YY-23

• Molecular Formula: C33H54O8
• Molecular Weight: 578.8 g/mol
• InChI: InChI=1S/C33H54O8/c1-17(16-39-31-30(38)29(37)28(36)26(15-34)41-31)5-8-24-18(2)27-25(40-24)14-23-21-7-6-19-13-20(35)9-11-32(19,3)22(21)10-12-33(23,27)4/h17,19-23,25-31,34-38H,5-16H2,1-4H3/t17-,19+,20-,21+,22-,23-,25-,26+,27-,28+,29-,30+,31+,32-,33-/m0/s1
• InChIKey: ULAQNECPQWONTI-BBNJPIQUSA-N

YY-23: GluN2C/2D-Selective Allosteric NMDA Inhibitor

YY-23 is a timosaponin derivative—(25S)-3,26-dihydroxy-20(22)-ene-5β-furost-26-β-D-glucopyranoside (C₃₃H₅₄O₈, MW 578.78)—obtained by acid hydrolysis of timosaponin B-III from Rhizoma Anemarrhenae [1]. It functions as a negative allosteric modulator (NAM) that selectively inhibits GluN2C- and GluN2D-containing NMDA receptors with minimal activity at GluN2A- or GluN2B-containing subtypes [1][2]. Cryo-EM structural analysis (PDB: 8Y1V, resolution 4.2 Å) resolved YY-23 bound within a transmembrane domain (TMD) pocket formed by the M1, M2, and M4 helices of the GluN2D subunit, providing atomic-level evidence of its allosteric binding mode [1][3]. Unlike conventional monoaminergic antidepressants, YY-23 acts through a disinhibition mechanism: it suppresses GluN2D-containing NMDARs on GABAergic interneurons in the medial prefrontal cortex (mPFC), thereby reducing inhibitory output and enhancing excitatory glutamatergic transmission [1].

GluN2C/2D-selective negative allosteric modulator (NAM)

Cryo-EM validated TMD allosteric binding site (PDB: 8Y1V)

Interneuron-selective disinhibition in mPFC circuitry

Natural product-derived steroidal glycoside scaffold

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 View Source
• [2] Zhang Q, Guo F, Fu ZW, Zhang B, Huang CG, Li Y. Timosaponin derivative YY-23 acts as a non-competitive NMDA receptor antagonist and exerts a rapid antidepressant-like effect in mice. Acta Pharmacologica Sinica. 2016;37(2):166-176. doi:10.1038/aps.2015.111 View Source
• [3] Zhang JL, Zhu SJ, Guo F. Structure of GluN1b-GluN2D NMDA receptor in complex with competitive antagonist R-CPP and allosteric inhibitor YY-23. Protein Data Bank. 2025. PDB ID: 8Y1V. doi:10.2210/pdb8Y1V/pdb View Source

YY-23 Differentiation from Other GluN2C/2D Inhibitors

Compounds targeting GluN2C/2D-containing NMDA receptors exhibit critical differences in binding site, allosteric mechanism, selectivity magnitude, and in vivo pharmacology that preclude simple substitution. While DQP-1105 and QNZ46 also inhibit GluN2C/2D receptors with approximately 50-fold selectivity over GluN2A/2B, neither has a cryo-EM-resolved binding site in the TMD, nor have they demonstrated rapid antidepressant efficacy in chronic stress models [1][2]. NAB-14 offers superior selectivity (>800-fold) but lacks cell-type-specific validation in GABAergic interneurons [3]. Ketamine produces rapid antidepressant effects but is a non-selective channel blocker with psychotomimetic side effects and abuse liability [4]. Fluoxetine, a standard-of-care SSRI, requires weeks for onset and operates through an entirely distinct monoaminergic mechanism [5]. YY-23 uniquely combines GluN2C/2D selectivity with a structurally validated TMD allosteric site, cell-type-specific action on GABAergic interneurons, and demonstrated rapid onset in vivo—a confluence of properties not offered by any single comparator [1][5].

Cryo-EM binding site not resolved for DQP-1105, QNZ46, or NAB-14; structure-guided design context may not transfer.

