BGPT: Paper Review: High-throughput mutagenesis reveals unique structural features of human ADAR1

Explore by Goal

Quick Answer Copied

Bottom line: Park et al. (Nat Commun 2020) combine saturation mutagenesis (Sat‑FACS‑Seq), ICP‑MS / native MS, XL‑MS and RosettaCM to identify and validate a second, surface zinc site (C1081, C1082, H1103 with H988 as 4th ligand candidate) in the human ADAR1 deaminase domain; mutation of these residues reduces Zn occupancy and editing activity in vitro and in cells — a robust, experimentally well-supported discovery that supplies a new, druggable structural feature unique to ADAR1 (full paper citation below).

Citation:

Long Answer

Visual paper analysis — Park et al., "High-throughput mutagenesis reveals unique structural features of human ADAR1" (Nat Commun, 2020)

Claim

Discovery of a second Zn site on hADAR1d (C1081/C1082/H1103 ± H988) that stabilizes the 5′ binding loop and is required for full deaminase activity.

Key evidence types

High-throughput saturation mutagenesis + yeast GFP reporter (Sat‑FACS‑Seq)
ICP‑MS and native mass spectrometry (metal stoichiometry)
In vitro kinetic assays on Gli1 RNA (kobs changes)
Full-length cellular editing assays (HEK293T endogenous targets)
RosettaCM constrained homology modeling + XL‑MS crosslinks (DSBU)

Practical implication

New, surface-exposed Zn site provides a structurally supported, ADAR1-specific pocket that could be exploited for selective small-molecule targeting / allosteric inhibition or regulation by redox/metal availability.

Figure 1 — In vitro deamination kinetics (k_obs) for MBP‑hADAR1d WT and C1082 mutants

Values from Park et al.: WT kobs = 0.480 ± 0.008 min^-1; C1082D = 0.084 ± 0.004 (≈5.7× slower); C1082E = 0.03 ± 0.01 (≈16× slower).

Figure 2 — Zinc per protein (ICP‑MS / native MS)

Park et al. measured ~1:0.90 Zn/protein for hADAR2d (single catalytic Zn), and ~1:1.93 Zn/protein for hADAR1d consistent with two Zn sites; MBP fusion reduced saturation (1.24) indicating fusion/solubilization effects.

Figure 3 — Structural model constraints & crosslink support

RosettaCM models (5,000 decoys → 10 lowest energy inspected) placed the long ADAR1 5′ binding loop folded back to ligate Zn2 via C1081/C1082/H1103 + H988 (mode distinct from ADAR2) — constraints used were: conserved 5′ loop residue positions (F972/D973/K996) and Zn coordination geometry from MetalPDB-derived distances.
DSBU XL‑MS identified reproducible crosslinks (e.g., S986–K1105, Cα–Cα ~14.6 Å) compatible with the proposed loop conformation, supporting folding-back of the 5′ loop near the second Zn site.

Modeling and crosslink evidence summarized from Park et al.

Critical appraisal — strengths, limitations, and blindspots

Major strengths

Multi-modal orthogonal evidence: functional mutagenesis (high-throughput), biochemical metal quantitation (ICP‑MS/native MS), kinetics, cellular editing assays and structural constraints (Rosetta + XL‑MS) converge on the same conclusion — increases confidence.
Data availability: raw sequencing (SRA PRJNA590991), MS datasets (PXD021052/PXD021175) and GitHub models enable reproducibility and reanalysis.
Robust functional readouts: both purified‑protein kobs and full‑length overexpression editing in cells were reduced by Zn‑site mutations.

Limitations & blindspots

Absence of a high‑resolution ADAR1d crystal/NMR structure: Rosetta models are well constrained but do not converge on a single loop conformation — the 5′ loop remains uncertain in atomistic detail.
MBP/His tags and yeast/insect expression can alter metal occupancy and folding (MBP fusion had sub‑saturating Zn), so metal stoichiometry could be influenced by construct/context.
In vivo relevance: cellular editing assays use overexpression in HEK293T — endogenous regulation, redox sensitivity, or tissue‑specific metal availability effects need further validation in native systems.
Potential alternative explanations: effects of cysteine mutation on local folding/aggregation vs direct metal ligation require careful separation (authors address this but residual uncertainty remains).

Falsification tests (what would overturn the conclusion)

High‑resolution crystal or cryo‑EM structure of ADAR1d + RNA showing no second Zn site or different ligation pattern (direct structural refutation).
Direct reconstitution of apo protein with careful metal chelation / titration showing no stoichiometric second Zn across multiple constructs/hosts and independent labs.
Mutants disrupting the proposed Zn ligands but restored activity by conservative engineering (e.g., second-site suppressors) indicating loss-of-activity was due to local destabilization not metal removal.

High‑priority experiments to advance/validate

Obtain high‑resolution ADAR1d structure (X‑ray or cryo‑EM) with native metal occupancy — definitive.
Measure Zn affinity (Kd) for the second site by titration (isothermal titration calorimetry or competition ICP‑MS) to quantify binding strength and ligand preferences.
Test redox sensitivity: treat cells/proteins with oxidizing agents (H2O2) and measure editing + Zn occupancy — links to regulation via cysteine oxidation.
Assess physiological relevance: knock‑in of C1082E (or H1103D) in endogenous ADAR1 alleles (CRISPR) in relevant cell types (e.g., primary immune cells or cancer lines) and measure global editing, IFN/ISG responses, and ADAR1-dependent phenotypes.
Structure‑guided small‑molecule/fragment screening against the Zn2+ pocket (both Zn chelators and allosteric binders) to test druggability and selective inhibition of ADAR1 vs ADAR2.

Primary source for all data and figures cited above:

Feedback:

Updated: February 12, 2026