Initiation of mtDNA transcription is followed by pausing, and diverges across human cell types and during evolution

Abstract

Mitochondrial DNA (mtDNA) genes are long known to be cotranscribed in polycistrones, yet it remains impossible to study nascent mtDNA transcripts quantitatively in vivo using existing tools. To this end, we used deep sequencing (GRO-seq and PRO-seq) and analyzed nascent mtDNA-encoded RNA transcripts in diverse human cell lines and metazoan organisms. Surprisingly, accurate detection of human mtDNA transcription initiation sites (TISs) in the heavy and light strands revealed a novel conserved transcription pausing site near the light-strand TIS. This pausing site correlated with the presence of a bacterial pausing sequence motif, with reduced SNP density, and with a DNase footprinting signal in all tested cells. Its location within conserved sequence block 3 (CSBIII), just upstream of the known transcription-replication transition point, suggests involvement in such transition. Analysis of nonhuman organisms enabled de novo mtDNA sequence assembly, as well as detection of previously unknown mtDNA TIS, pausing, and transcription termination sites with unprecedented accuracy. Whereas mammals (Pan troglodytes, Macaca mulatta, Rattus norvegicus, and Mus musculus) showed a human-like mtDNA transcription pattern, the invertebrate pattern (Drosophila melanogaster and Caenorhabditis elegans) profoundly diverged. Our approach paves the path toward in vivo, quantitative, reference sequence-free analysis of mtDNA transcription in all eukaryotes.

Figure 1.

Workflow of analysis. (A) GRO-seq and PRO-seq experiments generate genome-wide nascent transcript data. The extracted mtDNA sequences enable reconstruction of sample-specific mtDNA sequence, which is used in turn as a circular-like mapping reference. This allows counting the sequencing read coverage in a strand-specific manner. (B) Analysis of mtDNA transcription initiation and termination sites. Two steps were designed: (1) a crude step for candidate transcription initiation site (TIS) identification, identifying abrupt increase in read coverage in a nucleotide resolution within 200 bp sliding windows; and (2) fine analysis, focusing on highest scoring regions to identify the best TIS candidate. Notably, the identification of transcription termination sites utilizes the same approach, yet instead of an abrupt increase in reads, an abrupt decrease in read coverage is identified.

Figure 2.

Accurate identification of human mtDNA TIS. (A) Sequencing read coverage around the mtDNA TIS in two cell lines. (Top) Sequencing read coverage pattern of the mtDNA heavy strand (red); (bottom) sequencing read coverage of the light strand (blue). Putative TIS is designated by black triangle. (y-axis) Sequencing read coverage; (x-axis) mtDNA position. (Left) PRO-seq experiment of the K562 cell line. (Right) GRO-seq experiment from the U2OS cell line. (B) Ratio of sequencing coverage between the light and heavy strands. (y-axis) Ratio of read density between the light and heavy strands. Black dots correspond to the calculated ratios for each tested cell line (indicated near the dots). (C) Summary of mtDNA transcription pattern: PRO-seq and GRO-seq experiments in 11 human cell types: (TIS) Light-strand TIS was identified in all tested human cell types (N = 11) in positions 407–410. In most of the tested cells (seven of 11), the major heavy-strand TIS was mapped in positions 560–562 (TIS 1). In the remaining cell types (four), the major heavy-strand TIS was located in positions 634–689 (TIS 2). (Termination sites) Light-strand transcription termination was identified within the 16S rRNA gene, in the region encompassing positions 2612–3252 (dark blue arrow). Heavy-strand transcription termination was identified within the D-Loop, between positions 16,076 and 195 (dark red arrow).

Figure 3.

mtDNA transcription consistently pauses at distinct sites near the heavy- and light-strand TIS. (A) mtDNA transcriptional regulation elements. Presented is the complementary human sequence of the light mtDNA strand. The mtDNA sequence around the pausing peak is above the illustrated graph. (Square bracket) The bacterial pausing motif. The mandatory nucleotides within the motif are highlighted by a larger font size. (B) Coverage of the 3′ end of the PRO-seq experiment from K562 cell line. (x-axis) mtDNA nucleotide position; (y-axis) number of reads in the 3′. (Blue and red arrows) The direction of the light- and heavy-strand transcription, respectively. The “horizontal T” sign represents the pausing site. (C) DNase-seq experiment from K562 cell line. (x-axis) mtDNA nucleotide position; (y-axis) F-score of DNase-seq analysis. The lower the score, the more protected is the DNA by proteins. The black arrow points to the DGF site. (D) Pausing index across human cell types. (Left) Light strand; (right) heavy strand. (y-axis) Pausing index values. Dots correspond to the calculated pausing index for each tested cell line (indicated as numbers in brackets near the dots: [1] AC16; [2] CD4⁺; [3] GM12004; [4] GM 12750; [5] GM12878; [6] HeLa; [7] IMR90; [8] Jurkat; [9] K562; [10] MCF7; [11] U2OS). (E) Pausing site nucleotide position across human cell types. (Left) Light strand; (right) heavy strand; (x-axis) mtDNA position. Dots correspond to the pausing site nucleotide position of each tested cell line (numbering as in D). (F) Human population SNPs density. (x-axis) mtDNA position; (y-axis) SNPs density measured as variants per position (log scale).

Figure 4.

Identification of mtDNA nascent transcript across evolution: (A) PRO-seq experiment performed in four mammalian CD4⁺ cells: chimpanzee, rhesus macaque, rat, and mouse. x- and y-axes are identical to those in Figure 2, and “horizontal T” sign designates the pausing site. Filled arrowheads in all panels point to the calculated identified TIS. Notably, in three species (chimpanzee, rat, and mouse), the major heavy-strand TIS was identified downstream from the tRNA phenylalanine gene, similar to the human heavy-strand TIS 1. (B) Pausing site of light-strand transcription in mammals. (x-axis) Distance (in nucleotides) of the pausing site from the light-strand TIS; (y-axis) pausing index value of each species. The name of each species is indicated to the right of each dot. (C, left) Analysis of GRO-seq data from worm. In this species, there is a single TIS for a single transcription unit, present only at the heavy strand. (C, right) Analysis of PRO-seq data from Drosophila. Five candidate TISs were identified: two in the heavy strand and three in the light strand. Two minor additional TISs are marked by empty arrowheads.

Figure 5.

A model offering a mechanistic explanation for the role of light-strand transcription pausing. Presented are the stages right after transcription initiation of the light strand, as well as the suggested role for our discovered transcription pausing site. (Alternative products) Replication-based and transcription-based products (i.e., DNA and RNA products, respectively) of the light-strand promoter, as both require the same light-strand RNA primer (∼120 nt in length).