Although antibiotics are known to eliminate pathogens through a variety of mechanisms [1, 2], it is reported that antibiotics may induce the transition in the microbial phenotypes and promote the formation of persisters, which are dormant variants of regular cells that are highly tolerant to antibiotics . For example, Dr. Wood's group at Pennsylvania State University reported that E. coli pretreated with antibiotics such as rifampicin, tetracycline, and carbonyl cyanide m-chlorophenyl hydrazine increased the microbial persistence dramatically. They demonstrated that the bacterial persistence resulted from halted protein synthesis and a variety of environmental cues . The presence of persister-specific resistance is suggested to account for the recalcitrance of infectious diseases [5, 6]. It is thus important to study the mechanisms via which antibiotics cause pathogens to change their metabolism to form persister cells. Experimental approaches such as transposon insertion have been used to study the phenotype transition mechanisms during antibiotic treatment. Screening knockout libraries has not produced mutants that lack persisters, indicating that dormancy mechanisms are not regulated by single genes or enzymes [3, 7]. This understanding motivated us to develop a systems level approach that can incorporate the interaction of multiple genes/enzymes/metabolites to investigate potential persister-forming mechanisms triggered by antibiotics. Since genome-scale models contain the genes/enzymes/metabolites for metabolic pathways that determine the microbial growth rate, they can be used to pinpoint the genes/metabolites involved in the phenotypic transition. In particular, some metabolites in genome-scale models behave like signaling molecules in triggering persister formation. One example is acyl-homoserine lactone, which promotes the persister formation in Pseudomonas aeruginosa through quorum sensing . Studying the metabolic alternation between different phenotypes may provide cues to potential signaling metabolites and pathways to manipulate the formation of persister cells. Although research has been conducted on the reaction-flux distribution upon the antibiotic treatment , no genome-scale modeling approach has been published to quantify the effect of antibiotics on bacterial phenotype transition.
In this work, we present the first genome-scale modeling approach to quantify the effect of antibiotics on the metabolic flux redistribution, and thus investigate potential mechanisms for the persister formation that is triggered by antibiotics. Specifically, we first developed an approach to integrate gene expression data with metabolic models to identify the reactions that positively reflect the metabolic activities of microorganisms. On the basis of these metabolic-activity indicating reactions, we then developed the first genome-scale approach to predict microbial metabolic perturbation of P. aeruginosa upon antibiotic treatment. P. aeruginosa is selected as the example microorganism in this work, as it is one of the leading causes of nosocomial infections in hospitalized patients and it displays resistance to a wide array of antibiotics by forming a biofilm in chronic infectious processes [10, 11]. The results showed that 1110 metabolic reactions within the network, 171 reactions increased flux, 372 reactions decreased fluxes and 10 reversible reactions changed flux direction. Parts of our results were further validated by experiment data related to persister cell formation. The flux profile of cell with antibiotic treatment provides clues on the detection of novel signaling pathways that trigger cell persistence. The exchange rates of several metabolites with significant alternation were also proposed. Specifically, 14 metabolites including acetate, fumarate, gluconate, succinate, citrate, ethanol, glycerol, and pyruvate increased their exchange rates, while 6 metabolites including 4-hydroxybenzoate, orotate and allantoin decreased their exchange rates. These metabolites may be used in further controlling persister formation in nutrition supplement. Given the metabolic network information, this approach can be applied to study the metabolism adjustment in phenotype transition of other microorganisms.
See more of this Group/Topical: Topical Conference: Emerging Frontiers in Systems and Synthetic Biology