Derivatives of (R)-3-(5-furanyl)carboxamido-2-aminopropanoic acid as potent NMDA receptor glycine site agonists with GluN2 subunit-specific activity

Abstract

NMDA receptors mediate glutamatergic neurotransmission and are therapeutic targets due to their involvement in a variety of psychiatric and neurological disorders. Here, we describe the design and synthesis of a series of (R)-3-(5-fyranyl)carboxamido-2-aminopropanoic acid analogs 8a-s as agonists at the glycine (Gly) binding site in the GluN1 subunit, but not GluN3 subunits, of NMDA receptors. These novel analogs display high variation in potencies and agonist efficacies among the NMDA receptor subtypes (GluN1/2A-D) in a manner dependent on the GluN2 subunit. Notably, compound 8p is identified as a potent partial agonist at GluN1/2C (EC₅₀ = 0.074 μM) with agonist efficacy of 28% relative to activation by Gly and virtually no agonist activity at GluN1/2A, GluN1/2B and GluN1/2D. Thus, these novel agonists can modulate the activity of specific NMDA receptor subtypes by replacing the full endogenous agonists Gly or d-serine (d-Ser), thereby providing new opportunities in the development of novel therapeutic agents.

INTRODUCTION

L-Glutamate (Glu, 1, Figure 1) is the major excitatory neurotransmitter in the central nervous system, and is involved in most aspects of normal brain function including cognition, memory and learning.¹ Glutamatergic neurotransmission is mediated by ionotropic Glu receptors (iGluRs) that are ligand-gated ion channels^{2, 3} and metabotropic Glu receptors (mGluRs) that are G protein-coupled receptors (GPCRs).^4–6 The iGluRs are divided into N-methyl-d-aspartic acid (NMDA) receptors, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, and kainate receptors.^{2, 3} While AMPA and kainate receptors only bind Glu, the activation of NMDA receptors requires the simultaneous binding of both Glu and Gly or d-Ser.^{2, 3} However, NMDA receptors mainly rely on synaptic release of Glu for activation, since Gly (or d-Ser) is present at concentrations estimated at 0.5–10 μM^7–10, which will produce partial to full occupancy at GluN1 subunits^{11, 12}. NMDA receptors have been implicated in several psychiatric and neurological diseases, including Alzheimer’s disease, Parkinson’s disease, schizophrenia, epilepsy, depression and neuropathic pain.^{2, 3, 13–19}

Figure 1.

Chemical structures of L-Glu and agonists targeting the NMDA receptor Gly binding site.

NMDA receptors are tetrameric subunit complexes composed of two Gly-binding GluN1, and two Glu-binding GluN2 subunits (GluN2A-D) or two Gly-binding GluN3 subunits (GluN3A-B).^{2, 3} The GluN2 subunits have high sequence identity, and their Glu binding sites are highly conserved.^{2, 3} For this reason, the development of fully subtype-selective NMDA receptor ligands targeting the Glu binding site is profoundly challenging and has thus far been unsuccessful. Since synaptic NMDA receptors depend on Glu release for activation and the strength of Glu activation can be modulated by GluN1 agonists, therapeutic strategies to target the GluN1 agonist site have received considerable attention and have led to clinical trials with a number of Gly site ligands.^20–23

The endogenous agonists for GluN1, Gly (2, Figure 1) and d-Ser (3), are considered full agonists at GluN1/2A-D subtypes.^{11, 12} d-Cycloserine (DCS, 4), a Gly site agonist,²⁴ is a partial agonist at GluN1/2A, GluN1/2B, and GluN1/2D receptors with lower maximal response compared to Gly (R_max), whereas DCS has a higher R_max at GluN1/2C receptors compared to Gly (i.e. DCS is a superagonist at GluN1/2C).^25–27 Administration of DCS can enhance extinction of fear in both rodents and humans, and this effect might be relevant in the treatment of some psychiatric disorders.^28–30 (R)-2-Amino-3-(4-(2-ethylphenyl)-1H-indole-2-carboxamido)propanoic acid (AICP, 5) was reported by Urwyler et al. (2009) as a NMDA receptor Gly site agonist.³¹ Further investigation demonstrated superagonism at GluN1/2C (EC₅₀ = 0.0017 μM, R_max = 353%), full agonism at GluN1/2A (EC₅₀ = 0.066 μM, R_max = 100%), and partial agonist activity at GluN1/2B (EC₅₀ = 0.025 μM, R_max = 10%) and GluN1/2D (EC₅₀ = 0.025 μM, R_max = 27%).²⁶ Compared to DCS, AICP can reduce the spike frequency and burst firing of nucleus reticularis of the thalamus (nRT) neurons in wild type, but not GluN2C knockout mice.³² The nRT neurons are enriched with GluN2C-containing NMDA receptors that are proposed to regulate corticothalamic and thalamocortical communication.³² Moreover, the administration of DCS or AICP into the globus pallidus externa (GPe) was shown to improve motor function in a mouse model of Parkinson’s disease.³³ These observations demonstrate the potential for GluN2-specific Gly site agonists as modulators of NMDA receptors.

Apart from AICP, analogs 6 and 7 are GluN1/2C superagonists with moderate potencies (EC₅₀ = 1.97 μM and 0.32 μM at GluN1/2C, respectively) and different agonist efficacy profiles across NMDA receptor subtypes compared to AICP.¹¹

These results spurred our interest to develop new compound classes that would explore potency, efficacy, and GluN2 subunit-specific activity and facilitate further investigations of neuronal NMDA receptor subtypes. In this study, we maintained the essential α-amino acid part and amide part of AICP (5, Figure 2), whereas the N-acyl aromatic part was simplified to explore scaffold diversities. The fused phenyl ring was removed and the pyrrole ring was replaced by a furan ring, and substituted phenyl rings were introduced (8a-s). Thus, we exchanged the N-acyl aromatic scaffold in order to provide molecular flexibility that can enable new modes of agonist accommodation in the Gly binding pocket.

Figure 2.

Design of novel furanylamide derivatives as NMDA receptor Gly site agonists. R indicates substitutions on the phenyl ring.

RESULTS AND DISCUSSION

Chemistry

The synthesis of newly designed compounds 8a-p began with a Suzuki cross-coupling between 5-bromofuran-2-carbaldehyde 9 and a series of arylboronic acids/esters using the catalyst Pd(PPh₃)₄ and base Na₂CO₃ to give aldehydes 10d-p, which were then oxidized by AgNO₃ to yield key acid intermediates 11d-p (Scheme 1). For the synthesis of carboxylic acid 11q-s with substitutions on the ortho-position of phenyl ring, the route commenced with diazotization of 2-bromoaniline 13 to give the aldehyde 10q. After Suzuki coupling with styrylboronic acid or 3-fluorostyrylboronic acid, the aldehyde 10r-s were produced, which was then oxidized by AgNO₃ to give carboxylic acids 11q-s. The corresponding amides 12a-s were obtained by coupling of the corresponding carboxylic acid 11d-o, 11q-s or commercially available 11a-c and 11p with 3-amino-N-Boc-d-alanine. Carboxylic acid protection in this reaction would prevent byproduct formation,³⁴ but also add extra steps with concomitant material loss, resulting in minimal improvement to the overall yield. The Boc groups of 12a-s were removed with TFA and the final pure products 8a-s were obtained after purification by prep-HPLC.

Scheme 1.

Synthesis route of analogs 8a-s^a

^aReagents and conditions: (i) arylboronic acid/ester, Pd(PPh₃)₄, Na₂CO₃, N₂, Toluene, EtOH, H₂O, 85°C; (ii) AgNO₃, NaOH, EtOH, H₂O, 55°C; (iii) 3-amino-N-Boc-d-alanine, EDCI, HOBt, NEt₃, DCM, rt; (iv) TFA, DCM, (1:1 v/v), rt; (v) NaNO₂, HCl, H₂O, 0–5°C, then furfural, CuCl₂, 0°C to rt; (vi) styrylboronic acid/3-fluorostyrylboronic acid, Pd(PPh₃)₄, Na₂CO₃, N₂, Toluene, EtOH, H₂O, 85°C.

Pharmacological Evaluation

The pharmacological properties of analogs 8a-s were evaluated using two-electrode voltage-clamp (TEVC) recordings of responses from recombinant NMDA receptor subtypes expressed in Xenopus oocytes (Table 1, Figure 3). None of the analogs display discernable agonist activity at GluN1/2B. However, all compounds activate the GluN1/2C subtype with potencies in the low micromolar to nanomolar range. The agonists display partial, full, or superagonism ranging from <10% to 157% relative to the maximal response compared to Gly (R_max). In general, the compounds display a preference for GluN1/2C over the other three GluN1/2 NMDA receptor subtypes in terms of R_max (Table 1).

