Cshperspect-MIT-A021220 1..47

Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics and the Heteroplasmy Conundrum in Evolution and Disease

Douglas C. Wallace and Dimitra Chalkia

Center for Mitochondrial and Epigenomic Medicine, The Children’s Hospital of Philadelphia, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104 Correspondence: [email protected]

The unorthodox genetics of the mtDNA is providing new perspectives on the etiology of the common “complex” diseases. The maternally inherited mtDNA codes for essential energy genes, is present in thousands of copies per cell, and has a very high mutation rate. New mtDNA mutations arise among thousands of other mtDNAs. The mechanisms by which these “heteroplasmic” mtDNA mutations come to predominate in the female germline and somatic tissues is poorly understood, but essential for understanding the clinical variability of a range of diseases. Maternal inheritance and heteroplasmy also pose major chal- lengers for the diagnosis and prevention of mtDNA disease.

THE GENETIC CHALLENGES OF mtDNA most important bioenergetic genes. So mtDNA DISEASES defects impinge on a wide spectrum of cellular functions. t is has become increasingly clear that mito- A large number of pathogenic mtDNA mu- Ichondrial dysfunction lies at the nexus of a tations have been identiﬁed and the more severe wide range of metabolic and degenerative dis- mutations are frequently mixed with normal eases, cancer, and aging. Two major reasons for mtDNAs within the cell, a state known as het- why mitochondrial dysfunction has been over- eroplasmy. Heteroplasmic alleles can shift in looked in “complex” diseases is that subtle bio- percentage during both mitotic and meiotic energetic alterations can have major clinical cell division, leading to a potentially continuous consequences and mitochondrial defects can array of bioenergetic defects, a process known be generated by the unique quantitative genetics as replicative segregation. As the percentage of of the maternally inherited mitochondrial DNA mutant mtDNAs increases, the resulting bio- (mtDNA). energetic defect becomes increasingly severe. The mitochondrial genome encompasses Because different tissues have different bioener- between 1000 to 2000 nuclear DNA (nDNA) getic thresholds, as a patient’s bioenergetic ca- genes plus thousands of copies of the maternally pacity declines it eventually falls below the min- inherited mtDNA. The mtDNA codes for the imum threshold for that tissue and symptoms

Editors: Douglas C. Wallace and Richard J. Youle Additional Perspectives on Mitochondria available at www.cshperspectives.org Copyright # 2013 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a021220 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220

1 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

ensue. Because the tissues and organs with the tions have estimated that the incidence of clini- highest bioenergetic requirements are also those cal mitochondrial diseases is about one in 5000 that are primarily affected in the common met- (Schaefer et al. 2004, 2008). More surprising, abolic and degenerative diseases, it follows that a survey of newborn cord bloods revealed that mitochondrial dysfunction may be a major con- one in 200 infants harbored one of 10 common tributor to complex diseases. pathogenic mtDNA mutations (Elliott et al. Women that harbor deleterious heteroplas- 2008; Chinnery et al. 2012). Hence, pathogenic mic mutations have a high probability of hav- mtDNA mutations are very common and con- ing affected children, the nature and severity stantly arising. of the phenotype depending on the mtDNA mutation and the percentage of heteroplasmy. Human OXPHOS and the Range Cells and individuals can accumulate an array of Phenotypes: Conception to Old Age of different mtDNA mutations over time, the aggregate of which degrade the energetic ca- To understand the clinical implications of pacity of the cell. Such mutations are important mtDNA mutations, it is essential to understand in aging and cancer. Given the enormous po- the central role that mitochondrial oxidative tential explanatory power of heteroplasmic phosphorylation (OXPHOS) plays in cellular mtDNA mutations, it is striking that very little biology. The mitochondria oxidize the calories is known about the origin, genetics, and phe- in our diet with the oxygen that we breathe to notypic effects of heteroplasmic mtDNA mu- generate 90% of cellular energy. In OXPHOS, tations. electrons (reducing equivalents) derived from our food flow down the mitochondrial inner membrane electron transport chain (ETC) HUMAN mtDNA GENETICS from reduced to oxidized states, ultimately That mtDNA mutations could cause disease terminating with reduction of oxygen to water. was first reported at the molecular level in The ETC is initiated with oxidation of NADH 1988 with the demonstration that isolated pa- by complex I (NADH:CoQ oxidoreductase or tients with mitochondrial myopathy could har- NADH dehydrogenase) or succinate by com- bor heteroplasmic mtDNA deletions (Holt et al. plex II (succinate:CoQ oxidoreductase or succi- 1988); that the maternally inherited sudden on- nate dehydrogenase). The electrons are then set blindness disease, Leber hereditary optic transferred to coenzyme Q (CoQ), complex neuropathy (LHON), was caused by a homo- III, cytochrome c, complex IV (cytochrome c plasmic missense mutation in the ND4 gene at oxidase or COX), and finally to oxygen. As the nt 11778G.A (arginine codon 340 to histidine, electrons traverse complexes I, III, and IV, the R340H) (Wallace et al. 1988a); and that myo- energy released is used to pump protons from clonic epilepsy and ragged red fiber disease the mitochondrial matrix across the mito- (MERRF) was caused by a heteroplasmic muta- chondrial inner membrane to the intermem- tion in the tRNALys gene at nt 8344A.G (Wal- brane space (Wallace 2005, 2007, 2011). This lace et al. 1988b; Shoffner et al. 1990). These creates a transmembrane electrochemical gradi- discoveries set the stage for investigating and ent of 0.2 volts. This potential energy can then understanding a broad range of enigmatic fami- be used to drive OXPHOS complex V (Hþ- lial and age-related diseases. translocating ATP synthase) to condense ADP and phosphate (Pi) to generate ATP (Mitchell 1961), thus coupling oxidation by the ETC Incidence of mtDNA Mutations with phosphorylation by the ATP synthase. and Disease The mitochondrial ATP is then exported to Mutations in mtDNA are surprisingly common. the cytosol via the adenine nucleotide translo- Geneticepidemiological studies quantifying only cators (ANTs), where the ATPenergizes cellular the most common pathogenic mtDNA muta- reactions and drives work.