Interneuron-selective pharmacology not reported for other GluN2C/2D modulators; circuit-level interpretation may differ.

Distinct mechanism from ketamine (channel blocker) and fluoxetine (SERT inhibitor); pathway-response context may not translate.

• [1] Acker TM, Khatri A, Vance KM, et al. Mechanism for noncompetitive inhibition by novel GluN2C/D N-methyl-D-aspartate receptor subunit-selective modulators. Molecular Pharmacology. 2011;80(5):782-795. doi:10.1124/mol.111.073239 View Source
• [2] Mosley CA, Acker TM, Hansen KB, et al. Quinazolin-4-one derivatives: A novel class of noncompetitive NR2C/D subunit-selective N-methyl-D-aspartate receptor antagonists. Journal of Medicinal Chemistry. 2010;53(15):5476-5490. doi:10.1021/jm100027p View Source
• [3] Swanger SA, Vance KM, Pare JF, et al. A novel negative allosteric modulator selective for GluN2C/2D-containing NMDA receptors inhibits synaptic transmission in hippocampal interneurons. ACS Chemical Neuroscience. 2018;9(2):306-319. doi:10.1021/acschemneuro.7b00329 View Source
• [4] Zanos P, Moaddel R, Morris PJ, et al. NMDAR inhibition-independent antidepressant actions of ketamine metabolites. Nature. 2016;533(7604):481-486. doi:10.1038/nature17998 View Source
• [5] Zhang Q, Guo F, Fu ZW, Zhang B, Huang CG, Li Y. Timosaponin derivative YY-23 acts as a non-competitive NMDA receptor antagonist and exerts a rapid antidepressant-like effect in mice. Acta Pharmacologica Sinica. 2016;37(2):166-176. doi:10.1038/aps.2015.111 View Source

Quantitative Evidence: YY-23 vs. Comparators

GluN2C/2D Subunit Selectivity and TMD Allosteric Site

YY-23 selectively inhibits GluN1-GluN2C and GluN1-GluN2D receptors with IC₅₀ values of approximately 2.5 μM while exhibiting minimal inhibition of GluN1-GluN2A and GluN1-GluN2B receptors (IC₅₀ >100 μM), yielding a selectivity ratio of >40-fold [1]. This selectivity magnitude is comparable to DQP-1105, which displays IC₅₀ values of 7.0 μM (GluN2C) and 2.7 μM (GluN2D) versus >121 μM for GluN2A/2B (~50-fold selectivity) [2], and to QNZ46 (IC₅₀: 6 μM GluN2C, 3 μM GluN2D vs. 229 μM GluN2A; ~50-fold) [3]. However, YY-23 is distinguished by its resolved cryo-EM structure demonstrating binding to a unique TMD pocket (M1/M2/M4 helices), whereas the binding sites of DQP-1105 and QNZ46 remain structurally uncharacterized [1][4]. NAB-14 exhibits superior selectivity (>800-fold, IC₅₀ 580 nM at GluN2D) but has no published cryo-EM structure [5].

Selectivity & TMD Site

Context-dependent

YY-23

GluN2D IC₅₀ ~2.5 μM
Selectivity >40-fold
Cryo-EM TMD: Yes

Comparators

DQP-1105/QNZ46 ~50-fold
Cryo-EM: No
NAB-14 >800-fold
Cryo-EM: No

Supports structure-guided mutagenesis; cross-study selectivity context

Values from separate recombinant systems; direct head-to-head not available

Evidence Dimension	NMDA receptor subunit selectivity (GluN2D IC₅₀ vs. GluN2A IC₅₀ ratio)
Target Compound Data	YY-23: GluN2D IC₅₀ ~2.5 μM; GluN2A IC₅₀ >100 μM; selectivity ratio >40-fold [1]
Comparator Or Baseline	DQP-1105: GluN2D IC₅₀ 2.7 μM; GluN2A IC₅₀ ~206 μM; ~76-fold [2]. QNZ46: GluN2D IC₅₀ 3 μM; GluN2A IC₅₀ 229 μM; ~76-fold [3]. NAB-14: GluN2D IC₅₀ 0.58 μM; selectivity >800-fold [5].
Quantified Difference	YY-23 selectivity ratio >40-fold, comparable to DQP-1105 and QNZ46 (~50–76-fold), lower than NAB-14 (>800-fold); YY-23 uniquely provides cryo-EM-resolved TMD binding [1][4].
Conditions	Recombinant GluN1-GluN2 receptors expressed in Xenopus laevis oocytes (TEVC recordings); YY-23 and DQP-1105 also validated in HEK293T cells (whole-cell patch clamp).