Table 1.

Agonist Activities at Recombinant GluN1/2A-D NMDA Receptors

Compound	R	GluN1/2A		GluN1/2B		GluN1/2C		GluN1/2D
		EC50 (μM)	Rmax (%)	EC50 (μM)	Rmax (%)	EC50 (μM)	Rmax (%)	EC50 (μM)	Rmax (%)
Gly	-	0.99	100	0.24	100	0.21	100	0.091	100
d-Ser	-	1.0	96	0.62	100	0.19	111	0.15	93
8a		4.3	14	NRa		1.7	44	NRa
8b		ND		NRa		1.3	26	NRa
8c		4.8	22	NRb		2.9	77	4.3	16
8d		0.13	24	NRa		0.040	36	NRa
8e		23	20	NRb		4.2	74	31	24
8f		6.6	17	NRb		1.8	36	7.7	35
8g		4.9	27	NRb		1.3	68	5.0	38
8h		0.093	40	NRb		0.063	52	0.15	11
8i		3.1	32	NRb		1.6	114	3.2	11
8j		10	22	NRb		4.8	119	NRb
8k		0.18	29	NRb		0.11	53	0.23	12
8l		ND		NRb		3.9	17	NDb
8m		0.26	38	NRb		0.12	70	0.27	11
8n		0.22	58	NRc		0.14	103	0.26	17
8o		NRb		NRb		4.0	157	ND
8p		ND		NRa		0.074	28	NRa
8q		0.053	34	NRa		0.036	49	0.080	9
8r		ND		NRa		0.75	22	ND
8s		NRd		NRd		5.8	16	NRd

EC₅₀ values in the presence of 300 μM L-Glu measured using two-electrode voltage-clamp electrophysiology from Xenopus oocytes expressing recombinant rat NMDA receptor subtypes. Relative efficacy is the fitted maximal response to the agonist compared to the maximal response to 100 μM Gly in the same recording. Measurements of the full concentration-response data were limited by solubility for some compounds. Compound stock solutions were prepared in DMSO. The DMSO concentration was kept constant in all ligand solutions and never exceeded 0.5%. ND indicates that the EC₅₀ could not be determined, but some agonist activity was observed. NR indicates less than 5% response at ^a 100μM, ^b 300μM, ^c 10 μM, or ^d 30 μM of the agonist. Data are from 3–14 oocytes. See Table S1 in the Supporting Information for sample size, Hill slopes, and standard deviations.

Figure 3.

(A-B) Representative recordings of 8k and 8n concentration-response data in the continuous presence of 300 μM L-Glu at recombinant GluN1/2A-D receptors measured using two-electrode voltage-clamp electrophysiology. (C) Concentration-response data measured using two-electrode voltage-clamp recordings for Gly and novel agonists with nanomolar potencies (8h, 8k, 8m, 8n, 8p, 8r, and 8s) at recombinant GluN1/2A-D receptors. The fitted maximal response to agonist is normalized to the maximal response to Gly measured in the same recording. Data are from 4–14 oocytes and data points are mean ± SEM. See Table 1 for values.

Analogs 8b, 8l, 8o, 8p, 8r and 8s are selective GluN1/2C agonists with little to no agonist activity at GluN1/2A, GluN1/2B, and GluN1/2D (Table 1). Specifically, analog 8o is a superagonist (EC₅₀ = 4.0 μM, R_max = 157%) at GluN1/2C, while analogs 8p (EC₅₀ = 0.074 μM, R_max = 28%) and 8r (EC₅₀ = 0.75 μM, R_max = 22%) are partial agonists. Analog 8h is a potent GluN1/2A (EC₅₀ = 0.093 μM, R_max = 40%), GluN1/2C (EC₅₀ = 0.063 μM, R_max = 52%), and GluN1/2D (EC₅₀ = 0.15 μM, R_max = 11%) partial agonist with no agonist activity at GluN1/2B. Compared to 8h, analog 8q shows improved agonist potency, but similar agonist efficacy at GluN1/2A (EC₅₀ = 0.053 μM, R_max = 34%), GluN1/2C (EC₅₀ = 0.036 μM, R_max = 49%), and GluN1/2D (EC₅₀ = 0.080 μM, R_max = 9%). Compounds 8k, 8m and 8n display pharmacological profiles similar to 8h and 8q, but with slightly lower potencies. The ortho-CF₃ analog 8d is a partial agonist at GluN1/2A (EC₅₀ = 0.13 μM, R_max = 24%) and GluN1/2C (EC₅₀ = 0.040 μM, R_max = 36%) with no agonist activity at GluN1/2B and GluN1/2D.

The structure-activity relationship at GluN1/2A and GluN1/2C for chloro-substituted analogs 8a, 8g, 8m, and CF₃-substitued analogs 8d and 8f demonstrates that the introduction of ortho-substitution on the phenyl ring improves agonist potency (Table 1, Figure 4). This observation is consistent with the potencies of other ortho-substituted analogs 8h, 8n, and 8q. Similar relationship between ortho-substitution and agonist potency is observed at GluN1/2D (Table 1, Figure 4). However, larger substitutions on the ortho-position for compounds 8o, 8s and 8r result in a decrease in potency at GluN1/2C and is not tolerated in GluN1/2A and GluN1/2D. In addition, compounds 8i, 8j and 8l with hydrophilic groups on the ortho-position only show moderate potencies, compared to 8m and 8n. Therefore, the hydrophobic and small ortho-substitutions might be more favorable for this series of analogs and thereby obtain high potency at GluN1/2A and GluN1/2C.

Figure 4.

LogEC₅₀ values ± SEM for analogs with agonist activity are plotted for the indicated NMDA receptor subtypes. There is significant correlation for agonist potencies at GluN1/2A vs. GluN1/2C, GluN1/2D vs. GluN1/2C, and GluN1/2D vs. GluN1/2A. EC₅₀ values are lognormally distributed³⁵ and LogEC₅₀ values were therefore used in the calculation of the Pearson correlation. See Table 1 for values.

There is a significant correlation for agonist potencies at GluN1/2A, GluN1/2C, and GluN1/2D (Figure 4). That is, the most potent GluN1/2C agonists are also the most potent GluN1/2A (Pearson r = 0.9856, N = 13, P < 0.05) and GluN1/2D agonists (Pearson r = 0.9780, N = 10, P < 0.05). Similarly, the most potent GluN1/2A agonists are also the most potent GluN1/2D agonists (Pearson r = 0.9964, N = 10, P < 0.05). Agonist efficacies are not displaying any significant correlation among the NMDA receptor subtypes (2A vs. 2C, Pearson r = 0.335, N = 13, P > 0.05; 2D vs. 2C, Pearson r = −0.231, N = 10, P > 0.05; 2A vs. 2D, Pearson r = −0.470, N = 10, P > 0.05). Interestingly, there is a significant inverse correlation between agonist potencies and efficacies at GluN1/2A (Pearson r = −0.6264, N = 13, P < 0.05) such that the least potent agonists tend to have the highest agonist efficacy. By contrast, there is a significant correlation between agonist potencies and efficacies at GluN1/2D (Pearson r = 0.6811, N = 10, P < 0.05) such that the most potent agonists tend to have the highest agonist efficacy. Unlike GluN1/2A and GluN1/2D, there is no significant correlation between agonist potencies and efficacies at GluN1/2C (Pearson r = 0.2333, N = 19, P > 0.05). The strong correlation for agonist potencies suggest that the ligand-receptor interactions are shared among the NMDA receptor subtypes, consistent with agonists binding to the Gly binding site in the conserved GluN1 subunit. On the other hand, the lack of correlation for agonist efficacies among NMDA receptor subtypes and the different trends for agonist potencies and efficacies at individual subtypes suggest that the mechanisms governing agonist efficacy is vastly different among NMDA receptor subtypes in a GluN2 subunit-specific manner.

The apparent lack of agonist activity of all analogs at GluN1/2B receptors could be due to binding without any agonist efficacy (i.e. competitive antagonism), binding with very low agonist efficacy that produces no discernable current response, or no binding to GluN1 at all in GluN1/2B receptors. To determine if the analogs are binding the GluN1 subunit in GluN1/2B receptors, we evaluated whether compounds 8h and 8p inhibit responses to Gly (Figure 5). As expected, the analogs alone (10 μM) in the continues presence of 300 μM L-Glu were unable to activate any current responses at GluN1/2B. However, both 8h and 8p (10 μM) inhibited responses to Gly (100 μM) by 47 ± 6% (n =12) and 57 ± 2% (n =12), respectively. Using the Cheng-Prusoff relationship³⁶ and IC₅₀ values of ~10 μM, these inhibitions correspond to estimated binding affinities (K_i) of ~0.1 μM for 8h and 8p at the Gly binding site in GluN1/2B receptors. These results further corroborate that the analogs bind to the conserved Gly binding site in the GluN1 subunit of GluN1/2A-D subtypes, but that agonist efficacies are controlled by the GluN2 subunit.

