Variety of proline-specific peptidases in higher fungi N.A. Alkin1, Y.E. Dunaevsky2, M.A.Belozersky2, G.A.Belyakova1, V.F. Tereshchenkova3, I.Y.Filippova3, E.N. Elpidina2 1Department of Mycology and Algology, Faculty of Biology, Moscow State University, Moscow, Russia. [email protected] ; 2Department of Plant Proteins, A.N.Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow, Russia. [email protected], 3Division of Natural Compounds, Chemical Faculty, Moscow State University, Moscow, Russia. Introduction. Proline has a unique secondary amino group among other proteinogenic amino acids. This group provides resistance of peptide bonds formed by proline residues to peptidases with broad substrate specificity. To hydrolyze such bonds proline-specific peptidases (PSP) are used by the variety of living organisms. The most common roles of PSP are digestion of proline- rich proteins (for instance, wheat gliadins), processing and degradation of short bioactive peptides. Although PSP are found in all three domains of life, still the information about the distribution and variety of these enzymes among fungal species stays rather limited. Furthermore, fungal PSP are valuable for dairy and beer industry as it was demonstrated that proline-rich peptides including those of milk and barley may contribute a bitter taste to these food products [1]. The objective of this research is a search for genes of PSP in genomes of different species of dikaryotic fungi (Ascomycota and Basidiomycota) and subsequent analysis of predicted PSP sequences. Materials and methods. The search of sequenced and annotated fungal genomes was accomplished in NCBI public database (https://www.ncbi.nlm.nih.gov/). PSP amino acid sequences of chosen fungal species were obtained by Protein BLAST service (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The query used included 13 human PSP sequences, 2 sequences of nonspecific human exopeptidases capable of hydrolyzing the bonds formed by the carboxyl group of any amino acids including proline, and sequence of well characterized Aspergillus niger acid proline endopeptidase (PEP). Amino acid sequences of human PSP and information about their active sites was achieved in SWISS-PROT and TrEMBL databases (http://www.uniprot.org/), nucleotide sequence of acid PEP from A. niger was obtained from StrainInfo database (http://www.straininfo.net/) and then translated into amino acid sequence using the standard genetic code. The homologs with E-value lower than 1*10-20 and cover higher than 75% were used in subsequent analysis. Isoelectric point, number of amino acid residues and molecular weight were calculated via Isoelectric point calculator website (http://isoelectric.org/calculate.php) for each homolog found. Multiple alignment of amino acid sequences and phylogenetic tree building were performed via COBALT service (https://www.ncbi.nlm.nih.gov/tools/cobalt/cobalt.cgi). Fast minimal evolution algorithm was used for construction of dipeptidyl peptidase 4 (DPP4) phylogenetic tree. Mutations in the active sites of inspected PSP homologs were identified in comparison to the amino acid sequences of human PSP after the alignment. Potential transmembrane domains were analyzed using TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/) and potential signal peptides were predicted through SignalP-5.0 server (http://www.cbs.dtu.dk/services/SignalP). Batch CD- search service (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) was used to identify conserved domains in analyzed homologs. Results and discussion. We have analyzed PSP in 42 fungal species with sequenced genomes from every subphylum of the subkingdom Dikarya (2 species of Taphrinomycotina, 4 species of Saccharomycotina, 27 species of Pezizomycotina, 1 species of Pucciniomycotina, 3 species of Ustilaginomycotina and 6 species of Agaricomycotina). The chosen species represent different life forms (mycelial, yeast and dimorphic) and ecological niches (biotrophic and necrotrophic plant pathogens, xylotrophic and saprotrophic free-living fungi, animal parasites and nematophagous fungi, alkaliphilic fungi, etc.). All analysed fungal genomes lacked homologs for fibroblast activation protein (FAP), dipeptidyl peptidases 6, 8, 9, 10 and aminopeptidase 2 (APP 2). On the contrary, some PSP, such as aminopeptidase P1 (APP1), APP3 and DPP4 were present in all cases. The obtained data demonstrates that there may be a taxon-specific pattern of the PSP distribution among the investigated fungal species. Indeed, prolyl oligopeptidase (POP) and leucine aminopeptidase (LAP) homologs are present in the genomes of every studied species of phylum Basidiomycota and are absent in all studied species of phylum Ascomycota, with the exceptions of entomopathogen