Enhanced H2AK119 deubiquitinating activity, leading to secondary loss of histone H3 lysine 27 trimethylation (H3K27me3) marks and the widespread activation of genomic loci which are otherwise epigenetically silenced [58]. The mutant isoforms typically have a domain structure similar to ASXL-TF except that they Lenvatinib generally (but not always) lack the C-terminal EH[N/S]Y motif. Further, the mutant forms will presumably be translated at the normal level of ASXL (though the resulting mutant proteins may be unstable) whereas ASXL-TF should be less efficiently expressed and/or regulated. Thus we predict that disease-associated ASXL truncation mutants represent defective overexpressed versions of a natural functionally important ribosomal frameshifting product ASXL-TF.Conclusions In this study, we have demonstrated the existence of conserved overlapping ORFs in the +1 frames of the vertebrate ASXL1 and ASXL2 genes. These “TF ORFs” coincide with highly significant synonymous site conservation in the corresponding zero-frame sequences, as is typically observed for dual coding regions, where selective constraints operate on more than one reading frame [28]. The TF ORFs are not canonically translatable within any known splice variants of the human ASXL genes. However, the presence of highly conserved motifs at the 5 ends of the TF ORFs, identical to known +1/-2 PRF shift sites, suggests that the TF peptides may be expressed as transframe fusions with zero-frame ASXL via ribosomal frameshifting. The third paralogue, ASXL3, is more divergent and has neither a conserved long overlapping +1 frame ORF nor a conserved +1-frame EHxY motif. Methods The initial set of 37,257 human PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/16474207 mRNA RefSeqs was downloaded from NCBI in January 2010. To identify further orthologues of ASXL1 and ASXL2, tblastn searches of the nr/nt NCBI database were carried out, using selected reference species for each major vertebrate clade. For clades with poor representation in the nr/nt database, sequences from the transcriptome shotgun assembly (TSA) database were added, where available. To avoid the detection of non-orthologous genes, a minimum query coverage threshold of 80 was used to filter all results. Nucleotide sequences for each gene were retrieved from NCBI. RefSeq entries with missing or incomplete sequence data that precluded further analyses were discarded. Where multiple transcript isoforms were present for a given species, the lowest numbered isoform was used for analyses.Dinan et al. Biology Direct (2017) 12:Page PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/11719833 10 ofThe ASXL2 gene predictions of many sauropsids (reptiles and birds) have large ( 150 aa) N-terminal deletions relative to those of all other major vertebrate clades. Because we set a minimum query coverage threshold of 80 , these genes were excluded from our analyses. However, examination of the relevant genomic and transcriptomic sequences shows that all contain conserved copies of the frameshift site ?with the exception of lizards, as discussed above ?and large TF ORFs in the expected frames. For each of ASXL1 and ASXL2, full-length zero-frame coding nucleotide sequences were translated, aligned as amino acids with MUSCLE [64], and the amino acid alignments were used to guide codon-based nucleotide alignments using EMBOSS tranalign [65]. The zeroframe sequences of all mRNAs were scanned for the relevant +1 or -2 PRF sites. Sequence alignments were analysed for synonymous site conservation using synplot2 [28] with amino acid PhyML guide trees [66]; for thes.