2 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

In addition to generating ATP energy, the mtDNA Mutations mitochondria regulate cytosolic Ca2þ levels, which in turn modulate cellular and mitochon- The mtDNA genes have a very high sequence drial metabolic pathways, control the cellular evolution rate, on the order of 10–20 times REDOX state that regulates a wide array of cel- that of comparable nDNA genes (Brown et al. lular enzymatic reactions and transcription fac- 1982; Neckelmann et al. 1987; Wallace et al. tors via thiol-disulfide interconversion, regulate 1987). This is the product of both an exception- mitochondrial ROS production that is both a ally high mutation rate, perhaps 100- to 1000- signal transduction agent that impinges on fold higher than nDNA genes, times an mtDNA molecules such as HIF and RAS, and is the ma- mutant fixation rate: E ¼ mF where E is the jor source of oxidative stress that can activate sequence evolution rate, m the mutation rate, the innate immunity response through NF-kB and F the fixation rate. When a new mutation signaling. When mitochondria experience ex- arises in the mtDNA, it creates an intracellular treme stress (elevated Ca2þ and ROS, depleted heteroplasmic mixture of mutant and normal adenine nucleotides, and reduced membrane mtDNAs, but the mutant mtDNA is but one potential), this can activate the mitochondrial among thousands of nonmutant mtDNAs. In permeability transition pore (mtPTP) thus ini- some manner, the initial mutant mtDNA be- tiating apoptosis and necrosis (Wallace 2005, comes enriched within certain cells, ultimately 2011, 2012, 2013a,b; Wallace et al. 2013). coming to predominate and influence the cellular and patient phenotype. The mechanism by Mitochondrial Genetics which this enrichment occurs in either germline or somatic cells remains a mystery. The mtDNA Once an mtDNA mutation reaches an ap- The mtDNA (Fig. 1) codes for the 13 most im- preciable level within cells, the percentage of portant OXPHOS polypeptides. These include mutant mtDNAs can drift by replicative segre- seven of the 45 polypeptides of OXPHOS gation. For an embryo generated by the fertili- complex I (ND1-3, ND4L, ND4-6): one of the zation of a heteroplasmic oocyte, the percentage 11 polypeptides of complex III (cytochrome b, of mutant and normal mtDNAs in different de- cytb), three of the 13 polypeptides of complex scendant tissues and organs can have quite dif- IV (COI-III), and two of the 15 polypeptides ferent values. This genetic mosaicism results in of complex V (ATP6 and 8). In addition, the bioenergetic mosaicism and phenotypic com- mtDNA encodes the mitochondrial 16S and plexity. Added to the stochastic segregation of 18S rRNAs and 22 tRNAs for mitochondrial heteroplasmic mtDNA is the differential sensi- protein synthesis. The mtDNA also encompass- tivity of different organs to different mitochon- es an 1000 nt control region that contains an drial physiological alterations. The brain is the origin for replication of the G-rich heavy (H) most sensitive to partial bioenergetic defects stand and the promoters for transcription of followed by heart, muscle, kidney, and endo- both the H stand and the C-rich light (L) stand. crine systems (Wallace 2005). Hence, subtle sys- Both mtDNA stands are transcribed into large temic mitochondrial deficiencies can result in polycistronic transcripts in which the larger organ-specific symptoms. rRNA and mRNA transcripts are punctuated There are three classes of clinically relevant by tRNAs. The tRNAs are processed out and mtDNA variants: recent deleterious mutations, the larger RNA products are polyadenylated. ancient adaptive mtDNA mutations, and so- The mtDNA mRNAs are translated on mito- matic mtDNA mutations. chondrial-specific 55S ribosomes, which are Maternally Inherited Diseases. The most sensitive to bacterial ribosomal initiators chlor- clinically overt class of mtDNAvariants is newly amphenicol (CAP) and aminoglycosides and arising maternally inherited disease mutations. are initiated with an N-formylmethionine just Because of the high mtDNA mutation rate, like bacterial protein synthesis (Wallace 2007). new pathogenic mtDNA mutations are contin-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 3 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

Regulatory mutations: somatic, inherited?

DEAF A1555G F T CR V 12s Encephalomyopathy Cytb LHON T14484C rRNA mutations: inherited P LDYS G14459A 0/16589 Europe J2 ND6 16s E rRNA Europe J1 MELAS A3243G Asia-Am C L LHON G3460A Europe J ND6 ND1 Africa L Adaptive Europe J/T mutations: I Q Europe U Asia-Am A inherited L M S ADPD T4336C Europe T H ND2 Asia-Am D Prostate cancer A ND4 N Eurasia N,M mutations: LHON G11778A C Europe H inherited and somatic W Y ND4L Asia-Am B ND3 PC T6253C R S LHON T10663C PC G6261A C0I C0III G ATPase6 PC C6340T C0II PC A6663G D K MERRF A8344G ATPase8 NARP/Leigh’s T8993C/G