Figure 5.

(A) Representative recordings of responses to 8h and 8p in the absence and presence of Gly and in the continuous presence of 300 μM L-Glu at recombinant GluN1/2B receptors measured using two-electrode voltage-clamp electrophysiology. No responses to 8h and 8p (10 μM) are observed at GluN1/2B receptors, while both 8h and 8p inhibit responses to Gly. Data for each analog are from 12 oocytes and bar graphs are mean ± SEM. (B) Representative recordings of response to 8h and 8p in the absence and presence of Gly at recombinant GluN1^FATL/3A and GluN1^FATL/3B receptors that include F484A and T518L mutations in the GluN1 agonist binding site to prevent agonist binding to GluN1, thereby enabling evaluation of the pharmacology of GluN3 subunits. No responses to 8h and 8p are observed, and 8h and 8p did not inhibit responses to Gly. Data for each analog are from 12–13 oocytes and bar graphs are mean ± SEM.

The GluN3A and GluN3B NMDA receptor subunits also bind Gly,^{37, 38} but structural differences between GluN1 and GluN3 agonist binding sites can affect binding of Gly site agonists.^{39, 40} To determine if 8h and 8p bind GluN3A and GluN3B subunits, we evaluated these analogs at GluN1^FATL/3A and GluN1^FATL/3B receptors that include F484A and T518L mutations in the GluN1 agonist binding site (Figure 5). These mutations prevent agonist binding to GluN1 and enable evaluation of the pharmacology of GluN3 subunits.⁴⁰ Furthermore, agonist binding to GluN1 triggers strong desensitization of GluN1/3 receptors, while agonist binding to GluN3 mediates activation.^40–42 Both compounds 8h and 8p (10 μM) did not activate current responses by themselves and did not inhibit responses to Gly at GluN1^FATL/3A and GluN1^FATL/3B receptors (Figure 5). Thus, 8h and 8p (10 μM) selectively bind the agonist binding site in GluN1 over those in GluN3A and GluN3B subunits.

Molecular modeling

The structure-activity relationship was investigated in silico by induced-fit docking for compounds 8h, 8k and 8p into the Gly binding site of GluN1 using the crystal structure of the isolated GluN1/2A agonist binding domain (ABD) heterodimer (PDB ID: 5I57)⁴³ as template (See Supporting Information for PDB files with docking poses).

According to induced-fit docking, the key interactions of the α-amino acid part for analogs 8h, 8k and 8p (Figure 6A–D) are conserved compared to Gly and d-Ser bound structures. The amide spacers of these analogs form hydrogen bonds and are accommodated in the narrow tunnel shaped by GluN1 residues T518, S688 and D732, which is consistent with previous molecular dynamics simulations of compound 6 docked into the GluN1 ABD.¹¹ The substituted phenyl ring of these analogs interact with GluN1 residues Y692 and F754 and orient toward the interface of GluN1 and GluN2 subunits. These additional interactions with the surrounding residues might contribute to the improved binding affinity in the series. Furthermore, the docking poses are consistent with the observed selectivity of 8h and 8p for GluN1 over GluN3 subunits (Figure 5), since binding would be sterically occluded by GluN3A E871 and GluN3B E771 residues. The smaller GluN1 S756 at this position permits occupancy in the narrow tunnel toward the GluN1-GluN2 interface (Figure 6). Unlike 8h and 8k, the ortho-nitro substitution on the phenyl ring of 8p orients to the opposite direction and the phenyl ring shifts. Consequently, the para-Cl substituent on the phenyl ring is buried deeper into the GluN1-GluN2 subunit interface, thereby causing additional conformational change for surrounding residues. The distinct interactions formed by the ortho- and para- substitutions may, in part, mediate the remarkable GluN1/2C subtype preference of compound 8p.

Figure 6.

(A) Overview of GluN1/2A agonist binding domain heterodimer crystal structure (PDB ID: 5I57)⁴³ with 8h docked into Gly binding site in GluN1. (B-D) Detailed views of induced-fit docking models for compounds 8h, 8k, and 8p in the GluN1/2A agonist binding domain structure. GluN1 and GluN2A are depicted as grey and blue cartoons, respectively. The pink oval highlights a possible steric clash that could influence agonist activity in a GluN2-specific manner, since GluN2A V783 is a non-conserved residue (Phe in GluN2B and Leu in GluN2C/D). See Supporting Information for PDB files with docking poses.

Considerable movement of GluN1 L654, V689, Y692, F753 and F754 is required for binding of all furanylamide analogs, while the position of the GluN2A ABD is not shifted compared to the Gly bound GluN1/2A structure. We hypothesize that this conformational change for residues in the GluN1 ABD influences interactions with non-conserved residues in GluN2 subunits (e.g. GluN2A V783), and that the precise orientation of substituents on the furan ring dictates agonist potency and efficacy. To determine if the non-conserved residues at the position of GluN2A V783 influences GluN2-specific agonist activity, we mutated this residue in GluN2A and GluN2B and evaluated responses to 8h and 8p (Figure 7A,B). Increasing the sidechain bulk from valine to leucine (as in GluN2C/2D) or phenylalanine (as in GluN2B) reduced the maximal response of 8h and 8p relative to Gly in GluN1/2A-V783L and GluN1/2A-V783F. By contrast, reducing sidechain bulk from phenylalanine to leucine (as in GluN2C/2D) and valine (as in GluN2A) did not significantly increase the maximal response to 8h or 8p relative to Gly in GluN1/2B-F784L and GluN1/2B-F784V. Consistent with smaller relative agonist efficacies, the potency of 8h was reduced at GluN1/2A-V783L (EC₅₀ = 0.21 ± 0.07 μM, R_max = 30 ± 2%, n = 6) and GluN1/2A-V783F (EC₅₀ = 0.41 ± 0.06 μM, R_max = 9 ± 1%, n = 4) compared to GluN1/2A wild type (EC₅₀ = 0.11 ± 0.01 μM, R_max = 45 ± 7%, n = 7) (Figure 7C). These results demonstrate that the non-conserved residue at position GluN2A V783 in the GluN1-GluN2 ABD dimer interface (Figure 6) can influence GluN2-specific agonist efficacy. However, the superagonism of the novel agonist series at GluN1/2C combined with lack of responses at GluN1/2B suggest that the GluN2-specific agonist efficacy is primarily governed by additional GluN2-dependent mechanisms that presumably include other regions of the ABD or other receptor domains.

Figure 7.

(A) Maximal responses to 8h or 8p (10 μM) at GluN1/2A wild type (WT) and mutated GluN1/2A-V783L and GluN1/2A-V783F receptors measured using two-electrode voltage-clamp electrophysiology in the continuous presence of 300 μM L-Glu. Data are mean ± SEM from 6–8 oocytes for 8h and 3–6 oocytes for 8p. For each analog, * indicates significant difference from wild type (P < 0.05, two-way ANOVA with Tukey’s multiple comparisons test). (B) Maximal responses to 8h or 8p (10 μM) at GluN1/2B wild type (WT) and mutated GluN1/2B receptors in the continuous presence of 300 μM L-Glu. Data are mean ± SEM from 4–6 oocytes for 8h and 3–6 oocytes for 8p. Agonist efficacies relative to Gly are not significantly changed at mutated GluN1/2B receptors compared to wild type (P > 0.05, two-way ANOVA with Tukey’s multiple comparisons test). (C) Concentration-response data for 8h at GluN1/2A wild type (EC₅₀ = 0.11 ± 0.01 μM, R_max = 45 ± 7%, n = 7), GluN1/2A-V783L (EC₅₀ = 0.21 ± 0.07 μM, R_max = 30 ± 2%, n = 6), and GluN1/2A-V783F (EC₅₀ = 0.41 ± 0.06 μM, R_max = 9 ± 1%, n = 4). The fitted maximal response to 8h is normalized to the maximal response to Gly measured in the same recording.

