1. Volume 92, Issue 6, Pages 1324-1336 (December 2016)
Molecular Basis for Subtype Specificity and High-Affinity Zinc Inhibition in the GluN1- GluN2A NMDA Receptor Amino-Terminal Domain 
Annabel Romero-Hernandez, Noriko Simorowski, Erkan Karakas, Hiro Furukawa 
Neuron 
Volume 92, Issue 6, Pages (December 2016)
DOI: /j.neuron
Copyright © 2016 Elsevier Inc. Terms and Conditions

2. Figure 1
Domain Organization, Construct Design, and Purification of GluN1b-GluN2A ATD
(A) Domain organization of the GluN1b-GluN2A NMDA receptors. GluN1b-Cys22Ser (orange oval) is to prevent non-specific disulfide bond formation, whereas GluN1b-Asn61Gln and GluN1b-Asn371Gln (cyan oval) are to knock out putative Asn-linked glycosylation.
(B) The GluN1b-GluN2A ATD was expressed as a fusion protein where GluN1 is tethered to GluN2A by a 57-amino acid linker (57-link) flanked by two histidine tags and thrombin sites. The linker contains three artificial Asn-linked glycosylation sites to enhance secretion of the fusion protein.
(C) A chromatograph from Superdex200 size-exclusion chromatography (SEC) of GluN1b-GluN2A ATD digested with EndoF1 and thrombin and complexed with Fab. The peak with an asterisk corresponds to the GluN1-GluN2A ATD dimer in complex with Fab.
(D) Twelve percent SDS-PAGE gels stained with Coomassie Brilliant Blue showing purification levels and digestion patterns by thrombin and EndoF1.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

3. Figure 2 Crystal Structure of GluN1b-GluN2A ATD
GluN1b and GluN2A ATDs have bi-lobed architecture composed of R1 and R2 domains and are arranged as a heterodimer. Shown here is the structure of the Zn1-GluN1b-GluN2A ATD (see text). GluN1b R1 (residues 24–143 and 294–369), GluN1b R2 (residues 144–293 and 370–411), GluN2A R1 (residues 33–150 and 286–360), and GluN2A R2 (residues 151–285 and 361–390) are colored magenta, light pink, dark green, and light green, respectively. Zinc and N-acetylglucosamine at Asn-linked glycosylation sites are represented in gray sphere and sticks, respectively. Zinc anomalous difference Fourier map contoured at 5.0σ is shown in blue mesh. The heterodimeric structure is viewed from the “side” (left) or the “top” (right) of the N-terminal ends (NT). The Fab fragment contained in the asymmetric unit is omitted for clarity.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

4. Figure 3 Comparison of GluN1b-GluN2A and GluN1b-GluN2B ATDs
(A) Superposition of the GluN1b ATD from the Zn1-GluN1b-GluN2A ATD (magenta and light pink) and the ifenprodil-GluN1b-GluN2B (PDB: 3QEL; black) showing little or no difference in the overall structure.
(B) Superposition of the R1 lobes of the GluN2A ATD (dark green) and the GluN2B-ATD (PDB: 3JPY; gray) demonstrates more “open” bi-lobed architecture in GluN2A compared with GluN2B, characterized by the ∼13° rotation along the axis that runs through the hinge of the bi-lobe (black rod).
(C) Superposition of the R1 lobes of GluN2A (dark green) and GluN2B (PDB: 3QEL; gray) in the context of the GluN1-GluN2 heterodimers illustrates the difference in the GluN1b arrangement characterized by the ∼12° rotation along the axis that runs through the heterodimer interface (black rod).
(D) Model representing both the ∼13° rotation of R2 of GluN2A and the ∼12° rotation of GluN1.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

5. Figure 4
Validation of Heterodimeric Assembly of GluN1b and GluN2A ATDs
(A) Engineered cysteines at the subunit interface are shown as spheres for Site-I (GluN1b Leu341Cys-GluN2A Ser209Cys) and Site-II (GluN1b Phe113Cys-GluN2A Ala108Cys).
(B) Western blots in non-reducing (non-red.) (left) and reducing (red.) (right) conditions. In the reducing condition, samples contained 660 mM of beta-mercaptoethanol. The bands were recognized by both anti-GluN1 (top) and anti-GluN2A (bottom) monoclonal IgGs followed by anti-mouse secondary IgGs conjugated with peroxidase and detected by enhanced chemiluminescence.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

6. Figure 5
Comparison of the GluN1-GluN2A and GluN1-GluN2B Subunit Interfaces
(A) Ifenprodil-binding pocket at the GluN1b-GluN2B subunit interface (PDB: 3QEL).
(B) Close-up view of the ifenprodil-binding pocket shown in (A).
(C) Superposition of the Zn1-GluN1b-GluN2A ATD and the ifenprodil-GluN1b-GluN2B ATD at the R1 lobes of GluN2A and GluN2B show that the ifenprodil binding pocket is “filled in” by the residues from GluN1b in the GluN1b-GluN2A ATD heterodimer because of the smaller distance gap between the two subunits (arrow).
(D–F) Volume of the protein cavities at the subunit interface is calculated for the ifenprodil-GluN1b-GluN2B ATD (PDB: 3QEL) (D), the apo2-GluN1b-GluN2B ATD (E), and the Zn1-GluN1b-GluN2B ATD (F) by the software KVFinder (Oliveira et al., 2014). Protein cavities are represented as brown dots. The subunit interface cavity in the GluN1-GluN2B ATD is sufficiently large to accommodate ifenprodil (D and E; volume 697 and 550 Å3), whereas that in the GluN1b-GluN2A ATD is not (F; volume ∼130 Å3).
(G) Model representing differences in the inter-subunit cavities between the GluN1-GluN2A ATD and the GluN1-GluN2B ATD. The color code of the structural figures is as in Figure 3. Ifenprodil in (A) is shown as brown spheres and sticks.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

