Leishmania infantum (JPCM5) Transcriptome, Gene Models and Resources for an Active Curation of Gene Annotations

Abstract

Leishmania infantum is one of the causative agents of visceral leishmaniases, the most severe form of leishmaniasis. An improved assembly for the L. infantum genome was published five years ago, yet delineation of its transcriptome remained to be accomplished. In this work, the transcriptome annotation was attained by a combination of both short and long RNA-seq reads. The good agreement between the results derived from both methodologies confirmed that transcript assembly based on Illumina RNA-seq and further delimitation according to the positions of spliced leader (SAS) and poly-A (PAS) addition sites is an adequate strategy to annotate the transcriptomes of Leishmania, a procedure previously used for transcriptome annotation in other Leishmania species and related trypanosomatids. These analyses also confirmed that the Leishmania transcripts boundaries are relatively slippery, showing extensive heterogeneity at the 5'- and 3'-ends. However, the use of RNA-seq reads derived from the PacBio technology (referred to as Iso-Seq) allowed the authors to uncover some complex transcription patterns occurring at particular loci that would be unnoticed by the use of short RNA-seq reads alone. Thus, Iso-Seq analysis provided evidence that transcript processing at particular loci would be more dynamic than expected. Another noticeable finding was the observation of a case of allelic heterozygosity based on the existence of chimeric Iso-Seq reads that might be generated by an event of intrachromosomal recombination. In addition, we are providing the L. infantum gene models, including both UTRs and CDS regions, that would be helpful for undertaking whole-genome expression studies. Moreover, we have built the foundations of a communal database for the active curation of both gene/transcript models and functional annotations for genes and proteins.

Keywords: Leishmania; Mendeley data; Wikidata; gene expression; gene models; transcriptome.

Figure 1

Complex transcription patterns were observed at particular L. infantum loci. (A) According to Illumina RNA-seq reads coverage and position of SL-addition sites (black vertical arrows), a single transcript was deduced to be expressed from genes LINF_130006600 and LINF_130006600 (horizontal arrow shows the transcriptional direction). The existence of a complete Iso-Seq transcript (molecule shaded in red) derived from gene LINF_130006600 indicated that this gene is expressed into a mature individual RNA. However, no evidence of an individual mRNA for gene LINF_130006700 was found, suggesting that this gene may be expressed as a bicistronic transcript expanding both genes (transcript LINF_130006700-600-T). (B) Illumina RNA-seq reads coverage on gene LINF_240027300 suggested the existence of a sole mRNA molecule for this gene. However, the presence of two complete Iso-seq molecules indicated that two stable mRNAs are produced (LINF_240027300-T2 and LINF_240027300-T3) together with a third molecule expanding the complete gene (LINF_240027300-T1).

Figure 2

Allelic heterogeneity on a genomic region of L. infantum chromosome 32. (A) Two types of transcripts (Iso-Seq molecules) were identified: one derived from gene LINF_320009800 and the other showing a chimeric nature (derived from separated genomic regions). (B) The existence of a lower number of PacBio DNA-seq reads in the region-expanding genes LINF_320009700, and LINF_320009800 suggested an intrachromosomal deletion event affecting one allele. The hypothesized structure for both alleles is depicted together with the PacBio DNA-seq reads coverage on both regions. The existence of a direct repeat of 695 nt in length is denoted by black boxes. (C) Experimental testing by PCR of the deletion event. Two pairs of oligonucleotides (F-R and F-R’) were used to amplify the 3′-UTR of gene LINF_320009700 and the whole genomic region (genes LINF_320009700, LINF_320009800 and LINF_320009900), respectively. The PCR products were separated on a 1% agarose gel (panel C1). A band of around 8.7 kb was also observed after PCR amplification with oligonucleotides F and R’ (panel C2). At the boundaries of gels are shown the sizes (in nucleotides) of the HindIII-digested Ø29 phage DNA bands used as a molecular marker (M). Lanes 3 (gel C1) and 2 (gel C2) are controls in which PCR was performed in the absence of a DNA template.

Figure 3

Steps for creating gene models. (A) Transcripts are assembled from Illumina RNA-seq reads. (B) SL-containing RNA-seq reads are used to map the SL-addition sites (SASs); the main SAS (vertical arrow) is assigned to that supported by the largest number of SL-containing RNA-seq reads, and the rest as categorized as alternatives (see Supplementary File S1 for complete lists of main and alternative SASs). (C) Transcript is trimmed according to the position of the main SAS. (D) Coding sequences (CDS) are automatically annotated by the Companion tool. (E) If CDS surpasses transcript 5′-end, the CDS is manually shortened at the 5′-end. (F) After embedding the CDS into the transcript sequence, the gene model emerges with their 5′ and 3′ untranslated regions (UTRs).