Native R-loops persist throughout the mouse mitochondrial DNA genome

Abstract

Mammalian mtDNA has been found here to harbor RNA-DNA hybrids at a variety of locations throughout the genome. The R-loop, previously characterized in vitro at the leading strand replication origin (OH), is isolated as a native RNA-DNA hybrid copurifying with mtDNA. Surprisingly, other mitochondrial transcripts also form stable partial R-loops. These are abundant and affect mtDNA conformation. Current models regarding the mechanism of mammalian mtDNA replication have been expanded by recent data and discordant hypotheses. The presence of stable, nonreplicative, and partially hybridized RNA on the mtDNA template is significant for the reevaluation of replication models based on two-dimensional agarose gel analyses. In addition, the close association of a subpopulation of mtRNA with the DNA template has further implications regarding the structure, maintenance, and expression of the mitochondrial genome. These results demonstrate that variously processed and targeted mtRNAs within mammalian mitochondria likely have multiple functions in addition to their conventional roles.

FIGURE 1.

Map of mouse mtDNA. The circular genome is ∼16.3 kb. The two strand-specific origins (O_H and O_L) are designated by black arrows around the outside of the map. Displacement replication exposes a single H-strand between O_H and O_L containing the Cyt b, ND6, and COII genes. The control region encompassing O_H is shown in greater detail in Fig. 2. Genes with transcripts analyzed in this work include 16S rRNA, ND1, COII, ND6, and Cyt b and appear as bold red arrows. Other mitochondrial genes were excluded for clarity.

FIGURE 2.

Northern analysis and DNase and RNase sensitivity of mtDNA-bound nascent H-strand RNA and DNA at O_H. EtBr-CsCl-purified closed circular mtDNA was analyzed by Northern analysis to detect stable R-loops. RNA size markers are in lane 1. Poly(A)+-purified RNA is in the lane denoted by pA+. DNase I and RNase H sample treatments are indicated above the panels. A, probing for CSB-proximal RNA with T3#4 riboprobe (shown in D). B, less exposed film of view shown in A, revealing the increased intensity of ∼150-nt species in lane 3 after DNase treatment. C, probing for CSB-distal RNA with T3#1 riboprobe as shown in D. D, reference diagram showing the major noncoding region of mtDNA. This region is identified in Fig. 1 as the area encompassing O_H. The transcription start site is shown with a bent arrow followed by several relevant DNA sequence features, including CSBs III, II, and I. The termination-associated sequences (TAS) region is shown at the promoter distal end of the DNA. T3#4 and T#31 riboprobe positions are shown above. Below the DNA map are nucleic acids identified in the Northern blots shown in A-C. RNA is shown by thick black lines, and RNA primers in transition with DNA are shown in gray. DNA alone is in shown by thin black lines. The lines are to scale, with size interruptions shown by breaks.

FIGURE 3.

Northern analysis and RNase H and DNase I sensitivity of R-loops within mtDNA coding regions. Mitochondrial RNA remains bound to CsCl-purified mtDNA. Samples were prepared as described in the legend for Fig. 2 and treated with nucleases as indicated by (+) and (-). pA is poly(A)+-purified RNA. Gene-specific riboprobes were generated from PCR products as described under “Experimental Procedures.”

FIGURE 4.

R-loop RNA ends are unlike total RNA ends. The incidences of short 5′-ends, 3′-ends, and unmodified 3′-ends (3′-OH) are displayed. Incidences within total RNA are shown as gray bars, and those within the R-loop population are shown as black bars. Asterisks indicate a significant difference at p < 0.05 using ANOVA. Data are taken from Table 1.

FIGURE 5.

Bound RNA affects mtDNA conformation. A and B, AFM imaging of mtDNA before (A) and after (B) treatment with RNase A. Highly twisted plectoneme structures are seen in several regions of B only after removal of the bound RNA. DNA crossovers per molecule are shown in distributed form and as a relative measure of apparent writhe in C and D. The untreated crossover distribution is shown in C with gray bars, and the RNase-treated distribution is shown in D with black bars. Each bar represents data for an individual molecule. The median crossover numbers are 14 and 33 for C and D, respectively. ANOVA single factor analysis indicates a significant difference at p = 3.1 × 10^-5.