CONCLUSION

In the present work, a series of (R)-3-(5-furanyl)carboxamido-2-aminopropanoic acid derivatives (8a-s) were designed, synthesized and characterized in a functional assay at GluN1/2A-D NMDA receptor subtypes. Six analogs (8b, 8l, 8o, 8p, 8r and 8s) were identified as functionally selective GluN1/2C agonists with 8p and 8r displaying nanomolar potency. Compound 8h and 8q showed potent agonist activity at GluN1/2A, GluN1/2C and GluN1/2D, while the ortho-CF₃ substituted compound 8d was shown to be a dual partial agonist at GluN1/2A and GluN1/2C subtypes (EC₅₀ = 0.13 μM and 0.040 μM, respectively). Our in silico study suggested binding poses of 8h, 8k and 8p and illustrated possible structural mechanisms of agonist binding. However, more work is needed to appreciate the mechanisms by which these novel Gly site agonists engage in GluN2-specific interactions, thereby enabling variation in activity among NMDA receptor subtypes. The notable variation in potency, efficacy, and selectivity among these analogs at GluN1/2A-D receptor subtypes may provide new therapeutic strategies for neurological and psychiatric disorders, in which each NMDA receptor subtype can be distinctly modulated to balance therapeutic and potential adverse effects.

EXPERIMENTAL SECTION

Chemistry.

All reactions requiring inert conditions were performed in flame-dried glassware under an inert atmosphere of N₂ using standard Schlenk and syringe-septum technique. Dichloromethane (DCM), tetrahydrofuran (THF) and dimethylformamide (DMF) were dried using a glass contour solvent system (SG Water, USA LCC). All other solvents for reaction, extraction, and purification were purchased in HPLC grade and used as received. Normal phase column chromatography was performed on silica gel 60 (Merck, 0.015–0.040 mm). Reactions and chromatography fractions were monitored by analytical thin layer chromatography (TLC), which was carried out using silica gel 60 F₂₅₄ aluminum plates (Merck), and the compounds were visualized by using UV light (254 or 365 nm or both). All starting materials as well as the reagents were obtained from commercial sources and used without further purification.

Proton (¹H) and carbon (¹³C) NMR spectra were recorded at 300 K in the solvent indicated. Bruker 400 and Bruker 600 spectrometers were operated at 400 and 600 MHz for proton and at 100 and 150 MHz for carbon nuclei. Chemical shifts (δ) are given in parts per million (ppm) using signals of residual nondeuterated solvent as internal standards.

LC/MS was performed on an Agilent 1200 system with a Zorbax Eclipse XBD-C18 column (4.6 mm × 50 mm) and UV detector using a linear gradient of the binary solvent system of H₂O/MeCN/formic acid (A = 95/5/0.1 and B = 5/95/0.1) with a flow rate of 1 mL/min (0–100% B over 4 min, then 0.5 min 100% B) equipped with an API ion source (HP 1100 MSD).

Analytical HPLC was performed on a system consisting of an Ultimate 3000 pump and PDA detector, and a TSP AS-3000 autosampler with a Gemini-NX C₁₈ column (4.6 mm × 250 mm), or on an Agilent 1260 system with Innoval C₁₈ column (4.6 mm × 250 mm) equipped with a 254 nm UV detector using a linear gradient elution of the binary solvent system of H₂O/MeCN/TFA (v/v/v; A = 95/5/0.1 and B = 5/95/0.1) with the concentration of B from 0% to 100% over 20 min with a flow rate of 1.0 mL/min. Prep-HPLC for the purification of final compounds was performed on an Ultimate 3000 system with a Gemini-NX C₁₈ column (21 mm × 250 mm), or on a Waters instrument with an Innoval C₁₈ column (21.2 mm × 250 mm) equipped with a 254 nm UV detector with the same binary solvent system with a flow rate of 20 mL/min. Purities of the tested compounds were determined by analytical HPLC to be >95%.

Procedure A: General procedure for Suzuki coupling between arylboronic acids and aryl halogenides. 5-bromofuran-2-carbaldehyde (175.0 mg, 1.0 mmol), appropriate arylboronic acid (1.5 mmol), Pd(PPh₃)₄ (57.8 mg, 0.05 mmol) and K₂CO₃ (414.6 mg, 3.0 mmol) in toluene/EtOH/H₂O (10:3:2, v/v/v, 10 mL) were heated to 85°C overnight under a nitrogen atmosphere. The mixture was cooled to rt, diluted with EtOAc, filtered on Celite and evaporated to dryness. The crude was purified by column chromatography to afford the pure product.

Procedure B: General procedure for oxidation reaction from aldehyde to carboxyl acid. NaOH (320.0 mg, 4.0 mmol) was dissolved in H₂O/EtOH (1:1, v/v, 12 mL) and heated to 60°C. AgNO₃ (679.5 mg, 4.0 mmol) and the appropriate aldehyde (1.0 mmol) was added to the stirred solution. The reaction mixture was kept stirring for 2 h at 60°C and then filtered through Celite. The filtrate was acidified to pH~2 with 6 M HCl. The precipitate was then filtered off, washed with H₂O and dried to afford the product without further purification.

Procedure C: General procedure for acylation and Boc deprotection. The appropriate acid (1.0 mmol), EDCI (191.7 mg, 1.0 mmol), HOBt (135.1 mg, 1.0 mmol), NEt₃ (418.1 μL, 3.0 mmol) were dissolved in 10 mL anhydrous DCM and stirred for 20 min at rt. The solution of Boc-d-2,3-diaminopropionic acid (204.2 mg, 1.0 mmol) in 10 mL anhydrous DCM was added and stirred overnight at rt. The reaction mixture was washed with H₂O (2 × 25 mL) and the organic phase was then dried with MgSO₄. After concentration, the residue was treated with TFA/DCM (1:1, v/v, 10 mL). The resulting mixture was then stirred for another 4 h at rt and the solvent was evaporated to dryness. The final product was obtained as TFA salt after purification by prep-HPLC and 0.1% of TFA was added into the eluent. The yield and activity data were calculated considering the TFA salt.

5-(2-Ethylphenyl)furan-2-carbaldehyde (10h).

Procedure A. Yield: 75%. ¹H NMR (600 MHz, methanol-d₄) δ 1.21 (3H, t, J = 7.5 Hz), 2.87 (2H, q, J = 7.5 Hz), 6.83 (1H, d, J = 3.7 Hz), 7.27 (1H, td, J = 7.5, 1.6 Hz), 7.30–7.39 (2H, m), 7.50 (1H, d, J = 3.7 Hz), 7.66 (1H, dd, J = 7.7, 1.4 Hz), 9.59 (1H, s).

5-(2-(Methoxymethyl)phenyl)furan-2-carbaldehyde (10i).

Procedure A. Yield: 89%. ¹H NMR (600 MHz, methanol-d₄) δ 3.39 (3H, s), 4.58 (2H, s), 6.93 (1H, d, J = 3.7 Hz), 7.36–7.42 (2H, m), 7.47 (1H, d, J = 3.7 Hz), 7.49–7.53 (1H, m), 7.73– 7.81 (1H, m), 9.58 (1H, s).

2-(5-Formylfuran-2-yl)-N, N-dimethylbenzenesulfonamide (10j)

Procedure A. Yield: 93%. ¹H NMR (400 MHz, methanol-d₄) δ 2.62 (6H, s), 6.85–6.97 (1H, m), 7.34–7.45 (1H, m), 7.55–7.70 (3H, m), 7.91 (1H, d, J = 8.5 Hz), 9.56 (1H, s).

5-(Naphthalen-1-yl)furan-2-carbaldehyde (10k).

Procedure A. Yield: 63%. ¹H NMR (600 MHz, chloroform-d) δ 6.87 (1H, d, J = 3.6 Hz), 7.36 (1H, d, J = 3.6 Hz), 7.46–7.50 (2H, m), 7.51–7.54 (1H, m), 7.80 (1H, d, J = 7.2 Hz), 7.83–7.88 (2H, m), 8.33 (1H, d, J = 8.4 Hz), 9.67 (1H, s).

5-(o-Tolyl)furan-2-carbaldehyde (10n).

Procedure A. Yield: 68%. ¹H NMR (400 MHz, chloroform-d) δ 2.46 (3H, s), 6.65 (1H, d, J = 3.7 Hz), 7.18 (1H, d, J = 3.4 Hz), 7.21 (2H, d, J = 2.5 Hz), 7.25 (1H, d, J = 3.7 Hz), 7.70 (1H, dd, J = 7.3, 2.2 Hz), 9.58 (1H, s).

5-(2-(Benzyloxy)phenyl)furan-2-carbaldehyde (10o).

Procedure A. Yield: 85%. ¹H NMR (400 MHz, chloroform-d) δ 5.13 (2H, s), 6.97– 7.03 (3H, m), 7.23–7.44 (7H, m), 8.00 (1H, dd, J = 7.8, 1.7 Hz), 9.56 (1H, s).

5-(2-bromophenyl)furan-2-carbaldehyde (10q).