Figure 1. Human mitochondrial DNA map showing representative pathogenic and adaptive base substitution mutations. CR (control region) ¼ D-loop. The letters around the outside perimeter or on the inside circle indicate cognate amino acids of the tRNA genes. Other gene symbols are defined in the text. Arrows followed by continental names and associated letters on the inside of the circle indicate the position of defining polymorphisms of selected region-specific mtDNA lineages. Arrows associated with abbreviations followed by numbers around the outside of the circle indicate representative pathogenic mutations, the number being the nucleotide position of the mutation. The full array of pathogenic mtDNA mutations and polymorphisms are available through Mitomap.org (MITOMAP 2012). DEAF, deafness; MELAS, mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes; LHON, Leber hereditary optic neuropathy; ADPD, Alzheimer’s disease and Parkinson’s disease; MERRF, myoclonic epilepsy and ragged red fiber disease; NARP,neurogenic muscle weakness, ataxia, retinitis pigmentosum; LDYS, LHON þ dystonia; PC, prostate cancer. (From Wallace 2007; reproduced, with permission, from the author.)

uously being introduced into the human pop- Ancient Adaptive Polymorphisms. There is ulation. Hundreds of pathogenic mtDNA mu- also substantial mtDNA sequence diversity be- tations have now been documented (MITO- tween individuals and human populations. MAP 2012; Wallace et al. 2013) and these can These ancient mtDNA polymorphisms accu- affect virtually every tissue in the body, depend- mulated along radiating maternal lineages as ing on the mutation’s severity, nature, and het- women migrated out of Africa to colonize the eroplasmy level. Hence, pathogenic mtDNA globe. As new mtDNA mutations arose, new mutations and mitochondrial dysfunction re- branches of the mtDNA tree were generated. If sult in a wide range of multisystem degenerative a founder mutation changed mitochondrial diseases. physiology in a manner beneﬁcial to individuals

4 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

within that regional environment, then that ond, of all of the Asian mtDNAs, only three mtDNA lineage became enriched in that geo- mtDNA lineages (A, C, and D) moved to ex- graphic locality. Subsequent additional base treme northeast Siberia to found the Paleo-In- substitutions in descendant mtDNAs generated dians. Third, and most surprising, the mtDNA a group of related regional haplotypes designat- sequence evolution rate is such that it produced ed a haplogroup. Hence, each continent and important mtDNA evolutionary changes that geographical region is associated with a distinc- coincide with the major human geographic mi- tive array of mtDNA sequence types. grations. Such associations could not have oc- All African mtDNAs are related and encom- curred by chance. Rather, it is most likely that passed within one large continent-speciﬁc line- mtDNA variation permitted adaptation of our age designated “macrohaplogroup L.”Macroha- human ancestors to different regional environ- plogroup L arose between 130,000 and 200,000 ments, thus being the adaptive system that per- years before present (YBP) (Fig. 2), founded by mitted human colonization of the diverse envi- mtDNAs similar to haplogroup L0, which is ronments that they encountered around the common among the Khoi-San Bushmen of globe (Ruiz-Pesini et al. 2004; Mishmar et al. South Africa. In Ethiopia, 65,000 YBPAfrican 2006; Ruiz-Pesini and Wallace 2006; Wallace haplogroup L3 gave rise to two mtDNAs desig- 2013a). nated M and N. Only the mtDNA descendants The founding mtDNA for macrohaplo- from M and N mtDNAs left Africa to colonize group N, which moved directly from subtrop- the rest of the world, generating macrohaplo- ical Africa north into the European temper- groups M and N. From Africa, macrohaplo- ate zone, harbored two polypeptide variants, group M moved along tropical Southeast Asia, ND3 nt 10398G.A (A114T) and ATP6 nt ultimately reaching Australia. Later, M descen- 8701G.A (A59T) (Wallace et al. 1999, 2013). dants moved north outof Southeast Asiato form These variants have been associated with alter- a plethora of central and eastern Asian mtDNA ations in the mitochondrial membrane poten- haplogroups including C, D, G, and M1–M20. tial and Ca2þ metabolism (Kazuno et al. 2006). Out of Africa, macrohaplogroup N went in two Presumably, these variants reduced the coupling directions. In one, N moved through Southeast efﬁciency of mitochondrial OXPHOS (loose Asia to Australia and from southern Asia, north coupling), resulting in an increase in the num- into central Asia to generate haplogroups A and ber of calories burned by the mitochondria to Z. In the second, macrohaplogroup N moved generate the ATPrequired to perform work. Be- north out of Africa to form the European hap- cause a calorie is a unit of heat, burning more logroups I, X, and W. In western Eurasia, N also calories would increase core body-heat produc- gave rise to submacrohaplogroup R. R then gave tion, rendering these individuals more resistant risetotheremainingEuropeanhaplogroupsH,J, to the cold stress encountered in more northern Uk, T,U, and V.R also moved east to produce the environments. By contrast, macrohaplogroup Asian mtDNA haplogroups B and F (Fig. 3). M mtDNAs, which initially remained in the Of all the Asian mtDNAvariants, only A, C, tropics, did not acquire comparable functional and D became enriched in northeastern Siberia mtDNA mutations. Presumably, this mtDNA and were in a position to cross the Bering Land lineage retained the tight coupling of OXPHOS Bridge 20,000 YBP to establish the Paleo-In- found in Africa in which ATP production is dian populations. Later additional migrations maximized and heat production is minimized brought haplogroups B and X to join hap- (Ruiz-Pesini et al. 2004; Mishmar et al. 2006; logroups A, C, and D (Fig. 2) (Wallace et al. Ruiz-Pesini and Wallace 2006; Wallace 2013a). 1999, 2013). Consistent with the concept that mtDNAvaria- The regionality of the mtDNA haplogroups tion has permitted climatic adaptation, mtDNA is extraordinary in several respects. First, of all of variation but not nDNA variation has been the African diversity, only two mtDNA lineages found to correlate with climatic differences (M and N) colonized the rest of the world. Sec- (Balloux et al. 2009). Also, the basal metabolic

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 5 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

