Multiple G-quartet structures in pre-edited mRNAs suggest evolutionary driving force for RNA editing in trypanosomes

Abstract

Mitochondrial transcript maturation in African trypanosomes requires a U-nucleotide specific RNA editing reaction. In its most extreme form hundreds of U's are inserted into and deleted from primary transcripts to generate functional mRNAs. Unfortunately, both origin and biological role of the process have remained enigmatic. Here we report a so far unrecognized structural feature of pre-edited mRNAs. We demonstrate that the cryptic pre-mRNAs contain numerous clustered G-nt, which fold into G-quadruplex (GQ) structures. We identified 27 GQ's in the different pre-mRNAs and demonstrate a positive correlation between the steady state abundance of guide (g)RNAs and the sequence position of GQ-elements. We postulate that the driving force for selecting G-rich sequences lies in the formation of DNA/RNA hybrid G-quadruplex (HQ) structures between the pre-edited transcripts and the non-template strands of mitochondrial DNA. HQ's are transcription termination/replication initiation sites and thus guarantee an unperturbed replication of the mt-genome. This is of special importance in the insect-stage of the parasite. In the transcription-on state, the identified GQ's require editing as a GQ-resolving activity indicating a link between replication, transcription and RNA editing. We propose that the different processes have coevolved and suggest the parasite life-cycle and the single mitochondrion as evolutionary driving forces.

Figure 1. T. brucei mitochondrial genome organization and nucleotide propensities.

(a) Sketch of an insect-stage T. brucei cell highlighting the single, extended tubular mitochondrion (M) in maroon. N: nucleus; K: kinetoplast mitochondrial DNA. (b) Nucleotide content of the different classes of mitochondrial transcripts. Red: pan-edited RNAs in their unedited (un) and edited (ed) state. Green: marginally edited RNAs in their unedited (un) and edited (ed) state. Blue: never-edited transcripts. Grey: ribosomal RNAs. (c) Linear map of both strands of the coding region of the T. brucei mitochondrial DNA maxicircle. Genes are annotated as boxes and are colored as in (b). Ribosomal RNA genes: 9S, 12S; pan-edited genes: ND8, ND7, CO3, A6, CR3, RPS12, ND3, CR4, ND9; marginally edited genes: CYb, CO2, MURF II; never-edited genes: ND4, ND5, CO1, ND1, ND2, MURF V. The height of the individual boxes indicates the fractional G-content of the different ORF’s. Dashed line: average G-content of 22%. All pan-edited genes are above the mean. Lines in light blue above the individual genes indicate the nt-length of the fully edited transcripts (for details see Supplementary Table 1). (d) “U-centric” view of the unedited (top) and edited (bottom) transcript of CR4. Dark blue: U-nucleotides in the pre-edited mRNA. Light blue: Inserted U’s as a result of RNA editing. (e) “G-centric” view of the unedited (top) and edited (bottom) CR4-transcript emphasizing the conversion from a high number of G-runs (2 ≤ G ≤ 6) in the unedited state to a low number of G-tracts in the edited state. (f) Homopolymer cluster (n-mer)-analysis of all pan-edited, marginally edited and never-edited RNAs in T. brucei (Light blue: A-nt, dark blue: U-nt, red: G-nt, orange: C-nt). The editing-mediated reduction of homopolymer runs is unique to G-nt.

Figure 2. Experimental identification of GQ-folds.

(a) Structure of a G-tetrade consisting of four Hoogsteen-bonded (black dashed lines) G-residues (red) coordinated by a stabilizing monovalent cation (M⁺). (b) Parallel, intramolecular G-quadruplex formed by two stacked G-tetrades. Black line: phosphate/sugar backbone. Loop nucleotides are not shown for clarity. (c) Differential normalized reverse transcriptase (norm. ΔRT)-stop profiles (red) and predicted minimal free energy (MFE)-2D-structures of all pan-edited T. brucei transcripts in their pre-edited folding state. The individual structures are shown above and below the linear map of the coding region of the T. brucei maxicircle DNA as introduced in Fig. 1. Grey: ribosomal RNA genes; red: pan-edited genes; green: marginally edited genes; blue: never-edited genes. GQ-folds are drawn as “leaf-like” structures in red. AU: arbitrary unit. (d) Summary of the characteristics of all GQ-elements in the pan-edited transcripts.

Figure 3. RNA editing as a GQ-resolving process.

Gibbs free energy (ΔG/nt) changes (a) and cumulative G-score (∑G) changes (b) of the CR4-transcript over the course of seventeen gRNA-driven reaction steps. The colour transition from red to blue annotates the stepwise decrease in the number of G-tracts or inversely the stepwise increase in the U-content of the CR4-RNA. (c) MFE-2D-structures of the CR4-transcript from its unedited to its fully edited folding state (clockwise starting in the upper left corner). Unedited CR4-RNA contains 3 GQ-elements (“leaf-like” structures in red), which are progressively resolved to generate a fully edited, GQ-free mRNA. Guide RNA interaction sites are encircled in pink.

Figure 4. Switching between mitochondrial transcription and replication.

Artistic rendering of the transcription-on/replication-off and transcription-off/replication-on phases of the T. brucei maxicircle DNA. GQ-elements in the synthesized transcripts define the transcription-on/replication-off state and require RNA editing to structurally resolve the different GQ-elements to generate translatable mRNAs (upper panel). The formation of HQ-folds between the nascent transcripts and the non-template strand(s) of the DNA maxicircle are characteristic for the transcription-off/replication-on situation (lower panel). GQ- and HQ-folds are shown as connected “leaf-like” structures. RNA pol: RNA polymerase.