Scholarlink

<p>The stimulator of interferon genes (STING) protein plays a pivotal role in the human immune system, where agonist-activated STING induces type I interferons and pro-inflammatory cytokines to modulate immune responses. However, reported STING agonists often exhibit limitations that hinder their clinical utility. In this study, we de novo designed novel STING agonists by constructing a pharmacophore fragment library of 501 fragments and 12 core scaffolds, followed by fragment growing to expand chemical space and generate a library of 10<sup>6</sup> compounds. High-affinity candidates were progressively screened using multi-level molecular docking precision. To refine selection, we applied a multi-step free energy approach: initial MD simulations with MM/PBSA for preliminary Δ<i>G</i> assessment, clustering of trajectories to extract representative conformations, and extended MD simulations with weighted-average Δ<i>G</i> calculation to obtain reliable binding affinities. Complementary validation via the Boltz-2 deep learning model confirmed consistent affinity predictions. Compounds F3_181_62 and F11_94_105 emerged with superior affinities, with F11_94_105 outperforming the reference 2′,3′-cGAMP in both binding strength and drug-likeness. Binding mode analysis revealed analogous interactions with STING, stabilized by hydrogen bonds and hydrophobic contacts. ADMET profiling, using ADMETlab 2.0 indicated suboptimal physicochemical properties due to large molecular weights but highlighted enhanced human intestinal absorption (HIA) and extended half-lives for F3_181_62 and F11_94_105 compared to cGAMP’s poor ratings, alongside low toxicity risks and favorable safety. These profiles warrant further investigation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fragment-based discovery of novel STING agonists validated with free energy calculations

  • 期刊论文

  • 发表日期：

    2026年2月27日

查看全文
  • Meiting Wang,
  • Xiangbing Peng,
  • Baoyi Fan,
  • Yang Liu,
  • Di Han,
  • Jiarui Lu,
  • Taigang Liu,
  • Shaoli Cui,
  • Wei Xia,
  • Yongtao Xu

摘要

The stimulator of interferon genes (STING) protein plays a pivotal role in the human immune system, where agonist-activated STING induces type I interferons and pro-inflammatory cytokines to modulate immune responses. However, reported STING agonists often exhibit limitations that hinder their clinical utility. In this study, we de novo designed novel STING agonists by constructing a pharmacophore fragment library of 501 fragments and 12 core scaffolds, followed by fragment growing to expand chemical space and generate a library of 106 compounds. High-affinity candidates were progressively screened using multi-level molecular docking precision. To refine selection, we applied a multi-step free energy approach: initial MD simulations with MM/PBSA for preliminary ΔG assessment, clustering of trajectories to extract representative conformations, and extended MD simulations with weighted-average ΔG calculation to obtain reliable binding affinities. Complementary validation via the Boltz-2 deep learning model confirmed consistent affinity predictions. Compounds F3_181_62 and F11_94_105 emerged with superior affinities, with F11_94_105 outperforming the reference 2′,3′-cGAMP in both binding strength and drug-likeness. Binding mode analysis revealed analogous interactions with STING, stabilized by hydrogen bonds and hydrophobic contacts. ADMET profiling, using ADMETlab 2.0 indicated suboptimal physicochemical properties due to large molecular weights but highlighted enhanced human intestinal absorption (HIA) and extended half-lives for F3_181_62 and F11_94_105 compared to cGAMP’s poor ratings, alongside low toxicity risks and favorable safety. These profiles warrant further investigation.