷ͵͸ voltage sensitive phosphatases such as ciVSP (Yudin et al., 2011). ͷ͵͹

ͷ͵ͺ Stimulating cell surface receptors that couple to Gαq and activate PLC inhibits ͷ͵ͻ TRPM8 activity, but the mechanism is not fully elucidated (Liu and Qin, 2005; ͷͶͲ Premkumar et al., 2005; Zhang et al., 2012; Li and Zhang, 2013; Than et al., ͷͶͳ 2013). Based on the following arguments it was proposed that this inhibition

ͷͶʹ proceeds via direct binding of Gαq to the channel (Zhang et al., 2012). Application

ͷͶ͵ of Gαq and GTPγS inhibited TRPM8 activity in excised inside out patches in the

ͷͶͶ presence of PI(4,5)P2. Direct binding of Gαq to TRPM8 was demonstrated by ͷͶͷ immunoprecipitation. Co-expression of a constitutively active mutant (Q209L)

ͷͶ͸ chimera between Gαi and Gαq (3Gqiq) that does not activate PLC inhibited TRPM8 ͷͶ͹ activity (Zhang et al., 2012). ͷͶͺ

ͷͶͻ It was also proposed that decreasing PI(4,5)P2 levels does not contribute to

ͷͷͲ TRPM8 inhibition by Gαq-coupled receptors (Zhang et al., 2012) based on the

ͳͺ

ͷͷͳ following data: Histamine inhibited TRPM8 in the presence of the PLC inhibitor ͷͷʹ U73122. The inhibition however was less than without the drug, indicating ͷͷ͵ potential involvement of PLC activation. Two mutants in the putative TRP domain

ͷͷͶ were also tested, and the authors argued that since these “PI(4,5)P2 insensitive”

ͷͷͷ TRPM8 mutants were inhibited by GPCR activation, PI(4,5)P2 depletion is not

ͷͷ͸ involved. These mutants however are more sensitive to PI(4,5)P2 depletion, due ͷͷ͹ to their decreased apparent affinity for the lipid (Rohacs et al., 2005), and indeed ͷͷͺ the K995Q mutant was somewhat more inhibited by histamine than wild-type ͷͷͻ TRPM8, and the extent of inhibition for the R1008Q is difficult to evaluate due to ͷ͸Ͳ the very small current amplitudes (Zhang et al., 2012). Finally, the constitutively

ͷ͸ͳ active Q209L mutant of the 3Gqiq chimera that does not activate PLC, and does

ͷ͸ʹ not decrease PI(4,5)P2 levels, inhibited TRPM8 activity when co-expressed in ͷ͸͵ HEK cells. The inhibition however was substantially smaller than that induced by

ͷ͸Ͷ the constitutively active Gαq indicating PLC dependent mechanisms may also be ͷ͸ͷ involved (Zhang et al., 2012). In conclusion, most of the data discussed here is

ͷ͸͸ compatible with PI(4,5)P2 depletion contributing to TRPM8 inhibition. ͷ͸͹

ͷ͸ͺ To test the involvement of PI(4,5)P2 depletion, we supplemented the whole cell ͷ͸ͻ patch pipette with the lipid, and observed reduced inhibition of TRPM8 activity by

ͷ͹Ͳ Gαq-coupled receptor activation both in DRG neurons, and in an expression ͷ͹ͳ system. In DRG neurons, we used an inflammatory cocktail, similar to that ͷ͹ʹ described by (Zhang et al., 2012) for two reasons. First, DRG neurons are highly ͷ͹͵ heterogenous, and we could not find a single agonist that reliably induced a Ca2+ ͷ͹Ͷ signal (indicating PLC activation) in all TRPM8 positive neurons. Second, ͷ͹ͷ inflammation is mediated not by a single pro-inflammatory agonist, but rather a ͷ͹͸ combination of them, which is sometime referred to as an “inflammatory soup”. ͷ͹͹

ͷ͹ͺ While activation of PLC induces PI(4,5)P2 hydrolysis, it is debated to what extent

ͷ͹ͻ PI(4,5)P2 is decreased in native tissues upon activation of endogenous receptors ͷͺͲ (Nasuhoglu et al., 2002), and it is clearly cell type and agonist specific (van der

ͷͺͳ Wal et al., 2001). Liu et al., using a fluorescent PI(4,5)P2 sensor, the YFP-tagged

ͳͻ

ͷͺʹ PI(4,5)P2 binding domain of the tubby protein (Quinn et al., 2008), showed that

ͷͺ͵ bradykinin activated PLC in rat DRG neurons, but it did not decrease in PI(4,5)P2 ͷͺͶ levels (Liu et al., 2010). Our earlier data using the same sensor show that

ͷͺͷ bradykinin induced a small decrease in PI(4,5)P2 levels in mouse DRG neurons ͷͺ͸ (Lukacs et al., 2013b). Here we demonstrate that application of the same

ͷͺ͹ inflammatory Gαq-activating cocktail we used in our patch clamp measurements

ͷͺͺ decreased PI(4,5)P2 levels in most DRG neurons, using the tubby-based

ͷͺͻ PI(4,5)P2 sensor. Using a different, red PI(4,5)P2 sensor we also show that the

ͷͻͲ PI(4,5)P2 decrease was larger in TRPM8 positive neurons than in cells not ͷͻͳ expressing TRPM8 (Fig. 5D,F). ͷͻʹ

ͷͻ͵ We found that co-expressing the PLC deficient constitutively active 3GqiqQ209L ͷͻͶ inhibited menthol-induced Ca2+ signals in HEK cells expressing TRPM8, ͷͻͷ confirming the results of (Zhang et al., 2012). This decrease cannot be ͷͻ͸ accounted for by a decreased number of channels, as surface expression of

ͷͻ͹ TRPM8 did not decrease when 3GqiqQ209L was co-expressed. This is consistent ͷͻͺ with the findings of Zhang et al 2012, who showed that co-expressing

ͷͻͻ 3GqiqQ209L did not change the expression level of TRPM8 detected by Western

͸ͲͲ blot. We also found that the presence of 3GqiqQ209L markedly increased the

͸Ͳͳ sensitivity of inhibition of TRPM8 by decreasing PI(4,5)P2 levels with the voltage

͸Ͳʹ sensitive phosphoinositide 5-phosphatase ciVSP. Thus Gαq binding allows even

͸Ͳ͵ small decreases in PI(4,5)P2 concentrations, such as those occurring during ͸ͲͶ physiological receptor activation, to inhibit channel activity. ͸Ͳͷ

͸Ͳ͸ The increased sensitivity to PI(4,5)P2 depletion indicates that Gαq binding to the

͸Ͳ͹ channel decreases the apparent affinity of the channel for PI(4,5)P2 (Rohacs,

͸Ͳͺ 2007). It is estimated that basal plasma membrane PI(4,5)P2 levels correspond to

͸Ͳͻ the effect 40 μM diC8 PI(4,5)P2 applied to excised patches (Collins and Gordon,

͸ͳͲ 2013). This concentration of PI(4,5)P2 induced ~80% of the maximal current in ͸ͳͳ excised patches at 17 oC in the presence of 500 μM menthol at physiological ͸ͳʹ negative membrane potentials (Rohacs et al., 2005). This indicates that resting

