Redefining the spliceosomal introns of the sexually transmitted parasite Trichomonas vaginalis and its close relatives in columbid birds

Abstract

Trichomonas vaginalis infects the urogenital tract of men and women and causes the sexually transmitted infection trichomoniasis. Since the publication of its draft genome in 2007, the genome has drawn attention for several reasons, including its unusually large size, massive expansion of gene families, and high repeat content. The fragmented nature of the draft assembly made it challenging to obtain accurate metrics of features, such as spliceosomal introns. The number of introns identified varied over the years, ranging from 41 when first characterized in 2005, to 32 in 2018 when the repertoire was revised. In both cases, the results suggested that more introns could be present in the genome. In this study, we exploited our new T. vaginalis G3 chromosome-scale assembly and annotation and high-coverage transcriptome datasets to provide a definitive analysis of the complete repertoire of spliceosomal introns in the species. We developed a custom pipeline that distinguishes true splicing events from chimeric alignments by utilizing the extended motifs required by the splicing machinery and experimentally verified the results using transcript evidence. We identified a total of 63 active introns and 34 putative "inactive" intron sequences in T. vaginalis, enabling an analysis of their length distribution, extended consensus motifs, intron phase distribution (including an unexpected expansion of UTR introns), and functional annotation. Notably, we found that the shortest intron in T. vaginalis, at only 23 nucleotides in size, is one of the shortest introns known to date. We tested our pipeline on a chromosome-scale assembly of the bird parasite Trichomonas stableri, the closest known relative to T. vaginalis. Our results revealed some conservation of the main features (total intron count, sequence, length distribution, and motifs) of these two closely related species, although differences in their functional annotation and duplication suggest more specialized splicing machinery in T. vaginalis.

Fig 1.. Semi-automated pipeline for the identification and assembly of intron-containing genes in T. vaginalis using RNAseq and degenerate intron motifs.

In the first step, the RNAseq reads are mapped to the reference genome using STAR using default parameters (the mapping output file must be in a sorted BAM format and contain all SAM attributes). In the second step, SJs mapping to regions without the canonical intron motifs (degenerated intron motifs) are discarded. Finally, the transcript assembly is performed using the filtered version of the BAM file and StringTie using default parameters.

Fig 2.. An upset plot summarizing the intron content in T. vaginalis G3 described by different authors since 2005 and the present work.

The horizontal bars (dark cyan) indicate the total number of introns found by each author. Next to them, the intersection matrix shows the overlaps and unique elements among the different datasets. Finally, the height of the vertical bars (pink) indicates the number of elements in each intersection.

Fig 3.. The 35 new intron-containing transcripts found in the T. vaginalis G3 genome.

The new introns (delimited by vertical red lines) were identified by RNAseq (coverage shown in cyan) and validated by PCR (lane 1; low range ladder, lane 2; gDNA product and lane 3; cDNA product). The paralogs TVAGG3_0930320 and TVAGG3_0998030 were not included in this figure.

Fig 4.. Consensus sequence motif at the splice sites and intron features in T. vaginalis.

(A) The consensus sequence surrounding the exon-intron junction (+/− 20 bp) for the 63 intron-containing genes is shown; the heatmap shows the nucleotide frequency at each position; (B) length distribution of the new introns compared to those previously described and the complete repertoire; (C) the intron phase distribution (excluding the 9 UTR introns). In both cases (panels B and C), the violin plots display the distribution of the values, and the plot width represents the density of data points at that value.

Fig 5.. T. stableri BTPI-3 intron features.

Consensus sequence motifs and nucleotide frequency are shown in panel A, an histogtram of the intron length distribution is shown in panel B, and the PCR validation of the gene TsBTPI-3_004276 in panel C; lanes 1,2, and 3 are the replicates of the cDNA amplification, lane 4; RNA, lane 5 gDNA amplification, lane 6; negative control, and lane 7 the 1k Plus DNA ladder.