Structural basis of subunit selectivity for competitive NMDA receptor antagonists with preference for GluN2A over GluN2B subunits Genevieve E. Linda,b, Tung-Chung Moub,c, Lucia Tamborinid, Martin G. Pompere, Carlo De Michelid, Paola Contid, Andrea Pintof,1, and Kasper B. Hansena,b,1 aDepartment of Biomedical and Pharmaceutical Sciences, University of Montana, Missoula, MT 59812; bCenter for Biomolecular Structure and Dynamics, University of Montana, Missoula, MT 59812; cDivision of Biological Sciences, University of Montana, Missoula, MT 59812; dDepartment of Pharmaceutical Sciences, University of Milan, 20133 Milan, Italy; eRussell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins Medical School, Baltimore, MD 21205; and fDepartment of Food, Environmental and Nutritional Science, University of Milan, 20133 Milan, Italy Edited by Richard W. Aldrich, The University of Texas at Austin, Austin, TX, and approved July 5, 2017 (received for review May 10, 2017) NMDA-type glutamate receptors are ligand-gated ion channels that Considerable progress has been made in the development of contribute to excitatory neurotransmission in the central nervous subunit-selective allosteric modulators (7–13), but the development system (CNS). Most NMDA receptors comprise two glycine-binding of subtype-selective competitive NMDA receptor antagonists has GluN1 and two glutamate-binding GluN2 subunits (GluN2A–D). We been less successful. The competitive glutamate-site antagonist describe highly potent (S)-5-[(R)-2-amino-2-carboxyethyl]-4,5-dihy- NVP-AAM077 (hereafter NVP) was originally reported to have dro-1H-pyrazole-3-carboxylic acid (ACEPC) competitive GluN2 antag- 100-fold preference for GluN1/2A over GluN1/2B (14). As such, onists, of which ST3 has a binding affinity of 52 nM at GluN1/2A and this compound has been extensively used to investigate the role of 782 nM at GluN1/2B receptors. This 15-fold preference of ST3 for GluN2A-containing receptors in different brain regions and cel- GluN1/2A over GluN1/2B is improved compared with NVP-AAM077, lular processes. Subsequent studies using Schild analysis suggested a widely used GluN2A-selective antagonist, which we show has 11- that NVP only has a 5.4-fold preference for GluN1/2A over fold preference for GluN1/2A over GluN1/2B. Crystal structures of GluN1/2B (15). The modest 5.4-fold GluN2A preference of NVP the GluN1/2A agonist binding domain (ABD) heterodimer with has not discouraged its use in numerous published studies as a bound ACEPC antagonists reveal a binding mode in which the li- pharmacological tool compound to dissect the relative contribu- gands occupy a cavity that extends toward the subunit interface tions of GluN2A- and GluN2B-containing NMDA receptors to between GluN1 and GluN2A ABDs. Mutational analyses show that synaptic responses. The widespread use of competitive antagonists the GluN2A preference of ST3 is primarily mediated by four non- with only modest subunit preference as tool compounds highlights conserved residues that are not directly contacting the ligand, but a broad interest in GluN2A-selective antagonists and suggests a lack of structural and pharmacological understanding of competi- positioned within 12 Å of the glutamate binding site. Two of these tive antagonism at the glutamate site in NMDA receptors. Until residues influence the cavity occupied by ST3 in a manner that re- this study, crystal structures of NMDA receptor agonist binding sults in favorable binding to GluN2A, but occludes binding to domains (ABDs) in complex with competitive antagonists have GluN2B. Thus, we reveal opportunities for the design of subunit- been limited to the glycine site ligands—DCKA, cycloleucine, and selective competitive NMDA receptor antagonists by identifying a cavity for ligand binding in which variations exist between GluN2A and GluN2B subunits. This structural insight suggests that subunit Significance selectivity of glutamate-site antagonists can be mediated by mech- anisms in addition to direct contributions of contact residues to Despite decades of studies, the development of competitive binding affinity. glutamate-site antagonists that can distinguish between NMDA receptor subtypes based on GluN2 subunits has been unsuccess- synaptic transmission | Schild analysis | kinetic modeling | ful. The resulting lack of subunit-selective NMDA receptor ligands X-ray crystallography | PEAQX has led to the widespread use of competitive antagonists with only modest subunit preference in neurophysiological and be- havioral studies. This study describes competitive glutamate-site lutamate mediates fast excitatory neurotransmission in the antagonists with a binding mode in the GluN2A agonist binding mammalian CNS by binding to AMPA, kainate, and NMDA G domain that enables indirect engagement between ligands and receptors, which are ligand-gated ion channels involved in criti- nonconserved residues to achieve preferential binding to GluN1/ cal processes ranging from neuronal development to learning 2A over GluN1/2B. These findings are required for rational drug and memory (1, 2). In particular, dysfunction or dysregulation of design and suggest that glutamate-site competitive antagonists NMDA receptors has been implicated in numerous neurological with considerable subunit selectivity can be developed, despite and psychiatric disorders (1, 2). NMDA receptors are hetero- the highly conserved nature of the glutamate binding site. tetrameric subunit assemblies containing two GluN1 subunits that bind glycine or D-serine and two GluN2 subunits that bind Author contributions: G.E.L., T.