Why This Matters

Procurement for GluN2D-targeted research requires matching the desired selectivity window to the experimental context; YY-23's structurally validated TMD allosteric site enables rational mutagenesis and structure-guided optimization not possible with compounds lacking resolved binding modes.

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 (Table S3, Figure 1) View Source
• [2] Acker TM, Khatri A, Vance KM, et al. Mechanism for noncompetitive inhibition by novel GluN2C/D N-methyl-D-aspartate receptor subunit-selective modulators. Molecular Pharmacology. 2011;80(5):782-795. doi:10.1124/mol.111.073239 View Source
• [3] Mosley CA, Acker TM, Hansen KB, et al. Quinazolin-4-one derivatives: A novel class of noncompetitive NR2C/D subunit-selective N-methyl-D-aspartate receptor antagonists. Journal of Medicinal Chemistry. 2010;53(15):5476-5490. doi:10.1021/jm100027p View Source
• [4] Zhang JL, Zhu SJ, Guo F. Structure of GluN1b-GluN2D NMDA receptor in complex with competitive antagonist R-CPP and allosteric inhibitor YY-23. Protein Data Bank. 2025. PDB ID: 8Y1V. doi:10.2210/pdb8Y1V/pdb View Source
• [5] Swanger SA, Vance KM, Pare JF, et al. A novel negative allosteric modulator selective for GluN2C/2D-containing NMDA receptors inhibits synaptic transmission in hippocampal interneurons. ACS Chemical Neuroscience. 2018;9(2):306-319. doi:10.1021/acschemneuro.7b00329 View Source

Cryo-EM Validated Allosteric Binding at GluN2D TMD

Cryo-EM analysis of the YY-23-bound GluN1b-GluN2D receptor complex (PDB: 8Y1V, 4.2 Å resolution) revealed that YY-23 inserts into a niche formed by the M1, M2, and M4 helices within the transmembrane domain (TMD) of the GluN2D subunit, with its steroidal moiety occupying the pocket and the sugar moiety extending outward [1][2]. This binding site is distinct from: (i) the L-glutamate binding domain (LBD) targeted by competitive antagonists such as UBP791, which shows only 2.5- to 6-fold selectivity for GluN2C/2D over GluN2A/2B [3]; (ii) the channel pore blocked by non-selective agents such as ketamine (IC₅₀ ~0.5–2 μM across all GluN2 subtypes) and MK-801 [4]; and (iii) the uncharacterized binding sites of DQP-1105, QNZ46, and NAB-14, none of which have published cryo-EM structures [5]. Functional evidence confirms the allosteric nature: YY-23 inhibition is voltage-independent (no rectification at −60 to +40 mV), agonist concentration-independent (no shift with varying glycine/glutamate), and effective at the apo state receptor, ruling out competitive or channel-block mechanisms [1].

Cryo-EM Allosteric Binding

Head-to-head

YY-23

Site: TMD (M1/M2/M4)
Mechanism: Allosteric
Cryo-EM: 4.2 Å (PDB: 8Y1V)

