Discover millions of ebooks, audiobooks, and so much more with a free trial

Only $11.99/month after trial. Cancel anytime.

The Human Mitochondrial Genome: From Basic Biology to Disease
The Human Mitochondrial Genome: From Basic Biology to Disease
The Human Mitochondrial Genome: From Basic Biology to Disease
Ebook1,271 pages38 hours

The Human Mitochondrial Genome: From Basic Biology to Disease

Rating: 0 out of 5 stars

()

Read preview

About this ebook

The Human Mitochondrial Genome: From Basic Biology to Disease offers a comprehensive, up-to-date examination of human mitochondrial genomics, connecting basic research to translational medicine across a range of disease types. Here, international experts discuss the essential biology of human mitochondrial DNA (mtDNA), including its maintenance, repair, segregation, and heredity. Furthermore, mtDNA evolution and exploitation, mutations, methods, and models for functional studies of mtDNA are dealt with. Disease discussion is accompanied by approaches for treatment strategies, with disease areas discussed including cancer, neurodegenerative, age-related, mtDNA depletion, deletion, and point mutation diseases. Nucleosides supplementation, mitoTALENs, and mitoZNF nucleases are among the therapeutic approaches examined in-depth.

With increasing funding for mtDNA studies, many clinicians and clinician scientists are turning their attention to mtDNA disease association. This book provides the tools and background knowledge required to perform new, impactful research in this exciting space, from distinguishing a haplogroup-defining variant or disease-related mutation to exploring emerging therapeutic pathways.

  • Fully examines recent advances and technological innovations in the field, enabling new mtDNA studies, variant and mutation identification, pathogenic assessment, and therapies
  • Disease discussion accompanied by diagnostic and therapeutic strategies currently implemented clinically
  • Outlines and discusses essential research protocols and perspectives for young scientists to pick up
  • Features an international team of authoritative contributors from basic biologists to clinician-scientists
LanguageEnglish
Release dateJul 23, 2020
ISBN9780128226421
The Human Mitochondrial Genome: From Basic Biology to Disease

Related to The Human Mitochondrial Genome

Related ebooks

Biology For You

View More

Related articles

Reviews for The Human Mitochondrial Genome

Rating: 0 out of 5 stars
0 ratings

0 ratings0 reviews

What did you think?

Tap to rate

Review must be at least 10 words

    Book preview

    The Human Mitochondrial Genome - Giuseppe Gasparre

    Part 1

    Biology of human mtDNA

    Outline

    Chapter 1 mtDNA replication, maintenance, and nucleoid organization

    Chapter 2 Human mitochondrial transcription and translation

    Chapter 3 Epigenetic features of mitochondrial DNA

    Chapter 4 Heredity and segregation of mtDNA

    Chapter 1

    mtDNA replication, maintenance, and nucleoid organization

    Mara Doimo*, Annika Pfeiffer*, Paulina H. Wanrooij and Sjoerd Wanrooij,    Department of Medical Biochemistry and Biophysics, Umeå University, Umeå, Sweden

    Abstract

    Part of the genetic information in human cells resides in the mitochondria. Faithful maintenance of mitochondrial deoxyribonucleic acid (mtDNA) is crucial for the oxidative phosphorylation system that produces the majority of the cellular ATP, and therefore to life. This chapter provides an introduction into the characteristics of human mtDNA and summarizes the processes and factors required for the replication and maintenance of this small but essential genome. We also describe the organization of mtDNA in specialized nucleoprotein structures called nucleoids. Where applicable, we refer to human disease states that are caused by defects in the described factors or processes.

    Keywords

    Mitochondrial DNA; mtDNA; mtDNA maintenance; POL ɣ; Twinkle; mtSSB; POLRMT; TFAM; nucleoids; dNTPs

    1.1 Human mitochondrial DNA

    1.1.1 Characteristics of mitochondrial DNA

    The evidence for mitochondrial deoxyribonucleic acid (mtDNA) dates back to the early 1960s when fibers with characteristics of DNA were first observed in the mitochondrial matrix of chick embryo mitochondria [1,2]. The human mitochondrial genome is a closed circular double-stranded DNA molecule with a length of about 5 μm [3,4]. MtDNA molecules mostly occur as monomers, but can also exist as catenated circles, which are most often dimers [4–6]. In accordance with its small size, mtDNA amounts to less than one percent of the total cellular DNA in animal cells [7]. Early estimations suggested that a single animal mitochondrion contains between two and more than ten mtDNA genomes [8], and—because there are multiple mitochondria per cell—human cells in general harbor from several hundred to over ten thousand mtDNA molecules per cell, depending on the cell type [9–12]. However, the mtDNA copy number of all cell types does not fit in this range, as, for example, human oocytes contain hundreds of thousands of mtDNA copies per cell [13–15].

    Human mtDNA is maternally inherited [16]. A key characteristic of mtDNA that derives from its multicopy nature is that cells can contain more than one population of mtDNA genomes, a state referred to as heteroplasmy [17]. The level of heteroplasmy, that is, the frequency of a population of mutated mtDNA molecules, can determine the severity of symptoms in mitochondrial diseases that are caused by mutations in mtDNA [18]. Another peculiar feature that separates the mitochondrial genome from its nuclear counterpart is that mtDNA contains occasional embedded ribonucleotides, the building blocks of ribonucleic acid (RNA) [19,20].