ʹͲ

ͳ͵ plasma membrane PI(4,5)P2 concentrations do not fully saturate TRPM8,

ͳͶ therefore the decreased apparent affinity to PI(4,5)P2 upon Gαq binding is

ͳͷ expected to reduce TRPM8 current levels even without any decrease in PI(4,5)P2 ͳ͸ concentrations. This is consistent with reduced menthol induced Ca2+ signals and

ͳ͹ current amplitudes when GqiqQ209L was co-expressed. We show that TRPM8 ͳͺ surface expression did not decrease upon receptor activation, and when the

ͳͻ GqiqQ209L was co-expressed, therefore the decreased apparent affinity for

ʹͲ PI(4,5)P2 is likely to be responsible also for the reduced current amplitudes upon

ʹͳ direct Gαq binding. ʹʹ

ʹ͵ Overall we conclude that upon receptor activation, binding of Gαq not only directly

ʹͶ inhibits TRPM8, but also sensitizes it to PI(4,5)P2 depletion, and thus the two ʹͷ pathways converge and synergize in reducing channel activity. ʹ͸ ʹ͹ ʹͺ REFERENCES ʹͻ ͵Ͳ Almaraz, L., Manenschijn, J.A., de la Pena, E., and Viana, F. (2014). TRPM8. ͵ͳ Handb Exp Pharmacol 222, 547-579. ͵ʹ ͵͵ Bautista, D.M., Siemens, J., Glazer, J.M., Tsuruda, P.R., Basbaum, A.I., Stucky, ͵Ͷ C.L., Jordt, S.E., and Julius, D. (2007). The menthol receptor TRPM8 is the ͵ͷ principal detector of environmental cold. Nature 448, 204-208. ͵͸ ͵͹ Bhave, G., Hu, H.J., Glauner, K.S., Zhu, W., Wang, H., Brasier, D.J., Oxford, ͵ͺ G.S., and Gereau, R.W.t. (2003). Protein kinase C phosphorylation sensitizes but ͵ͻ does not activate the capsaicin receptor transient receptor potential vanilloid 1 ͶͲ (TRPV1). Proc Natl Acad Sci U S A 100, 12480-12485. Ͷͳ Ͷʹ Caterina, M.J., Leffler, A., Malmberg, A.B., Martin, W.J., Trafton, J., Petersen- Ͷ͵ Zeitz, K.R., Koltzenburg, M., Basbaum, A.I., and Julius, D. (2000). Impaired ͶͶ nociception and pain sensation in mice lacking the capsaicin receptor. Science Ͷͷ 288, 306-313. Ͷ͸ Ͷ͹ Cesare, P., and McNaughton, P. (1996). A novel heat-activated current in Ͷͺ nociceptive neurons and its sensitization by bradykinin. Proc Natl Acad Sci U S A Ͷͻ 93, 15435-15439. ͷͲ

ʹͳ

ͷͳ Colburn, R.W., Lubin, M.L., Stone, D.J., Jr., Wang, Y., Lawrence, D., D'Andrea, ͷʹ M.R., Brandt, M.R., Liu, Y., Flores, C.M., and Qin, N. (2007). Attenuated cold ͷ͵ sensitivity in TRPM8 null mice. Neuron 54, 379-386. ͷͶ ͷͷ Collins, M.D., and Gordon, S.E. (2013). Short-chain phosphoinositide partitioning ͷ͸ into plasma membrane models. Biophys J 105, 2485-2494. ͷ͹ ͷͺ Daniels, R.L., Takashima, Y., and McKemy, D.D. (2009). Activity of the neuronal ͷͻ cold sensor TRPM8 is regulated by phospholipase C via the phospholipid ͸Ͳ phosphoinositol 4,5-bisphosphate. J Biol Chem 284, 1570-1582. ͸ͳ ͸ʹ Dhaka, A., Murray, A.N., Mathur, J., Earley, T.J., Petrus, M.J., and Patapoutian, ͸͵ A. (2007). TRPM8 is required for cold sensation in mice. Neuron 54, 371-378. ͸Ͷ ͸ͷ Fujita, F., Uchida, K., Takaishi, M., Sokabe, T., and Tominaga, M. (2013). ͸͸ Ambient temperature affects the temperature threshold for TRPM8 activation ͸͹ through interaction of phosphatidylinositol 4,5-bisphosphate. J Neurosci 33, ͸ͺ 6154-6159. ͸ͻ ͹Ͳ Iwasaki, H., Murata, Y., Kim, Y., Hossain, M.I., Worby, C.A., Dixon, J.E., ͹ͳ McCormack, T., Sasaki, T., and Okamura, Y. (2008). A voltage-sensing ͹ʹ phosphatase, Ci-VSP, which shares sequence identity with PTEN, ͹͵ dephosphorylates phosphatidylinositol 4,5-bisphosphate. Proc Natl Acad Sci U S ͹Ͷ A 105, 7970-7975. ͹ͷ ͹͸ Knowlton, W.M., Bifolck-Fisher, A., Bautista, D.M., and McKemy, D.D. (2010). ͹͹ TRPM8, but not TRPA1, is required for neural and behavioral responses to acute ͹ͺ noxious cold temperatures and cold-mimetics in vivo. Pain 150, 340-350. ͹ͻ ͺͲ Li, L., and Zhang, X. (2013). Differential inhibition of the TRPM8 ion channel by ͺͳ Galphaq and Galpha 11. Channels (Austin) 7, 115-118. ͺʹ ͺ͵ Liu, B., Linley, J.E., Du, X., Zhang, X., Ooi, L., Zhang, H., and Gamper, N. ͺͶ (2010). The acute nociceptive signals induced by bradykinin in rat sensory ͺͷ neurons are mediated by inhibition of M-type K+ channels and activation of Ca2+- ͺ͸ activated Cl- channels. J Clin Invest 120, 1240-1252. ͺ͹ ͺͺ Liu, B., and Qin, F. (2005). Functional control of cold- and menthol-sensitive ͺͻ TRPM8 ion channels by phosphatidylinositol 4,5-bisphosphate. J Neurosci 25, ͻͲ 1674-1681. ͻͳ ͻʹ Lukacs, V., Rives, J.M., Sun, X., Zakharian, E., and Rohacs, T. (2013a). ͻ͵ Promiscuous activation of transient receptor potential vanilloid 1 channels by ͻͶ negatively charged intracellular lipids, the key role of endogenous ͻͷ phosphoinositides in maintaining channel activity. J Biol Chem 288, 35003- ͻ͸ 35013.