2-Bromoaniline (109.0 μL, 1.0 mmol) was dissolved in 30 mL 6 M HCl. The solution was heated until all material is dissolved completely. The solution was then cooled to 0°C with ice bath and NaNO₂ (75.89 mg, 1.1 mmol, 1.1 eq) in 10 mL H₂O was added into the solution dropwise while keeping the temperature lower than 5°C during addition. The reaction was stirred for 10 min at 0°C. Furfural (99.39 μL, 1.2 mmol, 1.2 eq) and CuCl₂ (40.34 mg, 0.3 mmol, 0.3 eq) dissolved in 10 mL H₂O was added into the reaction mixture and stirred for 48h at rt. The precipitation was then filtered, washed with H₂O and dried to afford the product without further purification. Yield: 11%. ¹H NMR (600 MHz, chloroform-d) δ 9.69 (s, 1H), 7.93 (dd, J = 7.9, 1.7 Hz, 1H), 7.69 (dd, J = 8.0, 1.2 Hz, 1H), 7.41 (td, J = 7.6, 1.3 Hz, 1H), 7.35 (d, J = 3.7 Hz, 1H), 7.34 (d, J = 3.8 Hz, 1H), 7.24 (td, J = 7.7, 1.7 Hz, 1H). ¹³C NMR (151 MHz, chloroform-d) δ 177.69, 156.71, 151.88, 134.49, 130.55, 130.07, 129.79, 127.83, 122.63, 120.92, 113.08.

(E)-5-(2-styrylphenyl)furan-2-carbaldehyde (10s).

5-(2-Bromophenyl)furan-2-carbaldehyde (10q, 251.1 mg, 1.0 mmol), styrylboronic acid (222.0 mg, 1.5 mmol), Pd(PPh₃)₄ (57.8 mg, 0.05 mmol) and K₂CO₃ (414.6 mg, 3.0 mmol) were dissolved in toluene, EtOH and H₂O (20 mL, 10:3:2, v/v/v). The reaction mixture was purged with N₂ for three times and heated to reflux overnight. The reaction mixture was then filtered through celite and further purified by column to afford the title product. Yield: 83%. ¹H NMR (600 MHz, chloroform-d) δ 9.71 (s, 1H), 7.83 (dd, J = 7.7, 1.3 Hz, 1H), 7.69 (dd, J = 7.7, 1.2 Hz, 1H), 7.52 (dd, J = 8.0, 1.1 Hz, 2H), 7.46 (td, J = 7.8, 1.4 Hz, 1H), 7.45 (d, J = 16.5 Hz, 1H), 7.39 (td, J = 7.7, 1.6 Hz, 2H), 7.35 (d, J = 3.7 Hz, 1H), 7.30 (dd, J = 7.4, 1.1 Hz, 2H), 7.05 (d, J = 16.2 Hz, 1H), 6.71 (d, J = 3.7 Hz, 1H). ¹³C NMR (151 MHz, chloroform-d) δ 177.65, 158.59, 152.29, 137.27, 136.60, 131.95, 129.90, 129.00, 128.55, 128.21, 127.95, 127.87, 127.57, 127.31, 126.87, 112.78, 112.75.

5-(3,4-Dichlorophenyl)furan-2-carboxylic acid (11e).

Procedure B. Yield: 85%. ¹H NMR (600 MHz, DMSO-d₆) δ 7.36–7.43 (2H, m), 7.80–7.87 (2H, m), 8.12 (1H, d, J = 2.0 Hz), 13.31 (1H, s).

5-(3-(Trifluoromethyl)phenyl)furan-2-carboxylic acid (11f).

Procedure B. Yield: 97%. ¹H NMR (400 MHz, DMSO-d₆) δ 7.38 (2H, q, J = 3.6 Hz), 7.72–7.79 (2H, m), 8.12 (2H, d, J = 7.3 Hz), 13.27 (1H, s).

5-(3-Chlorophenyl)furan-2-carboxylic acid (11j).

Procedure B. Yield: 92%. ¹H NMR (400 MHz, DMSO-d₆) δ 7.26–7.37 (2H, m), 7.45– 7.56 (2H, m), 7.77–7.90 (2H, m).

5-(2-Ethylphenyl)furan-2-carboxylic acid (11h).

Procedure B. Yield: 47%. ¹H NMR (600 MHz, methanol-d₄) δ 1.09–1.14 (3H, m), 2.78 (2H, q, J = 7.4 Hz), 6.61–6.65 (1H, m), 7.14–7.20 (1H, m), 7.19–7.29 (3H, m), 7.54 (1H, d, J = 7.5 Hz).

5-(2-(N,N-Dimethylsulfamoyl)phenyl)furan-2-carboxylic acid (11j).

Procedure B. Yield: 89%. ¹H NMR (400 MHz, methanol-d₄) δ 2.74 (6H, s), 6.88–7.05 (1H, m), 7.28–7.43 (1H, m), 7.65–7.86 (3H, m), 8.04 (1H, d, J = 7.7 Hz).

5-(Naphthalen-1-yl)furan-2-carboxylic acid (11k).

Procedure B. Yield: 50%. ¹H NMR (400 MHz, chloroform-d) δ 6.81 (1H, d, J = 3.5 Hz), 7.41– 7.58 (4H, m), 7.76–7.92 (3H, m), 8.30–8.38 (1H, m).

5-(2,3-Dihydrobenzo[b][1,4]dioxin-5-yl)furan-2-carboxylic acid (11l).

Procedure B. Yield: 51%. ¹H NMR (400 MHz, methanol-d₄) δ 4.18 (4H, s), 6.66 (1H, d, J = 3.6 Hz), 6.79 (1H, d, J = 8.1 Hz), 7.14–7.22 (3H, m).

5-(2-Chlorophenyl)furan-2-carboxylic acid (11m).

Procedure B. Yield: 82%. ¹H NMR (400 MHz, methanol-d₄) δ 7.15 (1H, d, J = 3.7 Hz), 7.22 (1H, d, J = 3.7 Hz), 7.26 (1H, td, J = 7.7, 1.7 Hz), 7.33 (1H, td, J = 7.7, 1.3), 7.42 (1H, dd, J = 7.9, 1.2 Hz), 7.86–7.90 (1H, m).

5-(o-Tolyl)furan-2-carboxylic acid (11n).

Procedure B. Yield: 64%. ¹H NMR (400 MHz, DMSO-d₆) δ 2.51 (3H, s), 6.93 (1H, dd, J = 3.5, 1.8 Hz), 7.33–7.36 (4H, m), 7.72 (1H, dt, J = 5.0, 2.3 Hz), 13.09 (1H, s).

5-(2-(Benzyloxy)phenyl)furan-2-carboxylic acid (11o).

Procedure B. Yield: 97%. ¹H NMR (400 MHz, chloroform-d) δ 5.13 (2H, s), 6.95 (1H, d, J = 3.7 Hz), 6.96–7.04 (2H, m), 7.24–7.41 (7H, m), 7.99 (1H, dd, J = 7.8, 1.6 Hz).

5-(2-bromophenyl)furan-2-carboxylic acid (11q).

Procedure B. Yield: 67%. ¹H NMR (600 MHz, chloroform-d) δ 7.92 (dd, J = 7.9, 1.7 Hz, 1H), 7.69 (dd, J = 8.0, 1.3 Hz, 1H), 7.42 (td, J = 7.8, 1.2 Hz, 1H), 7.41 (d, J = 3.7 Hz, 1H), 7.29 (d, J = 3.7 Hz, 1H), 7.23 (ddd, J = 8.1, 7.4, 1.7 Hz, 1H). ¹³C NMR (151 MHz, chloroform-d) δ 161.34, 156.00, 142.69, 134.41, 130.27, 130.05, 130.01, 127.81, 121.42, 120.74, 112.74.

(E)-5-(2-styrylphenyl)furan-2-carboxylic acid (11s).

Procedure B. Yield: 83%. ¹H NMR (600 MHz, DMSO-d₆) δ 13.16 (s, 1H), 7.83 (d, J = 7.7 Hz, 1H), 7.73 (dd, J = 7.7, 1.5 Hz, 1H), 7.62 (dd, J = 7.6, 1.5 Hz, 2H), 7.59 (dd, J = 16.5, 1.3 Hz, 1H), 7.47 (ddd, J = 22.6, 7.5, 1.5 Hz, 1H), 7.43 (td, J = 7.5, 1.3 Hz, 1H), 7.39 (t, J = 7.6 Hz, 2H), 7.37 (d, J = 3.6 Hz, 1H), 7.30 (t, J = 7.3 Hz, 1H), 7.21 (d, J = 16.2 Hz, 1H), 6.85 (d, J = 3.6 Hz, 1H). ¹³C NMR (151 MHz, DMSO-d₆) δ 159.31, 155.64, 144.43, 137.02, 135.07, 131.13, 129.32, 128.78, 128.76, 128.12, 127.92, 127.59, 127.01, 126.70, 126.60, 119.43, 111.77.