    1.1.2 Organization of the human mitochondrial genome

    The human mitochondrial genome was the first fully sequenced mitochondrial genome [21]. It is 16.5 kilo base pairs (kbp) in length and encodes for 2 ribosomal RNAs (rRNAs), 22 transfer RNAs (tRNAs), and 13 protein-coding genes with very few or no noncoding bases between neighboring genes (Fig. 1.1A) [21]. All 13 protein-coding genes encode indispensable subunits of the ATP-producing oxidative phosphorylation (OXPHOS) system [21–26]. The two strands of mtDNA are designated as heavy (H) and light (L) due to differences in base composition (GT/CA ratio) that leads to their different buoyant densities on alkaline cesium chloride gradients [27]. Twelve of the protein-coding genes, the 12S and 16S rRNAs as well as fourteen tRNA genes are encoded on the H-strand, while the L-strand encodes one protein-coding gene and eight tRNA genes [21]. The H-strand is G-rich and thereby contains several sequence motifs with the potential to form specific secondary DNA structures called G-quadruplexes (G4s) [28,29]. These motifs often colocalize with mtDNA deletion break-points [29], suggesting that they may hamper the mtDNA replication process.

    Figure 1.1 The human mitochondrial genome.

    (A) The mitochondrial genome has a size of 16,568 base pairs in humans. The two strands are termed heavy (H-strand) and light strand (L-strand). The location of the displacement loop (D-loop) is indicated in the noncoding region between tRNAPro and tRNAPhe, and the position of the genes for complex I (NADH dehydrogenase; ND), complex III (cytochrome b; CYTB), complex IV (cytochrome c oxidase; COX), as well as complex V (ATP synthase; ATPase) subunits of the respiratory system are shown. Highlighted with gray boxes are the sites of replication priming in the OH region (B) and at OL (C). (B) Characteristic features of the noncoding region are the promoters HSP and LSP, the highly conserved sequence blocks (CSB 1–3), the single-termination sequence (TAS), and the DNA stretch (7S DNA) that forms the D-loop. Priming in the OH region is initiated with a RNA primer and RNA-to-DNA transition sites were mapped to the CSB region. (C) Priming at OL involves the formation of a DNA stem-loop structure and the synthesis of a short RNA primer by POLRMT. The RNA-to-DNA transition sites were mapped downstream of the OL region.

    A circa 1100 bp long noncoding region (NCR), also referred to as the control region, is located between the tRNAPro and tRNAPhe genes [21]. The NCR contains the origin of replication for H-strand synthesis (OH) [21,30] as well as the H-strand promoter (HSP) and the L-strand promoter (LSP) for transcription of each mtDNA strand [31]. Although OH has traditionally been annotated at nucleotide position 191 [21], H-strand replication actually initiates at LSP, approximately 200 nucleotides (nt) upstream of OH [32]. We will therefore use the term OH region to refer to the region containing both LSP and OH [33]. The NCR also contains three highly conserved sequence blocks (CSB 1–3), which are considered to have a function in regulating replication [34], and a single termination–associated sequence (TAS) downstream of the three CSBs [35]. Early characterization of mtDNA in mouse [36–38] and later in human cells [39] revealed molecules with a short triple-stranded DNA region, the so-called displacement loop (D-loop), encompassing the OH region. The D-loop is the result of synthesis of a ~600 nt long DNA stretch, termed 7S DNA, that is complementary to the L-strand and is formed by replication initiated in the OH region followed by early termination at the TAS [35,40,41]. The second origin of replication on mtDNA, OL, directs L-strand synthesis. OL is located outside of the NCR, in a cluster of five tRNA genes two-thirds of the way around the genome from OH [42].

    1.2 The process of mtDNA replication

    1.2.1 Replication mechanisms

    In comparison to the nuclear DNA duplication process, human mitochondrial replication is distinct in terms of both the mechanism(s) and the proteins involved. The proteins of the mitochondrial replication fork duplicate the organelle’s DNA by an asymmetric mechanism in which both strands are synthesized continuously [43]. This so-called strand-displacement model of mtDNA replication is based on early pulse and pulse-chase labeling experiments, electron microscopy and 5′-end characterization [7,44], as well as more recent biochemical reconstitution of mtDNA replication in vitro using recombinant proteins [45,46]. This mode of DNA replication is strikingly different from that of mammalian nuclear DNA, but it is not unique in biology since it is similar to replication of ColE1 plasmid DNA [47].

    MtDNA replication starts with H-strand (leading strand) synthesis that initiates from the OH region. As the mtDNA replication machinery synthesizes a nascent H-strand, the parental H-strand is displaced, giving rise to the triple-stranded D-loop structure (Fig. 1.2). For yet unknown reasons, mtDNA replication often terminates around the TAS region and only a minority of the nascent H-strands are extended past this point. In the cases where H-strand replication continues beyond the TAS region, it continues unidirectionally, displacing the parental H-strand until it reaches OL. Unwinding of the DNA duplex at OL allows the formation of a characteristic hairpin structure that leads to activation of this origin [48]. L-strand (lagging strand) replication thus initiates and continues unidirectionally from OL in the opposite orientation to H-strand synthesis, and the synthesis of both strands proceeds until they have reached full-circle.

    Figure 1.2 The model of mitochondrial DNA (mtDNA) replication.

    Mitochondrial DNA is replicated by a unique enzymatic machinery that includes the heterotrimeric DNA polymerase POL ɣ (pink), the mitochondrial single-stranded DNA-binding protein mtSSB (green), and the DNA helicase Twinkle (blue). (1) After initiation at the leading strand origin (OH), the replisome consisting of POL ɣ and Twinkle proceeds to unidirectionally replicate one new strand. The displaced, second strand is bound and stabilized by mtSSB (green). (2) When the replication machinery passes the second strand (lagging-strand) origin (OL), a stem-loop structure is formed. A short primer is synthesized at the stem-loop, and is used to initiate the second strand (lagging strand) DNA synthesis. (3) Both the leading and lagging strand replication machineries proceed full-circle around the mtDNA molecule. (4) After completion of mtDNA strand synthesis, replication is terminated at either OH or OL, depending on where DNA synthesis was initiated.

