Unsupervised Identification of Isotope-Labeled Peptides

Area of Science:

• Metabolomics
• Proteomics
• Systems Biology

Background:

• In vivo isotopic labeling combined with high-resolution proteomics is crucial for investigating primary metabolism through techniques like stable isotope probing (protein-SIP) and peptide-based metabolic flux analysis (PMFA).
• The distribution of isotopes within amino acids, determined by carbon substrate enrichment and intracellular metabolism, leads to convoluted amino acid mass distributions (AMDs) into peptide mass distributions (PMDs) during protein synthesis.
• Current peptide identification software relies on prior knowledge of PMDs, which are unknown for nonuniformly labeled samples, hindering accurate flux analysis.

Purpose of the Study:

• To develop an automated, unsupervised framework for identifying and quantifying peptide mass distributions (PMDs) in nonuniformly labeled samples.
• To overcome limitations in current peptide identification software that require a priori knowledge of PMDs.
• To enable high-throughput metabolic flux analysis and other complex labeling experiments using proteomics.

Main Methods:

• Developed an unsupervised method utilizing discrete deconvolution of mass distributions from identified peptides.
• Applied feature reconstruction and deconvolution algorithms to uniformly (13)C-labeled Escherichia coli protein for testing.
• Validated peptide identification by comparing MS(2)-identified peptides with those identified from PMDs of unlabeled E. coli protein.
• Demonstrated the technology on nonuniformly labeled Glycine max protein for flux analysis.

Main Results:

• The developed algorithms successfully reconstructed and deconvoluted mass distributions for peptide identification.
• Automatic peptide identification and quantification performance was comparable or superior to manual extraction.
• The method effectively identified and quantified PMDs in both uniformly and nonuniformly labeled protein samples.

Conclusions:

• The novel unsupervised method enables accurate identification and quantification of peptide mass distributions without prior knowledge.
• This technology significantly advances proteomics-based approaches for high-throughput metabolic flux analysis.
• The developed framework unlocks the potential of complex isotopic labeling experiments in biological research.