Beauveria bassiana (Sordariomycetes) and saprobic soil yeast Saitoella complicata (Taphrinomycetes), whereas the other examined species of those groups (Cordyceps confragosa, Metarhizium anisopliae (Sordariomycetes), Taphrina deformans (Taphrinomycetes)) lack homologs for both POP and LAP. Homologs of A. niger acid PEP are detected in all examined species excluding those, which are members of Saccharomycotina subphylum (Saccharomyces cerevisiae, Candida albicans, Wickerhamomyces anomalus and Yarrowia lipolytica), characterized by yeast or dimorphic yeast-mycelial life forms. Close homolog of human prolyl carboxypeptidase (PRCP) is found only in genome of nematophagous fungus Pochonia chlamydosporia (Clavicipitaceae, Hypocreales, Sordariomycetes), while genomes of the other examined species lack such homologs. Prolidase (XPD) homologs are detected in 36 of 42 examined fungal species, however the distribution of this enzyme among taxa is rather mosaic: homologs of XPD are found in genomes of both ascomycetes and basidiomycetes, still both taxa include species which lack this enzyme (for instance, Taphrina deformans from phylum Ascomycota, Ustilago maydis and Tilletia indica from phylum Basidiomycota). Human dipeptidyl peptidase 2 (DPP2) demonstrates low level of homology with fungal proteins as only 2 possible fungal homologs are detected in B. bassiana and Colletotrichum tofieldiae (Glomerellaceae, Glomerellales, Sordariomycetes). Most of the PSP sequences have the same active site composition as corresponding human PSP. The conservative amino acid triad of serine peptidases (serine, histidine, and aspartic acid) is present in each discovered fungal homolog of DPP4 and POP; metal binding sites are found in every analyzed APP1 and APP3 homolog, which is the evidence in favor of enzyme activity of these peptidases. Every fungal homolog of human PSP includes the same set of conserved domains as the query sequence, which may be another argument for functionality of the homologs. A wide occurrence and distinct amino acid sequence conservatism of DPP4 and PSP among fungal species may indicate a high biological significance of these PSP. Isoelectric point (pI) of the vast majority of detected PSP ranges within pH 5-6, though a third part of acid PEP homologs is characterized by pI below 5 and one of two APP1 homologs of every examined Agaricomycotina species has no electric charge at pH 8-9. Analysis of the fungal PSP on the presence of potential transmembrane domains and signal peptides demonstrates that the most of obtained homologs are more likely localized inside the cell. It should be noted that many found homologs with potential transmembrane domain lack signal peptide and vice versa and it remains unclear how these peptidases reach cell membrane. It is noticed that fungal genomes may contain both transmembrane and secreted homologs of the PSP known in humans as transmembrane only. Thus, it may be expected that these homologs are paralogs and perform different functions. The phylogenetic analysis of fungi and human DPP4 related sequences showed that fungal part of phylogenetic tree consists of three well-resolved clades: secreted DPP4 homologs (found in ascomycetes only), transmembrane DPP4 homologs of basidiomycetes, and transmembrane DPP4 homologs of ascomycetes with the last two forming a cluster. When examining each of these three clades separately, remarkable accordance with the results gained in the last reviews on fungal systematics is achieved [2-3]. One of the possible interpretations may be that the ancestral form of higher fungi S9B peptidases was transmembrane and its gene was duplicated in the early ascomycetes soon after the Ascomycota-Basidiomycota divergence, so that the second copy became secreted. It is revealing that the matching between the phylogenetic tree of DPP4 sequences and the independent data from aforementioned works is evident both at high-level taxa (phyla) and at much lower taxa like families and even genera. This result demonstrates a potential opportunity of using the most widespread fungal PSP (APP1, DPP4) as additional markers in phylogenetic analysis. The absence of FAP and DPP6 homologs in fungal genomes alongside with abundance of DPP4 homologs can be explained by diversification of those three related animal peptidases from the common ancestral one belonging to S9B family. This point of view is supported by the phylogenetic tree of DPP4 homologs, in which human DPP4, porcine DPP4, human FAP, bovine FAP and human DPP6 form a sister outgroup to all studied fungal homologs of S9B proteases. Conclusions 1. A systematic analysis of PSP variety in fungal genomes was performed for the first time. 2. APP1, APP3, DPP4, CND and PEP were present in the majority of 42 studied