    An alternative interpretation of the strand-displacement mode of replication has been made based on two-dimensional agarose gel electrophoresis experiments. However, the two models are very similar with the principal difference that in this so-called bootlace/RITOLS model [49,50], the displaced H-strand is coated with preformed RNA in contrast to the mitochondrial single-stranded DNA-binding protein (mtSSB) suggested in the strand-displacement model [51,52]. Both models agree on that mtDNA replication results in extensive ssDNA replication intermediates, which are indeed the dominant replication intermediates observed in dividing cells [53]. In nondividing cells, however, the majority of mtDNA replication intermediates are double-stranded in nature [53]. This could potentially be the consequence of enhanced promiscuous (OL-independent) priming of L-strand replication [54], or, as proposed by others [55], the result of an alternative bidirectional mode of replication carried out by an as-yet unidentified set of proteins.

    1.2.2 Priming

    The first step of DNA replication is the synthesis of an RNA primer that can subsequently be elongated by the mtDNA polymerase γ (POL γ). A mitochondrial primase activity was isolated from human mitochondria as early as 1985 [56,57], but the identity of the protein was not elucidated at the time. Subsequent work has revealed that the primers at both origins of replication are in fact synthesized by the mitochondrial RNA polymerase (POLRMT) that is related to the RNA polymerases of the T-odd lineage of bacteriophages [58].

    The primer for leading strand replication has a 5′ end at the LSP, the same site where L-strand transcripts initiate [32]. Hence, transcription by POLRMT from the LSP creates not only the near-genome-length transcripts that are processed to liberate tRNAs and mRNAs [59], but also the much shorter primer for leading strand synthesis (Fig. 1.1B). A G4 structure that forms at CSB2 causes premature termination of transcription about 120 nt downstream of LSP [60,61]. This prematurely terminated transcript remains stably hybridized to the DNA template as an RNA–DNA hybrid referred to as the mitochondrial R-loop [62,63] and is assumed to be used as the H-strand replication primer. In support of this model of H-strand priming, RNA-to-DNA transition sites have been mapped to the CSB region [32,60,64,65]. In an in vitro transcription assay based on purified recombinant proteins, the level of premature transcription termination—and thus the switch from transcription to replication—is modulated by the mitochondrial transcription elongation factor (TEFM) that facilitates the bypass of obstacles such as DNA secondary structures and oxidative lesions by POLRMT [66,67]. In accordance with TEFM’s role in transcription elongation, it is essential in mice and heart-specific knockout causes a dramatic decrease in promoter-distal transcripts [68]. Because the majority of LSP-initiated transcripts fail to be elongated far enough to reach CSB2 in the absence of TEFM, de novo mtDNA replication is dramatically reduced in TEFM knockout hearts [68].

    As described above, POLRMT synthesizes the H-strand primer using double-stranded mtDNA as a template. In contrast, L-strand replication is initiated when H-strand synthesis has reached two-thirds of the way around the genome and exposes OL in single-stranded form (Fig. 1.2). Once single-stranded, OL adopts a hairpin structure that directs POLRMT to initiate primer synthesis within a stretch of six Ts within the loop (nucleotide position 5747–5751) (Fig. 1.1C [48]). On this ssDNA template, POLRMT is nonprocessive and generates only short RNAs that can be elongated by POL γ [46,48]. In accordance with this model, in vivo RNA-to-DNA transition sites have been mapped just downstream of the OL hairpin [48].

    In contrast to the L-strand replication primer, the RNA primer generated in the OH region cannot be directly elongated by POL γ. Instead, in vitro studies suggest that the H-strand primer may require processing by RNase H1 before elongation by POL γ is possible [65]. As discussed below, RNase H1 also plays a defined role in removal of the RNA primers at both mitochondrial origins [69]. Accordingly, mutations that decrease the activity of RNase H1 cause mtDNA depletion and deletions, manifesting as chronic progressive external ophthalmoplegia (CPEO) [70].

    1.2.3 Elongation of mtDNA replication

    After priming of mtDNA synthesis, elongation of the replication primer is catalyzed by the nuclear-encoded POL γ, which was the first polymerase isolated from mitochondria of human HeLa cells [71]. POL γ needs the help of the mtDNA helicase Twinkle and mtSSB to replicate mtDNA in human cells. Together, POL γ, Twinkle and mtSSB constitute the minimal mtDNA replisome (Fig. 1.3) and are essential for reconstitution of mitochondrial replication in vitro [72].

    Figure 1.3 The core mitochondrial DNA (mtDNA) replication fork proteins.

    The TWINKLE helicase (blue) moves in a 5′ to 3′ direction while unwinding dsDNA. The mtSSB protein (dark green) stabilizes the single-stranded conformation of DNA and stimulates DNA synthesis by POLγ (red (A) and gray (B)). POLRMT (light green) synthesizes the RNA primer (yellow line) needed for lagging strand DNA synthesis.

    1.2.3.1 The mitochondrial DNA polymerase POL γ

    POL γ was first described as an RNA-dependent DNA polymerase [73]. The holoenzyme consists of one 140-kDa POL γA catalytic subunit and two 55-kDa POL γB accessory subunits [74–76]. The catalytic subunit possesses DNA polymerase and 3′–5′ exonuclease activities in addition to 5′-deoxyribose phosphate lyase activity [77,78]. In general, the intrinsic exonuclease activity of POL γ functions as a proofreading mechanism during replication, which ensures fidelity of the polymerase through high nucleotide selectivity and slow extension of mismatches [79]. POL γ can switch between polymerase and exonuclease activities without dissociating from the DNA template because it can transfer the nascent DNA from the polymerase site to the exonuclease site and back through a mechanism of intramolecular strand transfer [80]. With an error rate of less than 1 × 10−6 per nucleotide, POL γ is one of the most accurate polymerases [79].