-C.M., L.T., M.G.P., C.D.M., P.C., A.P., and K.B.H. designed glutamate (3, 4). Four different GluN2 subunits (GluN2A–D) research; G.E.L., T.-C.M., L.T., P.C., A.P., and K.B.H. performed research; G.E.L., T.-C.M., L.T., exist that have distinct regional and developmental expression P.C., A.P., and K.B.H. analyzed data; and G.E.L., T.-C.M., L.T., M.G.P., C.D.M., P.C., A.P., and patterns and endow NMDA receptor subtypes with strikingly K.B.H. wrote the paper. different biophysical and pharmacological properties (5, 6). The The authors declare no conflict of interest. GluN2 subunits therefore determine the physiological roles of This article is a PNAS Direct Submission. NMDA receptor subtypes, and, for this reason, they have re- Data deposition: The atomic coordinates and structure factors have been deposited in the Protein Data Bank, www.pdb.org [PDB ID codes 5VIJ (ST1), 5VII (ST3), 5VIH (ST6), and ceived considerable interest as potential therapeutic targets. To 5DEX (FRA-19)]. this end, ligands that distinguish NMDA receptor subtypes based 1 To whom correspondence may be addressed. Email: [email protected] or on GluN2 subunits are desirable due to their obvious utility as [email protected]. pharmacological tools and as potential therapeutic agents for the This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. treatment of CNS disorders (7, 8). 1073/pnas.1707752114/-/DCSupplemental. E6942–E6951 | PNAS | Published online July 31, 2017 www.pnas.org/cgi/doi/10.1073/pnas.1707752114 Downloaded by guest on October 1, 2021 TK40 (16-20) —and the glutamate site ligands—2-amino-5- that a compound in the ACEPC series of competitive antagonists, ST3 PNAS PLUS phosphonopentanoic acid (D-AP5), (−)-PPDA, and NVP (18, 21). {(S)-5-[(R)-2-amino-2-carboxyethyl]-1-[4-(3-fluoropropyl)phenyl]- In this study, we explore the structural and pharmacological 4,5-dihydro-1H-pyrazole-3-carboxylic acid}, has a 15-fold prefer- properties of a series of ligands based on (S)-5-[(R)-2-amino-2- ence for GluN1/2A over GluN1/2B. To facilitate the design of H carboxyethyl]-4,5-dihydro-1 -pyrazole-3-carboxylic acid (ACEPC) novel competitive antagonists, we use a combination of pharma- (22). Competitive NMDA receptor antagonists in this series have cological, crystallographic, and mutational experiments to describe been evaluated as potential neuroprotective and radioligand imag- the structural determinants of binding and subunit selectivity for ing agents (22, 23). Furthermore, preliminary functional results the ACEPC ligands. suggested that addition of halogen substituents to one of these li- S R gands, FRA-19 {( )-5-[( )-2-amino-2-carboxyethyl]-1-phenyl-4,5- Results dihydro-1H-pyrazole-3-carboxylic acid}, resulted in modest pref- erence for GluN1/2A over GluN1/2B receptors, as seen for Pharmacology of ACEPC Compounds at NMDA Receptors. We per- compounds ST1 {(S)-5-[(R)-2-amino-2-carboxyethyl]-1-(4-fluo- formed Schild analyses to determine binding affinities for the rophenyl)-4,5-dihydro-1H-pyrazole-3-carboxylic acid} and ST6 ACEPC ligands at GluN1/2A and GluN1/2B receptors (Materials {(S)-5-[(R)-2-amino-2-carboxyethyl)-1-(4-bromophenyl)-4,5-dihy- and Methods). Schild analyses of glutamate concentration–response dro-1H-pyrazole-3-carboxylic acid} (23). The GluN2A preference relationships in the absence and presence of FRA-19, ST1, ST6, prompted the synthesis of additional analogs, and we show here or ST3 revealed variation in selectivity between GluN1/2A and [antagonist] (µM) A GluN1/2A se control 100 K =23nM 0.1 i 80 0.3 60 1 R-1) D ed respon 3 40 g( GluN1/2B FRA-19 %control) 10 GluN1/2A GluN1/2B lo ( 20 Ki =93nM 0 Normaliz 0.1 1 10 100 1000 B 100 100 GluN1/2A K =24nM 80 80 i sponse 1) - trol) 60 60 on ed re 40 40 g(DR GluN1/2B ST1 GluN1/2A GluN1/2B lo (% c 20 20 Ki = 111 nM 0 0 Normaliz 0.1 1 10 100 1000 0.1 1 10 100 1000 C e 100 2.0 GluN1/2A K =32nM 80 1.5 i 60 1.0 (DR-1) 40 g GluN1/2B ST6 GluN1/2A GluN1/2B lo (% control) 0.5 20 Ki = 271 nM 0 0.0 Normalized respons 0.1 1 10 100 1000 0.1 1 10 D 100 GluN1/2A HOOC 2NH 80 Ki =52nM ol) COOH r N t 60 N 40 ST3 GluN1/2A GluN1/2B GluN1/2B (% con 20 malized response K = 782 nM r i 0 No 0.1 1 10 100 1000 F [glutamate] (µM) E 0.1 µM – 1 mM glutamate 0.3 µM – 3 mM glutamate 0.1 µM – 1 mM glutamate 0.3 µM – 3 mM glutamate GluN1/2A GluN1/2B GluN1/2B GluN1/2A 1µM ST3 1µM ST3 Fig.