ʹʹ

ͻ͹ ͻͺ Lukacs, V., Yudin, Y., Hammond, G.R., Sharma, E., Fukami, K., and Rohacs, T. ͻͻ (2013b). Distinctive changes in plasma membrane phosphoinositides underlie ͲͲ differential regulation of TRPV1 in nociceptive neurons. Journal of Neuroscience Ͳͳ 33, 11451-11463. Ͳʹ Ͳ͵ Nasuhoglu, C., Feng, S., Mao, Y., Shammat, I., Yamamato, M., Earnest, S., ͲͶ Lemmon, M., and Hilgemann, D.W. (2002). Modulation of cardiac PIP2 by Ͳͷ cardioactive hormones and other physiologically relevant interventions. Am J Ͳ͸ Physiol Cell Physiol 283, C223-234. Ͳ͹ Ͳͺ Numazaki, M., Tominaga, T., Toyooka, H., and Tominaga, M. (2002). Direct Ͳͻ phosphorylation of capsaicin receptor VR1 by protein kinase Cepsilon and ͳͲ identification of two target serine residues. J Biol Chem 277, 13375-13378. ͳͳ ͳʹ Premkumar, L.S., Raisinghani, M., Pingle, S.C., Long, C., and Pimentel, F. ͳ͵ (2005). Downregulation of transient receptor potential melastatin 8 by protein ͳͶ kinase C-mediated dephosphorylation. J Neurosci 25, 11322-11329. ͳͷ ͳ͸ Quinn, K.V., Behe, P., and Tinker, A. (2008). Monitoring changes in membrane ͳ͹ phosphatidylinositol 4,5-bisphosphate in living cells using a domain from the ͳͺ transcription factor tubby. J Physiol 586, 2855-2871. ͳͻ ʹͲ Rohacs, T. (2007). Regulation of TRP channels by PIP2. Pflugers Arch 453, 753- ʹͳ 762. ʹʹ ʹ͵ Rohacs, T. (2014). Phosphoinositide regulation of TRP channels. Handb Exp ʹͶ Pharmacol 233, 1143-1176. ʹͷ ʹ͸ Rohacs, T. (2016). Phosphoinositide signaling in somatosensory neurons. Adv ʹ͹ Biol Regul 61, 2-16. ʹͺ ʹͻ Rohacs, T., Lopes, C.M., Michailidis, I., and Logothetis, D.E. (2005). PI(4,5)P2 ͵Ͳ regulates the activation and desensitization of TRPM8 channels through the TRP ͵ͳ domain. Nat Neurosci 8, 626-634. ͵ʹ ͵͵ Stratiievska, A., Nelson, S., Senning, E.N., Lautz, J.D., Smith, S.E., and Gordon, ͵Ͷ S.E. (2018). Reciprocal regulation among TRPV1 channels and phosphoinositide ͵ͷ 3-kinase in response to nerve growth factor. Elife 7. ͵͸ ͵͹ Takashima, Y., Daniels, R.L., Knowlton, W., Teng, J., Liman, E.R., and McKemy, ͵ͺ D.D. (2007). Diversity in the neural circuitry of cold sensing revealed by genetic ͵ͻ axonal labeling of transient receptor potential melastatin 8 neurons. J Neurosci ͶͲ 27, 14147-14157. Ͷͳ

ʹ͵

Ͷʹ Tewson, P.H., Quinn, A.M., and Hughes, T.E. (2013). A multiplexed fluorescent Ͷ͵ assay for independent second-messenger systems: decoding GPCR activation in ͶͶ living cells. J Biomol Screen 18, 797-806. Ͷͷ Ͷ͸ Thakur, M., Crow, M., Richards, N., Davey, G.I., Levine, E., Kelleher, J.H., Agley, Ͷ͹ C.C., Denk, F., Harridge, S.D., and McMahon, S.B. (2014). Defining the Ͷͺ nociceptor transcriptome. Front Mol Neurosci 7, 87. Ͷͻ ͷͲ Than, J.Y., Li, L., Hasan, R., and Zhang, X. (2013). Excitation and modulation of ͷͳ TRPA1, TRPV1, and TRPM8 channel-expressing sensory neurons by the ͷʹ pruritogen chloroquine. J Biol Chem 288, 12818-12827. ͷ͵ ͷͶ Tominaga, M., Wada, M., and Masu, M. (2001). Potentiation of capsaicin ͷͷ receptor activity by metabotropic ATP receptors as a possible mechanism for ͷ͸ ATP-evoked pain and hyperalgesia. Proc Natl Acad Sci U S A 98, 6951-6956. ͷ͹ ͷͺ van der Wal, J., Habets, R., Varnai, P., Balla, T., and Jalink, K. (2001). ͷͻ Monitoring agonist-induced phospholipase C activation in live cells by ͸Ͳ fluorescence resonance energy transfer. J Biol Chem 276, 15337-15344. ͸ͳ ͸ʹ Varnai, P., Thyagarajan, B., Rohacs, T., and Balla, T. (2006). Rapidly inducible ͸͵ changes in phosphatidylinositol 4,5-bisphosphate levels influence multiple ͸Ͷ regulatory functions of the lipid in intact living cells. J Cell Biol 175, 377-382. ͸ͷ ͸͸ Velisetty, P., Borbiro, I., Kasimova, M.A., Liu, L., Badheka, D., Carnevale, V., and ͸͹ Rohacs, T. (2016). A molecular determinant of phosphoinositide affinity in ͸ͺ mammalian TRPV channels. Sci Rep 6, 27652. ͸ͻ ͹Ͳ Xing, H., Chen, M., Ling, J., Tan, W., and Gu, J.G. (2007). TRPM8 mechanism of ͹ͳ cold allodynia after chronic nerve injury. J Neurosci 27, 13680-13690. ͹ʹ ͹͵ Yao, J., and Qin, F. (2009). Interaction with phosphoinositides confers adaptation ͹Ͷ onto the TRPV1 pain receptor. PLoS Biol 7, e46. ͹ͷ ͹͸ Yudin, Y., Lukacs, V., Cao, C., and Rohacs, T. (2011). Decrease in ͹͹ phosphatidylinositol 4,5-bisphosphate levels mediates desensitization of the cold ͹ͺ sensor TRPM8 channels. J Physiol 589, 6007-6027. ͹ͻ ͺͲ Yudin, Y., Lutz, B., Tao, Y.X., and Rohacs, T. (2016). Phospholipase C δ4 ͺͳ regulates cold sensitivity in mice. J Physiol 594, 3609-3628. ͺʹ ͺ͵ Zakharian, E., Cao, C., and Rohacs, T. (2010). Gating of transient receptor ͺͶ potential melastatin 8 (TRPM8) channels activated by cold and chemical agonists ͺͷ in planar lipid bilayers. J Neurosci 30, 12526-12534. ͺ͸

