Structure-guided design of a peripherally restricted chemogenetic system

Abstract

Designer receptors exclusively activated by designer drugs (DREADDs) are chemogenetic tools for remotely controlling cellular signaling, neural activity, behavior, and physiology. Using a structure-guided approach, we provide a peripherally restricted Gi-DREADD, hydroxycarboxylic acid receptor DREADD (HCAD), whose native receptor is minimally expressed in the brain, and a chemical actuator that does not cross the blood-brain barrier (BBB). This was accomplished by combined mutagenesis, analoging via an ultra-large make-on-demand library, structural determination of the designed DREADD receptor via cryoelectron microscopy (cryo-EM), and validation of HCAD function. Expression and activation of HCAD in dorsal root ganglion (DRG) neurons inhibit action potential (AP) firing and reduce both acute and tissue-injury-induced inflammatory pain. The HCAD chemogenetic system expands the possibilities for studying numerous peripheral systems with little adverse effects on the central nervous system (CNS). The structure-guided approach used to generate HCAD also has the potential to accelerate the development of emerging chemogenetic tools for basic and translational sciences.

Keywords: GPCR; chemogenetics; peripheral nervous system.

Figure 1.. Structure-based design of HCAD

(A) Mutated residues mapped onto the HCA2 structure (PDB: 7XK2). Represented by side view and top view, respectively. (B) The difference between log(Emax/EC50) of WT human HCA2 and log(Emax/EC50) of each mutant activated by MK-0354 or niacin is depicted in a heatmap. For raw values, please refer to Table S1. N.D. means “not determined” since the value is lower than 4.25. (C) Characterization of WT and R111K by niacin and MK-0354 using Gi-mediated cAMP GloSensor assay. Data are mean ± SEM of n = 3 biological replicates in triplicate, unless otherwise indicated, in which case the number of biologically independent experiments is indicated in parentheses next to the receptor. (D) Chemical structures of MK-0354, ‘088, ‘089, FCH-2296413, niacin, MK-1903, and MK-6892. (E and F) Dose-response curve of MK0-0354, ‘088, and ‘089 against hHCA2 WT (E) and hR111K (F). (G) Comparison of ‘088-induced cAMP response between human WT (or human R111K) and mouse WT (or mouse R108K). (H) Comparison of niacin-induced cAMP response between human WT (or human R111K) and mouse WT (or mouse R108K). (I) The effect of receptor expression levels on ‘088 activity of mouse HCA2 WT (gray) and mouses HCAD (red). As can be observed, DNA concentration is strongly connected to agonist efficacy in both WT HCA2 and HCAD. Data are mean ± SEM of n = 4 biological replicates in triplicate, unless otherwise indicated, in which case the number of biologically independent experiments is indicated in parentheses next to the compound. (J) The effect of ‘088 and its racemic mixture FCH-2296413 at mouse HCAD (mR108K). Data mean ± SEM of n = 3 biological replicates in triplicate. See also Figures S1, S2, and S4.

Figure 2.. Identification and structural analysis of FCH-2296413-mHCA2-Gi complex in comparison to hHcar

(A) Overall cryo-EM structure mHCAD:Gai1:Gβ1γ2:scFv16 in complex with FCH-2296413 (magenta) (PDB: 7XK2). (B) Detailed interactions between FCH-2296413 (magenta) and mHCARD. (C) Effects of mutations of the ligand-binding pocket residues of mHCARD on changes in ΔpEC50 in response to stimulation with FCH-2296413, evaluated using a GloSensor cAMP assay. All data are presented as mean ± SEM of three independent experiments for the WT and mutants (n = 3). *p < 0.0332; **p < 0.0021; ***p < 0.0002; ****p < 0.0001; ns, not significant; and nd, not determined (one-way ANOVA). (D) Center: superposition of the 7TMs of mHCAD and human HCA2:MK-6892 complex (PDB: 7XK2) revealed a root-mean-square deviation (RMSD) of 1.25Å of all receptor C_α with many of the differences between the two structures being in the extracellular loops. Coloring scheme the same as in Figure 3. Top right: noticeably is the extension of the N terminus which adopts a twisted β-hairpin that lays over the top of the extracellular cavity and is secured in place with a pair of disulfide bonds connecting the N-terminal β-hairpin to ECL2 and TM7, ECL2 additionally adopts a β-hairpin structure with these two extracellular β-hairpins interacting with each other through side-chain aliphatic residues effectively sealing off the extracellular cavity. Right middle: R111^3.36 in the hHCA2 structure is involved in coordinating the carboxylate moiety of the MK-6892 ligand putatively mimicking the coordination of niacin but would sterically clash with the FCH-2296413’s tetrazole moiety in the mHCAD structure. Conversely, K108^3.36 in the mHCAD structure would be unable to effectively coordinate MK-6892 in the hHCA2 structure with putative distances of 3.7–4.4Å. Yellow lines indicate putative interactions. Right bottom: the unstructured portion of ECL2 in which residues S176 and F177 delve 1–1.3 Å deeper into the orthosteric pocket. Top left: coordination of mHCAD K108^3.36 with residues in the orthosteric site. Bottom left: coordination of HCA2 R111^3.36 with residues in the orthosteric site. (E) Probability density of ECL2 calculated from three independent simulations (1 μs each) for the respective HCARD structure, K108R (WT), and K108A. Right middle: calculated RMSF (root mean square fluctuation) plots per residue. Highlighted in yellow is ECL2 for each respective set of simulations while intracellular loop (ICL)3-TM6 is tinted blue. Right bottom: ECL2 and ICL3-TM6 colored by the normalized RMSF across the three sets of simulations. See also Figures S3, S5, and S6.