<8000 ~20,000 39,000– H,J A,D Z X~15,000 51,000 T,U,Uk,V Y 12,000– A I,W,X C,D,G A A C,D R 15,000 7000– M1-M40 9000 B B 65,000– N F 70,000 ~3000 M B L2 L3 M

L1 B C,D 130,000– L0 A 170,000 M42 Q S P 48,000

Figure 2. Diagram of the migratory history of the human mtDNA haplogroups. Homo sapiens mtDNAs arose in Africa 130,000–200,000 years before present (YBP), with the first African-specific haplogroup branch being L0, followed by the appearance in Africa of lineages L1, L2, and L3. In northeastern Africa, L3 gave rise to two new lineages, M and N. Only M and N mtDNAs successfully left Africa 65,000 YBPand colonized all of Eurasia and the Americas. In Eurasia and the Americas, M and N gave rise to a diverse array of mtDNA lineages designated macrohaplogroups M and N. The founders of macrohaplogroup M moved out of Africa through India and along the Southeast Asian coast down along the Malaysian peninsula and into Australia, generating haplogroups Q and M42 48,000 YBP. Subsequently, M moved north out of Southeast Asia to produce a diverse array of Central Asian mtDNA lineages including haplogroups C, D, G, and manyother M haplogroup lineages. In northeast Asia, haplogroup C gave rise to haplogroup Z. The founders of macrohaplogroup Nalso moved through Southeast Asia and into Australia, generating haplogroup S. In Asia, macrohaplogroup N mtDNAs also moved north to generate central Asian haplogroup A and Siberian haplogroup Y. In western Eurasia, macrohaplogroup N founders also moved north to spawn European haplogroups I, W, and X, and in western Eurasia, gave rise to submacrohaplogroup R. R moved west to produce the European haplogroups H, J, Uk, T, U, and V and also moved east to generate Australian haplogroup Pand eastern Asian haplogroups F and B. By 20,000 YBP,mtDNA haplogroups C and D from M, and A from N, were enriched in northeastern Siberia and thus were positioned to migrate across the Bering land bridge (Beringia) to give rise to the first Native American populations, the Paleo-Indians. Haplogroups A, C, and D migrated throughout North America and on through Central America to radiate into South America. Haplogroup X, which is most prevalent in Europe but is also found in Mongolia though not in Siberia, arrived in North America 15,000 YBP but remained in northern North America. Haplogroup B, which is not found in Siberia but is prevalent along the coast of Asia, arrived in North America 12,000–15,000 YBPand moved through North and Central America and into South America, combining with A, C, D, and X to generate the five dominant Paleo-Indian haplogroups (A, B, C, D, X). A subsequent migration of haplogroup A out of the Chukotka peninsula 7000–9000 YBP gave rise to the Na-Deńe´ (Athabaskins, Navajo, Apache, etc.). Subsequent movement acrossthe Bering Strait, primarilycarrying haplogroups AandD after6000 YBP,produced the Eskimo and Aleut populations. Most recently, eastern Asian haplogroup B migrated south along the Asian coast through Micronesia and out into the Pacific to colonize all of the Pacific islands. Ages of migrations are approximated using mtDNA sequence evolution rates determined by comparing regional archeological or physical anthropological datawith corresponding mtDNA sequencediversity.Because selection may have limited the accumulation of diversity in certain contexts, ages for regional migrations were estimated from the diversity encompassed within an individual regional or continental haplogroup lineages. This is because selection would have acted onthe haplogroup mtDNAbut most subsequentmutations would accumulate by random geneticdrift and thus be “clock-like.” (From Wallace 2013a,b; reproduced, with permission, from the author.)

rate of Siberian populations is higher than that one example of the importance of functional of more southern populations (Leonard et al. variants in founding and deﬁning the branches 2002; Snodgrass et al. 2005, 2008). of the mtDNA haplogroup trees (Fig. 3). The The portion of the European mtDNA tree J-T lineages were founded by two polypeptide encompassing haplogroups J and T provides gene amino acid substitution variants, ND1 nt

6 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

A 4216-ND1 Nonsynonymous (NS) mutation 15452A-cytb Synonymous (S) mutation Overlapping NS/S mutation J 10398-ND3 Overlapping NS/NS mutation 13708-ND5 4917-ND2 T RNA mutation Noncoding mutation Pathological mutation J1 J2 15257-cytb T1

14798-cytb

B Hplgr Gene npΔ CI% Function T ND2 4917A 90 ? J1 Cytb 14798C 79 Qi J2 Cytb 15257A 95 Qo

Figure 3. (A) Phylogeny of the haplogroups J and T demonstrating that each new branch of the mtDNA phylogeny is founded by a functionally signiﬁcant polypeptide variant that is subsequently transmitted to all downstream descendants. Key internal replacement mutations are designated by the gene name and the nucleotide substitution. (B) The table provides function information and interspeciﬁc sequence conservation (conservation index ¼ CI) for selected polymorphic amino acid sites. (From Ruiz-Pesini et al. 2004; reproduced, with permission, from the author and the American Association for the Advancement of Science # 2004.)