ʹͶ

͹ͺ͹ Zakharian, E., Thyagarajan, B., French, R.J., Pavlov, E., and Rohacs, T. (2009). ͹ͺͺ Inorganic polyphosphate modulates TRPM8 channels. PLoS One 4, e5404. ͹ͺͻ ͹ͻͲ Zhang, X., Li, L., and McNaughton, P.A. (2008). Proinflammatory mediators ͹ͻͳ modulate the heat-activated ion channel TRPV1 via the scaffolding protein ͹ͻʹ AKAP79/150. Neuron 59, 450-461. ͹ͻ͵ ͹ͻͶ Zhang, X., Mak, S., Li, L., Parra, A., Denlinger, B., Belmonte, C., and ͹ͻͷ McNaughton, P.A. (2012). Direct inhibition of the cold-activated TRPM8 ion ͹ͻ͸ channel by Gαq. Nat Cell Biol 14, 851-858. ͹ͻ͹ ͹ͻͺ ͹ͻͻ FIGURE LEGENDS ͺͲͲ 2+ ͺͲͳ Figure 1: Ca responses of DRG neurons to agonists of Gαq-coupled receptors ͺͲʹ and TRPM8 channels. A, Venn diagram showing the number of DRG neurons ͺͲ͵ responding to 500 μM menthol, 2 μM bradykinin (BK), 1 μM PGE2, and 500 nM ͺͲͶ Endothelin (ET). Experiments were performed on neurons from 3 independent ͺͲͷ preparations from 3 mice both for DRG and TG neurons. B, Venn diagram ͺͲ͸ showing the number of Trigeminal neurons (TG) responding to the same stimuli. ͺͲ͹ C, Representative traces of calcium imaging measurements in TRPM8-GFP ͺͲͺ mouse DRG neurons. TRPM8 channels were activated by 10 μM WS12. A ͺͲͻ cocktail containing 100 μM ADP, 100 μM UTP, 500 nM bradykinin, 100 μM ͺͳͲ histamine, 100 μM serotonin, 10 μM PGE2, 10 μM PGI2 was applied to activate

ͺͳͳ various Gαq-coupled receptors. Measurements were conducted in 2 mM Ca2+ ͺͳʹ NCF solution. D, Summary data of calcium imaging measurements on n=12 ͺͳ͵ neurons that responded to both the cocktail and WS12 displayed as mean ± ͺͳͶ SEM. E, GFP positive neurons constitute 11% (23 neurons) of all DRG neurons ͺͳͷ (214 neurons, 2 independent preparations from 2 mice). F, Within these GFP ͺͳ͸ positive neurons, twelve neurons (57%) respond to both WS12 and cocktail, four ͺͳ͹ neurons (12%) respond to neither WS12 nor cocktail, six neurons (26%) only ͺͳͺ respond to cocktail and one neuron ͺͳͻ only respond to WS12. ͺʹͲ

ʹͷ

ͺʹͳ Figure 2. Intracellular dialysis of PI(4,5)P2 alleviates inhibition of TRPM8 activity

ͺʹʹ by Gαq-coupled receptors in DRG neurons. A, B, Representative whole-cell ͺʹ͵ voltage-clamp traces of inward currents recorded at -60 mV from DRG neurons ͺʹͶ isolated from TRPM-GFP mice. Measurements were conducted in Ca2+-free

ͺʹͷ NCF solution. Control-no lipid supplement (A), 100 μM diC8 PI(4,5)P2 supplement ͺʹ͸ (B). WS12 (10 μM) was applied to activate TRPM8 channels for 4 times as ͺʹ͹ indicated by the horizontal lines. The inflammatory cocktail containing 100 μM ͺʹͺ ADP, 100 μM UTP, 500 nM bradykinin, 100 μM histamine, 100 μM serotonin, 10 ͺʹͻ μM PGE2, 10 μM PGI2 was applied between the 2nd and 3rd application of WS12, ͺ͵Ͳ as indicated by the horizontal line. C, D, Statistical analysis of n=7 neurons from ͺ͵ͳ 5 different DRG neuron preparations from 5 mice (control group) and n=8

ͺ͵ʹ neurons (diC8 PI(4,5)P2 group) displayed, from 5 different DRG neuron ͺ͵͵ preparations from 5 mice. Bars represent mean ± SEM; statistical significance ͺ͵Ͷ was calculated with two-way analysis of variance for D, F(7, 52)=16.82, p = 2.3 x ͺ͵ͷ 10-11, and with t-test for C. Four peaks of current responses in each group were ͺ͵͸ analyzed. Raw current densities are shown in panel C, current values normalized ͺ͵͹ to the peak of the second current response in each group are displayed in panel ͺ͵ͺ D. ͺ͵ͻ

ͺͶͲ Figure 3. Receptor-induced decrease of PI(4,5)P2 in the plasma membrane of ͺͶͳ DRG neurons. A, Confocal images of wild type mouse DRG neurons transfected ͺͶʹ with YFP-tagged R322H Tubby domain as a reporter of plasma membrane

ͺͶ͵ PI(4,5)P2. Images are representatives of reporter distribution before and after ͺͶͶ application of either inflammatory cocktail or control solution (2 mM Ca2+ NCF). ͺͶͷ Measurements were conducted in 2 mM Ca2+ NCF solution. Neurons were ͺͶ͸ serum-deprived in 2 mM Ca2+ NCF solution for 20 min. B, Graphs correspond to ͺͶ͹ fluorescence intensities plotted along the lines indicated in each image in A ͺͶͺ before (red) and after (black) application of cocktail or control. C, D, Time ͺͶͻ courses of fluorescence density changes for YFP-tagged R322H Tubby in ͺͷͲ response to control solution (C) or inflammatory cocktail (D) in plasma membrane ͺͷͳ and cytosol of neuron cell body. E, Statistical analysis of fluorescence density

ʹ͸

ͺͷʹ right before and 1 min after application of inflammatory cocktail (n=14 neurons ͺͷ͵ from 2 independent DRG preparations from 4 mice) and control solution (n=3) ͺͷͶ displayed. Bars represent mean ± SEM; statistical significance was calculated ͺͷͷ with two-way analysis of varianceF(3, 30)=12.37, p=1.9 x 10-5. Fluorescence ͺͷ͸ density values were normalized to the time point of right before application of ͺͷ͹ inflammatory cocktail or control solution for each group. ͺͷͺ

ͺͷͻ Figure 4. Receptor-induced decrease of PI(4,5)P2 in DRG neurons and HEK ͺ͸Ͳ cells. A, Confocal images of wild type mouse DRG neurons tranducted with

ͺ͸ͳ Green PI(4,5)P2 sensor BacMam and human muscarinic cholinergic receptor 1 ͺ͸ʹ (M1). Images are representatives of reporter distribution before and after ͺ͸͵ application of either cocktail or control solution (2 mM Ca2+ NCF). Measurements ͺ͸Ͷ were conducted in 2 mM Ca2+ NCF solution. B, Graphs correspond to ͺ͸ͷ fluorescence intensities plotted along the lines indicated in each image in A ͺ͸͸ before (red) and after (black) application of cocktail or control. C, Representative

ͺ͸͹ traces of time course changes for Green PI(4,5)P2 sensor BacMam in response ͺ͸ͺ to cocktail followed by 100 μM carbachol in DRG neurons. D, Averaged time

ͺ͸ͻ course of fluorescence density changes for Green PI(4,5)P2 sensor BacMam in ͺ͹Ͳ response to cocktail followed by 100 μM carbachol in DRG neurons. Four GFP ͺ͹ͳ positive neurons showed no changes in red fluorescence in response to cocktail ͺ͹ʹ application, and only one of those four responded to carbachol; these cells were ͺ͹͵ excluded from the summary. E, Statistical analysis of fluorescence density at ͺ͹Ͷ baseline, 40 seconds after application of cocktail, right before application of ͺ͹ͷ carbachol and 40 seconds after application of carbachol in neurons (n=31) from 2 ͺ͹͸ different DRG neuron preparations from 2 mice. Bars represent mean ± SEM; ͺ͹͹ statistical significance was calculated with one-way analysis of variance, F(3, ͺ͹ͺ 124) = 21.07, p=4.5 x 10-11. Fluorescence density values were normalized to