7. Figure 6
The High-Affinity Zinc-Binding Site of GluN2A in Comparison with the Micro-molar-Affinity Zinc Binding of GluN2B
(A and B) The high-affinity zinc-binding site of GluN2A where zinc (gray sphere) is coordinated by GluN2A-His44, -His128, -Glu266, and -Asp282. The blue mesh (left) represents the zinc anomalous difference Fourier map contoured at 5.0σ. The green mesh (right) represents the Fo-Fc omit map contoured at 3.0σ (B).
(C and D) Side-by-side comparison of the high-affinity zinc-binding site in GluN2A (C) and the micro-molar-affinity zinc-binding site in GluN2B (PDB: 3JPY) where zinc is coordinated by GluN2B-His127 and -Glu284 (D).
(E) In GluN2A, GluN2A-His44 contained in Zn-loop specifically present in the primary sequence of GluN2A orients to directly coordinate zinc.
(F) The high-affinity zinc inhibition is abolished by mutation of GluN2A-Asp282.
(G) Shown here are representative recordings of zinc inhibition of the WT GluN2A and GluN2A-Asp282His co-expressed with the WT GluN1-1a by two-electrode voltage clamp (TEVC) held at −60 mV and the data plot done as in Figure S2.
Error bars represent ±SD for data obtained from 17 and 4 different oocytes per experiment for the wild-type and the mutants, respectively.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

8. Figure 7
Residues at the R1-R2 Interface Are Critical for Functional Regulation in Both GluN2A and GluN2B
(A) Shown here is the GluN2A subunit from the Zn1-GluN1b-GluN2A ATD structure. The inter-lobe space of the GluN2A ATD and the zinc coordinating residues (left) are highlighted. The Fo-Fc omit map contoured at 3.3σ shows clear electron density for the inter-lobe residues GluN2A-Asp105 in R1, GluN2A-Lys233 in R2, and GluN2A-Asn264 in R2 (right). Black dots indicate polar interactions.
(B) Close-up view of the zinc-bound GluN2B ATD (PDB: 3JPY) where the interaction between the conserved residues GluN2B-Asp104 in R1 and GluN2B-Lys234 in R2 stabilizes the closure of the GluN2B ATD bi-lobe.
(C) The mutants that disrupt the inter-R1-R2-lobe polar interactions, GluN2A-Lys233Ala (n = 4), GluN2A-Lys233Arg (n = 6), GluN2A-Asn264Ala (n = 4), and GluN2A-Asn264Trp (n = 3), significantly alter zinc sensitivity compared with the WT (GluN2A-WT: n = 17) as monitored by TEVC recordings. TEVC recordings were conducted as in Figure 6.
(D) Estimation of open probability of the inter-lobe mutants by measurement of MTSEA potentiation in the GluN1-1a Ala652Cys-GluN2A NMDA receptors. Disruption of the inter-lobe polar interaction by mutagenesis shows an increased open probability, as indicated by decreases in IMTSEAIagonist (GluN2A-WT: 4.79 ± 0.56 [n = 17], GluN2A-Lys233Ala: 3.80 ± 0.23 [n = 6], GluN2A-Lys233Arg; 2.47 ± 0.31 [n = 8], GluN2A Asn264Ala: 2.76 ± 0.47 [n = 8], GluN2A-Asn264Trp: 3.54 ± 0.32 [n = 7]).
(E) The analogous mutation, GluN2B-Lys234Ala, which disrupts the similar polar interaction, also shows an increased open probability (GluN2B-WT: 43.50 ± 5.89 [n = 10] and GluN2B-Lys234Ala: 20.63 ± 6.51 [n = 8]).
∗p < 0.05 and ∗∗p < as assessed by the Kolmogorov-Smirnov test. Error bars represent ±SD.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions

9. Figure 8
Plausible Mechanism of Gating Control in GluN1-GluN2A NMDA Receptors
(A) We assume GluN1-GluN2A NMDA receptors form hetero-tetramers composed of a dimer of GluN1-GluN2A heterodimer in a similar manner to the GluN1-GluN2B NMDA receptors.
(B) The opening of the GluN2A ATD bi-lobe result in activation of the channel as speculated by our data in Figure 7, which demonstrates that perturbation of the R1-R2 inter-lobe interaction favors channel opening. We predict that the movement in ATD affects subunit arrangement in LBD to control channel gating, as previously shown for GluN1-GluN2B NMDA receptors. Zinc stabilizes the GluN2A ATD bi-lobe and favors channel closure.
Neuron  , DOI: ( /j.neuron )
Copyright © 2016 Elsevier Inc. Terms and Conditions