(R)-2-Amino-3-(5-(4-chlorophenyl)furan-2-carboxamido)propanoic acid (8a).

Procedure C. Yield for two steps: 5%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.84 (t, J = 5.8 Hz, 1H), 7.92 (d, J = 8.4 Hz, 2H), 7.55 (d, J = 8.4 Hz, 2H), 7.21 (d, J = 3.7 Hz, 1H), 7.17 (d, J = 3.6 Hz, 1H), 3.78 (td, J = 11.2, 9.3, 4.8 Hz, 1H), 3.72 (q, J = 4.6 Hz, 1H), 3.64 (dt, J = 14.1, 6.9 Hz, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 169.35, 158.72, 153.83, 147.45, 133.56, 129.47, 128.65, 126.45, 116.53, 108.83, 53.63, 40.00. MS calcd for C₁₄H₁₃ClN₂O₄: 308.1, found: 309.5 [M+H]⁺.

(R)-2-Amino-3-(5-(4-methoxyphenyl)furan-2-carboxamido)propanoic acid (8b).

Procedure C. Yield for two steps: 5%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.74 (t, J = 6.0 Hz, 1H), 8.41 (s, 2H), 7.84 (d, J = 8.8 Hz, 2H), 7.18 (d, J = 3.5 Hz, 1H), 7.03 (d, J = 8.8 Hz, 2H), 6.96 (d, J = 3.6 Hz, 1H), 4.07 (dd, J = 7.0, 4.9 Hz, 1H), 3.81 (s, 3H), 3.80 – 3.67 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 169.69, 160.13, 159.05, 155.41, 146.34, 126.44, 122.58, 116.80, 114.83, 106.41, 55.75, 52.98, 39.96. MS calcd for C₁₅H₁₆N₂O₅: 304.1, found: 305.5 [M+H]⁺.

(R)-2-Amino-3-(5-phenylfuran-2-carboxamido)propanoic acid (8c).

Procedure C. Yield for two steps: 6%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.76 (t, J = 6.0 Hz, 1H), 8.38 (s, 2H), 7.96 – 7.87 (m, 2H), 7.48 (dd, J = 8.4, 6.9 Hz, 2H), 7.43 – 7.35 (m, 1H), 7.22 (d, J = 3.6 Hz, 1H), 7.13 (d, J = 3.6 Hz, 1H), 4.11 (dd, J = 6.8, 5.0 Hz, 1H), 3.82 – 3.68 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 169.66, 159.00, 155.14, 146.99, 129.74, 129.37, 129.16, 124.79, 116.73, 108.14, 52.83, 39.99. MS calcd for C₁₄H₁₄N₂O₄: 274.1, found: 275.5 [M+H]⁺.

(R)-2-Amino-3-(5-(2-(trifluoromethyl)phenyl)furan-2-carboxamido)propanoic acid (8d).

Procedure C. Yield for two steps: 10%. ¹H NMR (600 MHz, DMSO-d₆) δ 8.84 (t, J = 5.9 Hz, 1H), 8.61 (s, 2H), 7.99 (d, J = 7.8 Hz, 1H), 7.89 (d, J = 7.8 Hz, 1H), 7.80 (t, J = 7.6 Hz, 1H), 7.67 (t, J = 7.7 Hz, 1H), 7.38 (d, J = 3.6 Hz, 1H), 6.89 (d, J = 3.6 Hz, 1H), 4.09 (t, J = 5.7 Hz, 1H), 3.78 (t, J = 5.9 Hz, 2H). ¹³C NMR (151 MHz, DMSO-d₆) δ 169.50, 158.51, 151.64, 148.01, 133.29, 131.37, 130.07, 128.33, 127.25, 127.21, 127.18, 127.14, 127.00, 116.24, 112.50, 112.48, 52.73, 39.10. MS calcd for C₁₅H₁₃F₃N₂O₄: 342.1, found: 343.5 [M+H]⁺.

(R)-2-Amino-3-(5-(3,4-dichlorophenyl)furan-2-carboxamido)propanoic acid (8e).

Procedure C. Yield for two steps: 41%. ¹H NMR (600 MHz, methanol-d₄) δ 8.10 (d, J = 2.0 Hz, 1H), 7.78 (dd, J = 8.4, 2.0 Hz, 1H), 7.60 (d, J = 8.4 Hz, 1H), 7.27 (d, J = 3.6 Hz, 1H), 7.06 (d, J = 3.7 Hz, 1H), 4.17 (dd, J = 6.6, 3.9 Hz, 1H), 4.00 (dd, J = 14.7, 3.9 Hz, 1H), 3.89 (dd, J = 14.7, 6.7 Hz, 1H). ¹³C NMR (151 MHz, methanol-d₄) δ 168.62, 160.50, 153.73, 146.67, 132.78, 132.04, 130.77, 129.74, 125.95, 123.88, 116.83, 108.51, 53.73, 39.18. MS calcd for C₁₄H₁₂Cl₂N₂O₄: 342.0, found: 343.4 [M+H]⁺.

(R)-2-Amino-3-(5-(3-(trifluoromethyl)phenyl)furan-2-carboxamido)propanoic acid (8f).

Procedure C. Yield for two steps: 32%. ¹H NMR (600 MHz, methanol-d₄) δ 8.09 (s, 1H), 7.99 (d, J = 6.8 Hz, 1H), 7.54 (d, J = 7.3 Hz, 2H), 7.17 (d, J = 3.6 Hz, 1H), 6.99 (d, J = 3.6 Hz, 1H), 3.85 (dt, J = 11.9, 3.6 Hz, 2H), 3.76 (dt, J = 15.3, 7.1 Hz, 1H). ¹³C NMR (151 MHz, methanol-d₄) δ 160.60, 146.75, 131.43, 131.22, 131.00, 130.79, 130.46, 126.76, 124.96, 124.81, 124.79, 124.76, 124.73, 123.16, 121.36, 120.79, 116.76, 108.35, 54.65, 39.58. MS calcd for C₁₅H₁₃F₃N₂O₄: 342.1, found: 343.5 [M+H]⁺.

(R)-2-Amino-3-(5-(3-chlorophenyl)furan-2-carboxamido)propanoic acid (8g).

Procedure C. Yield for two steps: 46%. ¹H NMR (600 MHz, methanol-d₄) δ 7.91 (t, J = 1.9 Hz, 1H), 7.76 (dt, J = 7.7, 1.3 Hz, 1H), 7.41 (t, J = 7.9 Hz, 1H), 7.35 (ddd, J = 8.0, 2.1, 1.0 Hz, 1H), 7.26 (d, J = 3.6 Hz, 1H), 6.99 (d, J = 3.6 Hz, 1H), 4.28 (dd, J = 6.8, 4.0 Hz, 1H), 4.02 (dd, J = 14.7, 4.0 Hz, 1H), 3.91 (dd, J = 14.7, 6.9 Hz, 1H). ¹³C NMR (151 MHz, methanol-d₄) δ 168.57, 160.53, 154.72, 146.38, 134.66, 131.26, 130.12, 128.27, 124.03, 122.58, 116.85, 108.01, 53.45, 39.12. MS calcd for C₁₄H₁₃ClN₂O₄: 308.1, found: 309.4 [M+H]⁺.

(R)-2-Amino-3-(5-(2-ethylphenyl)furan-2-carboxamido)propanoic acid (8h).

Procedure C. Yield for two steps: 41%. ¹H NMR (600 MHz, methanol-d₄) δ 7.63 – 7.58 (m, 1H), 7.27 – 7.21 (m, 2H), 7.20 – 7.15 (m, 2H), 6.62 (d, J = 3.6 Hz, 1H), 3.89 (dd, J = 6.8, 3.8 Hz, 1H), 3.83 (dd, J = 14.6, 3.9 Hz, 1H), 3.75 (dd, J = 14.7, 6.9 Hz, 1H), 2.77 (q, J = 7.6 Hz, 2H), 1.11 (t, J = 7.5 Hz, 3H). ¹³C NMR (151 MHz, methanol-d₄) δ 169.58, 160.76, 156.16, 145.87, 141.89, 129.15, 128.97, 128.41, 128.39, 125.75, 116.38, 110.18, 54.66, 39.55, 26.58, 14.12. MS calcd for C₁₆H₁₈N₂O₄: 302.1, found: 303.5 [M+H]⁺.