4216T.C (Y304H) and cytb nt 15452C.A an amino acid conserved to Caenorhabditis ele- (L236I). The lineages then subdivided into the gans (CI ¼ 79%), and the J2 cytb 15257 variant T and J haplogroups, the T haplogroup being alters an amino acid that is conserved down to founded by an ND2 amino acid substitution bacteria (CI ¼ 95%). Thus, mtDNA polymor- at nt 4917A.G (N150D) and haplogroup J phic variants have accumulated within our founded by the reversion of the out-of-Africa species that change amino acids highly con- ND3 10398A.G (T114A) variant and the ac- served throughout evolution. This phenome- quisition of a new ND5 variant at nt 13708G.A non, in which conserved amino acids among (A458T). Haplogroup J then subdivided into species are polymorphic within a species, is in- two lineages: J1 founded by a 16S rRNA consistent with classical neo-Darwinian theory. 3010G.A variant followed by cytb variant at However, it can be explained through mito- nt 14798T.C (F18 L) and J2 founded by a chondrial physiology and the high mtDNA se- cytb variant at nt 15257G.A (D171N). Each quence evolution rate. Once a species arises and of the founding polypeptide substitutions begins to expand its range, it encounters envi- changes an evolutionarily highly conserved ronmental variation that favors bioenergetic al- amino acid. The haplogroup T nt 4917 variant terations in central OXPHOS functions. These changes an amino acid that has been conserved demands are met by the high mutation rate from the first metazoans to man, having an in- of the energetically important genes of the terspecific conservation index (CI) ¼ 90%. The mtDNA. However, when a new species arises, subhaplogroup J1 cytb 14798 variant changes it likely may require a more efficient mitochon-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 7 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

drial bioenergetic system. The high mtDNA Somatic mtDNA Mutations. Finally, addi- mutation rate accommodates this by reverting tional clinically relevant mtDNA mutations ac- the regional variants back to the more efficient cumulate over time in tissues. These somatic and universal energy production system (Wal- mtDNA mutations arise in tissues as well as in lace 2013a). stem cell lineages with age and progressively That mtDNA variants alter mitochondrial erode mitochondrial function, generating the function has been confirmed by OXPHOS anal- aging clock (Wallace 2005). De novo mtDNA ysis of European haplogroups H and Uk (Rol- mutations can accumulate at anytime through- lins et al. 2009; Gomez-Duran et al. 2010) and out life from the oocyte to the cells of the elderly. Asian haplogroups B and F (Ji et al. 2012). The earlier in development the mutation oc- The strict maternal inheritance of the curs, the more broadly it will be distributed. mtDNA means that mtDNAs can never mix Hence, mtDNA mutations that arise in the em- and thus do not recombine. Hence, mtDNA bryonic period can be dispersed throughout the single-nucleotide variants that have accumulat- body, while those that arise in an adult organ ed throughout human history remain in total will be tissue specific (Holt et al. 1988; Coskun linkage disequilibrium. The significance of an et al. 2010). mtDNA variant is strongly influenced by the preexisting mtDNA variants on which it arose. nDNA This is particularly clearly demonstrated for the mtDNA variant in ND1 at nt 3394C, which Mitochondrial diseases can also result from mu- causes amino acid substitution Y30H (Ji et al. tations in any one of the hundreds of nDNA- 2012). When this variant arises on macroha- coded mitochondrial genes. Most of the .200 plogroup N mtDNAs, it reduces mitochondrial pathogenic nDNA mitochondrial mutations complex I activity by 15%–28% and markedly that have been reported to date are highly dele- increases the penetrance of the milder LHON terious and result in severe childhood disease mutations (Brown et al. 1995; Liang et al. (Koopman et al. 2012; Wallace et al. 2013). 2009). However, if the mutation arises on a macrohaplogroup M mtDNA and this mtDNA re- nDNA–mtDNA Interaction sides in high altitude, then this same variant is associated with maximum complex I activity Mild nDNA mitochondrial gene variants can and adaptation to high altitude (Ji et al. 2012). also become clinically relevant when combined Consistent with their functional impor- with an incompatible mtDNA. Severe encepha- tance, mtDNA haplogroups have been correlat- lomyopathy associated with a complete com- ed with predisposition to a wide range of plex I deficiency was observed in the boys of metabolic and degenerative diseases, various one family. Genetic analysis revealed that their cancers, and longevity (Khusnutdinova et al. disease was the result of inheriting from their 2008; Wallace 2008; Gomez-Duran et al. 2010; mother an X-linked NDUFA1 gene mutation Wallace et al. 2013). For example, Asian hap- (G32R) that caused a 30% reduction in complex logroup F, which has been correlated with pre- I activity. This occurred in the context of inher- dilection to diabetes and obesity (Fuku et al. iting their mother’s mtDNA, which harbored 2007), is associated with a 30% lower complex two additional complex I gene mutations, ND1 I activity relative to other macrohaplogroup N (M21T) and ND5 (M88T), and which also mtDNAs (Ji et al. 2012). In addition to meta- caused a 25% deficiency in complex I activity. bolic and degenerative diseases, mtDNA hap- In the mother, the NDUFA1 G32R mutation logroups have been associated with the severity was masked by her normal X-chromosome of sepsis (Baudouin et al. 2005) and the out- gene. However, in her son, the hemizygous come of ischemia, strokes (Chinnery et al. NDUFA1 G32R was unmasked and interacted 2010), and trauma (Gomez et al. 2009; Zhang directly with the mother’s mtDNA ND1 et al. 2010a; Krysko et al. 2011). (M21T) and ND5 (M88T) mtDNA variants to

