RADICAL: a rationally designed ion channel activated by ligand for chemogenetics

Graphical abstract

Dear Editor,

Manipulating neuronal activities by exogenous means is always much desired in the field of neuroscience and beyond. Though many optogenetics and chemogenetics tools, such as channelrhodopsins and DREADDs (Armbruster et al., 2007; Sternson and Roth, 2014), have been developed and widely used, there are several limitations in these tools. For optogenetics, the invasiveness of implanted optical fiber is unavoidable. For chemogenetics, the DREADDs are G-protein coupled receptors that rely on complex cellular signaling networks to indirectly modulate ion channels to influence neuronal activity (Armbruster et al., 2007), therefore, their pharmacokinetics are slow (Lerchner et al., 2007). Chemogenetic tools that utilize engineered chimeric ligand-gated ion channels, like the nicotinic acetylcholine receptor (nAChR), are designed to modify neuronal excitability directly (Magnus et al., 2011). However, since the nAChR is abundantly expressed in the brain, how the subunits in the engineered nAChR interact with the endogenous subunits, as well as the potential effects of these interactions, remain to be investigated. Therefore, a noninvasive chemogenetic tool that directly manipulates neuronal potential with little endogenous expression of its native form in the brain is needed.

To develop such a chemogenetic tool, we targeted the transient receptor potential melastatin 8 (TRPM8), which is a calcium-permeable ion channel with limited expression regions in the brain (Jang et al., 2012; McKemy et al., 2002; Mohandass et al., 2020; Peier et al., 2002). We rationally designed a mutant TRPM8 channel to be activated by the chemical cyclohexanol (CHXOL) (Fig. 1A). CHXOL is analogous to the classic TRPM8 agonist menthol, but it lacks the isopropyl and methyl groups as compared to menthol (Fig. 1B). CHXOL (1–3 mmol/L) did not affect activities of a range of ion channels including hERG and many TRP channels (Fig. S1). Furthermore, CHXOL did not alter the resting membrane potential (RMP) or action potential in neurons and brain slices of mice (Fig. S2). As we found that the interaction between isopropyl group in menthol and residue I846 in TRPM8 is essential for binding (Fig. 1B) (Xu et al., 2020), CHXOL lacking this group did not bind well with TRPM8 as the top 10 scored CHXOL binding models with best binding energy were highly diversified in their binding configuration as shown in the molecular docking (Fig. 1B).

CHXOL cannot activate the wildtype TRPM8 even at a high concentration of 10 mmol/L in patch-clamp recordings (Fig. 1E). Therefore, we hypothesized that by increasing the sidechain bulkiness in I846, binding of CHXOL to TRPM8 mutant can be re-established to open the mutant channel. To test this hypothesis, we first docked CHXOL to the TRPM8-I846F mutant with a bulky sidechain. We observed that the top 10 scored models of CHXOL exhibited better convergence and larger binding energy (Fig. 1B–D). Indeed, in patch-clamp recordings, we found CHXOL concentration-dependently activated TRPM8-I846F with the EC₅₀ of 1.13 ± 0.07 mmol/L (Fig. 1F and 1G). On the other hand, when residue I846 was mutated to residues with smaller side chains such as glycine (G, no side chain), alanine (A), and leucine (L), the mutated channels could not respond to CHXOL (Fig. S3).

However, as the current elicited by CHXOL at hyperpolarizing membrane potential of −80 mV was small (Fig. 1F), we further performed calcium imaging to test whether CHXOL can induce calcium influx. We observed that 1 mmol/L CHXOL indeed elicited calcium influx in cells expressing TRPM8-I846F (Fig. 1H and 1J), while no such increase in intracellular calcium was observed by CHXOL in cells expressing the wildtype TRPM8 (Fig. 1H and 1I). These results suggested we indeed engineered a TRPM8 mutant to respond to CHXOL as we designed.

