Identification of 14-3-3 Proteins Phosphopeptide-Binding Specificity Using an Affinity-Based Computational Approach

Area of Science:

• Proteomics
• Computational Biology
• Biochemistry

Background:

• 14-3-3 proteins are crucial in eukaryotic cells, with the 14-3-3σ isoform specifically implicated in epithelial cancers.
• Understanding peptide-protein interactions with 14-3-3 isoforms is vital for cancer research.

Purpose of the Study:

• To develop a novel computational method for identifying peptide motifs that bind to the 14-3-3σ isoform.
• To predict the binding affinity of peptides to 14-3-3σ and differentiate binding preferences.

Main Methods:

• A new predictor was built using sampling criteria and nine physicochemical properties of amino acids.
• Auto-cross covariance was employed to capture correlative amino acid properties.
• Elastic net regression, incorporating ridge regression and LASSO, was used for affinity prediction.

Main Results:

• The method achieved high accuracy, with Pearson correlation coefficients (PCC) of 0.84 (N-terminal) and 0.77 (C-terminal) for 14-3-3σ.
• Predicted affinity values for 16,000 peptide sequences closely matched experimental data.
• Identified phosphopeptides with preferential binding to 14-3-3σ and revealed substrate specificity patterns.

Conclusions:

• The developed computational method is fast, reliable, and effective for predicting peptide-14-3-3σ binding.
• This approach offers a general tool for peptide-protein binding identification in proteomics.
• Findings contribute to understanding 14-3-3σ's role in cancer and identifying potential therapeutic targets.