Extensive Expansion of A1 Family Aspartic Proteinases in Fungi Revealed by Evolutionary Analyses of 107 Complete Eukaryotic Proteomes

Abstract

The A1 family of eukaryotic aspartic proteinases (APs) forms one of the 16 AP families. Although one of the best characterized families, the recent increase in genome sequence data has revealed many fungal AP homologs with novel sequence characteristics. This study was performed to explore the fungal AP sequence space and to obtain an in-depth understanding of fungal AP evolution. Using a comprehensive phylogeny of approximately 700 AP sequences from the complete proteomes of 87 fungi and 20 nonfungal eukaryotes, 11 major clades of APs were defined of which clade I largely corresponds to the A1A subfamily of pepsin-archetype APs. Clade II largely corresponds to the A1B subfamily of nepenthesin-archetype APs. Remarkably, the nine other clades contain only fungal APs, thus indicating that fungal APs have undergone a large sequence diversification. The topology of the tree indicates that fungal APs have been subject to both “birth and death” evolution and “functional redundancy and diversification.” This is substantiated by coclustering of certain functional sequence characteristics. A meta-analysis toward the identification of Cluster Determining Positions (CDPs) was performed in order to investigate the structural and biochemical basis for diversification. Seven CDPs contribute to the secondary structure of the enzyme. Three other CDPs are found in the vicinity of the substrate binding cleft. Tree topology, the large sequence variation among fungal APs, and the apparent functional diversification suggest that an amendment to update the current A1 AP classification based on a comprehensive phylogenetic clustering might contribute to refinement of the classification in the MEROPS peptidase database.

Aspartyl Proteinases of Eukaryotic Microbial Pathogens: From Eating to Heating

“…Aspartyl (or aspartic) proteinases (APs) are a class of proteinases (or proteases) highly conserved from retroviruses, including the HIV-1 protease, to mammals, including pepsins, cathepsins, and renins [1]. APs of eukaryotic pathogens usually have two-domain structures, each of which provides a catalytic Asp residue to the active enzymatic site.…”

“…The relevance of APs for the success of eukaryotic pathogens as infectious agents is reflected in the APs’ redundancy and organization in protein families with distinctive but genetically related members [7,8], an evolutionary expansion that appears to have reached a particularly high level of diversification in some fungal organisms [1]. Functionally, this organization enables the pathogen to select the right AP at the right time and in the right place to exploit synergistic effects or to use alternative APs when one is lost or inactive, thereby compensating for the biological cost of having many copies of the same gene [1,7–9].…”

“…Functionally, this organization enables the pathogen to select the right AP at the right time and in the right place to exploit synergistic effects or to use alternative APs when one is lost or inactive, thereby compensating for the biological cost of having many copies of the same gene [1,7–9]. …”

Substrate binding and specificity appear as major forces in the functional diversification of eqolisins

“…The seed HMMER profile was made by means of hmmbuild using default settings and the MEROPS [15] G1 alignment of holozymes (HMMER Version 3.0 [16] ). This was used to 100 iteratively screen the 107 complete proteome dataset previously used to identify Aspartic Proteases [17] and the HMMER NR database (W1). All sequences identified with an E-value smaller than the HMMER exclusion threshold were considered as Eqolisin homologues.…”

“…A profile made from the holotype sequences of MEROPS [38] was used as a seed. Initially, a collection previously used for a study of aspartic proteases [16], of mostly fungal, 7 140 145 150 155 but also other taxonomically well distributed eukaryotic complete proteomes were used. This resulted in few sequences, hence, the NR database was scanned and all identified sequences were combined and searched iteratively until data convergence.…”