Loop Nucleotide Chemical Shifts as a Tool to Characterize DNA G-Quadruplexes.

Area of Science:

• Biochemistry and Molecular Biology
• Structural Biology
• Genomics

Background:

• DNA G-quadruplexes are crucial in cellular functions and have diverse scientific applications.
• Solution-state NMR spectroscopy is vital for characterizing DNA G-quadruplex structures.
• Recent advancements enable rapid backbone topology identification using ¹³C/¹H chemical shifts.

Purpose of the Study:

• To demonstrate that loop conformations can be identified using ¹³C/¹H chemical shifts.
• To integrate loop information with backbone topology for complete 3D structural characterization.
• To explore the utility of machine learning in predicting DNA G-quadruplex loop conformations.

Main Methods:

• Utilizing ¹³C/¹H chemical shifts of propeller, lateral, and diagonal loops for topology discrimination.
• Applying random forest and k-nearest neighbors machine learning algorithms to predict loop conformation from nucleotide-based ¹H shifts.
• Combining predicted loop information with traditional NMR methods for comprehensive structural analysis.

Main Results:

• ¹³C/¹H chemical shifts effectively differentiate between various loop conformations within a given backbone fold.
• Machine learning models achieved modest accuracy in predicting loop conformations based on ¹H shifts.
• The integration of predicted loop data and NMR methods provides a convenient and reliable approach to characterizing DNA G-quadruplex folds.
• Analysis revealed significant conformational flexibility in DNA G-quadruplex loops.

Conclusions:

• ¹³C/¹H chemical shifts, coupled with machine learning, offer a powerful tool for DNA G-quadruplex structural characterization.
• This approach facilitates a more complete understanding of DNA G-quadruplex folds, including loop dynamics.
• The findings motivate further detailed investigations into the conformational flexibility and dynamics of DNA G-quadruplex loops.