In silico characterization of hypothetical proteins from Paracoccidioides lutzii

Abstract

Nearly 60% of Paracoccidioides lutzii genes encode products annotated as hypothetical or predicted proteins (HPs). In this study, we describe the global detection and functional inference of HPs, using computational methods based on sequence similarity, identification of targeting signals, presence of known protein domains, and use of the Gene Ontology functional classification scheme. Our analysis enabled a high-throughput characterization of predicted cellular localization and presence of protein domains, clustering HPs into different functional categories including metabolism, localization, cell cycle, response to stimulus, and signaling. To investigate P. lutzii HP expression profiles, we used data obtained from the expressed sequence tag database (dbEST). These analyses revealed 2364 HPs expressed in different situations, namely in mycelial and yeast forms, during the transition from mycelium to yeast, and under conditions mimicking infection. Based on this transcriptomic data, we performed a functional enrichment analysis according to the domains present in the HPs expressed in each condition. The most overrepresented functional domains were those involved in the regulation of gene expression, suggesting important and as yet undescribed roles for these HPs in the adaptation of P. lutzii to environmental conditions. In addition, the expression profiles of six randomly selected HPs were analyzed by quantitative real-time polymerase chain reaction in order to verify their expression in the complementary DNA libraries analyzed in this investigation. The approach used in this study should improve functional characterization of P. lutzii HPs.