Tier 4 — mechanistic

Deciphering the allosteric activation mechanism of SIRT6 using molecular dynamics simulations

Zhiyuan Zhao, Jintong Du, Yu Du, Yuan Gao, Mingxuan Yu, Yingkai Zhang, Hao Fang, Xuben Hou
Journal of Chemical Information and Modeling 2023 Volume 63, issue 18, pages 5896–5902

Bibliography

PubMed
PMID 37653718
PubMed Central
PMC10530556
DOI
10.1021/acs.jcim.3c00227
Funding
National Natural Science Foundation of China grants 82003590 and 92053105; Natural Science Foundation of Shandong Province grants ZR2020QH342 and ZR2022QH209; US National Institutes of Health grant R35-GM127040 to Y.Z.
Competing interests
The authors declare no competing financial interest.

Study snapshot

DesignExtensive molecular dynamics simulations of SIRT6 complexes in the presence of different substrates and activators. Seven simulation systems constructed, each run in duplicate for 2 microseconds (14 microseconds total). Systems compared: SIRT6 with acetyl-lysine substrate versus myristoyl-lysine substrate; SIRT6 with acetyl-lysine plus docked myristic acid in two orientations; SIRT6 with acetyl-lysine plus known synthetic activators UBCS039, MDL-801, or 12q.
ModelFull-length human SIRT6 protein in atomic-resolution simulation using the AMBER14 force field. Initial coordinates from published crystal structures (PDB IDs 3ZG6, 4I5I, 7CL1, 5MF6). Molecular dynamics performed with ff14SB force field for protein, TIP3P water model, Zinc AMBER Force Field for zinc parameters.
SampleSeven simulation systems, each run in duplicate for 2 microseconds. Total simulation time 14 microseconds. Two independent runs per system with periodic boundary conditions.
InterventionComparison of substrate binding (acetyl versus myristoyl lysine), fatty acid activator binding (myristic acid in two docking orientations), and synthetic activator binding (UBCS039, MDL-801, 12q). Free energy analysis of the catalytic distance between substrate acyl oxygen and NAD⁺ C1 atom.
Duration2 microseconds per simulation run (in silico time).
EndpointsDistance between substrate acyl oxygen and NAD⁺ C1 atom (proxy for catalytic proximity); Dihedral angle between substrate and cofactor (proxy for reaction geometry); Estimated free energy barrier for the first catalytic step; Structural conformation of the NAD⁺ ribose ring; Position of Phe62 residue relative to the hydrophobic pocket; Binding poses of small-molecule activators relative to the SIRT6 hydrophobic channel

What the study showed, in plain terms

This is a computer simulation study, not an experiment on cells or animals. The authors — an interdisciplinary team from Shandong University and New York University — used molecular dynamics simulations to answer a long-standing puzzle in SIRT6 biology: why is SIRT6 much better at removing long fatty tags from proteins than short acetyl tags, and why do certain small molecules stimulate the enzyme?

By running seven separate atomic-resolution simulations totalling fourteen microseconds of simulated protein motion, they observed something striking. When SIRT6 was given a long-chain fatty substrate, the enzyme's active site held a chemically productive shape. When it was given a short acetyl substrate, the same site collapsed into a loose, misaligned configuration — the NAD⁺ cofactor's sugar ring rotated away, and the geometry became unfavourable for catalysis.

The authors then tested what happens when a small-molecule activator is added. Whether the activator was a natural fatty acid (myristic acid), a synthetic activator (UBCS039, MDL-801, or the more recent compound 12q), or the long-chain substrate itself, all four bound to the same hydrophobic pocket at the far end of SIRT6's active site. All four blocked the collapse of the ribose ring. All four restored a catalytically productive geometry.