    The accessory subunit POL γB is also known as the processivity factor because it increases the processivity of POL γA by increasing its DNA binding affinity. In addition, POL γB stimulates both the exonuclease and polymerase activities, and it can also improve nucleotide binding and incorporation, thus increasing the polymerization rate of the holoenzyme [81,82]. In solution, dimers of POL γB form a heterotrimer with the catalytic subunit POL γA through the tight binding of the POL γB dimer to the POL γA monomer [75]. Each POL γB unit in the holoenzyme has a specific function: the monomer proximal to POL γA in the holoenzyme increases the interaction with DNA, whereas the distal POL γB monomer is responsible for enhancing the reaction rate [83]. Modeling of POL γ binding to DNA suggests that POL γA binds approximately 10 bp of template DNA and that interaction with POL γB increases the footprint of POL γA on the DNA to 25 bp. Although POL γB has dsDNA-binding activity [84], it does not interact with the primer-template DNA [85]. While it is not essential for the stimulation of POL γA, the dsDNA-binding activity of POL γB is required for the function of the mtDNA replisome on a dsDNA template through coordination of POL γ and Twinkle at the replication fork [86].

    1.2.3.2 The mitochondrial DNA helicase Twinkle

    Twinkle (T7 gp4-like protein with intramitochondrial nucleoid localization) is the nuclear-encoded mitochondrial replicative helicase. It shares structural similarity with phage T7 primase/helicase and colocalizes with mtDNA in nucleoprotein complexes [87]. Twinkle is required for unwinding of the double-stranded DNA template ahead of POL γ that can only use ssDNA as a template [72]. It is an NTP-dependent DNA helicase that unwinds DNA in a 5′−3′ orientation and needs a fork-like structure with a single-stranded 5′-DNA loading site and a short 3′-tail to initiate unwinding [88]. The helicase activity of Twinkle is stimulated considerably by mtSSB [88], but it can only unwind longer dsDNA stretches in the presence of POL γ [72]. Earlier reports suggested that Twinkle is hexameric [87,89], while more recent analysis by electron microscopy revealed evidence of both hexameric and heptameric forms [90,91].

    1.2.3.3 The mitochondrial single-stranded DNA-binding protein

    Human mtSSB is a 15 kDa protein that shares sequence similarity with Escherichia coli SSB [92], and forms a tetramer that binds 59 nt of ssDNA [93]. Single-stranded DNA wraps once around the mtSSB tetramer [94,95]. MtSSB stimulates POL γ polymerase activity as well as Twinkle helicase activity [88,96,97]. The stimulatory effect on POL γ polymerase activity has been suggested to occur without any direct interaction between the proteins, through the ability of mtSSB to organize the ssDNA template and remove any secondary DNA structures [98]. Studies on Drosophila mtSSB showed that mtSSB increases Drosophila POL γ synthesis primarily by increasing primer recognition and binding and, as a result, increases the rate of initiation of DNA synthesis [99,100]. In addition to stimulating POL γ polymerase activity, Drosophila mtSSB can even stimulate the exonuclease activity of POL γ [101].

    1.2.3.4 The mitochondrial DNA replisome

    In biochemical assays, POL γ alone is unable to use dsDNA as a template, but the addition of Twinkle dramatically improves polymerization by POL γ on a primed mini-circle template and allows synthesis of extensive stretches of DNA in a rolling-circle replication mode [72]. The Twinkle-POL γ interaction also allows Twinkle to unwind longer stretches of dsDNA although the proteins do not seem to form a stable complex [72]. The addition of mtSSB further improves DNA synthesis allowing the formation of up to genome-length DNA products [72]. In conclusion, the mtDNA polymerase POL γ together with Twinkle and mtSSB constitute the core machinery necessary to replicate the mtDNA in human cells. The importance of these proteins for mtDNA maintenance in vivo is underscored by the fact that defects in these core components of the mtDNA replisome lead to mtDNA instability and mitochondrial disease [87,102–106].

    1.2.4 Termination of mtDNA replication

    Once the replisome has synthetized the mtDNA strands in their entirety, three sequential steps need to occur in order to give rise to the two daughter mtDNA molecules: (I) primer removal, (II) ligation, and (III) separation of the newly synthetized molecules.

    1.2.4.1 Primer removal

    As described previously, the process of mtDNA replication at both origins of replication starts with the synthesis of RNA primers by POLRMT. Long stretches of ribonucleotides are expected to impair mtDNA stability [107] and interfere with the replisome function [108]. Therefore these RNA primers need to be removed before the end of the replication process. The main enzyme implicated in removal of the primers at both OH and OL is RNase H1, as implicated by the finding that the loss of this enzyme causes the retention of the RNA primers at both origins [69]. In vitro, RNase H1 processes the RNA strand in RNA–DNA hybrid substrates and, although it does not exhibit any sequence specificity, it requires four consecutive ribonucleotides flanking the cleavage site [109]. Cleavage by RNase H1 leaves two ribonucleotides attached to the 5′-end of the nascent DNA strand [109]. Consequently, a second nuclease is required to remove these last ribonucleotides.

    The removal of the primer at OL has been reconstituted in vitro [110]. According to this model, RNase H1 processes the primer at OL and leaves 1–3 ribonucleotides at the 5′-end of the nascent DNA that are subsequently removed by the flap-structure specific endonuclease 1 (FEN1). FEN1 is 5′−3′specific endonuclease with dual mitochondrial and nuclear localization [111]. It preferentially cuts short RNA or DNA 5′-flaps flanked by dsDNA. Among its other functions, FEN1 is implicated in the processing of nucleic acid intermediates that form during lagging strand replication in the nucleus [112]. However, its function inside the mitochondria is still debated because the variant of the protein that is imported into the mitochondrial matrix is truncated and, in vitro, cannot cleave 5′-ssDNA flaps [113]. Therefore another nuclease with Fen1-like activity might be involved in the removal of the primer at OL in vivo.

