Incidences of infection by K. pneumoniae have grown in frequency to become the leading agents of CRE infection among hospitalized patients in the United States and Europe. Transcriptomic meta-analysis of data collected from both laboratory and clinical isolates indicated significant shifts in expression of key transcription factors related to metabolism. Metabolic network reconstructions have previously proven effective for quickly identifying potential targets in silico, therefore we combined these approaches by integrating the transcriptomic data from each isolate type into a well-curated GENRE of K. pneumoniae to predict emergent metabolic patterns. Leveraging this systems-biology approach we found discordant patterns of active metabolism between clinical and laboratory isolates, with a striking difference in L-valine catabolism. Exogenous valine is known to increase macrophage phagocytosis, and our results may support immunomodulatory activity in K. pneumoniae evolved to avoid host clearance.

Below is a visual representation for the general procedure of generating context-specific models of metabolism from transcriptomic data. All 56 datasets from the transcriptome meta-analysis were used to generate distinct context-specific models of K. pneumoniae metabolism.

In our new pre-print, we ultimately found that clinical isolate-associated metabolic models were more dependent on environmental valine and consumed it at much faster rates that lab isolates in silico.

You can read the pre-print here!

Also, you can find all the analysis code here!

Our paper on modeling context-specific metabolism of C. difficile is finally accepted in mSystems! In it we generated and extensively curated a C. difficile GENRE for a hypervirulent isolate (str. R20291). In silico validation revealed high degrees of agreement with experimental gene essentiality and carbon source utilization data sets. Then in collaboration with Dr. Rita Tamayo at UNC, whose research focuses on an evolutionary strategy in some bacteria for phenotypic heterogeneity and virulence known as phase variation. We utilized transcriptome sequencing from distinct phases of C. difficile, and in combination with our previously published integration algorithm, we created context-specific models of metabolism to identify possible metabolite signals that drive differences in virulence.

Using this GENRE-based analyses we found that glucose utilization through the pentose phosphate pathway is essential in the smooth phase variants of str. R20291. Here we show (A) Gene and reaction essentiality results for glycolysis and the pentose phosphate pathway across both the rough and smooth phase variant context-specific models. Components were deemed essential if models failed to generate <1% of optimal biomass flux. (B and C) Colony morphologies resulting from smooth and rough variants of C. difficile str. R20291 grown on either BHIS or BDM ± glucose after 48hr of growth. Defined medium colonies were then subcultured onto BHIS medium for an additional 24hr as indicated. Increased colony perimeter was found to be the defining characteristic of the rough colony morphology. This feature was quantified for multiple colonies under each permutation of colony variant and growth medium. (D) Colony perimeter for smooth and rough progenitor colony variants grown on BHIS. (E and F) Smooth (E) or rough (F) colony variant perimeter during subculture onto each of the BDM carbon source medium formulations.

Our results support that differential C. difficile virulence is associated with distinct metabolic programs related to use of carbon sources and provide a platform for identification of novel therapeutic targets.

You can read the whole open-access paper here!

Also, you can find all the analysis code here!