High-throughput proteomics of Streptomyces

The Streptomyces genus is well known for its ability to produce numerous and diverse bio-active molecules useful to human health. Despite numerous important scientific contributions over the past 40 years, a systemic understanding of the biosynthesis of these bio-active metabolites characterizing the producing bacteria remains incomplete.

To progress on the field, a label-free shotgun comparative proteomic analysis was carried out in the model strains S. coelicolor and S. lividans. These species are closely related (95% of their genes have orthologs) but exert different abilities to produce secondary metabolites. The high antibiotic production of S. coelicolor was correlated with a low lipid content whereas S. lividans showed a poor ability to synthetize antibiotics that was correlated with higher lipid content.

A total of 4372 proteins were identified using LC-MS/MS technology. To this date, this numbers represent the largest dataset for S. coelicolor and S. lividans with >52% of their theoretical proteomes. The R pipeline that I implemented used generalized linear and mixed-effects models. After data and statistical treatment (multiple ANOVA), 1040 proteins showed significant abundance variation according to strain, medium and/or time.

This work is pioneering in the elucidation of the basis of the metabolic differences underlying the drastically different abilities of S. coelicolor and S. lividans to produce antibiotics. A novel view of the role of the antibiotics in the physiology of the producing bacteria was proposed.

# ggplot2 #  glm #  lme #  mass spec #  R