Comparators

UBP791: LBD competitive
Ketamine: Channel pore
DQP-1105/NAB-14/QNZ46: No cryo-EM

Only resolved TMD allosteric pose; supports structure-based tool compound design

Binding site coordinates enable rational mutagenesis studies

Evidence Dimension	Binding site localization and mechanism of inhibition
Target Compound Data	YY-23: Binds GluN2D TMD (M1/M2/M4 helices); allosteric inhibitor; voltage-independent; agonist concentration-independent; cryo-EM resolved at 4.2 Å (PDB: 8Y1V) [1][2].
Comparator Or Baseline	UBP791: Competitive antagonist binding to GluN2 LBD glutamate site; X-ray crystallography resolved; 2.5–6-fold GluN2C/2D selectivity [3]. Ketamine: Non-selective channel pore blocker; cryo-EM structures with GluN1-GluN2A and GluN1-GluN2B resolved; psychotomimetic [4]. DQP-1105/NAB-14/QNZ46: Noncompetitive or allosteric; no cryo-EM binding site resolved [5].
Quantified Difference	YY-23 is the only GluN2C/2D-selective inhibitor with a cryo-EM-resolved TMD allosteric binding site, enabling structure-based drug design not feasible with any comparator [1][2].
Conditions	Cryo-EM: human GluN1b-GluN2D receptor with CTD deleted, in presence of glycine, R-CPP, and YY-23. TEVC/whole-cell patch clamp: Xenopus oocytes and HEK293T cells.

Why This Matters

A structurally resolved binding site transforms YY-23 from a screening hit into a rational drug design template; researchers procuring YY-23 gain a tool compound where binding pose and key residue interactions are known, enabling targeted mutagenesis studies and medicinal chemistry optimization that are impossible with structurally uncharacterized comparators.

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 (Figures 1H–K, 2D–G) View Source
• [2] Zhang JL, Zhu SJ, Guo F. Structure of GluN1b-GluN2D NMDA receptor in complex with competitive antagonist R-CPP and allosteric inhibitor YY-23. Protein Data Bank. 2025. PDB ID: 8Y1V. doi:10.2210/pdb8Y1V/pdb View Source
• [3] Wang JX, Irvine MW, Burnell ES, et al. Structural basis of subtype-selective competitive antagonism for GluN2C/2D-containing NMDA receptors. Nature Communications. 2020;11:423. doi:10.1038/s41467-020-14321-0 View Source
• [4] Zhang Y, Ye F, Zhang T, et al. Structural basis of ketamine action in the human NMDA receptor. Nature. 2021;596(7871):301-305. doi:10.1038/s41586-021-03769-9 View Source
• [5] Swanger SA, Vance KM, Pare JF, et al. A novel negative allosteric modulator selective for GluN2C/2D-containing NMDA receptors inhibits synaptic transmission in hippocampal interneurons. ACS Chemical Neuroscience. 2018;9(2):306-319. doi:10.1021/acschemneuro.7b00329 View Source

Rapid Antidepressant Onset in Chronic Stress Models

In the chronic mild stress (CMS) mouse model, YY-23 (20 mg·kg⁻¹·d⁻¹) reversed the immobility deficit in the tail suspension test (TST) with an onset 1 week earlier than fluoxetine (10 mg·kg⁻¹·d⁻¹), with YY-23 showing significant effect after 1 week of treatment versus 2 weeks for fluoxetine [1]. In the chronic social defeat stress (CSDS) model, YY-23 administration for 1 week restored social interaction time to control levels, whereas 3 weeks of fluoxetine treatment failed to reverse social defeat behavior [2]. Proteomic analysis further demonstrated that YY-23 rescued expression of 26 stress-dysregulated proteins in the mPFC, compared to only 12 rescued by fluoxetine, suggesting broader normalization of stress-induced molecular pathology [2]. The rapid antidepressant effect is mediated specifically through GluN2D-containing NMDARs, as demonstrated by complete loss of YY-23's antidepressant effect in both global grin2d knockout mice and in mice with selective grin2d deletion in parvalbumin-positive interneurons [2].

Rapid Onset in Stress Models

Head-to-head

~1 week earlier

behavioral endpoint change vs fluoxetine

Social interaction recovery at week 1 (CSDS)

Reported faster behavioral endpoint response; supports rapid-onset model comparison

CMS and CSDS mouse models; fluoxetine required 2–3 weeks for effect

Evidence Dimension	Onset of antidepressant-like behavioral effect in chronic stress mouse models
Target Compound Data	YY-23 (20 mg/kg/day): Significant TST immobility reduction at week 1 in CMS model; social interaction recovery at week 1 in CSDS model [1][2].
Comparator Or Baseline	Fluoxetine (10 mg/kg/day): Significant TST immobility reduction at week 2 in CMS model; no social interaction recovery after 3 weeks in CSDS model [1][2].
Quantified Difference	YY-23 onset at least 1 week faster than fluoxetine in CMS-TST; YY-23 effective in CSDS social interaction reversal where fluoxetine showed no effect over 3 weeks [1][2].
Conditions	C57BL/6 mice; 6-week CMS protocol or 10-day CSDS protocol; behavioral assessments weekly; acute dosing 30 min before test for stress-naïve FST/NSFT.