8 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

cause severe complex I deficiency and disease from genetic, epigenetic, and environmental (Potluri et al. 2009). Similarly, the severity of factors. Alterations in nDNA-coded mitochon- the cardiomyopathy of members of a 13-gener- drial genes could impair energy metabolism by ation Mennonite pedigree whose members are inactivating an OXPHOS polypeptide, perturb- homozygous for a frameshift mutation in the ing antioxidant defenses, altering mtDNA rep- heart muscle ANT1 isoform gene was found to lication and repair, or affecting mitochondrial be determined by their mtDNA haplogroup in- quality control through alterations in mito- herited from their mothers. Homozygous chondrial fission and fusion or in mitophagy ANT1 mutant individuals that harbored hap- (Chen et al. 2010; Youle and van der Bliek logroup H mtDNAs had mild cardiomyopathy, 2012; Jokinen and Battersby 2013). Mitochon- while those who harbored haplogroup U drial OXPHOS could also be perturbed by mod- mtDNAs presented with severe cardiomyopathy ulation of the expression of the nDNA-coded leading to heart failure (Strauss et al. 2013). mitochondrial genes through variation in the epigenome (Wallace and Fan 2010). Mitochondrial function could also be per- Mitochondrial Pathophysiology of Complex turbed by mtDNA variation, either by recent Diseases deleterious mtDNA mutations or ancient The complexity of the genetics and pathophys- adaptive mtDNA polymorphisms. Finally, mi- iology of common diseases can now be reinter- tochondrial energy production could be per- preted in the context of mitochondrial bioener- turbed by the nature and availability of calo- getic and genetic principles (Fig. 4) (Wallace ries; the demands made on cellular energy for 2011, 2013b). Assuming common diseases re- growth, maintenance, and reproduction; and sult from partial mitochondrial deficiencies, the acute sensitivity of mitochondrial OXPHOS this could then perturb an array of physiological to a broad range of environmental toxins. As processes including energy production, REDOX mitochondrial energetics declines, the organs state, Ca2þ homeostasis, ROS production, ana- with the highest energy requirements would be bolic and catabolic metabolic pathways, etc. the first to show functional alterations. Evenvery (Wallace et al. 2010). Alteration in mitochondri- subtle bioenergetic defects can adversely affect al bioenergetics would increase mtDNA damage the central nervous system with its high mito- and mutation rate, perturb mtDNA replication chondrial energetic demand. Other sensitive and mitophagy, and lead to the accumulation of organs include the heart, muscle, and kidney. somatic mtDNA mutations. This would result in Alterations in the nature of available cal- the progressive decline in mitochondrial func- ories (carbohydrates, fats, proteins) would be tion with age. Cells that are sufficiently energet- differentially metabolized by individuals with ically impaired would malfunction and ulti- different mtDNA haplogroups. Hence, mito- mately undergo apoptosis and necrosis. Thus, chondrial alterations that perturb the flux of the accumulation of somatic mtDNA mutations calories through the system could result in met- becomes the aging clock in individuals born abolic diseases such as diabetes, obesity, hyper- with a normal mitochondrial function. For tension, and cardiovascular disease. individuals born with partial mitochondrial The mitochondria are also the most com- dysfunction, the accumulation of mtDNA mu- mon bacterium in the human body, our bodies tations and mitochondrial damage could ac- harboring on the order of 1017 mitochondria. count for the delayed onset and progressive Hence, damage to cells can release into the ex- course of their diseases. The stochastic nature tracellular space and bloodstream mitochon- of this process could also explain variable ex- drial N-formylmethionine-bearing polypep- pressivity and/or penetrance of disease. tides, mtDNA fragments, cardiolipin, and vari- Perturbation of mitochondrial bioenerge- ous other mitochondrial breakdown products, tics can predispose to a wide range of “complex” knownasdamage-associatedmolecularpatterns diseases. Bioenergetic perturbations can result (DAMPs), which can elicit an inflammatory re-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 9 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

A mitochondrial etiology of complex disease mtDNA variants Ancient adaptive polymorphisms Environmental factors Recent deleterious mutations Energy sources nDNA variation Carbohydrates, Mutations fats, amino acids Deleterious mutations, Energy uses OXPHOS mitochondrial gene polymorphisms Growth, dysfunction Epigenomics Histone modifications, maintenance, ↓energy, ↑ROS, signal transduction, reproduction Δ REDOX, Δ Ca2+ Toxins REDOX controls

mtDNA damage and Neuropsychological Metabolic somatic mutations Blindness, deafness, Type II diabetes, obesity, AD, PD, depression, hypertension, CVD Progressive muscle myalgia, Stress bioenergetic fatigability Thermal, trauma decline Cardiomyopathy Renal failure Apoptosis Inflammation, immunity MS, type I diabetes Cancer (DAMPs) Aging Energy production, Infection predisposition ROS, REDOX Sepsis, AIDS Penetrance and expressivity, delayed onset, progression

Figure 4. Integrated mitochondrial paradigm to explain the genetic and phenotypic complexities of metabolic and degenerative disease, aging, and cancer. The top three arrows show the three types of variation that impact on individual mitochondrial OXPHOS robustness and hence risk for developing disease symptoms. These include nuclear DNA (nDNA) variation encompassing DNA sequence changes and epigenomic modification of gene regulation and signal transduction pathways, mitochondrial DNA (mtDNA) variation including recent deleterious mutations and ancient adaptive polymorphisms, and environmental influences encompassing the availability and demand for calories and inhibition of mitochondrial function by environmental insults. The central oval encompasses the pathophysiological basis of disease processes and the basis of disease progression. The primary defect is reduction in the energy-transformation capacity of OXPHOS. This can result in reduced energy output, increased reactive oxygen species (ROS) production, altered REDOX status, and altered calcium homeostasis. The decline in OXPHOS efficiency can, in turn, perturb mitochondrial biogenesis, increase ROS production, impair mitophagy, etc., resulting in progressive increase in mtDNA damage and somatic mutations and further decline in mitochondrial function. Once mitochondrial function falls below the bioenergetic threshold of a tissue, symptoms ensue. Continued energetic failure can initiate cell destruction by apoptosis or necrosis. The lower five arrows summarize the disease categories and the phenotypic outcomes of perturbed mitochondrial energy transformation. The bottom arrow shows the effect of the stochastic accumulation of somatic mtDNA mutations resulting in delayed onset and a progressive course of diseases and aging. The right arrow indicates clinical problems that can result from reduced energy production in the most energetic tissues: the brain, heart, muscle, and kidney. The number and severity of symptoms in these organs reflect the degree and specific nature of the mitochondrial defect. The left arrow indicates the metabolic effects of mitochondrial dysfunction, which result in the perturbation of the body’s energy balance. This results in the symptoms of the metabolic syndrome. The lower right arrow indicates that mitochondrial alterations are critical for cancer initiation, promotion, and metastasis. The lower left arrow outlines the hypothesized inflammatory and autoimmune responses that may result from the chronic introduction of the mitochondria’s bacteria-like DNA and N-formylmethionine proteins into the bloodstream (Wallace 2011).