(R)-2-Amino-3-(5-(2-(methoxymethyl)phenyl)furan-2-carboxamido)propanoic acid (8i).

Procedure C. Yield for two steps: 44%. ¹H NMR (400 MHz, methanol-d₄) δ 7.91 – 7.86 (m, 1H), 7.53 (dd, J = 7.1, 1.9 Hz, 1H), 7.46 – 7.39 (m, 2H), 7.30 (t, J = 3.5 Hz, 1H), 6.85 (d, J = 3.6 Hz, 1H), 4.60 (s, 2H), 4.26 (dd, J = 6.7, 4.0 Hz, 1H), 4.02 (dd, J = 14.7, 4.1 Hz, 1H), 3.91 (dd, J = 14.7, 6.7 Hz, 1H), 3.41 (s, 3H). ¹³C NMR (101 MHz, methanol-d₄) δ 168.59, 160.70, 154.99, 146.19, 134.83, 130.08, 129.06, 128.52, 128.01, 127.88, 116.61, 110.94, 72.62, 56.83, 53.43, 39.08. MS calcd for C₁₆H₁₈N₂O₅: 318.1, found: 319.5 [M+H]⁺.

(R)-2-Amino-3-(5-(2-(N,N-dimethylsulfamoyl)phenyl)furan-2-carboxamido)propanoic acid (8j).

Procedure C. Yield for two steps: 29%. ¹H NMR (600 MHz, methanol-d₄) δ 7.98 (d, J = 7.9 Hz, 1H), 7.81 (d, J = 7.6 Hz, 1H), 7.74 (t, J = 7.5 Hz, 1H), 7.68 (t, J = 7.7 Hz, 1H), 7.28 (d, J = 3.6 Hz, 1H), 6.98 (d, J = 3.6 Hz, 1H), 4.24 (dd, J = 6.6, 4.0 Hz, 1H), 4.00 (dd, J = 14.7, 3.9 Hz, 1H), 3.88 (dd, J = 14.7, 6.7 Hz, 1H), 2.74 (s, 6H). ¹³C NMR (151 MHz, methanol-d₄) δ 168.53, 160.54, 152.85, 146.68, 136.71, 132.42, 132.21, 129.50, 129.22, 128.88, 116.03, 113.16, 53.44, 39.05, 36.18. MS calcd for C₁₆H₁₉N₃O₆S: 381.1, found: 382.5 [M+H]⁺.

(R)-2-Amino-3-(5-(naphthalen-1-yl)furan-2-carboxamido)propanoic acid (8k).

Procedure C. Yield for two steps: 51%. ¹H NMR (400 MHz, methanol-d₄) δ 8.22 (d, J = 8.1 Hz, 1H), 7.84 (d, J = 7.9 Hz, 2H), 7.76 (d, J = 7.2 Hz, 1H), 7.45 (tt, J = 7.6, 4.1 Hz, 3H), 7.27 (d, J = 3.5 Hz, 1H), 6.84 (d, J = 3.6 Hz, 1H), 4.15 (dd, J = 6.7, 4.1 Hz, 1H), 3.90 (dd, J = 14.7, 4.0 Hz, 1H), 3.80 (dd, J = 15.1, 7.2 Hz, 1H). ¹³C NMR (101 MHz, methanol-d₄) δ 168.50, 160.77, 155.92, 146.43, 134.02, 130.18, 129.54, 128.40, 127.00, 126.84, 126.83, 125.92, 124.90, 124.55, 116.55, 111.26, 53.44, 39.08. MS calcd for C₁₈H₁₆N₂O₄: 324.1, found: 325.5 [M+H]⁺.

(R)-2-Amino-3-(5-(2,3-dihydrobenzo[b][1,4]dioxin-5-yl)furan-2-carboxamido) propanoic acid (8l).

Procedure C. Yield for two steps: 39%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.72 (t, J = 5.8 Hz, 1H), 7.45 (d, J = 2.1 Hz, 1H), 7.37 (dd, J = 8.4, 2.1 Hz, 1H), 7.16 (d, J = 3.6 Hz, 1H), 6.99 – 6.92 (m, 2H), 4.29 (s, 4H), 3.88 (dd, J = 7.3, 4.9 Hz, 1H), 3.77 – 3.61 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 169.40, 158.92, 154.95, 146.51, 144.48, 144.18, 123.29, 118.24, 118.04, 116.67, 113.61, 106.90, 64.74, 64.61, 53.41, 40.68. MS calcd for C₁₆H₁₆N₂O₆: 332.1, found: 333.5 [M+H]⁺.

(R)-2-Amino-3-(5-(2-chlorophenyl)furan-2-carboxamido)propanoic acid (8m).

Procedure C. Yield for two steps: 34%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.84 (t, J = 5.6 Hz, 1H), 8.11 (d, J = 7.8 Hz, 1H), 7.60 (d, J = 7.9 Hz, 1H), 7.50 (t, J = 7.6 Hz, 1H), 7.43 (t, J = 7.7 Hz, 1H), 7.26 (q, J = 3.8 Hz, 2H), 3.72 (dd, J = 9.6, 4.5 Hz, 2H), 3.67 – 3.62 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 169.19, 158.57, 151.07, 131.32, 130.44, 130.17,129.10, 128.08, 127.98, 116.12, 113.18, 53.67, 40.44. MS calcd for C₁₄H₁₃ClN₂O₄: 308.1, found: 309.5 [M+H]⁺.

(R)-2-Amino-3-(5-(o-tolyl)furan-2-carboxamido)propanoic acid (8n).

Procedure C. Yield for two steps: 22%. ¹H NMR (400 MHz, methanol-d₄) δ 7.86 (td, J = 7.4, 3.5 Hz, 1H), 7.31 (td, J = 8.7, 4.5 Hz, 4H), 6.84 – 6.78 (m, 1H), 4.25 (tt, J = 6.5, 3.8 Hz, 1H), 4.02 (dq, J = 15.0, 4.4 Hz, 1H), 3.92 (dq, J = 14.5, 6.9 Hz, 1H), 2.53 (s, 3H). ¹³C NMR (101 MHz, methanol-d₄) δ 168.58, 160.78, 156.12, 145.62, 135.31, 130.90, 128.83, 128.55, 127.43, 125.82, 116.54, 110.46, 53.56, 39.12, 20.47. MS calcd for C₁₅H₁₆N₂O₄: 288.1, found: 289.5 [M+H]⁺.

(R)-2-Amino-3-(5-(2-(benzyloxy)phenyl)furan-2-carboxamido)propanoic acid (8o).

Procedure C. Yield for two steps: 40%. ¹H NMR (400 MHz, methanol-d₄) δ 7.39 (dd, J = 7.7, 1.7 Hz, 1H), 7.33 – 7.15 (m, 8H), 6.99 (d, J = 7.4 Hz, 1H), 6.96 – 6.92 (m, 1H), 4.17 (dd, J = 6.9, 3.8 Hz, 1H), 3.96 (d, J = 4.7 Hz, 1H), 3.92 (d, J = 3.7 Hz, 1H), 3.85 (s, 2H). ¹³C NMR (101 MHz, methanol-d₄) δ 168.56, 160.95, 155.19, 150.67, 145.32, 140.31, 130.71, 130.48, 128.31, 128.00, 125.76, 124.21, 119.22, 117.81, 117.15, 115.79, 53.60, 39.00, 31.08. MS calcd for C₂₁H₂₀N₂O₅: 380.1, found: 381.6 [M+H]⁺.

(R)-2-amino-3-(5-(4-chloro-2-nitrophenyl)furan-2-carboxamido)propanoic acid (8p).

Procedure C. Yield for two steps: 42%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.79 (t, J = 6.0 Hz, 1H), 8.43 (s, 3H), 8.16 (d, J = 2.1 Hz, 1H), 8.02 (d, J = 8.5 Hz, 1H), 7.86 (dd, J = 8.5, 2.2 Hz, 1H), 7.26 (d, J = 3.7 Hz, 1H), 6.90 (d, J = 3.7 Hz, 1H), 4.10 (t, J = 5.7 Hz, 1H), 3.75 (ddt, J = 25.3, 14.3, 6.3 Hz, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 169.61, 158.53, 148.66, 148.60, 147.83, 134.65, 133.06, 131.33, 124.56, 121.32, 116.33, 112.36, 52.67, 39.29. MS calcd for C₁₄H₁₂ClN₃O₄: 353.0, found: 354.4 [M+H]⁺.

(R)-2-Amino-3-(5-(2-bromophenyl)furan-2-carboxamido)propanoic acid (8q).