ͺ͹ͻ baseline. F, Time course of fluorescence density changes for Green PI(4,5)P2 ͺͺͲ sensor BacMam in response to two applications of control solution (2 mM Ca2+ ͺͺͳ NCF) in DRG neurons (n=12). G, Confocal images of HEK cells transduced with

ͺͺʹ Green PI(4,5)P2 sensor BacMam as reporters of PI(4,5)P2 and human muscarinic

ʹ͹

ͺͺ͵ receptor 1 (M1) receptor. Images are representatives of reporter distribution ͺͺͶ before and after application of either 100 μM carbachol or control solution (2 mM ͺͺͷ Ca2+ NCF). Measurements were conducted in 2 mM Ca2+ NCF solution. H, ͺͺ͸ Graphs correspond to fluorescence intensities plotted along the lines indicated in ͺͺ͹ each image in G before (red) and after (black) application of 100 μM carbachol or

ͺͺͺ control. I, Representative traces of time course changes for Green PI(4,5)P2 ͺͺͻ sensor BacMam in response to carbachol. J, K, Average time course of

ͺͻͲ fluorescence density changes for Green PI(4,5)P2 sensor BacMam in response ͺͻͳ to carbachol or control (2 mM Ca2+ NCF) in HEK cells. L, Statistical analysis of ͺͻʹ fluorescence density at baseline and after application of carbachol (n=18) or ͺͻ͵ control (n=11). Bars represent mean ± SEM; statistical significance was ͺͻͶ calculated with one-way analysis of variance, F(2,44)=60.03, p=2.7 x 10 -13. ͺͻͷ Fluorescence density values were normalized to baseline ͺͻ͸

ͺͻ͹ Figure 5. Receptor induced decrease of PI(4,5)P2 in TRPM8 positive and ͺͻͺ TRPM8 negative DRG neurons. A,B,C, Confocal images of TRPM8-GFP mouse

ͺͻͻ DRG neurons transduced with Red PI(4,5)P2 sensor BacMam. Images are ͻͲͲ representatives of fluorescence density before and after application of either ͻͲͳ cocktail or control solution (2 mM Ca2+ NCF). Graphs correspond to fluorescence ͻͲʹ intensities plotted along the lines indicated in each image before (red) and after ͻͲ͵ (black) application of cocktail or control. (A, GFP-positive DRG neurons with red

ͻͲͶ PI(4,5)P2 sensor; B, GFP-negative DRG neurons with red PI(4,5)P2 sensor; C, ͻͲͷ bleaching controls of red fluorescence and GFP fluorescence in neurons.) ͻͲ͸ Measurements were conducted in 2 mM Ca2+ NCF solution. D, Averaged time

ͻͲ͹ course of fluorescence density changes for Red PI(4,5)P2 sensor BacMam and ͻͲͺ GFP fluorescence in response to cocktail in DRG neurons. Four GFP-negative ͻͲͻ neurons with red fluorescence were excluded due to no changes of red ͻͳͲ fluorescence density after application of cocktail. E, Averaged time course of

ͻͳͳ fluorescence density changes for Red PI(4,5)P2 sensor BacMam and GFP ͻͳʹ fluorescence in response to control solution in DRG neurons. F, Statistical ͻͳ͵ analysis of red fluorescence density at baseline, and at the peak response after

ʹͺ

ͻͳͶ application of cocktail or control. Data are shown from 3 different neuron ͻͳͷ preparations from 3 mice. Bars represent mean ± SEM; statistical significance ͻͳ͸ was calculated with one-way analysis of variance, F(3,28)=31.02, p=4.8 x 10-9. ͻͳ͹ Fluorescence density values were normalized to baseline. ͻͳͺ

ͻͳͻ Figure 6. Intracellular dialysis of PI(4,5)P2 alleviates inhibition of TRPM8 activity

ͻʹͲ by Gαq-coupled receptors in HEK cells. A, B, C, Representative whole-cell ͻʹͳ voltage-clamp traces of inward currents recorded at -60 mV and + 60 mV in Ca2+- ͻʹʹ free NCF solution for HEK cells transfected with TRPM8 channels and M1 ͻʹ͵ receptors. Control-no lipid supplement (A), whole cell patch pipette

ͻʹͶ supplemented with 100 μM diC8PI(4,5)P2 (B), cells treated with the PKC inhibitor ͻʹͷ BIM IV (1 μM) (C). Menthol (200 μM) was applied to activate TRPM8 channels ͻʹ͸ for 5 times, carbachol (100 μM) was applied to activate M1 receptor. D, Statistical ͻʹ͹ analysis at -60 mV and + 60 mV corresponding to panels A-C. Bars represent ͻʹͺ mean ± SEM; statistical significance was calculated with one-way analysis of ͻʹͻ variance, +60 mV: F(3,36)=20.08, p=8.1 x 10-8; -60 mV: F(3,36) = 13.80, p=3.8 x ͻ͵Ͳ 10 -6. Current values were normalized to the 3rd current response to menthol. E, ͻ͵ͳ F, G, Representative whole-cell voltage-clamp traces of inward currents recorded ͻ͵ʹ at – 60 mV in Ca2+-free NCF solution for HEK cells transfected with TRPV1 ͻ͵͵ channels; the applications of 20 nM capsaicin, 100 μM ATP and 1 μM BIM were ͻ͵Ͷ indicated by the horizontal lines. H, Statistical analysis of control (n=6 cells), ATP ͻ͵ͷ (n=6 cells) and BIM+ATP (n=6 cells) both at -60 mV (corresponding to panels E- ͻ͵͸ G). Bars represent mean ± SEM; statistical significance was calculated with two- ͻ͵͹ way analysis of variance, F(5,30)=13.03, p=8.8 x 10-7. Current values of the 2nd ͻ͵ͺ peak were normalized to the 1st peak induced by capsaicin. I, J, Representative ͻ͵ͻ whole-cell voltage-clamp traces of inward currents recorded at – 60 mV and + 60 ͻͶͲ mV in 2 mM Ca2+ NCF solution for HEK cells transfected with TRPM8 channels ͻͶͳ and M1 receptors; the applications of 200 μM Menthol and 100 μM carbachol

ͻͶʹ were indicated by the horizontal lines; 100 μM diC8 PI(4,5)P2 was added to the ͻͶ͵ whole cell patch pipette solution in J, no lipid was added in I. K, Statistical

ͻͶͶ analysis of n=14 cells (control) and n=13 cells (diC8 PI(4,5)P2) both at -60 mV

ʹͻ

ͻͶͷ and +60 mV (corresponding to panels I, J) displayed. Bars represent mean ± ͻͶ͸ SEM, two-way analysis of variance: +60 mV: F(3,50) = 106.8, p=0 and -60 mV ͻͶ͹ F(3,50)=104.0, p=0. Current values at the time point of 1.5 min after application ͻͶͺ of carbachol were analyzed. Current values were normalized to the peak current ͻͶͻ value under menthol application before carbachol. ͻͷͲ