Figure 3.. ‘088 screen on 318 GPCRs and 97 kinases and its pharmacokinetic study

(A and B) ‘088 (A) or ‘089 (B) screening throughout the GPCRome (at 318 receptors) using 10 μM each on the PRESTO-Tango platform. Black dashed line indicated 3-fold of basal levels. Quinpirole, a full D2R (Dopamine receptor D2) agonist, was used as a positive control in GPCRome screening assays under D2R receptor-expressing conditions. Data represent mean ± SEM of fold over basal for each receptor (n = 4 technical replicates). (C) Single concentration (10 μM) of ‘088 screen against 97 kinases. Black dashed line indicated 70% inhibition. (D) Dose-response curve of ‘088 at three kinases AURKA (pink), LKB1 (orange), and MLK1 (green), respectively. (E and F) FCH-2296413 (E) or dechloroclozapine (DCZ) (F) concentration-time profiles in the plasma and brain after a single intraperitoneal injection to male C57BL/6 mice (10 mg/kg).

Figure 4.. mHCAD modulation of peripheral neuron physiology

(A) Confocal images of DRG sections from Avil^Cre/+ and Avil^+/+ mice (left; scale bar, 50 μm) and dissociated DRG neurons from Avil^Cre/+ (right; scale bar, 100 μm) mouse injected with PHP.S-CAG-FLEx-mHCAD-mCitrine. (B) Representative traces of voltage-activated calcium currents in mHCAD-mCitrine-positive (top) and mHCAD-mCitrine-negative (Ctrl) (bottom) neurons before (gray) and after (mHCAD-mCitrine-positive, green; mHCAD-mCitrine-negative, black) 10 μM FCH-2296413 application. (C) Percent change in peak current amplitude following FCH-2296413 application compared with baseline. Data are presented as mean ± SEM. n = 9–24. ***p < 0.001, one-way ANOVA followed by Tukey’s post hoc test. (D) Percent change in rise time (20%–80%) following 10 μM FCH-2296413 application compared with baseline. Data are presented as mean ± SEM. n = 12–24. ***p < 0.001, unpaired Student’s t test. (E) Representative AP firing at baseline, following 20 μM FCH-2296413 application (90 s), and after wash (90 s) in mHCAD-mCitrine-positive (top, green) and mHCAD-mCitrine-negative (bottom, gray) neurons. (F) Left: resting membrane potential (RMP) at baseline in mHCAD-mCitrine-positive and mHCAD-mCitrine-negative neurons. Right: change in membrane potential from baseline following 20 μM FCH-2296413 application, and after wash. Data are presented as mean±SEM. n = 14–17. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 two-way ANOVA followed by Fisher’s LSD (Least Significant Difference) post hoc test. (G) Number of APs evoked during the current injection protocol at baseline, following 20 μM FCH-2296413 application, and after wash. Data are presented as mean ± SEM. n = 14–16. **p ≤ 0.01, two-way ANOVA followed by Tukey’s post hoc test. (H) A representative AP trace from a mHCAD-mCitrine-expressing cell. Inset: a representative AHP trace from a mHCAD-mCitrine-expressing cell. Analyzed AP waveform properties are indicated. (I–M) Comparisons of AP waveform properties between mHCAD-mCitrine-positive and mHCAD-mCitrine-negative (Ctrl) neurons: voltage threshold (I), overshoot (J), half-width (K), AHP peak magnitude (L), and AHP decay time (M). Data are presented as mean ± SEM. n = 17–34. ***p < 0.001, unpaired Student’s t test. (N) Number of APs during 500-ms indicated current injections for mHCAD-mCitrine-positive (green) and mHCAD-mCitrine-negative (Ctrl) (gray) neurons. Data are presented as mean ± SEM. n = 17–26. Two-way ANOVA followed by Sidak’s post hoc test. (O) RNA levels in the DRG as measured by RT-qPCR, n = 3/group. For each gene, no differences in mean were found between control and mHCAD treated groups, p > 0.3, unpaired Student’s t test.

