A Network Approach of Gene Co-expression in the Zea mays/Aspergillus flavus Pathosystem to Map Host/Pathogen Interaction Pathways.

Bryan M Musungu,Greg Obrian,Robert L Brown,Gary A Payne,Matt Geisler,Ahmad M Fakhoury,Deepak Bhatnagar

doi:10.3389/fgene.2016.00206

Abstract

A gene co-expression network (GEN) was generated using a dual RNA-seq study with the fungal pathogen Aspergillus flavus and its plant host Zea mays during the initial 3 days of infection. The analysis deciphered novel pathways and mapped genes of interest in both organisms during the infection. This network revealed a high degree of connectivity in many of the previously recognized pathways in Z. mays such as jasmonic acid, ethylene, and reactive oxygen species (ROS). For the pathogen A. flavus, a link between aflatoxin production and vesicular transport was identified within the network. There was significant interspecies correlation of expression between Z. mays and A. flavus for a subset of 104 Z. mays, and 1942 A. flavus genes. This resulted in an interspecies subnetwork enriched in multiple Z. mays genes involved in the production of ROS. In addition to the ROS from Z. mays, there was enrichment in the vesicular transport pathways and the aflatoxin pathway for A. flavus. Included in these genes, a key aflatoxin cluster regulator, AflS, was found to be co-regulated with multiple Z. mays ROS producing genes within the network, suggesting AflS may be monitoring host ROS levels. The entire GEN for both host and pathogen, and the subset of interspecies correlations, is presented as a tool for hypothesis generation and discovery for events in the early stages of fungal infection of Z. mays by A. flavus.

Highlights

A systems biology approach can greatly aid in understanding the dynamics of fast changing systems like stress and defense responses in plants, and allows for both discovery and novel hypotheses generation
The individual reads were competitively mapped to a combined gene model set of both A. flavus (13775 gene models) and Z. mays (181929 gene models). 44264 gene models had unique mapped reads in at least one pooled kernel sample, while 151440 gene models did not have any expression (0 unique mapped reads) across all the samples. 6035 genes were selected for co-expression analysis based upon their significant level of differential expression
When visualized using centroid graphs from K-means clustering to observe expression patterns there was a high level of variation between Z. mays and A. flavus at 12 hpi, 48 hpi, and 72 hpi for genes in both systems

Summary

Introduction

A systems biology approach can greatly aid in understanding the dynamics of fast changing systems like stress and defense responses in plants, and allows for both discovery and novel hypotheses generation. “Omics” based techniques, such as gene co-expression networks (GENs), have aided in. Dual RNA-seq of A. flavus/Z. mays the discovery and functional analysis of genes and pathways involved in the response to biotic and abiotic stresses in plants (Sekhon et al, 2013). Gene expression analysis has been done individually on plants and fungi such as Zea mays, Glycine max (Libault et al, 2010), Arabidopsis thaliana (Mao et al, 2009), and Saccharomyces cerevisiae (Carlson et al, 2006). Co-expression networks, especially those that have relied on microarrays, have been limited by the size of the array used to generate the data, which can lead to a loss in information. The more extensive RNA-sequencing based studies far have often been limited to either the host or the pathogen

Methods

Results

Discussion

Conclusion