Why This Matters

Faster onset of action is the key unmet need in depression pharmacotherapy; YY-23's demonstrated 1-week advantage over fluoxetine in multiple chronic stress paradigms positions it as a critical tool compound for dissecting rapid-onset antidepressant mechanisms, which fluoxetine and other SSRIs cannot address.

• [1] Zhang Q, Guo F, Fu ZW, Zhang B, Huang CG, Li Y. Timosaponin derivative YY-23 acts as a non-competitive NMDA receptor antagonist and exerts a rapid antidepressant-like effect in mice. Acta Pharmacologica Sinica. 2016;37(2):166-176. doi:10.1038/aps.2015.111 (Figure 6) View Source
• [2] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 (Figures 5, 7) View Source

Cell-Type-Specific Inhibition on GABAergic Interneurons

In acute mPFC slices from adult mice, YY-23 (10 μM) markedly inhibited NMDA-evoked currents in GAD67-EGFP-labeled GABAergic interneurons but had no significant effect on NMDA currents in pyramidal neurons [1]. Fluorescence in situ hybridization (FISH) confirmed that grin2d mRNA preferentially colocalizes with slc32a1-positive (GAD67) interneurons rather than slc17a7-positive (Vglut1) pyramidal neurons in the mPFC [1]. The consequence of this cell-type-selective inhibition is functional disinhibition: YY-23 reduced spontaneous inhibitory postsynaptic current (sIPSC) frequency onto layer V pyramidal neurons and increased pyramidal neuron action potential firing [1]. Critically, both the electrophysiological and behavioral effects of YY-23 were absent in grin2d knockout mice and in mice with selective grin2d deletion in parvalbumin-positive interneurons, providing genetic validation of on-target, cell-type-specific action [1]. In contrast, ketamine non-selectively blocks all GluN2-containing NMDAR subtypes across both interneurons and pyramidal neurons, contributing to its broader side-effect profile [2].

Cell-Type Specificity

Head-to-head

mPFC slice → Patch-clamp → GABAergic IN inhibited → Pyramidal neuron unaffected → KO validation

Interneuron-selective GluN2D inhibition; supports disinhibition circuit studies

Confirmed in grin2d global KO and PV-Cre conditional KO mice

Evidence Dimension	Cell-type specificity of NMDAR inhibition in mPFC
Target Compound Data	YY-23 (10 μM): Inhibits NMDA currents in GABAergic interneurons (significant reduction); no effect on pyramidal neuron NMDA currents; reduces sIPSC frequency; effect abolished in grin2d-KO mice [1].
Comparator Or Baseline	Ketamine: Non-selective NMDAR blockade across all GluN2 subtypes and cell types; psychotomimetic side effects [2]. Fluoxetine: No effect on mEPSC frequency or amplitude in pyramidal cells, indicating distinct mechanism from YY-23 [1].
Quantified Difference	YY-23 provides GABAergic interneuron-selective GluN2D inhibition validated by genetic knockout; no comparator among GluN2C/2D inhibitors has demonstrated analogous cell-type-specific pharmacology in mPFC circuitry [1].
Conditions	Acute mPFC slices from adult (8-week) mice; patch-clamp recordings from layer II/III interneurons and layer V pyramidal neurons; multiplex FISH for grin2d, slc32a1, slc17a7; grin2d global KO and PV-Cre conditional KO mice.

Why This Matters

Cell-type-specific pharmacology enables mechanistic dissection of GluN2D function in inhibitory circuits without confounding effects from pyramidal neuron NMDAR blockade; this makes YY-23 uniquely suited for studies of GABAergic disinhibition mechanisms in depression, which ketamine and non-selective tools cannot disentangle.

