Integrating deep learning for post-translational modifications crosstalk on Hsp90 and drug binding

Area of Science:

• Biochemistry and Molecular Biology
• Proteomics
• Cancer Biology

Background:

• Post-translational modifications (PTMs) increase proteome complexity and regulate cellular proteostasis.
• Heat shock protein-90 (Hsp90) is a crucial molecular chaperone involved in proteostasis and cancer signaling, making it a therapeutic target.
• PTM crosstalk significantly contributes to protein functional diversity.

Purpose of the Study:

• To investigate the impact of histone deacetylase 3 (HDAC3) and histone deacetylase 8 (HDAC8) deletion on Hsp90 PTMs and drug binding.
• To decipher PTM crosstalk on Hsp90 using both experimental and computational approaches.
• To evaluate the efficacy of a deep-learning AI model in predicting PTM crosstalk.

Main Methods:

• Utilized human cell lines with deleted HDAC3 and HDAC8.
• Analyzed Hsp90 binding to ATP and the inhibitor Ganetespib via mass spectrometry.
• Employed a deep-learning artificial intelligence (AI) prediction model for PTM analysis.

Main Results:

• HDAC3 and HDAC8 deletion increased Hsp90 binding to ATP and Ganetespib.
• Hsp90 from knockout cells showed similar phosphorylation and acetylation PTMs when bound to Ganetespib.
• A common proteomic network signature was observed in Hsp90 from both knockout cell types.
• AI model predictions aligned with mass spectrometry data for deciphering PTM crosstalk.

Conclusions:

• HDAC3 and HDAC8 play a role in regulating Hsp90 PTMs and drug interactions.
• Deep-learning AI models provide an efficient and rapid method for deciphering PTM crosstalk on complex proteins like Hsp90.
• Understanding Hsp90 PTM crosstalk can inform cancer therapeutic strategies.