System-level examination of algicidal mechanisms in bacteria
Elias Azar
Mentor: Dr. Ali Navid, Biochemical and Biophysical Systems Group, Lawrence Livermore National Laboratory.
Date/Time: August 23rd, 2024 at 1:20pm.
Abstract: Algicidal bacteria have ecological and industrial significance; they control algal bloom dynamics and are used as intermediaries in algal lipid extraction for biofuel synthesis, the latter contributing to carbon-neutrality. Algicidal bacteria are diverse in both their modes and compounds utilized in algal killing. The project sought to explain this diversity through a systems-level analysis of metabolism.
Towards this end, preliminary genome-scale metabolic models of Aeromonas hydrophila AD9, Kordia algicida OT-1, and Sacchariphagus degradans 2-40 – model organisms of various observed modes of algicidal activity – were drafted through the Dept. of Energy Systems Biology Knowledge Base’s (KBase) metabolic model builder. We sought to broaden functional coverage by developing an annotation ranking scheme that integrates annotations from KEGG, Rapid Annotation using Subsystem Technology toolkit (RASTtk), and Distilled & Refined Annotations for MAGs (DRAM) – an in-house annotation pipeline. We hypothesized that algicidal bacteria have access to the same algicidal toolkit, but environmental context coupled to metabolic constraints dictates the mode of action. To substantiate this claim, candidate proteins from each model were submitted for structural analysis of their active sites.
The model of Aeromonas hydrophila AD9 was selected for further refinement based on its algal efficiency and significance as an opportunistic human pathogen. The model was curated using phenotype data from Aeromonas hydrophila subsp. M800, which relates growth capacity to different carbon substrates. The model was first gap-filled per positive and negative phenotype data, allowing for judicious decision-making in the gap-filling algorithm. This resulted in an initial classification accuracy of 83% with respect to positives and negatives. Uptake rates of substrate were then constrained based on uptake predicted from parsimonious flux balance analysis (pFBA), maximizing growth while minimizing net flux through the network. Model predictions of bacterial growth in Luria-Bertani medium were in agreement with experimental rates. The metabolic burden (reduction of biomass production) of the algicidal metabolite clavulanate was then predicted in this medium. In the near future, we plan to integrate the results of our computational analyses into a population dynamics model that accounts for algal lipid consumption and detriment of algicidal compound synthesis on growth.