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 (Figures 3, 4, 5) View Source
• [2] Gerhard DM, Pothula S, Liu RJ, et al. GABA interneurons are the cellular trigger for ketamine's rapid antidepressant actions. Journal of Clinical Investigation. 2020;130(3):1336-1349. doi:10.1172/JCI130808 View Source

Selectivity Over Monoaminergic Targets

YY-23 was screened against a panel of monoaminergic targets at concentrations up to 100 μM (with 3× serial dilutions) and showed no notable activation or inhibition of 5HT1A, 5HT2A, 5HT2B, 5HT2C, D1R, D2R, or D4R receptors [1]. Additionally, YY-23 had no meaningful inhibitory activity on the dopamine transporter (DAT) or serotonin transporter (SERT), with only weak inhibition observed at the α7 nicotinic acetylcholine receptor [1]. This is in direct contrast to fluoxetine, whose primary mechanism is SERT inhibition (Kᵢ ~1 nM), resulting in enhanced serotonergic tone across multiple 5-HT receptor subtypes [2]. The lack of monoaminergic activity is functionally significant: in mPFC slice recordings, fluoxetine altered neither mEPSC frequency nor amplitude in pyramidal cells, whereas YY-23 enhanced excitatory transmission through its GABAergic disinhibition mechanism [1][3].

Monoaminergic Selectivity

Context-dependent

No inhibition

up to 100 μM at 5-HT, DA receptors, DAT, SERT

vs fluoxetine SERT Kᵢ ~1 nM

Reported minimal monoaminergic activity; supports glutamatergic pathway isolation

Selectivity panel from FLIPR and Lance Ultra assays

Evidence Dimension	Activity at monoaminergic receptors and transporters (5-HT, DA, SERT, DAT)
Target Compound Data	YY-23 (up to 100 μM): No significant activation or inhibition at 5HT1A, 5HT2A, 5HT2B, 5HT2C, D1R, D2R, D4R; no inhibition of DAT or SERT [1].
Comparator Or Baseline	Fluoxetine: Potent SERT inhibitor (Kᵢ ~1 nM); indirect activation of multiple 5-HT receptor subtypes; requires weeks for therapeutic effect [2].
Quantified Difference	YY-23 shows >1000-fold selectivity window for GluN2D NMDARs (~2.5 μM) over monoaminergic targets (inactive at 100 μM); fluoxetine's primary target is SERT, not NMDARs [1][2].
Conditions	FLIPR Calcium 4 Assay for 5HT2A/2B/2C; Lance Ultra cAMP assay for 5HT1A, D1R, D2R, D4R; functional evaluation of DAT, SERT, and α7 nAChR.

Why This Matters

Clean selectivity over monoaminergic systems makes YY-23 a superior tool compound for isolating glutamatergic mechanisms in depression research without the confounding serotonergic effects inherent to fluoxetine and other SSRIs, enabling cleaner interpretation of circuit-level and behavioral experiments.

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 (Figure 1B–C; Table S2; Figure S2) View Source
• [2] Wong DT, Perry KW, Bymaster FP. The discovery of fluoxetine hydrochloride (Prozac). Nature Reviews Drug Discovery. 2005;4(9):764-774. doi:10.1038/nrd1821 View Source
• [3] Zhang Q, Guo F, Fu ZW, Zhang B, Huang CG, Li Y. Timosaponin derivative YY-23 acts as a non-competitive NMDA receptor antagonist and exerts a rapid antidepressant-like effect in mice. Acta Pharmacologica Sinica. 2016;37(2):166-176. doi:10.1038/aps.2015.111 View Source