To further improve the CHXOL sensitivity in TRPM8-I846F at the physiologically relevant hyperpolarizing membrane potential, we introduced the I985K mutation to TRPM8-I846F as this mutation enhances the voltage sensitivity of TRPM8 (Taberner et al., 2014). As expected, the G–V curve of I846F-I985K mutant showed a dramatic left shift with half-maximal activation voltage (V_1/2) of 117.2 ± 13.2 mV as compared with WT (208.2 ± 3.4 mV) or I846F mutant (152.4 ± 10.8 mV). And the addition of CHXOL decreased the V_1/2 to 123.6 ± 13.5 mV for I846F mutant and 91.9 ± 13.7 mV for I985K-I846F (Fig. 1K–M). As a result, the TRPM8-I846F-I985K double mutant exhibited a much larger current activated by CHXOL at −80 mV (Fig. 1N). Though the EC₅₀ to CHXOL on TRPM8-I846F-I985K double mutant remains unchanged at +80 mV (1.17 ± 0.11 mmol/L) as compared to TRPM8-I846F mutant, these results suggested that the introduction of I985K indeed enhanced the voltage sensitivity in this channel. In single-channel recordings, CHXOL also elicited opening events in the double mutant at −80 mV (Fig. 1O). In calcium imaging, 100 µmol/L CHXOL, which barely induced calcium influx with TRPM8-I846F, elicited robust calcium signal in cells transfected with TRPM8-I846F-I985K mutant (Fig. 1P and 1Q). Moreover, prolonged exposure of CHXOL over 50 s almost decreased the intracellular calcium fluorescence to baseline (Fig. S4A and S4B). Such a decrease in calcium fluorescence was not due to photobleaching of the loaded Fluo-4 AM dyes because we observed a higher fluorescence intensity upon the application of ionomycin. In addition, the presence of 3 mmol/L CHXOL did not increase the cell death ratio in HEK293T cells expressing WT TRPM8 channels or I846F-I985K mutants (Fig. S4C and S4D), indicating the extent of CHXOL-induced activation of the rationally designed ion channel activated by ligand (RADICAL) was insufficient to induce cell death. These results suggested the influx of calcium was controllable and the amount of Ca²⁺ influx was limited. Moreover, the calcium selectivity of the TRPM8-I846F-I985K mutant was maintained as the wildtype TRPM8 channel (Fig. 1R–U).

We further test the ligand activation and temperature activation on the TRPM8-I846F-I985K mutant. I846F-I985K mutant and WT TRPM8 showed similar response to icilin with EC50 of 53.8 ± 15.3 nmol/L and 51.4 ± 12.9 nmol/L, respectively (Fig. S5A and S5B). In addition, we also calculated the change of enthalpic (ΔH) and entropic (ΔS) due to the temperature-driven transition of the TRPM8 channel. ΔH and ΔS exhibited similar behavior among WT TRPM8, I846F and I846F-I985K mutants, suggesting the temperature-related property of RADICAL remains unchanged (Fig. S5D–G).

We next employed virus-based delivery techniques to assess the in vivo effects of activating the CHXOL-TRPM8 mutant system in mice. We used lentivirus, which allows packaging of plasmid with 4.5–6 bp in length, together with the human synapsin promoter to deliver eGFP tagged mutant TRPM8 channel specifically in neurons (both pLenti-hSyn-TRPM8 (I846F) and pLenti-hSyn-TRPM8 (I846F-I985K)) (Fig. S6). We chose to express the virus in the infralimbic cortex (IL) (Fig. S7), which is necessary for the retrieval of fear memory extinction in rodents (Duvarci and Pare, 2014; Orsini and Maren, 2012). Twenty-one days after bilateral injection into the IL, pLenti-mediated viral transfection led to the expression of mTRPM8 (I846F) in neurons throughout the IL (Fig. 2A). To confirm the efficacy of the CHXOL-mTRPM8 I846F mutant system, we performed whole-cell patch clamp recordings from virally transfected IL neurons (Fig. 2B–I). We found that while 0.1 mmol/L CHXOL did not affect any physiological characteristics of the viral transfected neurons (Fig. 2B–E), 1 mmol/L CHXOL depolarized RMP, increased the intrinsic excitability on pLenti-hSyn-TRPM8 (I846F) transfected neurons (Fig. 2F–I).

To assess if in vivo activation of the CHXOL-TRPM8 (I846F) system affects IL-related behavioral consequences, we injected CHXOL (i.p. 10 mg/kg and 100 mg/kg) or the vehicle control in mice expressing pLenti-hSyn-TRPM8 (I846F). As expected, in the fear conditioning test (Fig. 2J–T), activation of IL TRPM8 (I846F) transfected neurons by 100 mg/kg but not 10 mg/kg CHXOL enhanced the retrieval of fear extinction memory without affecting locomotion levels compared to the virus and vehicle controls (Fig. 2J–T).

As the TRPM8-I846F-I985K double mutant exhibits a higher sensitivity to CHXOL (Fig. 1N–Q), we further tested the in vivo efficacy of pLenti-hSyn-TRPM8 (I846F-I985K) system (Fig. 2U). We found that a much lower concentration of CHXOL (0.1 mmol/L) significantly depolarized RMP, increased the intrinsic excitability on pLenti-hSyn-TRPM8 (I846F-I985K) transfected neurons (Fig. 2V–Y). In the fear conditioning test (Fig. 2Z), a lower concentration of 10 mg/kg CHXOL significantly enhanced the retrieval of extinction memory of TRPM8 (I846F-I985K) mice compared to the virus and vehicle controls (Fig. 2AA–AE).

