Abstract
Cryptococcus neoformans is a deadly fungal pathogen. Upon entering a mammalian host, it deploys a voluminous polysaccharide capsule that is necessary for it to survive host defenses and maintain an infection. Capsule expansion is regulated transcriptionally, as deletion of many transcription factors (TFs) alters the capsule. Thus, we set out to map the transcriptional regulatory network of C. neoformans-that is, to identify the TFs that directly regulate each gene in the genome. First, we carried out RNA-seq of 120 single-TF-deletion strains, together with wild-type controls. We then applied NetProphet3, a TF network mapping algorithm, to predict the direct functional targets of each TF. Unexpectedly, analysis of this network indicated that there are no TFs that primarily regulate genes involved in capsule formation. Rather, the TFs that play a role in deploying the capsule also regulate many other genes and processes. Comparison to a TF network map we built for Saccharomyces cerevisiae, a distantly related model yeast, identified pairs of TFs that are functionally orthologous-that is, their targets are enriched for orthologous genes. In many cases, these pairs are different from the ones identified by sequence homology alone. We suggest that network analyses should be used to complement sequence comparison when searching for functionally orthologous TFs. Our network map can be searched and visualized at https://cryptococcus.net/ . IMPORTANCE: Cryptococcus neoformans is a fungus that can cause life-threatening infections, in part by producing a protective capsule around itself. In this study, we analyzed how cryptococcal genes are turned on and off by its many transcription factors (TFs), the proteins that control gene activity. By studying mutant strains lacking 120 TFs and applying a powerful network analysis method, we found that no single TF is dedicated primarily to controlling capsule formation. Instead, the TFs that affect the capsule also influence many other processes. We also compared the cryptococcal network to that of a well-studied model yeast. We found that the yeast TF whose predicted protein sequence is most similar to a cryptococcal TF often regulates completely unrelated sets of target genes, while TFs with less sequence similarity often have more shared targets. This work shows the value of network-based approaches for uncovering hidden biological relationships important for infection and disease.