Figure 5.. Validation of mHCAD function and antinociceptive activity in vivo

(A) Schematic of intrathecal injections of PHP.S-CAG-FLEx viral constructs in Trpv1^Cre mice for the expression of mHCAD-mCitrine or tdTomato (Ctrl group) in DRG nociceptors. (B) Demonstration mHCAD-mCitrine or tdTomato expression in DRG sections using confocal imaging. (C) Antinociceptive effect of FCH-2296413 (10 mg/kg subcutaneous) against acute heat pain in the hotplate test (52C) in mice expressing mHCAD-mCitrine but not in mice expressing tdTomato. Bar plot representations of the sum of nocifensive behaviors and the latency to first nocifensive behavior. (D) In the complete Freund’s adjuvant (CFA) model of tissue injury and chronic inflammatory pain, FCH-2296413 (10 mg/kg subcutaneous) significantly reduced tissue injury-induced mechanical and heat hypersensitivity in the von Frey and Hargreaves tests, respectively, and only in mice expressing mHCAD. (E) Schematic of intrathecal injections in Trpv1^Cre mice to express mHCAD-mCitrine in DRG nociceptors with a PHP.S-CAG-FLEx-mHCAD-mCitrine viral construct. Mice received a subcutaneous injection of either vehicle or of FCH-2296413 30 min prior to behavior assessment. (F) Antinociceptive effect against acute heat pain in the hotplate test (52C) following mHCAD activation with FCH-2296413 (10 mg/kg subcutaneous). Bar plot representations of the sum of nocifensive behaviors and the latency to first nocifensive behavior. (G) In CFA model, FCH-2296413 (10 mg/kg subcutaneous) significantly reduced tissue-injury-induced mechanical and heat hypersensitivity in the von Frey and Hargreaves tests, respectively, while vehicle administration did not affect sensitivity. (H) In wild-type animals, FCH-2296413 (10mg/kg subcutaneous) did not change mechanical sensitivity in the von Frey test. (I–L) FCH-2296413 (10 mg/kg subcutaneous) did not change heat sensitivity in the tail immersion (I), hot plate (J and K), and Hargreaves (L) tests. (M–O) FCH-2296413 (10 mg/kg subcutaneous) had no effect on locomotion, anxiety, or motor coordination in the open field (M and N) and the rotarod (O) assays. (P) FCH-2296413 (10 mg/kg subcutaneous) does not alter voiding and defecation behaviors. All the values are mean ± SEM (n = 8 for Ctrl, n = 11 for mHCAD-mCitrine, n = 10 for vehicle and FCH-2296413). Student’s t test (C, F, I, K, L, and N–P); one-way ANOVA and Holm’s Sidak’s test (D, G, and H); *p < 0.05; **p < 0.01; ***p < 0.001 vs. Ctrl or vehicle. See also Figure S7.

Figure 6.. mHCAD expression in DRG nociceptor axons and cell bodies with associated FCH-2296413 antinociceptive effects

(A) Confocal images of DRG sections (12 μm) from Trpv1^Cre mice intrathecally injected with AAV-PHP.S-CAG-FLEx-mHCAD-mCitrine. Axons and cell body are shown in the left and right panels, respectively. Scale bar, 100 μm. (B–E) Bar plot representations of the antinociceptive effects of FCH-2296413 (10 mg/kg subcutaneous) in Trpv1^Cre mice intrathecally injected either with AAV-PHP.S-CAG-FLEx-tdTomato (Ctrl, magenta) or with AAV-PHP.S-CAG-FLEx-mHCAD-mCitrine (mHCAD-mCitrine, green), showing the number of hindpaw licks (B), the number of forepaw licks (C), the duration of hindpaw licking (D), and the duration of forepaw licking (E). (F–I) Bar plots representations of the effects of vehicle (gray) or FCH-2296413 administration (10 mg/kg subcutaneous, green) in wild-type mice, representing the number of hindpaw licks (F), the number of hindpaw licks (G), the duration of hindpaw licking (H), and the duration of forepaw licking (I). All the values are means ± SEM (n = 8 for Ctrl, n = 11 for mHCAD-mCitrine, n = 10 for vehicle and FCH-2296413). Student’s t test (B–I); *p < 0.05; **p < 0.01; ***p < 0.001 vs. Ctrl.