Procedure C. Yield for two steps: 3%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.77 (t, J = 5.8 Hz, 1H), 8.02 (dd, J = 7.9, 1.7 Hz, 1H), 7.79 (dd, J = 8.0, 1.2 Hz, 1H), 7.54 (td, J = 7.6, 1.3 Hz, 1H), 7.35 (td, J = 7.7, 1.7 Hz, 1H), 7.30 (d, J = 3.7 Hz, 1H), 7.26 (d, J = 3.6 Hz, 1H), 3.82 (dd, J = 7.5, 4.9 Hz, 1H), 3.73 (dt, J = 13.9, 5.0 Hz, 1H), 3.66 (dt, J = 14.0, 7.0 Hz, 1H). ¹³C NMR (151 MHz, DMSO-d₆) δ 169.30, 158.64, 152.34, 147.43, 134.65, 130.77, 129.99, 129.98, 128.51, 119.92, 115.92, 112.83, 53.46, 40.00. MS calcd for C₁₄H₁₃BrN₂O₄: 352.0, found: 353.5 [M+H]⁺.

(R,E)-2-Amino-3-(5-(2-(3-fluorostyryl)phenyl)furan-2-carboxamido)propanoic acid (8r).

Procedure C. Yield for two steps: 4%. ¹H NMR (400 MHz, DMSO-d₆) δ 8.80 (t, J = 5.8 Hz, 1H), 7.90 – 7.83 (m, 1H), 7.81 – 7.74 (m, 1H), 7.59 (d, J = 16.2 Hz, 1H), 7.46 (dd, J = 8.4, 3.8 Hz, 5H), 7.29 (d, J = 3.6 Hz, 1H), 7.20 (d, J = 16.2 Hz, 1H), 7.13 (tt, J = 6.7, 2.7 Hz, 1H), 6.75 (d, J = 3.6 Hz, 1H), 3.92 – 3.88 (m, 1H), 3.78 (dt, J = 15.8, 7.9 Hz, 1H), 3.68 (dt, J = 13.7, 6.7 Hz, 1H). ¹³C NMR (151 MHz, DMSO-d₆) δ 169.68, 164.36, 161.84, 158.90, 153.86, 147.42, 140.09, 140.01, 135.15, 131.18, 131.10, 130.61, 130.59, 129.50, 128.72, 128.57, 128.42, 128.39, 127.71, 123.49, 123.47, 116.31, 115.12, 114.91, 113.54, 113.33, 112.55, 53.41, 39.97. MS calcd for C₂₂H₁₉FN₂O₄: 394.1, found: 395.6 [M+H]⁺.

(R,E)-2-Amino-3-(5-(2-styrylphenyl)furan-2-carboxamido)propanoic acid (8s).

Procedure C. Yield for two steps: 36%. ¹H NMR (600 MHz, DMSO-d₆) δ 8.73 (t, J = 5.9 Hz, 1H), 7.84 (dd, J = 7.4, 1.8 Hz, 1H), 7.80 (dd, J = 7.5, 1.6 Hz, 1H), 7.61 (dd, J = 8.5, 1.2 Hz, 2H), 7.52 (d, J = 16.2 Hz, 1H), 7.46 (dt, J = 7.5, 2.8, 1.6 Hz, 2H), 7.40 (td, J = 7.7, 1.5 Hz, 2H), 7.30 (tt, J = 7.2, 2.4, 1.3 Hz, 1H), 7.28 (d, J = 3.5 Hz, 1H), 7.20 (d, J = 16.2 Hz, 1H), 6.75 (d, J = 3.5 Hz, 1H), 3.96 (dd, J = 7.1, 5.0 Hz, 1H), 3.75 (dt, J = 14.1, 5.2 Hz, 1H), 3.69 (dt, J = 13.9, 6.8 Hz, 1H). ¹³C NMR (151 MHz, DMSO-d₆) δ 169.05, 158.47, 153.58, 146.86, 136.92, 135.11, 131.34, 129.11, 128.83, 128.01, 127.97, 127.83, 127.09, 126.75, 126.57, 115.92, 112.01, 52.70, 40.06. MS calcd for C₂₂H₂₀N₂O₄: 376.1, found: 377.6 [M+H]⁺.

Pharmacological evaluation

DNA Constructs.

cDNAs encoding GluN1–1a (Genbank accession number U11418 and U08261), GluN2A (D13211), GluN2B (U11419), GluN2C (M91563), GluN2D (L31611), GluN3A (U29873), and GluN3B (AF440691) were generously provided by Dr. S. Heinemann (Salk Institute, La Jolla, CA), Dr. S. Nakanishi (Osaka Bioscience Institute, Osaka, Japan), Dr. P. Seeburg (University of Heidelberg, Germany), and Dr. D. Zhang (Sanford-Burnham Medical Research Institute, La Jolla, CA). The GluN2B open reading frame was edited without changing the amino acid sequence to remove a T7 RNA polymerase termination site.⁴⁴

Two-Electrode Voltage-Clamp Recordings.

For expression in Xenopus laevis oocytes, cDNAs were linearized using restriction enzymes and used as templates to synthesize cRNA using the mMessage mMachine kit (Ambion, Life Technologies, Paisley, UK). Xenopus oocytes were obtained from Rob Weymouth (Xenopus 1, Dexter, MI). The oocytes were injected with cRNAs encoding GluN1 and GluN2 in a 1:2 ratio and maintained as previously described.⁴⁵

Two-electrode voltage-clamp recordings were performed 2–4 days following cRNA injection at room temperature and at a holding potential of −40 mV essentially as previously described.⁴⁵ Oocytes were perfused with extracellular recording solution comprised of 90 mM NaCl, 1 mM KCl, 10 mM HEPES, 0.5 mM BaCl₂ and 0.01 mM EDTA (pH 7.4 with NaOH). Oocytes expressing GluN1/2A or GluN1/2B were injected with 50 nL and 30 nL, respectively, of 50 mM BAPTA approximately 10–30 min before recordings to prevent activity-dependent increases in response amplitude.⁴⁶ Compounds were dissolved in DMSO to make 20–100 mM stock solutions and the concentration of DMSO was kept constant (< 0.5%) in all recording solutions.

Data were analyzed using GraphPad Prism (GraphPad Software, La Jolla, CA). Agonist concentration-response data for individual oocytes were fitted to the Hill equation, I = R_max/(1+10^((logEC₅₀-log[A])*nH)), where R_max is the maximum current in response to the agonist relative to the maximal response to 100 μM Gly in the same recording, nH is the Hill slope, [A] is the agonist concentration, and EC₅₀ is the agonist concentration that produces half-maximum response.

Molecular modeling

Marvin and JChem for Office were used for drawing, characterizing, and organizing a subset of the chemical structures, Marvin 20.15.0 and JChem 20.14.0.668, ChemAxon (https://www.chemaxon.com). All modeling was carried out in the Schrodinger Maestro program, Maestro version 11.9.011, Release 2019–1. The GluN1/2A ABD crystal structure (PDB ID: 5I57) was prepared using the Protein Preparation Wizard with default parameters. The protein bond order was reassigned with CCD database. Protein hydrogens were added, all H₂O molecules were deleted, and the protonation state for the ligands and receptors was set to pH 7.0. Restrained minimization was carried out with OPLS3e force field, and the resulting protein structure was used for next step induced-fit docking. Compounds 8h, 8k and 8p were prepared by LigPrep with default parameters. Ligands were generated at pH 7.0 with OPLS3e force field to neutralize the molecules. For each ligand, the highest scoring conformation in LigPrep was used for docking.

Induced-fit docking of compound 8h was performed with default parameters. Standard protocol was used and up to 20 docking poses were generated. The Gly binding site in the GluN1/2A ABD complex (PDB ID: 5I57) was identified as the binding box center and the box size was set as similar in size to 8h. The Glide redocking precision was set as XP Glide Redocking Precision and the other parameters was set as default. The docking model of 8h was used for the induced-fit docking of 8k and 8p with the box center as 8h and the same docking parameters. The binding model of each compound was selected based on the overlay with Gly and the IFDScore. All docking results were treated with energy minimization.

ABBREVIATIONS USED

AMPA

α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

ABD

agonist binding domain

Boc

tert-butyloxycarbonyl

DCM

dichloromethane

DIPEA

N, N-diisopropylethylamine

d-Ser

, d-serine

EDCI

N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride

EtOH

ethanol

Glu

glutamate

Gly

glycine

HOBt

hydroxybenzotriazole

NMDA

N-methyl-d-aspartic acid

Pd(PPh₃)₄

Tetrakis(triphenylphosphine)palladium

prep-HPLC

preparative-high performance liquid chromatography

rt

room temperature

TEVC

two-electrode voltage-clamp