ͻͷͳ Figure 7. Receptor-induced depletion of PI(4,5)P2 in the plasma membrane of ͻͷʹ HEK cells in the presence or absence of calcium. A, Confocal images of HEK ͻͷ͵ cells transfected with YFP-tagged wild type Tubby domain as reporters of plasma ͻͷͶ membrane phosphoinositides and M1 receptor in 2 mM Ca2+ NCF or Ca2+ -free ͻͷͷ NCF solution. Images are representatives of reporter distribution before and after ͻͷ͸ application of either 100 μM carbachol or control solution (2 mM Ca2+ NCF). B, ͻͷ͹ Graphs correspond to fluorescence intensities plotted along the lines indicated in ͻͷͺ each image in A before (red) and after (black) application of carbachol or control. ͻͷͻ C, D, Time courses of fluorescence density changes for YFP-tagged WT Tubby ͻ͸Ͳ in response to carbachol (C) or control (D) in plasma membrane and cytosol of ͻ͸ͳ HEK cells. E, Statistical analysis of fluorescence density at baseline and at the ͻ͸ʹ end after application of carbachol (n=25 for 2 mM Ca2+ NCF , n=28 for Ca2+ -free ͻ͸͵ NCF) and control (n=8) displayed. Bars represent mean ± SEM; statistical ͻ͸Ͷ significance was calculated with one-way analysis of variance; membrane: ͻ͸ͷ F(3,85)=89.12, p=0; cytosol: F(3,85)=45.20, p=0. Fluorescence density values ͻ͸͸ were normalized to the baseline for each group. F,G, Similar experiments to ͻ͸͹ those shown in C and D, in HEK293 cell transfected with YFP-tagged R322H ͻ͸ͺ Tubby domain and M1 receptors. H, Statistical analysis of fluorescence density ͻ͸ͻ at baseline and at the end after application of carbachol (n=25 for 2 mM Ca2+ ͻ͹Ͳ NCF , n=24 for Ca2+ -free NCF) and control (n=11) displayed. Bars represent ͻ͹ͳ mean ± SEM; one-way analysis of variance; membrane: F(3,80)=57.41, p=0; ͻ͹ʹ cytosol: F (3,80)=12.97, p=5.5 x 10-7. Fluorescence density values were ͻ͹͵ normalized to the baseline for each group. ͻ͹Ͷ

͵Ͳ

ͻ͹ͷ Figure 8: Muscarinic receptor activation does not reduce cell surface TRPM8 ͻ͹͸ levels. TIRF measurements were performed as described in the Methods section ͻ͹͹ in HEK293 cells transfected either with M1 muscarinic receptors and GFP-

ͻ͹ͺ TRPM8 (M8-GFP), or M1 muscarinic receptors and the tubby YFP PI(4,5)P2 ͻ͹ͻ sensor. A-C representative TIRF images before and after the application of ͻͺͲ carbachol (A,C) or no application of carbachol (B). D time course of TIRF ͻͺͳ fluorescence in the three groups, the application of 100 μM carbachol is indicated ͻͺʹ by the arrow; mean ± SEM is plotted, fluorescence was normalized to the point ͻͺ͵ before the application of carbachol. E. Summary of the three groups 50 seconds ͻͺͶ after the application of carbachol, fluorescence was normalized to the point ͻͺͷ before the application of carbachol, one way analysis of variance, F(2,70) = ͻͺ͸ 258.2, p=0. ͻͺ͹

ͻͺͺ Figure 9. A PLC-deficient Gαq chimera inhibits TRPM8 activity, but does not ͻͺͻ decrease cell surface expression A, Calcium imaging measurements (mean ±

ͻͻͲ SEM) of HEK cells transfected with TRPM8, plus 3GqiqQ209L or 3Gqiq in ͻͻͳ response to two applications of 200 μM Menthol. B, Pooled data of A normalized ͻͻʹ to baseline. C, D, statistical summary of A and B, respectively. Bars represent ͻͻ͵ mean ± SEM; statistical significance was calculated with two-way analysis of ͻͻͶ variance, F(8, 30) = 25.55, p=2.2 x 10-11 for C, and F(5, 20)=11.10, p=3.2 x 10-5 ͻͻͷ for D. Fluorescence values of 340/380 ratio were normalized to the baseline for ͻͻ͸ each group. Two peaks of menthol responses were analyzed. TRPM8 group

ͻͻ͹ (n=4 slides, including 103 cells), TRPM8+3GqiqQ209L group (n=5 slides,

ͻͻͺ including 84 cells), TRPM8 + 3Gqiq group (n=4 slides, including 103 cells). E, ͻͻͻ Representative TIRF images from HEK293 cells co-transfected with GFP-

ͳͲͲͲ TRPM8, or GFP-TRPM8 plus 3GqiqQ209L. TIRF images were obtained 24-36 ͳͲͲͳ hours after transfection, as described in the methods section. F, Summary for ͳͲͲʹ data obtained from 5 and 6 coverslips, from 3 different transfections. The ͳͲͲ͵ experimenter blinded to the transfection collected similar number of random ͳͲͲͶ images from each cover slide. The average fluorescence intensity of cells from ͳͲͲͷ one cover slide was taken as one individual data point. The total number of cells

͵ͳ

ͳͲͲ͸ analyzed was 139 for TRPM8 and 148 for TRPM8 plus 3GqiqQ209L. Statistical ͳͲͲ͹ comparison was performed with two-sample t-test. ͳͲͲͺ

ͳͲͲͻ Figure 10. TRPM8 is inhibited more efficiently by PI(4,5)P2 depletion in the

ͳͲͳͲ presence of Gαq. A, B, Representative whole-cell voltage-clamp traces of inward ͳͲͳͳ currents recorded at -60 mV from HEK cells transfected with TRPM8 and the ͳͲͳʹ voltage sensitive phosphatase CiVSP (A), or with ciVSP, TRPM8 and 2+ ͳͲͳ͵ 3GqiqQ209L (B). Measurements were conducted in Ca -free NCF solution. ͳͲͳͶ TRPM8 channels were activated by 500 μM menthol. CiVSP was activated by ͳͲͳͷ +100 mV steps for 0.1, 0.2, 0.3 and 1 second. C, Summary of raw current ͳͲͳ͸ densities before the application of voltage steps (peak) and after voltage steps to ͳͲͳ͹ +100 mV for 0.1, 0.2, 0.3 and 1 second for each group; n=7 for TRPM8 + CiVSP,

ͳͲͳͺ and n=9 for TRPM8 + CiVSP + 3GqiqQ209L, two-way analysis of variance, ͳͲͳͻ F(9,70) = 5.40, p = 1.3 x 10-5, D. Summary of the same data expressed as ͳͲʹͲ current inhibition induced by voltage steps. Bars represent mean ± SEM; ͳͲʹͳ statistical significance was calculated with two-way analysis of variance, F(7,56) ͳͲʹʹ = 13.66, p=3.7 x 10-10. ͳͲʹ͵ ͳͲʹͶ ͳͲʹͷ

