Metabolic modeling predicts specific gut bacteria as key determinants for Candida albicans colonization levels

Mohammad H Mirhakkak,Glen J Weiss,Hortense Slevogt,Ruben V Uribe,Gianni Panagiotou,Morten O A Sommer,Slavena Vylkova,Sascha Brunke,Tilman E Klassert,Philipp Brandt,Daniel Loos,Sascha Schäuble,Yueqiong Ni,Felipe Senne De Oliveira Lino,Bernhard Hube

doi:10.1038/s41396-020-00848-z

Abstract

Candida albicans is a leading cause of life-threatening hospital-acquired infections and can lead to Candidemia with sepsis-like symptoms and high mortality rates. We reconstructed a genome-scale C. albicans metabolic model to investigate bacterial-fungal metabolic interactions in the gut as determinants of fungal abundance. We optimized the predictive capacity of our model using wild type and mutant C. albicans growth data and used it for in silico metabolic interaction predictions. Our analysis of more than 900 paired fungal–bacterial metabolic models predicted key gut bacterial species modulating C. albicans colonization levels. Among the studied microbes, Alistipes putredinis was predicted to negatively affect C. albicans levels. We confirmed these findings by metagenomic sequencing of stool samples from 24 human subjects and by fungal growth experiments in bacterial spent media. Furthermore, our pairwise simulations guided us to specific metabolites with promoting or inhibitory effect to the fungus when exposed in defined media under carbon and nitrogen limitation. Our study demonstrates that in silico metabolic prediction can lead to the identification of gut microbiome features that can significantly affect potentially harmful levels of C. albicans.

Highlights

We either selected five bacterial species that significantly correlated with C. albicans by relative abundance and two that significantly correlated by Growth Rate InDex (GRiD) value, or selected bacteria based on whether they increase performance as described in the main text
To develop a C. albicans genome-scale metabolic models (GSMMs), we started with a model automatically generated by the Comparative
Resolving energy-generating cycles (EGCs) reduced the flux through the biomass reaction towards 1.4, which is closer to the biomass reaction flux for other fungal models such as the yeast consensus model [35, 36] (Fig. 1A)

Summary

Introduction

The fungus Candida albicans is found on the mucosal surfaces of at least 50–70% of healthy adults [1] and is a classic opportunistic pathogen. Neither the dynamics of Candida species in the human gut nor the specific microbial contributors to the observed reduction have been studied, Based on these diverse findings, the commensal status of C. albicans appears to be related to the global taxonomy and functionality of the host microbiome. Recent pioneering studies have developed highquality GSMMs for gut bacteria that enable in silico analysis of gut metabolic functions and interactions [18, 19] These resources have advanced the study of gut microbes and their respective pairwise interactions but have not yet been used to study interactions with opportunistic fungal pathogens such as C. albicans. We assessed fungal growth in bacterial spent media experiments

Materials and methods

Results and discussion

B Western diet

Compliance with ethical standards