Mechanisms of replication and repair in mitochondrial DNA deletion formation

Abstract

Deletions in mitochondrial DNA (mtDNA) are associated with diverse human pathologies including cancer, aging and mitochondrial disorders. Large-scale deletions span kilobases in length and the loss of these associated genes contributes to crippled oxidative phosphorylation and overall decline in mitochondrial fitness. There is not a united view for how mtDNA deletions are generated and the molecular mechanisms underlying this process are poorly understood. This review discusses the role of replication and repair in mtDNA deletion formation as well as nucleic acid motifs such as repeats, secondary structures, and DNA damage associated with deletion formation in the mitochondrial genome. We propose that while erroneous replication and repair can separately contribute to deletion formation, crosstalk between these pathways is also involved in generating deletions.

Figure 1.

Schematic representation of human mtDNA, a double-stranded circular molecule of 16.5 kb. The H-strand and L-strand significantly differ in their base composition, with the H-strand containing a higher proportion of guanines. The mtDNA encodes for 13 proteins (colored boxes), the micropeptide Humanin (pink box), 2 rRNAs (yellow boxes) and 22 tRNAs (black boxes). With the exception of Humanin, all proteins encoded by the mtDNA genes are key components of complexes I, II and IV of the electron transport chain. Non-coding regions are depicted as white or grey boxes. A non-coding control region (expanded) contains crucial regulatory regions: three promoters (two H-strand promoters, HSPs and a L-strand promoter, LSP, shown as blue arrows), three conserved boxes (CSB 1–3, red boxes) and a termination-associated sequence (TAS, purple box). This non-coding region also contains the origin of replication of the H-strand (OriH, black arrow) and a triple-stranded displacement-loop (D-loop), a structure formed during premature termination of replication. The hyper-variable regions (HVS1–2, gray boxes) are highly polymorphic sequences and hotspots of germline and somatic mtDNA mutations. A second minor non-coding region is localized at approximately two-thirds of the genome and contains the origin of replication of the L-strand (OriL, black arrow). The mtDNA regions comprised between OriH and OriL shown as the major arc (red dotted line) and minor arc (blue dotted line) are depicted.

Figure 2.

(A) Schematic of the strand displacement model (SDM) for mtDNA replication, which starts from two origins of replication, OriH and OriL, dedicated respectively to the replication of the H- and L-strand. In the SDM, mtDNA replication initiates from OriH. The mitochondrial RNA polymerase POLRMT synthesizes a short RNA sequences that primes the subsequent replication catalyzed by DNA Polγ. The Twinkle helicase progresses in front of DNA Polγ, unwinding DNA in an ATP-dependent manner. The exposed single-stranded H-strand is bound by mtSSB to prevent spurious replication events. Once the replisome reaches OriL, a single-stranded stem-loop structure is formed, blocking mtSSB binding and promoting the initiation of replication of the L strand. The stem-loop is recognized by POLRMT, which synthesizes a short RNA primer in the OriL region. DNA Polγ replaces POLRMT, and replication of the two strands proceeds unidirectionally and continuously to form two full double-stranded mtDNA daughter molecules. (B) Replication fork stalling is suggested to promote the mis-annealing of single-stranded mtDNA regions containing direct repeats. The loop generated during replication-slippage is extruded from the mtDNA molecule, resulting in deletions. (C) Deletion formation from copy-choice recombination involves replication starting from OriL. During replication, secondary structures and loops could be formed by the exposed single-stranded regions containing direct repeats. If those structures persist, the sequences comprised in this aberrant mtDNA conformation is deleted during the second round of replication.

Figure 3.

(A) Schematic representation of DNA repair pathways. While BER activity in mitochondria is highly characterized, MMR and DSB are still lacking substantive evidence. Repair proteins found in mitochondria are depicted in black, with proteins reported to play a role in deletion formation indicated in red. BER (left) repairs oxidized bases and abasic sites and comprises both short-patch (SP) and long-patch (LP) sub-pathways. MMR (middle) still lacks convincing evidence regarding its activity in mitochondria. YB-1, so far, is the only mitochondrial MMR protein identified. DSB (right) also lacks convincing evidence regarding efficient mitochondrial function. Proteins involved in mitochondrial non-homologous end-joining (NHEJ), microhomology-mediated repair (MMEJ) and homology-directed repair (HDR) were identified and appear as shared between the nucleus and the mitochondria; however, the mechanistic aspects of mitochondrial DSB repair, as well as the similarities or divergences with the nuclear pathways, are still largely unknown. In addition, mtDNA molecules harboring damaged bases and/or DSBs could also be degraded. (B) Hypothesized models concerning error-prone repair of mtDNA DSBs that lead to deletion formation, either by inter- or intra-molecular rearrangements. Potential mechanisms for the formation of small- and/or large-scale deletions could result from faulty re-joining of DNA ends during NHEJ (top), MMEJ (middle) occurring in presence of mis-annealed direct repeats, and unproductive HDR (bottom).

Figure 4.

(A) Linear representation of human mtDNA. Genes are presented as colored boxes and non-coding regions as white boxes. tRNAs, regulatory elements and the distinction between genes localized on the H- or L-strand were omitted for clarity. (B) G4-quadruplex forming potential (QFP) sequences where computationally identified and mapped on the H- (light blue) and L- (dark blue) strands. Direct (orange) and indirect (green) repeats are indicated on the mtDNA as well as (C) 248 unique mtDNA deletions (accessed from Mitomap August 2020). Deletions are ordered by increasing size from top to bottom, ranging from 4 to 12 807 bp. Deletions flanked by direct repeats are indicated as orange lines, while deletion flanked by indirect repeats are indicated as green lines and deletions not yet reported to be flanked by direct or indirect repeats are instead depicted as gray lines. The top pie chart indicates the relative percentages of deletions flanked by direct, indirect, and flanking regions not yet reported to display repeats. The middle pie chart reports mtDNA deletion abundance based on size, while the bottom pie chart shows the percentages of deletions affecting the major and minor arcs or both.

Figure 5.

The main molecular determinants that participate in generating mtDNA deletions. mtDNA motifs, such as direct repeats and G4 quadruplexes, as well as mtDNA replication and repair pathways have been directly implicated in the formation of small- and large-scale deletions in the mitochondrial genome. In addition, crosstalk between replication and repair pathways (indicated as dotted red lines) have been described, particularly in formation of the CD where a replication-coupled mtDNA repair mechanism has been shown to cause accumulation of deleted mtDNA molecules. Mis-annealing of direct repeats has been linked to the perturbations in mtDNA repair and replication that ultimately lead to deletion formation. Similarly, replication fork stalling caused by G4 quadruplexes structures within the mtDNA could cause impairments in replication, leading to formation of loops and ultimately to deletions.