This gave the field a unifying picture: SIRT6 is naturally in an inactive-favouring conformation with a short-chain substrate, and activators of any chemical class work by occupying an allosteric pocket that stabilises the productive shape. This unifies fatty acid activation, polyphenol activation (fucoidan, cyanidin), and small-molecule drug activation into a single mechanistic framework — one that is directly informing the medicinal chemistry effort to develop clinical SIRT6 activators.

Key findings

  • In molecular dynamics simulations, the SIRT6 hydrophobic pocket between the zinc-binding and Rossmann fold domains is well-occupied by myristoylated substrate but sits empty with an acetylated substrate. This packing defect drives the NAD⁺ ribose ring to flip out of the catalytically productive orientation.
  • The distance between substrate acyl oxygen and NAD⁺ C1 atom — the proxy for catalytic proximity — averaged 4.63 ± 0.85 Ångstroms in the myristoyl system versus 6.19 ± 0.87 Ångstroms in the acetyl system. Estimated free energy barrier for the first catalytic step was accordingly much lower in the myristoyl system.
  • Docking of myristic acid into the empty hydrophobic pocket of the SIRT6-acetyl complex in two orientations produced clearly different results. One orientation (MYA-I) rotated during simulation to anchor at the α3 helix with its terminal group close to nicotinamide, compensating the packing defect and restoring the productive NAD⁺ geometry (average distance 5.22 ± 0.68 Ångstroms). The other orientation (MYA-II) failed to correct the packing defect.
  • Three synthetic SIRT6 activators for which structural or docking data were available (UBCS039, MDL-801, 12q) all bound to the same allosteric pocket at the distal hydrophobic channel, formed by the SIRT6 N-terminal tail and α3 helix. All three rotated the α3 helix, stabilised Phe62 in a productive position, and restored the NAD⁺ ribose to a catalytically favourable orientation with dihedral angles clustering around 90–100 degrees.
  • The proposed mechanism unifies activation by long-chain substrates, endogenous fatty acids, and synthetic small molecules into a single allosteric framework: all activators occupy the hydrophobic pocket, all block ribose flipping, all restore the hydrogen bond between NAD⁺ ribose and catalytic His131.

What this study can and cannot tell us

This is a computational simulation study. No in vitro biochemistry, cellular experiments, or animal models are performed. All conclusions rest on the accuracy of the molecular dynamics force field (AMBER ff14SB, ZAFF for zinc) and on the starting crystallographic coordinates. Results should be treated as high-resolution mechanistic hypotheses rather than direct experimental observations.

The 2-microsecond simulation timescale, while extensive by molecular dynamics standards, may miss slower conformational transitions relevant to SIRT6 regulation, including those involving the intrinsically disordered C-terminal region (residues 295-355) that had not been crystallographically solved as of the study's writing.

The docking of myristic acid into the SIRT6 hydrophobic pocket generates two candidate binding orientations, one of which (MYA-I) supports the activation hypothesis and the other of which (MYA-II) does not. The choice to interpret MYA-I as the biologically relevant pose is based on the simulation outcome — activator produces the geometry consistent with activation — but no experimental validation of this docking pose was provided in the study.

The mechanism proposed here explains activation of SIRT6 deacetylase activity but does not address whether the same allosteric pocket occupation affects the mono-ADP-ribosyltransferase or long-chain defatty-acylase activities of SIRT6. Empirically, some activators (myristic acid, CL5D) competitively inhibit long-chain demyristoylation while stimulating deacetylation — a nuance not directly modelled in this simulation study.

Citation network

Articles citing this research paper

1 article

Biohack Blueprint analyses that reference this study in their evidence base.

Editorial review

Reviewed by the Biohack Blueprint research team

Last verified

Dr. Khaled Mahmoud, MBBChCritical Care Medicine, Longevity Science & Editorial Direction Dr. Amr Yasser, PhDCardiometabolic, Musculoskeletal, Gut & Bioavailability Dr. Abdulla Mostafa, PhDCellular Senescence, Neurological, Mitochondrial & Brain-Ageing Biology