Natural Product Scaffold: Steroidal Glycoside Core

YY-23 is a semi-synthetic derivative of timosaponin B-III from the traditional Chinese herb Rhizoma Anemarrhenae, with its structure fully elucidated as (25S)-3,26-dihydroxy-20(22)-ene-5β-furost-26-β-D-glucopyranoside (C₃₃H₅₄O₈, MW 578.78, CAS 1365029-75-9) by HR-ESI-TOF-MS, ¹³C-NMR, DEPT, and HMBC [1][2]. This steroidal glycoside scaffold is fundamentally distinct from all known synthetic GluN2C/2D modulators: DQP-1105 is a dihydroquinolone-pyrazoline [3], QNZ46 is a quinazolin-4-one [4], NAB-14 is an N-aryl benzamide [5], and UBP791 is a phenanthrene-dicarboxylic acid derivative [6]. The natural product origin provides a unique chemical space for exploring GluN2D allosteric modulation and may confer distinct physicochemical properties, including the glycosidic moiety that contributes to binding site interactions at the TMD [1]. YY-23 can be obtained by acid hydrolysis of timosaponin B3 or by chemical synthesis from timosaponin B2, providing reproducible sourcing routes [1].

Natural Product Scaffold

Class-level

Chemotype Steroidal furostanol glycoside

Origin Timosaponin B-III derivative

Source Rhizoma Anemarrhenae (semi-synthetic)

Comparators Synthetic quinazolinones, benzamides

Natural product-derived chemotype; supports scaffold-hopping and pharmacognosy studies

Dual sourcing routes (acid hydrolysis or synthesis from B2) provide supply flexibility

Evidence Dimension	Chemical scaffold class and molecular framework
Target Compound Data	YY-23: Steroidal furostanol glycoside (timosaponin derivative); C₃₃H₅₄O₈; MW 578.78; semi-synthetic from Rhizoma Anemarrhenae [1][2].
Comparator Or Baseline	DQP-1105: Dihydroquinolone-pyrazoline (synthetic) [3]. QNZ46: Quinazolin-4-one (synthetic) [4]. NAB-14: N-aryl benzamide (synthetic) [5]. UBP791: Phenanthrene-piperazine dicarboxylic acid (synthetic) [6].
Quantified Difference	YY-23 is the only GluN2C/2D-selective NMDAR inhibitor derived from a natural product scaffold; all comparators are fully synthetic small molecules with distinct chemotypes [1].
Conditions	Structural elucidation: HR-ESI-TOF-MS, ¹³C-NMR, DEPT, HMBC; [α]D −66.0 (c 0.05, pyridine).

Why This Matters

For labs pursuing natural product-inspired drug discovery or exploring chemical diversity beyond synthetic small-molecule libraries, YY-23 offers access to a validated steroidal glycoside chemotype targeting GluN2D with full structural characterization, enabling scaffold-hopping and pharmacophore studies unavailable with synthetic-only comparators.

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 (Figure 1A; Figure S1; Table S1) View Source
• [2] Zhang Q, Guo F, Fu ZW, Zhang B, Huang CG, Li Y. Timosaponin derivative YY-23 acts as a non-competitive NMDA receptor antagonist and exerts a rapid antidepressant-like effect in mice. Acta Pharmacologica Sinica. 2016;37(2):166-176. doi:10.1038/aps.2015.111 View Source
• [3] Acker TM, Khatri A, Vance KM, et al. Mechanism for noncompetitive inhibition by novel GluN2C/D N-methyl-D-aspartate receptor subunit-selective modulators. Molecular Pharmacology. 2011;80(5):782-795. doi:10.1124/mol.111.073239 View Source
• [4] Mosley CA, Acker TM, Hansen KB, et al. Quinazolin-4-one derivatives: A novel class of noncompetitive NR2C/D subunit-selective N-methyl-D-aspartate receptor antagonists. Journal of Medicinal Chemistry. 2010;53(15):5476-5490. doi:10.1021/jm100027p View Source
• [5] Swanger SA, Vance KM, Pare JF, et al. A novel negative allosteric modulator selective for GluN2C/2D-containing NMDA receptors inhibits synaptic transmission in hippocampal interneurons. ACS Chemical Neuroscience. 2018;9(2):306-319. doi:10.1021/acschemneuro.7b00329 View Source
• [6] Wang JX, Irvine MW, Burnell ES, et al. Structural basis of subtype-selective competitive antagonism for GluN2C/2D-containing NMDA receptors. Nature Communications. 2020;11:423. doi:10.1038/s41467-020-14321-0 View Source