͵ʹ

Figure 2 A B Cocktail Cocktail 100 WS12 50 WS12 0 0 -100 -50 -200 -100 -300 -150

current (pA) 3 current (pA) -400 -200 -500 4 -250 1 234 1 2 -600 0 CaϸЀ Control -300 0 CaϸЀ PIP2 0 200 400 600 800 0 200 400 600 800 me (s) me (s) CD 1234 0 1.50 p = 1.8 x 10- 6 1.25 -10 p = 3.0 x 10 - 6 1 -20 0.75 -30 0.50 0.25 -40 p = 0.00012 Current Density (pA/pF) Density Current Normalized Current Density Current Normalized 0 -50 1234 PIP2 (8 cells) Control (7 cells) PIP2 (8 cells) Control (7 cells) Figure 3 AB m m m m DRG cell body ʅ Cocktail ʅ Cocktail ʅ ʅ 80 Before 5 5 5 5 A er 60

Fluorescence Density Fluorescence 0 Before A er Distance (ʅm) Cocktail Cocktail DRG Axon 80 Before A er 60

20 m m m m Fluorescence Density Fluorescence ʅ ʅ ʅ ʅ 10 10 10 10 0 Before A er 0123 Distance ʅ( m) m m m m Control

ʅ 80 ʅ Control ʅ Control ʅ 5 5 5 5 Before A er 60

20 Fluorescence Density Fluorescence 0 Before A er Distance (ʅm) CDE - 6 p = 4.6 x 10 1.1 InŇammatory Cocktail Control 1.1 p = 0.0016 e 1 1 1 0.9 0.9 0.9 0.8 0.8 0.8 (n=14) Membrane Cytosol (n=3) Membrane Cytosol 0.7 0.7 0.7 Before A er Normalized Fluor Density in DRG cell body Normalized Normalized DensityFluor in membran 024 cell body DRG in DensityFluor Normalized 024 Time (min) Time (min) Control Cocktail Figure 4 AB DRG DRG DRG 100 Cocktail Before m m m ʅ ʅ ʅ 5 5 80 A er 60 40 20 0 Fluorescence Density Fluorescence BeforeCocktail A er Cocktail Dsitance (ʅm) DRG DRG DRG Before m m5 m 100 Control ʅ ʅ ʅ 5 5 5 80 A er 60 40 20 0 Fluorescence Density Fluorescence Before Control A er Control Distance (ʅm) CDEF

Cocktail Cocktail y 1.1 Control 1 p = 9.2 x 10 -9 Carb 1.1 Carb 1.1 1200 1 p = 0.000054 Control 2 1000 1 1 0.9 800 0.9 0.8 0.9 600 0.8 400 0.8 Fluo Dens Normalized 0.7

Fluorescence Density Fluorescence (n=31) (n=12) Normalized Fluo Density Normalized

200 Fluo Density Normalized 0.7 0.7 0 1.5 3 4.5 6 0 1.5 3 4.5 6 0 1.5 3 4.5 6 Time (min) Time (min) Time (min) GH

HEK HEK HEK 100 Carb Before A er m m m 80 ʅ ʅ ʅ 10 10 10 60 40 20 0 Fluorescence Density Fluorescence Distance (ʅm) Before CarbA er Carb HEK HEK HEK y 100 Control Before A er m m

m 80 ʅ ʅ ʅ 10 10 10 60 40 20

Fluorescence Dens Fluorescence 0 Before ControlA er Control Distance (ʅm) IJKL p = 3.5 x 10- 10 Carb Carb Control 1400 1 p = 3.6 x 10- 13 1200 1 1 0.9 1000 0.9 0.8 800 0.8 0.6 0.8 600 0.4 0.7 0.7 400 0.2 Normalized Fluo Density

Fluorescence Density Fluorescence (n=18) (n=11) Normalized Fluo Density Normalized

Normalized Fluo Density Normalized 0.6 200 0.6 0 0123 0123 01 2 3 baseline Carb Control Time (min) Time (min) Time (min) Figure 5

A Cocktail RFP Cocktail RFP DRG GFP-posi ve DRG neurons 100 BEFORE 80 AFTER 60 40 20 0 Red Fluorescence Red density m m m ʅ ʅ ʅ 5 5 Before 5 A er Distance (ʅm) Cocktail GFP Cocktail GFP DRG GFP-posi ve DRG neurons 100 BEFORE 80 AFTER 60 40 20 0 m m m ʅ ʅ ʅ 5 5 Green Fluorescence density Fluorescence Green Before 5 A er Distance (ʅm) Cocktail Merge Cocktail Merge DRG m m m ʅ ʅ ʅ 5 5 Before 5 A er Cocktail DRG GFP-nega ve DRG neurons Cocktail RFP RFP 100 B BEFORE 80 AFTER 60 40 20 0 Red Fluorescence Fluorescence density Red m m m ʅ ʅ ʅ 5 5 Before 5 A er Distance (ʅm) Control RFP Control RFP DRG C 100 Control BEFORE 80 AFTER 60 40 20 0 m m m Red Fluorescence Red density m m ʅ ʅ ʅ ʅ ʅ 5 5 5 Distance (ʅm) 5 Before 5 A er Cocktail Merge Cocktail Merge Control GFP Control GFP DRG 100 Control 80 BEFORE AFTER 60 40 20 0 m m m m m m m ʅ ʅ ʅ ʅ ʅ Green Fluorescence density Fluorescence Green ʅ ʅ 5 5 5 5 Before 5 A er Distance (ʅm) D E F - 5 InŇammatory Cocktail Control P = 5.2 x 10 1.2 - 8 1.1 1.2 P = 1.1 x 10 1.1 1 1 P = 0.0062 1 0.9 0.8 0.9 0.8 0.6 0.8 0.7 0.7 0.4 0.6 0.6 0.2 Normalized Fluo Density Fluo Normalized 0.5 0.5 0

0.4 Density Fluorescence Normalized 0.4 Normalized Fluorescence Density Density Fluorescence Normalized 01234 01234 Time (min) Time (min) Red ŇƵŽŝŶ GFP-posi ve (n=9) RĞĚŇƵŽ Control (n=6) RĞĚŇƵŽ in GFP-nega ve (n=8) GFP ŇƵŽControl (n=8) GFP ŇƵŽŝŶGFP-posi ve (n=9) Figure 6

AB C D + 60mV - 60mV BIM Carb - 6 2 2 Carb p = 3.2 x 10 Ment Ment Carb - 5 Ment p = 1.1 x 10 1.2 - 7 1.5 3 p = 6.2 x 10 1.5 1 1 2.5 0.8 2 0.6 1 0.5 0.4 1.5 0.2 0 1 0 0.5 Before CON PIP2 BIM Current (nA) Current Current (nA) Current 0.5 -0.2 Carb (n=13) (n=8) (n=6) -0.5 (nA) Current -0.4 0 0 -0.6

-1 density current Normalized -0.5 -0.8 2+ 2+ 0 Ca2+ BIM -0.5 0 Ca Control 0 Ca PIP2 -1 -1 -1.5 - 5 01 2345678 012345678 0123456789 -1.2 p = 1.4 x 10 Ɵme (min) Ɵme (min) Ɵme (min) p = 9.3 x 10 - 5 p = 9.6 x 10 - 5 EFG H BIM ATP ATP Caps Caps Caps Caps Caps Caps CON ATP BIM+ATP 0 -0.5 -1 -1.5 50 pA 50 1 min 50 pA 50 -2 1 min 1st Peak

