Complete minicircle genome of Leptomonas pyrrhocoris reveals sources of its non-canonical mitochondrial RNA editing events

Abstract

Uridine insertion/deletion (U-indel) editing of mitochondrial mRNA, unique to the protistan class Kinetoplastea, generates canonical as well as potentially non-productive editing events. While the molecular machinery and the role of the guide (g) RNAs that provide required information for U-indel editing are well understood, little is known about the forces underlying its apparently error-prone nature. Analysis of a gRNA:mRNA pair allows the dissection of editing events in a given position of a given mitochondrial transcript. A complete gRNA dataset, paired with a fully characterized mRNA population that includes non-canonically edited transcripts, would allow such an analysis to be performed globally across the mitochondrial transcriptome. To achieve this, we have assembled 67 minicircles of the insect parasite Leptomonas pyrrhocoris, with each minicircle typically encoding one gRNA located in one of two similar-sized units of different origin. From this relatively narrow set of annotated gRNAs, we have dissected all identified mitochondrial editing events in L. pyrrhocoris, the strains of which dramatically differ in the abundance of individual minicircle classes. Our results support a model in which a multitude of editing events are driven by a limited set of gRNAs, with individual gRNAs possessing an inherent ability to guide canonical and non-canonical editing.

Figure 1.

(A) A scheme of a typical L. pyrrhocoris H10 minicircle with two conserved regions (CRs). The zero coordinate is placed at the conserved sequence block 1 (CSB1) within the CR of the left unit. Fifty-five of 67 minicircles are asymmetric in that they possess a left unit encoding a gRNA and a right unit lacking one as shown. The left monomer almost always bears a gRNA and a downstream motif as shown, based on alignments to edited mRNAs and small RNA read coverage. The right monomer is usually inactive (gray) based on small RNA read coverage, but sometimes contains a putative gRNA locus and/or a highly diverged downstream motif (potential locations indicated). Universal minicircle CSB1 and CSB3 are shown. (B) Scatterplot comparing the number of shotgun DNA sequencing reads mapped (a proxy of copy number) to the number of total RNA-seq reads, mapped for each minicircle, and the linear regression line for these values with its R² value. Each data point on the plot reflects the dimeric minicircle, each dot size is proportional to the level of small RNA sequencing reads mapped on monomeric unit with highest expression (usually the left). (C) Relative individual minicircle class abundances in the kDNA of various L. pyrrhocoris strains, as determined by the number of shotgun DNA sequencing reads mapped on each minicircle. The Y-axis shows read counts for a particular minicircle class divided by the number of reads mapped to the most abundant minicircle class. Each bar shows the coverage of a single minicircle class, with its placement along the X-axis determined by its relative abundance in strain H10 from the highest (left) to the lowest (right). From top to bottom panels show L. pyrrhocoris strains H10, F19, F165, 25EC, 324RV and P59.

Figure 2.

Phylogeny of minicircle monomeric units of Leptomonas pyrrhocoris and Leptomonas seymouri. The background colors encompass clades of monomeric units. The L. seymouri minicircles are composed of one monomer each from the yellow and green clades. The L. pyrrhocoris minicircles are composed either of one monomer each from blue (B) and gray (g) clades (‘Bg- type’), or of both monomers from blue clade (‘2B-type’). Operational taxonomic units that contribute to 2B minicircles are marked with red circles. Pink bar heights for each L. pyrrhocoris monomeric unit are proportional to the negative logarithm of the e-value of finding the 66-mer motif on the respective monomer: higher bars represent higher confidence of motif detection.

Figure 3.

The canonical edited mRNA of the Leptomonas pyrrhocoris RPS12 cryptogene showing the positions of putative gRNA:mRNA alignments. In the alignment string, ‘|’ indicates match; ‘:’ indicates a G:U base pair; ‘∧’ indicates mismatch. Edited RPS12 mRNA sequence is highlighted in blue. Three gRNAs (H10_mc_15:L, H10_mc_12:L and H10_mc_22:L) can also guide the editing of the canonical mRNA of the ND9 cryptogene, alignments to the relevant sequence of ND9 are highlighted in red. Primary gRNAs are highlighted in green. Us that correspond to Ts present at the DNA level are marked with red (deletions) and blue arrows. T-Aligner predicted start and stop codons are boxed in red.

