Comparison of statistical methods of inference of cooccurrence networks within micro- bial ecosystems from metagenomics data - Archive ouverte HAL
Poster De Conférence Année : 2017

Comparison of statistical methods of inference of cooccurrence networks within micro- bial ecosystems from metagenomics data

Résumé

Post translational modifications (PTM) enhance the functional diversity of proteins. Among these PTM, glycosylation represents one of the most abundant and complex modification of proteins and N-glycosylation is determinant for cell survival. For 96% of N-glycosylproteins, glycosylation occurs at the canonical motif [N]-!P-[S/T] known as sequon [1]. In the laboratory, we study the evolutionary relationships and molecular evolution of glycosyltransferases that are Golgi enzymes implicated in the last steps of glycosylation [2] like B4GALNT2 implicated in the synthesis of blood group antigen Sda [3]. This enzyme is a glycoprotein and our working hypothesis is that the gain or loss of an ancestrally conserved N-glycosylation site in this enzyme might have resulted in the evolution of protein structure, subcellular localization and functional modifications with impact on the phenotype [4-7]. As a first towards answering this question, we identified over 150 B4GALNT-related sequences (37 B4GALNT1 paralogues and 85 B4GALNT2 orthologues) from public databases (NCBI, Ensembl, CAZy) using a BLAST approach. We carried out sequence-based analysis using various multiple sequence alignment algorithms (MUSCLE, ClustalΩ, MAFFT, T-Coffee, ClustalW) to delineate informative regions and define conserved sequence motifs. Furthermore, we used NetNGlyc 1.0 server [8] and GlycoMine [9] to predict potential N-glycosylation sites. We found differentially conserved N-glycosylation sites in the homologous B4GALNT sequences: a canonical sequon, [N]-!P-[S/T], present in 11 mammalian B4GALNT2 sequences, which has disappeared in the human B4GALNT2 sequence [10]. This N-glycosylation site is also present in some ancestral B4GALNT sequences and has disappeared in the paralogous B4GALNT1 sequences suggesting no major role for this N-glycan. Interestingly, two atypical N-glycosylation sites [N]-X-[C] and [N]-[G] were predicted in mammalian B4GALNT2 sequences. Despite the fact that the [N]-[G] site is present in almost all the mammalian B4GALNT2 sequences, we found using PNGase F, an enzyme that remove N-Glycans, that the human B4GALNT2 enzyme is not glycosylated at this position, but at a second unusual [N]-X-[C] glycosylation site. This latter was found to be highly conserved in nearly all B4GALNT sequences except the teleostean B4GALNT2 sequences further suggesting an important functional role, which is currently investigated.
Fichier principal
Vignette du fichier
comparison-jobim_2017_1.pdf (19.71 Mo) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte
Loading...

Dates et versions

hal-01563604 , version 1 (02-06-2020)

Licence

Identifiants

  • HAL Id : hal-01563604 , version 1
  • PRODINRA : 398242

Citer

Sophie Schbath, Mahendra Mariadassou, Julie Lao. Comparison of statistical methods of inference of cooccurrence networks within micro- bial ecosystems from metagenomics data. JOBIM 2017 - Journées Ouvertes Biologie Informatique Mathématiques, Jul 2017, Lille, France. , pp.234, 2017, JOBIM 2017. ⟨hal-01563604⟩
158 Consultations
496 Téléchargements

Partager

More