Furthermore, we also tested the in vivo efficacy of RADICAL in glial cells in the ventral tegmental area (VTA). The astrocytes in VTA played a central role in the control of locomotor activity (Requie et al., 2022). We used lentivirus, together with the astrocyte-specific GfaABC1D promoter to deliver mCherry tagged mutant TRPM8 channel specifically in VTA astrocytes. After bilateral VTA injection of pLenti-GfaABC1D-mCherry-P2A-mTRPM8 (I846F-I985K)-WPRE virus, locomotion of mice was tested (Fig. 2AF). As we expected, 10 mg/kg CHXOL induced locomotor hyperactivity in TRPM8 (I846F-I985K) mice as compared to control mice (Fig. 2AG–AH).

Therefore, we have developed a rationally designed ion channel activated by ligand (RADICAL) for chemogenetics manipulation of animal behaviors. As RADICAL is based on calcium permeable TRPM8-I846F-I985K mutant, neuronal processes where elevation of intracellular calcium level is involved, such as learning and memory, can be tuned by this tool. TRPM8 channel is barely expressed in the brain (Jang et al., 2012; McKemy et al., 2002; Peier et al., 2002), so virus injection and expression of our RADICAL tool in the brain are expected to less perturb the subunit assembly and physiological function of endogenous ion channels.

Though our RADICAL is currently hosted on a lentiviral vector, such a system is compatible with the Cre recombinase (Nie et al., 2018; Sumegi et al., 2012). It is feasible to specifically express RADICAL in different cell types of Cre mice. We acknowledge that due to the size of RADICAL being too large, it cannot be packaged into an adeno-associated virus (AAV) vector so the application of the current version of RADICAL may be partially limited. Therefore, to broaden the usage of RADICAL we are dedicated to engineer an improved version of this chemogenetics tool in future, which can be packaged into an AAV vector by decreasing the size of TRPM8 without changing its channel activity.

References

[1] Armbruster BN, Li X, Pausch MH et al Evolving the lock to fit the key to create a family of G protein-coupled receptors potently activated by an inert ligand. Proc Natl Acad Sci U S A 2007;104:5163–5168.

[2] Duvarci S, Pare D. Amygdala microcircuits controlling learned fear. Neuron 2014;82:966–980.

[3] Jang Y, Lee Y, Kim SM et al Quantitative analysis of TRP channel genes in mouse organs. Arch Pharm Res 2012;35:1823–1830.

[4] Lerchner W, Xiao C, Nashmi R et al Reversible silencing of neuronal excitability in behaving mice by a genetically targeted, ivermectin-gated Cl-channel. Neuron 2007;54:35–49.

[5] Magnus CJ, Lee PH, Atasoy D et al Chemical and genetic engineering of selective ion channel–ligand interactions. Science 2011;333:1292–1296.

[6] McKemy DD, Neuhausser WM, Julius D. Identification of a cold receptor reveals a general role for TRP channels in thermosensation. Nature 2002;416:52–58.

[7] Mohandass A, Krishnan V, Gribkova ED et al TRPM8 as the rapid testosterone signaling receptor: implications in the regulation of dimorphic sexual and social behaviors. FASEB J 2020;34:10887–10906.

[8] Nie X, Kitaoka S, Tanaka K et al The innate immune receptors TLR2/4 mediate repeated social defeat stress-induced social avoidance through prefrontal microglial activation. Neuron 2018;99:464–479.e7.

[9] Orsini CA, Maren S. Neural and cellular mechanisms of fear and extinction memory formation. Neurosci Biobehav Rev 2012;36:1773–1802.

[10] Peier AM, Moqrich A, Hergarden AC et al A TRP channel that senses cold stimuli and menthol. Cell 2002;108:705–715.

[11] Requie LM, Gomez-Gonzalo M, Speggiorin M et al Astrocytes mediate long-lasting synaptic regulation of ventral tegmental area dopamine neurons. Nat Neurosci 2022;25:1639–1650.

[12] Sternson SM, Roth BL. Chemogenetic tools to interrogate brain functions. Annu Rev Neurosci 2014;37:387–407.

[13] Sumegi M, Fukazawa Y, Matsui K et al Virus-mediated swapping of zolpidem-insensitive with zolpidem-sensitive GABA(A) receptors in cortical pyramidal cells. J Physiol 2012;590:1517–1534.

[14] Taberner FJ, Lopez-Cordoba A, Fernandez-Ballester G et al The region adjacent to the C-end of the inner gate in transient receptor potential melastatin 8 (TRPM8) channels plays a central role in allosteric channel activation. J Biol Chem 2014;289:28579–28594.

[15] Xu L, Han Y, Chen X et al Molecular mechanisms underlying menthol binding and activation of TRPM8 ion channel. Nat Commun 2020;11:3790.

RIGHTS & PERMISSIONS

The Author(s) 2024. Published by Oxford University Press on behalf of Higher Education Press.