    The process of primer removal at OH remains poorly understood. Initially, the mitochondrial genome maintenance exonuclease 1 (MGME1) was proposed to be involved [114]. This ssDNA-specific exonuclease, which localizes exclusively in the mitochondria, can process DNA flap substrates [115]. In vitro MGME1 works in combination with POL γ to remove DNA flaps that are originated by the strand-displacement activity of POL γ [116]. However, MGME1 cannot process short RNA flaps [116], indicating that an additional nuclease must be recruited to remove the ribonucleotides left by RNase H1. The involvement of MGME1 in mtDNA maintenance is nevertheless demonstrated by the fact that mutations in the MGME1 gene are associated with a multisystemic mitochondrial disorder characterized by mtDNA depletion and the accumulation of multiple mtDNA deletions [115].

    Recently, also the endonuclease/exonuclease G (EXOG) has been proposed to have a role in the removal of the ribonucleotides left behind by RNase H1 [117]. EXOG localizes to the mitochondria [118] and can process RNA–DNA hybrids in vitro [117]. Nonetheless, the in vivo involvement of EXOG in primer removal has not been clarified so far.

    1.2.4.2 Ligation

    After the removal of the primers, the nascent mtDNA strands are ligated by DNA ligase 3 (LIG3). Due to the presence of two in-frame ATGs in the coding sequence, the LIG3 gene encodes both a nuclear and a mitochondrial variant of the enzyme [119]. Depletion of the protein in human cells results in reduced mtDNA content and in the accumulation of nicked mtDNA [120], indicating that LIG3 is essential for the maintenance of mtDNA.

    1.2.4.3 Separation

    At the end of the replication process, the two daughter molecules are still interlinked and need to be separated. Topoisomerase TOP3α was recently found to be the main enzyme involved in this process [121]. TOP3α is a type 1A topoisomerase that localizes both in the nucleus and in the mitochondria [122]. Enzymes belonging to this class of topoisomerases separate two catenated molecules of DNA by cleaving one of the two strands to allow the passage of the other strand [123]. In the mtDNA, at the end of replication, the newly synthetized molecules remain interlinked in a region close to OH [121]. TOP3α is essential to separate these molecules and its depletion causes the accumulation of catenated mtDNA structures. The same abnormal mtDNA topology is detected also in the skeletal muscle of patients with CPEO caused by bi-allelic mutations in TOP3A gene [121].

    1.2.5 Other proteins involved in mtDNA replication

    Besides the above-described proteins, additional proteins whose precise role remains elusive have been implicated in mtDNA maintenance. TOP1MT is a mitochondria-specific topoisomerase [124] that is nonessential for mtDNA replication under basal conditions. However, TOP1MT binds to the NCR of mtDNA [125] and is required for proper mtDNA replication under specific stress conditions in mice [126]. Moreover, the chemical inhibition of yet another topoisomerase, TOP2β, was shown to reduce mtDNA replication initiation in human cells [127]. Although the mechanistic details of this defect were not addressed, it suggests that TOP2β has an important function in mtDNA replication.

    The human nuclease/helicase DNA2 can be found inside the mitochondria and specifically localizes to nucleoids after induction of mtDNA replication stalling [128]. The implication of DNA2 in mtDNA maintenance is supported by the observation that mutations in the DNA2 gene lead to mtDNA instability in patients with mitochondrial myopathies [129,130]. Despite valuable biochemical characterization of the purified human DNA2 [131], the precise mitochondrial import mechanism and function is not resolved. A second helicase that can be found in the nucleus and mitochondria of human cells and that has many alleged functions in DNA metabolism is PIF1 (petite frequency integration 1) [132]. Translation initiation at an alternative ATG of the PIF1 gene produces a strictly mitochondrial isoform [133]. PIF1 knockout mice accumulate mtDNA alterations and develop mitochondrial myopathy [134], however, the exact role of mammalian PIF1 in mtDNA stability remains to be determined.

    The DNA polymerase/primase PrimPol has roles in both nuclear and mtDNA maintenance [135]. PrimPol functionally interacts with mtSSB and Twinkle, but nothing is known about the mechanism by which it is targeted to the mitochondrion [136,137]. PrimPol was shown to be able to reinitiate stalled mtDNA replication by priming mtDNA replication at nonconventional replication origins [54]. However, many regulatory and mechanistic details of this mtDNA replication reinitiation still need to be disclosed.

    1.3 The mitochondrial dNTP supply

    The faithful maintenance of mtDNA requires not only a functional replication machinery, but also a sufficient and balanced supply of mitochondrial deoxyribonucleoside triphosphates (dNTPs), the building blocks of DNA. The importance of an appropriate dNTP supply is highlighted by the fact that defects in mitochondrial nucleotide metabolism give rise to mitochondrial diseases such as PEO and mtDNA depletion syndrome [138–142]. There are two principal ways to synthesize dNTPs: by producing them from ribonucleoside diphosphates (NDPs) through the de novo pathway, or by phosphorylating existing deoxyribonucleosides (dNs) by action of the salvage pathway. Both the de novo and the salvage pathways contribute to maintenance of mitochondrial dNTP pools (Fig. 1.4).

    Figure 1.4 Nucleotide synthesis pathways that provide mitochondrial dNTPs.