YY-23 Research and Procurement Applications

Structure-Guided Drug Design at GluN2D Allosteric Pocket

YY-23 is the only GluN2C/2D-selective inhibitor with a cryo-EM-resolved binding pose (PDB: 8Y1V), revealing occupancy within the M1/M2/M4 helix pocket of the GluN2D TMD [1]. This makes YY-23 the sole viable template for structure-based drug design (SBDD) and molecular docking campaigns aimed at developing improved GluN2D-negative allosteric modulators. Computational chemists can use the YY-23 binding pose for pharmacophore modeling and virtual screening; medicinal chemists can perform structure-guided derivatization of the steroidal core or sugar moiety. No other GluN2C/2D-selective compound (DQP-1105, QNZ46, NAB-14, UBP791) offers comparable structural information for rational design [1].

GABAergic Disinhibition Mechanisms in PFC Circuitry

YY-23's validated cell-type specificity—inhibition of GluN2D-containing NMDARs on GABAergic interneurons without affecting pyramidal neuron NMDARs—has been confirmed by both electrophysiology and grin2d genetic knockout experiments in the mPFC [1]. This enables circuit-level studies that require selective suppression of interneuron NMDAR function without confounding pyramidal neuron effects. Researchers investigating the disinhibition hypothesis of rapid antidepressant action can use YY-23 to isolate the contribution of GABAergic interneuron GluN2D to network oscillations, synaptic plasticity, and behavioral outcomes [1][2].

Rapid Antidepressant Mechanism Comparator Studies

YY-23 demonstrates antidepressant-like behavioral effects with onset at least 1 week faster than fluoxetine in chronic mild stress (CMS) models and reverses social defeat behavior in the CSDS paradigm where fluoxetine is ineffective [1][2]. This makes YY-23 a critical positive control or comparator compound for preclinical studies evaluating novel rapid-onset antidepressant candidates. Its glutamatergic mechanism—operating through GluN2D-mediated disinhibition rather than monoamine reuptake inhibition—provides a mechanistically orthogonal benchmark to fluoxetine and other SSRIs [2].

Natural Product-Derived GluN2D Pharmacology

As the only natural product-derived GluN2C/2D-selective inhibitor, YY-23 provides a steroidal glycoside chemotype (timosaponin scaffold) that expands chemical diversity beyond synthetic small-molecule libraries [1][2]. This is valuable for: (i) natural product screening laboratories seeking a validated GluN2D-active reference standard; (ii) chemical biology studies exploring how glycosidic moieties contribute to TMD allosteric site recognition; (iii) pharmacognosy and ethnopharmacology programs investigating the molecular basis of Rhizoma Anemarrhenae's traditional antidepressant use [1]. The compound's dual sourcing routes (acid hydrolysis of timosaponin B3 or synthesis from timosaponin B2) provide supply chain flexibility [1].

Application

Selection Property

Validation Focus

GluN2D allosteric pocket tool design

Cryo-EM resolved TMD binding pose

Structure-based mutagenesis and docking

PFC GABAergic disinhibition mechanism

Interneuron-selective GluN2D inhibition

Circuit-level electrophysiology and behavior

Rapid-onset behavioral model comparator

Faster behavioral endpoint change vs fluoxetine

Chronic stress model endpoint comparison

Natural product-based GluN2D chemotype exploration

Steroidal glycoside scaffold

Scaffold-hopping and pharmacognosy studies

• [1] Zhang J, Duan J, Li W, et al. An antidepressant mechanism underlying the allosteric inhibition of GluN2D-incorporated NMDA receptors at GABAergic interneurons. Science Advances. 2025;11(10):eadq0444. doi:10.1126/sciadv.adq0444 View Source
• [2] Zhang Q, Guo F, Fu ZW, Zhang B, Huang CG, Li Y. Timosaponin derivative YY-23 acts as a non-competitive NMDA receptor antagonist and exerts a rapid antidepressant-like effect in mice. Acta Pharmacologica Sinica. 2016;37(2):166-176. doi:10.1038/aps.2015.111 View Source