sponse (Zhang et al. 2010a,b; Krysko et al. 2011; chondrial disease would foster the release of mi- Oka et al. 2012). Because apoptosis, which de- tochondrial antigens. This, in turn, would result stroys the mitochondria before they are released in local inflammation in degenerative diseases into the bloodstream, is an energy-dependent and systemic inflammation in autoimmune process, the chronic energy deficiency of mito- diseases.

10 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

Finally, the growth of cancer cells is directly this mutation is high, it can present as lethal limited by energetics. Hence, cancer cells must childhood Leigh syndrome, MELAS, chronic acquire modifications in their mitochondrial progressive external ophthalmoplegia (CPEO), physiology to optimize energy production to cardiomyopathy, migraines, diabetes mellitus, their changing environments (Wallace 2012). and deafness. In pedigrees with high heteroplasmy members, meiotic segregation of the SEGREGATION OF HETEROPLASMIC mutant mtDNAs can result in the full range of GERMLINE mtDNA MUTATIONS phenotypes from asymptomatic to lethal disease (see example in Fig. 1 in Brown et al. While the causes of mitochondrial dysfunction 2001). Yet in other pedigrees, when the 3243A can be genetically complex because of the inter- .G mutation is present in 10%–30% hetero- action of the large number of nDNA and plasmy, this same mutation results in only ma- mtDNA genes and variants, the most unpredict- ternally inherited diabetes and deafness (see ex- able aspect of mitochondrial genetics is the seg- ample in Fig. 1 in van den Ouweland et al. 1992). regation of mtDNA heteroplasmy. Heteroplas- Variability in clinical presentation as a result mic mutants segregate along both the female of heteroplasmic variation can also be observed germline and in somatic tissues. Therefore, un- in mtDNA polypeptide missense mutation ped- derstanding the quantitative genetics of mtDNA igrees. The mtDNA ATP6 nt 8993T.G muta- segregation is essential if we are to understand tion, which causes the amino acid substitution the mitochondrial etiology of complex diseases. L156R, is the most common example (Holt et al. 1990). This mutation was originally asso- Germline Segregation of mtDNA ciated with the clinical designation of neuro- Heteroplasmy genic muscle weakness, ataxia, and petinitus pigmentosum (NARP), but also can cause le- Human mtDNA Disease Segregation thal childhood Leigh syndrome, olivopontocer- Familial Transmission of Heteroplasmic ebellar atrophy, cerebellar ataxia, and/or retini- mtDNA Mutations. Maternal transmission of tis pigmentosa when present in 70% to 100% heteroplasmic mtDNA mutations is now well of the mtDNAs (Tatuch et al. 1992; Ortiz et al. documented for both tRNA and polypeptide 1993). mutations (Wallace et al. 2013). The first exam- However, heteroplasmic variation is not the ple of the interaction between mtDNA hetero- only cause of clinical variation in mtDNA dis- plasmy variation and phenotype was the report ease. In Leber hereditary optic neuropathy, ped- that the tRNALys nt 8344A.G (Fig. 5A) muta- igrees that are essentially homoplasmic for one tion that causes myoclonic epilepsy and ragged of the common causal LHON mtDNA muta- red fiber (MERRF) disease was heteroplasmic tions, males are three to four times more likely (Wallace et al. 1988b; Shoffner et al. 1990). In to manifest mid-life subacute blindness than this initial family, the phenotypic variability females. Other factors also affect the onset of ranged from severe MERRF (III-1) to mild mi- blindness including mtDNA haplogroup back- tochondrial myopathy and electrophysiological ground and environmental stressors such as aberrations (II-4, III-2, 3, and 4) (Fig. 5B). The smoking and alcohol abuse (Brown et al. 1997, severity of the clinical phenotypes varied with 2002; Torroni et al. 1997; Sadun et al. 2003, the percentage of mutant mtDNAs after the in- 2011). dividuals were stratified by age (Fig. 5C). Variability of mtDNA Heteroplasmy in Mater- The effect of mtDNA mutant heteroplasmy nal Oocytes. One reason for the high variabil- on phenotype is even more striking with the ity observed in the mtDNA heteroplasmy levels tRNALeu(UUR) nt 3243A.G mutation associat- of maternal relatives is variation in the percent- ed with mitochondrial encephalomyopathy, lac- age of mutant mtDNAs in the oocytes of tic acidosis, and stroke-like episodes (MELAS) heteroplasmic women. Variability in oocyte het- (Goto et al. 1990). When the heteroplasmy of eroplasmy has been most intensively investigat-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 11 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

3′ OH A - B A 5′ P CG-- I 1 AT- CG- TA- II 123 4 GC- TA- III 1 2 34567 AT- A C A C A T TCTC A

A - - - - - A A TCG AAG AG C VER 3+ 2+ 2+ + + + + + + – ND ND

C - - - - A A C A TAGC T T T EEG 5+ 4+ + 4+ 4+ + 2+ 3+ 2+ –NDND A A T - A G A G TA- Mito. myop. B 3+ 2+ 2+ 2+ 2+ + + + – – – – DHU loop A - T TψC loop AT- Deafness C +++++––+–––– -C G C A ME D 2+ + – – – – – – – – – – T A TTT Anticodon Dementia +––––––––––– E AAA +––––––––––– AAG Codon Hypovent