    Enzyme names are in gray type. The synthesis of dNTPs can occur through the de novo or the salvage pathways. In the de novo pathway, the cytosolic enzyme ribonucleotide reductase (RNR) catalyzes the reduction of NDPs to dNDPs that can subsequently be phosphorylated to dNTPs by nucleoside diphosphate kinases (NDPKs) that are present in both the cytosol (left-hand side) and the mitochondria (right-hand side). Alternatively, dNTPs can be produced through the salvage pathway that recycles bases or nucleosides (dNs) that are taken up from outside the cell or that derive from the breakdown of endogenous nucleotides or nucleic acids. The rate-limiting step of dNTP salvage synthesis is the phosphorylation of dNs to deoxyribonucleoside monophosphates (dNMPs). This step is carried out by thymidine kinase 1 (TK1) and deoxycytidine kinase (dCK) in the cytosol, and thymidine kinase 2 (TK2) and deoxyguanosine kinase (dGK) in the mitochondria. The dNMPs thus produced can be further phosphorylated to deoxyribonucleoside diphosphates (dNDPs) and then to dNTPs through the sequential action of nucleoside monophosphate kinases (NMPKs) and NDPKs. The synthesis of dNTPs is counteracted by catabolic enzymes such as 5′-deoxynucleotidases (5′-dNs) that dephosphorylate dNMPs to dNs in both the cytosol and the mitochondria, as well as nucleoside phosphorylases (thymidine phosphorylase [TP] and purine nucleoside phosphorylase [PNP]) that degrade dNs to bases in the cytosol. Also the dNTP phosphohydrolase SAMHD1 limits the levels of dNTPs in cells outside of S phase. Due to the presence of nucleotide carriers such as SLC25A33, SLC25A36 and as-yet unidentified carrier(s), the cytosolic and mitochondrial pools of deoxynucleotides (dNMPs, dNDPs, and/or dNTPs) are in rapid exchange. Similarly, equilibrative nucleoside transporters (ENTs) ensure the exchange of dNs between the two cellular compartments.

    The main regulatory step of de novo dNTP synthesis is catalyzed by the enzyme ribonucleotide reductase (RNR) that is a heterotetramer consisting of two large RRM1 catalytic subunits and two small RRM2 accessory subunits [143]. RNR carries out the reduction of NDPs to dNDPs in the cytosol; the dNDPs can then be phosphorylated to dNTPs by nucleoside diphosphate kinases (NDPKs) that are present in both the cytosol and the mitochondria [144–146]. The synthesis of dTTP requires the contribution of additional enzymes. The activity of the de novo pathway follows the cell cycle, peaking in S phase when the demand for dNTPs is high due to the replication of nuclear DNA. This cell-cycle-phase-dependent regulation is mainly achieved by modulating the levels of the small RRM2 subunit of RNR, while the levels of the RRM1 subunit remain stable [143]. RRM2 is undetectable in nondividing cells [147] where RNR activity instead relies on the alternative small subunit RRM2B to provide the dNTPs required for DNA repair and mtDNA replication [148–151]. However, the levels of the RRM1/RRM2B complex in nondividing cells are clearly lower than those of the RRM1/RRM2 complex in S phase, resulting in ~18-fold lower dNTP pools in nondividing cells [152]. Despite the low activity of de novo dNTP synthesis in nondividing cells, its contribution is nonetheless critical for mtDNA maintenance as evidenced by the severe mtDNA depletion observed in patients with RRM2B defects [141].

    Unlike de novo dNTP synthesis, mitochondrial dNTP salvage is considered to produce a low but constitutive level of mitochondrial dNTPs throughout the cell cycle [153]. The rate-limiting step of mitochondrial salvage is the phosphorylation of dNs to deoxyribonucleoside monophosphates (dNMPs), a step that is catalyzed by thymidine kinase 2 (TK2; phosphorylates dT, dU, and dC) and deoxyguanosine kinase (dGK; phosphorylates dG and dA) [153]. Accordingly, defects in either of these nucleoside kinases manifest as mtDNA depletion [138,139], and TK2 mutations can also result in multiple mtDNA deletions [142]. The dNMPs produced by TK2 and dGK are phosphorylated by mitochondrial nucleoside monophosphate kinases (NMPKs) to dNDPs and further by NDPKs to dNTPs (Fig. 1.4).

    Both of the described pathways of mitochondrial dNTP synthesis involve import of either dNs or deoxyribonucleotides (dNMPs, dNDPs, or dNTPs) into mitochondria. In fact, mitochondrial dN and deoxyribonucleotide pools are believed to be in exchange with cytosolic pools [154–156] due to the presence of carrier proteins in the inner mitochondrial membrane. dNs, which constitute the substrates of mitochondrial salvage, can originate from the degradation of existing mitochondrial deoxyribonucleotides or be imported from the cytosol by equilibrative nucleoside transporters (ENTs) [157,158]. In contrast, the products of cytosolic de novo dNTP synthesis are imported into mitochondria in the form of dNMPs, dNDPs, and/or dNTPs. So far, a comprehensive understanding of mitochondrial nucleotide carriers and their individual contributions to maintenance of the intramitochondrial deoxyribonucleotide pool remains elusive. However, two solute carrier 25 family proteins, SLC25A33 and SLC25A36, are likely responsible for the exchange of mitochondrial and cytosolic pyrimidine nucleotides (U, T, C) [159–161].

    The overall size of the cellular dNTP pool is determined not only by the anabolic pathways that synthesize dNTPs, but also by the activity of catabolic enzymes. These include 5′-deoxynucleotidases (5′-dNs) that dephosphorylate dNMPs to dNs in both the cytosol and the mitochondria as well as nucleoside phosphorylases that degrade dNs in the cytosol [162]. Defects in one such phosphorylase, thymidine phosphorylase (TP), imbalance cellular dNTP pools and manifest as mtDNA deletions and consequent mitochondrial disease [140,163]. Furthermore, the SAMHD1 (SAM domain and HD domain-containing protein 1) dNTP phosphohydrolase converts dNTPs to dNs in the cytosol and thereby limits cellular dNTP levels [164–166]. Due to the exchange of cytosolic and mitochondrial pools, the activity of SAMHD1 also reduces mitochondrial dNTP pools [167].