Figure 4.

Comparison of gRNA characteristics of three species: Leptomonas pyrrhocoris (L. pyr), Leishmania tarentolae (L. tar; datasets for the analysis were taken from (25)), Trypanosoma brucei (T. bru; datasets for analysis were taken from (24)), and Leishmania tarentolae where characteristics were derived by utilizing our algorithm and parameters on data from (25), noted as ‘Lt(tal)’ (T-aligner derived) violin plot in all subfigures. (A) Distributions of templated gRNA lengths in these species. For L. pyrrhocoris the boxplot on the far left shows the distribution of gRNA length calculated using the lengths as determined by minicircle coverage of short RNA sequencing reads; boxplots 2, 3 and 4 are calculated based on the length of minicircle:mRNA alignments as in (B) and (C). (B) Species-specific distribution of percent of the G:U base pairs of the total pairings per gRNA for all gRNAs. (C) Species-specific distribution of percent of mismatches of the total pairings per gRNA for all gRNAs.

Figure 5.

Scenario for multiple alternative editing pathways guided by the same gRNA. (A) U insertion patterns observed in sequences from reads recovered by alignment to a single RPS12 gRNA with an algorithm permitting a degree of mismatch. Read support for each pattern is given in the column on the right. The anchor region is boxed. A rainbow color scheme was used to highlight homologous As between gRNA, pre-sequence, and the edited mRNA reads in each pattern for tracking differences in editing. For example, if the rightmost A (red) pairs with the first U of a gRNA upstream of the anchor, then the AGG nucleotides of the gRNA act as guiding nucleotides, directing the insertion of three Us after this A. However, a group of alternative patterns could have resulted from the rightmost A pairing with the first G upstream of the anchor (which would be considered a mismatch pairing, excluding this G from guiding). (B) Translation of the editing patterns in (A) to the T-Aligner editing state matrix. The X coordinate is equal to the number of the A/G/C cryptogene reference nucleotide, the Y coordinate is the number of inserted or deleted Us after this A/G/C nucleotide (where Y = 0 is the reference level) that were observed in any of the read sequences in (A). Dots representing reference As from (A) are colored in the same way. For example, after the second A (orange), sequences were recovered from the reads in which either 0, 1, 2 or 4 Us were inserted. The canonical editing pattern is shown as lines connecting each respective canonical editing state. (C) The T-Aligner editing state matrix for cryptogene RPS12. The scheme represents the numbers of inserted\deleted Us as dots, located above (insertion) or below (deletion) the reference line, with its X coordinate representing an A/G/C nucleotide position in the cryptogene sequence. Black line shows the path of the canonical edited mRNA for the cryptogene. The editing states observed only in the total RNA sequencing read mapping appear as black dots. The editing states that are also supported by gRNA:read alignments are circled in blue. The red box indicates the region of the RPS12 cryptogene that is edited with mc_48:RPS12 gRNA with alternative outcomes shown in (A).

Figure 6.

A visualization of proposed canonical and alternative editing. Two sequential editing steps are presented. The RPS12 editing states dot matrix (as in Figure 5C) is shown at the bottom of the figure, depicting the number of editing states observed from raw read mapping that can be attributed to editing with one of our identified gRNAs as blue-circled editing events. Similar dot matrices for additional cryptogenes are found in Supplementary Figure S7. The red square on the dot matrix indicates the region edited with canonical RPS12 gRNA of minicircle #48, and all patterns produced with that gRNA are depicted on the right panel as in Figure 5A. The green square and left panel shows three possible alternatives the next gRNA to bind and guide transcript editing, which was canonically edited with mc_48:L gRNA. Rainbow coloring highlights homologous nucleotides in each pattern. In each case cryptogene reference DNA is aligned with the most supported pattern. Approximately 85% of reads support editing with canonical gRNA coming from minicircle #15 that has primary alignment with RPS12. Alternative patterns are generated with non-canonical gRNAs that have primary alignments with A6 and ND8. Black boxes indicate the anchoring nucleotides used by mc_48:L and mc_15:L gRNAs from canonical RPS12 pathway.