50 pA 50 -2.5 1 min 2nd Peak -3 -3.5 -4 P=0.0000014

Normalized Current Density Current Normalized -4.5 -5 P=0.000059 P=0.000032 P=0.031

IJ K

Ment Ment

.4 Carb Carb 0 1 0.8 2 0.6 0.8 p = 0.0037 0.4 1.5 0.6 0.2 + 60mV 0 1 0.4 -0.2 - 60mV 0.2 -0.4 current (nA) 0.5

current (nA) -0.6 0 p = 0.0010 0 -0.8

-0.2 current density Normalized 2+ 2 Ca2+ PIP2 -1 -0.5 2 Ca Control Ment Ment+Carb 0 1 234 0 1234 Ɵme (min) Ɵme (min) CON (14 cells) PIP2 (13 cells) Figure 7 AB Before 0 Ca2+ A er 0 Ca2+ 100 2 Ca2+ Before A er 80

20 Fluorescence Density Fluorescence 0

Tubby-YFPϭϬʅŵ Tubby-YFP ϭϬʅŵ ϭϬʅŵ DistanceDistance (ʅŵʅŵ)) Before 2 Ca2+ A er 2 Ca2+ 100 0 Ca2+ Before A er 80

20 Fluorescence Density Fluorescence 0 ϭϬʅŵ Tubby-YFP Tubby-YFP ϭϬʅŵ ϭϬʅŵ Distance (ʅŵ) Before Control A er Control 100 Control Before A er 80

20 Fluorescence Density Fluorescence

Tubby-YFP ϭϬʅŵ Tubby-YFP ϭϬʅŵ ϭϬʅŵ Distance (ʅŵ) CDE p = 1.2 x 10 - 8 Carb Control - 6 1.6 p = 6.6 x 10 1.6 p = 6.9 x 10 - 18 1.5 1.5 1.6 1.4 1.4 1.5 - 4 1.4 p = 1.4 x 10 1.3 - 5 1.3 1.3 p = 2.9 x 10 1.2 1.2 p = 1.3 x 10 - 8 ŵeŵbrane 2 Ca2+ 1.2 - 14 1.1 - 27 1.1 ŵeŵbrane 0 Ca2+ 1.1 p = 9.6 x 10 p = 8.7 x 10 - 6 1 cytosol 2 Ca2+ 1 1 cytosol 0 Ca2+ 0.9 0.9

alized Fluo Density Fluo alized 0.9 alized Fluo Density alized 0.8 alized Fluo Density Fluo alized ŵ e brane control 0.8 0.8 ŵ ŵ ŵ 0.7 ŵ cytosol control Nor 0.7 0.7 0.6 Nor 0.6 Nor 0.6 0.5 Tubby-YFP Tubby-YFP membrane cytosol 0.5 0.5 01234 01234 baseline 2 Ca2+ (n=25) Tiŵe (ŵin) Tiŵe (ŵin) 0 Ca2+ (n=28) control (n=8)

FGH p = 2.9 x 10 - 7 Carb Control 1.4 - 5 1.4 1.4 p = 8.1 x 10 - 16 1.2 p = 4.0 x 10 1.2 1.2 p = 7.1 x 10 - 17 Density 0 Ca2+ cytosol 1 0 Ca2+ ŵeŵbrane Fluo 1 2 Ca2+ cytosol 1 0.8

2 Ca2+ ŵeŵbrane Fluo Density alized alized Fluo Density Fluo alized

alized alized cytosol ŵ 0.6 ŵ ŵ 0.8 0.8 ŵeŵbrane Nor Nor Nor Tubby-YFP-R322H Tubby-YFP-R322H 0.4 0.6 0.6 membrane cytosol 0 0.5 1 1.5 2 0 0.5 1 1.5 2 baseline 2 Ca2+ (n=25) Tiŵe (ŵŝŶͿ Tiŵe (ŵŝŶͿ 0 Ca2+ (n=24) control (n=11) Figure 8 ABC M8-GFP+M1 Carb M8-GFP+M1 Control TubbyYFP+M1 Carb m m m ʅ ʅ ʅ 5 5 Before 5 Before Before M8-GFP+M1 Carb M8-GFP+M1 Control TubbyYFP+M1 Carb m m m ʅ ʅ ʅ 5 5 A er A er 5 A er DE + Carb or control p = 0.75

- 28 1.0 M8-GFP + M1 control 1.2 p = 2.2 x 10 1 M8-GFP + M1 Carb 0.8 0.8 0.6 0.4 0.2 0.6 Density Fluorescence Tubby-YFP + M1 Carb 0

0.4

150 200 250 300 Time (s) Figure 9 AB Ment Ment Ment Ment M8 o M8 2 M8+3GqiqQ209L 2.8 M8+3GqiqQ209L M8+3Gqiq o 1.8 M8+3Gqiq 2.4 1.6 1.4 2 1.2 1.6 1 340/380 nm ra nm 340/380 1.2 0.8

0.6 nm ra Normalized 340/380 0.8 0123456 0123456 Time (min) Time (min) CD p = 0.0013 p = 0.00043 p = 0.00015 p = 0.00008 2.4 o p = 0.013 p = 0.0083 3 p = 0.00098 p = 0.00036

2 o 1.6 2.6 1.2 2.2 0.8 1.8 0.4 1.4 340/380 nm ra

0 340/380 ra nm 1 baseline 1st peak 2nd peak baseline to Normalized 1st peak 2nd peak M8 (n=4) M8+3GqiqQ209L(n=5)M8+3Gqiq (n=4) M8 M8+3GqiqQ209L M8+3Gqiq EF M8-GFP M8-GFP + 3GqiqQ209L P = 0.13 1000 800 600 400 200 (n=5) (n=6) Fluorescence Fluorescence Density 0 M8-GFP M8-GFP + ϱʅŵ ϱʅŵ 3GqiqQ209L Figure 10 AB TRPM8+CiVSP TRPM8+CiVSP+3GqiqQ209L 0.1 0.2 0.3 1.0s 0.1 0.2 0.3 1.0s +100mV +100mV 0.2 0.2 -0.8 -0.3 -1.8 -2.8 -0.8 -3.8 Current (nA) Current -4.8 (nA) Current -1.3 -5.8 Menthol -1.8 Menthol C D p = 0.076 Peak 0.1s 0.2s 0.3s 1.0s p = 0.0008 0 1 p = 0.00001 -40 - 6 0.8 p = 2.4 x 10 on -80 p = 0.044 0.6 -120 0.4 -160 p = 0.048

Percent inhibi Percent 0.2 -200 p = 0.0052 -240 0 p = 0.00048 Peak 0.1s 0.2s 0.3s 1.0s Current Density (pA/pF)Current p = 0.0036 -280 M8+CiVSP (7 cells) M8+CiVSP+3GqiqQ207L (9 cells) M8+CiVSP+3GqiqQ209L (9 cells) M8+CiVSP (7 cells)