    A peculiar feature of mtDNA is that it contains frequent ribonucleotides, the building blocks of RNA [19,20,168]. Owing to the great excess of free NTPs over dNTPs [169] in the cell, ribonucleotides are occasionally inserted in place of dNTPs during replication and repair of the nuclear and mitochondrial genomes [108,170–173]. However, while ribonucleotides incorporated in nuclear DNA are removed by a dedicated repair pathway [174–176], the ones in mtDNA are not effectively repaired and therefore persist [177,178]. Owing to the lack of repair, the main factor defining the frequency and identity of mtDNA ribonucleotides is the ratio of NTPs to dNTPs in the cell [177,179,180]. At this stage it is unclear whether the ribonucleotides in mtDNA serve a specific purpose or if they are merely tolerated. They are in any case neither essential for mtDNA maintenance nor exceedingly harmful because a dramatic reduction in ribonucleotide content has no striking effect on mtDNA stability during mouse lifespan [180].

    1.4 Mitochondrial nucleoids

    MtDNA localizes in specific areas of the mitochondrial matrix and is organized in nucleoprotein complexes called mitochondrial nucleoids. The concept of mitochondrial nucleoids dates back to the 1960s, when the term was used, in analogy with the bacterial nucleoids, to indicate specific, electron-transparent, DNA-containing small rod-like structures in the mitochondrial matrix [181]. In the following three decades, mtDNA-protein complexes were isolated from the slime mold Physarum polycephalum [182], HeLa cells [183], Xenopus laevis oocytes [184,185], yeast cells [186,187], and rat liver [188]. In more recent years, methods were developed to directly visualize nucleoids in cells using either fluorescent nucleic acid-binding dyes like 4′,6-diamidino-2-phenylindole (DAPI) [12,189] and PicoGreen [190,191] or antibodies against nucleoid-associated proteins [87,192].

    The exact protein composition of the mitochondrial nucleoids is still unknown, but comprises the packaging factor TFAM, mtSSB, and other proteins involved in DNA replication, transcription, and mitochondrial metabolism. However, recent observations have pointed out that, with the exception of TFAM, the protein composition of nucleoids is not uniform and different subsets of nucleoids coexist within human cells [193], highlighting the extremely dynamic nature of these structures.

    The organization of mtDNA into nucleoids might have a dual function: on one hand, the association with proteins protects the DNA and renders it more resistant to damaging agents such as reactive oxygen species produced by the respiratory chain complexes [194]. On the other hand, it provides the proper microenvironment for the different DNA transactions including mtDNA replication, transcription, and repair [195]. Therefore the mtDNA nucleoids are essential for the maintenance of mtDNA (Box 1.1).

    Box 1.1

    Approaches to study nucleoids in cells and tissues

    Nucleoids can be detected by staining the nucleic acid component with fluorescent dyes like DAPI [12] that binds to AT-rich regions of the genome and PicoGreen that selectively binds to dsDNA [191]. Conversely to DAPI, PicoGreen is permeable to cell membranes and can be used to visualize nucleoids in living cells [190].

    Alternatively, nucleoids can be visualized by detecting the proteic-component with immunocyto- and histochemistry techniques employing specific antibodies [87,192].

    Finally, the subpopulation of replicatively active nucleoids can be visualized either with antibodies against proteins of the mtDNA replisome (POL γ, Twinkle, or mtSSB) [196] or by treating the cells with thymidine analogs like BrdU (5-bromo-2′-deoxyuridine) or EdU (5-Ethynyl-2′-deoxyuridine) that will be incorporated during the replication process and can be detected with specific antibodies [106,192].

    Because conventional fluorescence microscopy allows to reach resolution of 200–350 nm, and the size of a nucleoid is around 100 nm, high-resolution techniques must be used in order to visualize single nucleoids. These include STED (stimulation emission depletion) microscopy [197] and PALM (photoactivated localization) microscopy [198]. While based on different physical principles, both methods overcome the diffraction barrier and allow to resolve structures smaller than 50 nm [199].

    1.4.1 Nucleoid composition

    1.4.1.1 mtDNA

    In relaxed form, the 16-kb-long human mtDNA molecule has a contour length of 5 μm [200]. In the nucleoids, the mtDNA is highly packaged into a condensed form that is only 100 nm in diameter [197]. Several groups have quantified the number of nucleoids and mtDNA copies in cells and found that in rapidly dividing cells, each nucleoid contains around 2–10 mtDNA molecules [201]. However, these observations were done using conventional fluorescence microscopy whose resolution does not allow resolving structures smaller than 200 nm. More recently, superresolution microscopy techniques have indicated an average of 1.1 mtDNA molecules per nucleoid [202]. Notably, two decades earlier, Satoh and Kuroiwa reached the same conclusion by calculating the fluorescence intensity of the ethidium bromide signal in the nucleoids [12]. A rise in the mtDNA copy number leads to an increased number of nucleoids rather than to a change in their morphology, further corroborating the ~1:1 ratio of mtDNA and nucleoids [202]. However, this ratio might not be applicable to all cell types and tissues, since mtDNA is not always present as single-monomeric molecule, but can also exist as catenated molecules containing multiple complete genomes [203,204].

    1.4.1.2 Nucleoid-associated proteins

    The first nucleoid-associated protein identified was the Saccharomyces cerevisiae ARS-binding factor 2 protein (Abf2p), originally named HM [205,206]. This 20-kDa protein contains two high mobility group (HMG) domains and is homologous to the DNA-binding HMG proteins located at the nuclear chromatin [207]. The protein is highly abundant in yeast and coats the entire mtDNA in a ratio of one molecule for every 15 bp [206]. The deletion of Abf2p causes the loss of mtDNA in strains grown on a fermentable carbon source (e.g., glucose), indicating the importance of this protein for the maintenance of the mtDNA. However, when abf2− cells are grown on nonfermentable carbon sources where the mitochondrial genome is not dispensable, they maintain normal levels of mtDNA [206], suggesting that also other factors are involved in yeast mtDNA maintenance. Nonetheless, abf2− cells grown under these conditions are more sensitive to damaging agents [208] and have altered nucleoid morphology and segregation [209]. The mtDNA instability due to loss of Abf2p can be rescued by expression of the 25-kDa human mitochondrial transcription factor A (TFAM) [210]. In mouse, TFAM is an essential gene, since its loss causes embryonic lethality and mtDNA depletion [211]. Like the yeast counterpart, it contains two HMG domains and it is sufficiently abundant to fully cover the mtDNA [212]. In vitro, the protein coordinates the packaging of DNA [213], and it colocalizes with mtDNA in vivo [192]. The crystal structure of the protein complexed with the LSP region of mtDNA revealed that TFAM, through its HMG domains, bends the DNA 180°, thus inducing a U-turn [214]. The same DNA distortion into a U-turn also occurs when TFAM binds to nonspecific DNA sequences [215]. A similar DNA bending-mode was detected also for Abf2p, indicating that it represents a common mechanism for mtDNA compaction by HMG proteins [216]. However, unlike Abf2p, TFAM plays an additional role in the initiation of mitochondrial transcription [217]. The two functions are ascribed to two different domains of the protein. Besides the HMG-box domains, TFAM contains an additional C-terminal tail that accounts for the specific binding to the LSP promoter and the role of TFAM in transcription initiation [218]. Accordingly, a chimeric Abf2p protein fused with the C-terminal tail of TFAM can activate transcription from the LSP promoter [219]. The mechanism by which TFAM coordinates both DNA packaging and transcription initiation is still unknown;[220] however, the amount of protein seems to be implicated in the switch between the two functions as high TFAM concentrations result in increased genome compaction and reduced mtDNA transcription and replication in vitro (Fig. 1.5B) [221].

    Figure 1.5 mtDNA nucleoids.

    (A) In vivo imaging of HeLa cells stained with MitoTracker Red CMXRos (red), a fluorescent dye that accumulates into mitochondria depending on the membrane potential, and PICOGREEN (green), that specifically stains nucleic acids. In living cells, nucleoids appear like a punctate pattern inside the mitochondrial network. Cells were imaged with the Leica SP8 FALCON Confocal instrument equipped with a HC PL APO 63x/1.20 Water objective. (B) Schematic representation of the dual function of TFAM (green dots) in mtDNA metabolism. TFAM is implicated in the initiation of mtDNA transcription and in the compaction of mtDNA. Low protein abundance is linked in vitro to increased levels of mtDNA transcription, while higher TFAM concentrations induce the compaction of the mitochondrial genome. (C) Schematic representation of nucleoids segregation within the mitochondrial network. Actively replicating nucleoids (green circles), for example, containing POL γ (purple), are located in close proximity to a subset of ER-mitochondria contact site (acid green). Upon replication and generation of two novel mtDNA molecules, a constriction of the mitochondrial membranes occurs at these sites triggering the assembly of the fission machinery (violet dots). This results in the segregation of the two mtDNA molecules across the mitochondrial network.

    Recently, it was proposed that TFAM alone can package single mtDNA molecules and form a nucleoid unit [202]. This is further corroborated by the findings that the components of the mtDNA replication machinery, such as Twinkle and mtSSB, colocalize in situ only with a subset of nucleoids, suggesting that this association is transient and that nucleoids do not have a uniform composition [193]. Besides the mtDNA replication factors, several other proteins have over the years been found to be associated with nucleoids. These factors were identified either by in situ colocalization with mtDNA or by purification or immuno-precipitation of nucleoids with antibodies against DNA, and known nucleoid-associated proteins, followed by mass-spectrometry analysis [222]. More recently, also proximity-dependent labeling methods have been used to find mtDNA-associated proteins [223]. These methods have identified several components of the mtDNA replication and transcription machineries as nucleoid proteins, but also proteins involved in RNA processing and translation, as well as other factors like the Lon protease and the ATPase family AAA-domain-containing 3A protein (ATAD3A) [224]. The Lon protease is an ATP-dependent protease that binds to mtDNA in living cells [225] and has role in the regulation of cellular levels of TFAM through degradation [226,227]. ATAD3A is a mitochondrial inner membrane (MIM) protein [228] implicated in mtDNA maintenance [229,230]. ATAD3A downregulation caused altered nucleoid distribution and the protein is proposed to be part of a protein platform that connects the nucleoids to the MIM, as discussed below [231].

    1.4.2 Nucleoid topology

    In human cells, mitochondrial nucleoids appear as spherical, ellipsoid structures with a diameter of around 100 nm [197,202]. When visualized with fluorescent dyes, nucleoids look like a punctate pattern inside the mitochondrial network (Fig. 1.5A). The total number of nucleoids per cell has been reported for several human and mouse cell types and ranges from 500 to a few thousand [232].

    The organization of the nucleoprotein complex is largely unknown and, as mentioned above, there is no consensus on what defines a nucleoid unit and which proteins are part of it. This uncertainty can be ascribed partly to the transient nature of the interactions of many mtDNA-associated factors with the mtDNA [193] and partly to the fact that the methods for nucleoid purification used by different groups differ in experimental conditions, and, as result, give different read-outs [222]. In 2008, Bogenhagen and coauthors [233] compared two different methods: a stringent purification that implied formaldehyde-cross-linking of the proteins to the mtDNA before nucleoid purification, and a native purification, in which nucleoids were immunoprecipitated using antibodies against TFAM or mtSSB after lysis of mitochondria with nonionic detergent. After comparing the cohort of proteins detected with the two methods, a layered model for mtDNA structure was proposed. According to this model, the proteins of the replication and transcription machineries are part of the central core of the nucleoids and are located in close proximity to the mtDNA, while components of the translation machinery as well as the RNA processing and respiratory chain complexes are located in the peripheral area of the

    Enjoying the preview?
    Page 1 of 1