C I 1 II CL 1 23 4 A BC III 1 2 3456 %wt 0 2 22 16 6 6 27 3 4 4 10 15 100 100 100 %mt 100 98 78 84 94 94 73 97 96 96 90 85 0 0 0

88 bp wt

48 mt

mt 40

Figure 5. MERRF tRNALys A8344G pedigree showing variable clinical expression in association with variable mtDNA mutant heteroplasmy modified by age. (A) Structure of tRNALys showing position of A8344G transition in TCC loop. (B) Pedigree of proband (III-1) showing that all maternal relatives had some clinical manifestations (filled symbols), though highly variable, while the three paternal sons were symptom free. VER, visual evoked response; EEG, electroencephalograph; Mito. myop., mitochondrial myopathy with ragged red fibers and abnormal mitochondria; deafness, sensory neural hearing loss; ME, myoclonic epilepsy; dementia, progressive cognitive decline; hypovent, hypoventilation. (C) Pedigree showing variable proportions of mutant- type (mt) and wild-type (wt) mtDNA along the maternal pedigree. A 183-nt PCR fragment was digested with CviJI. The wild-type (8344A) gave an 88-nt uncut fragment, whereas the mutant (8344G) created a new site cutting the 88-nt fragment into 48 and 40 nt fragments. “CL” is a cloned mutant fragment. Cases (A), (B), and (C) are independent pedigrees. Individual (C) is the maternal aunt of proband III-1 in (A), which manifested MERRF.All of the maternal relatives of the pedigree are heteroplasmic for the mutant mtDNA and the severity of the phenotype correlated with the percentage of heteroplasmy when corrected for age. (From Wallace et al. 1988b; reproduced, with permission, from the author and from Shoffner et al. 1990; reproduced, with permission, from the author and the National Academy of Sciences # 1990.)

12 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

ed in women harboring the tRNALeu(UUR) able mutant ,2%. Among the 35 embryos 3243A.G and ATP6 8993T.G mutations. A from the 27% heteroplasmy woman, 83% of 41-year-old mother who harbored the 3243A the embryos harbored the 3243A.G mtDNA, .G mutation in 18.1% of her quadriceps mus- with a heteroplasmy range of 5% to 77%, but cle and 7.2% of her leukocyte mtDNAs had two none of her embryos were pure mutant. sons, one of whom was found at 15 years of age Germline transmission of the heteroplasmy to harbor 11.7% 3243A.G mutant in his blood mtDNA ATP6 8993T.G gene missense muta- cells. The woman underwent a hysterectomy for tion has also been found to result in oocytes endometriosis and 82 oocytes were recovered with widely different hetroplasmy levels. In and analyzed for their 3243A.G mutation lev- one case, an asymptomatic mother with 50% els, which ranged from 0% to 45% with a mean blood mutant mtDNAs had three boys; one heteroplasmy of 12.6% and a median of 8.2%. died of Leigh syndrome with 98% mutant Eight of the oocytes (9.8%) lacked detectable mtDNAs in muscle and ﬁbroblasts, one died mutant mtDNA and 35 of the oocytes had of sudden infant death syndrome (SIDS) with heteroplasmy levels ranging from 1% to 45%. 92% mutant in blood, and one was affected with To model the distribution of oocyte genotypes, Leigh syndrome and harbored 87% mutant in the researchers assumed that the distribution of blood. The mother was superovulated and seven oocyte heteroplasmylevels wouldapproximate a oocytes could be recovered and genotyped. One binomial distribution, with the initial maternal of the oocytes had no detectable mutant mtDNA mutant allele frequency ( p0) being rep- mtDNA, whereas the remaining six oocytes resented by the mean of the allele frequencies of had .95% mutant (Blok et al. 1997). the oocytes. Furthermore, it was hypothesized Low-Level Maternal Germline Hetero- that the extent of the variation in mtDNA het- plasmy. While the maternal transmission of eroplasmy levels was determined by a “bottle- biallelic heteroplasmy of mtDNA disease has neck.” This bottleneck was envisioned to reduce been extensively studied, the advent of next- the number of mtDNA segregating units “N” generation sequencing (NGS) is now provid- within the maternal germline over “g” germline ing the capacity to determine whether the ma- cell divisions, to be followed by expansion of the ternal germline might also harbor a spectrum mtDNA copy number back to an inﬁnite size. of mtDNAvariants each at a very low percentage The variance (V ) was calculated by the formula of heteroplasmy. This is possible because NGS permits sequencing a region of the human g V ¼ p0ð1 p0Þf1 ½1 ð1=NÞ g: mtDNA from a human sample more than a thousand times revealing rare variants. When From the mean oocyte heteroplasmy level of the mtDNA sequence of the control region hy- 12.6%, p0 ¼ 0.126, and with a variance of V ¼ pervariable region 2 (MT-HV2 from nt 162– 0.0143, the authors estimated that the number 455) and the COIII region (MT-CO3 from nt of segregating units of the bottleneck (N) would 9307–9591) were analyzed using the Roche be 173 if maintained over 24 germline cell divi- 454 sequencing platform from blood and skel- sions or eight if the bottleneck lasted for one cell etal muscle samples of seven subjects, every generation (Brown et al. 2001). subject was reported to harbor heteroplasmic Studies on the germline segregation of het- variants in one or more bases at .0.2% hetero- eroplasmic 3243A.G mutant mtDNAs was ex- plasmy in both tissues. Overall, the number of tended to 38 preimplantation embryos from variants per base position was greater in skeletal two women. The oral mucosa heteroplasmy of muscle than in blood and also was greater in one woman was 30% and the mean hetero- MT-HV2 than in MT-CO3. Heteroplasmy lev- plasmy level of her embryos was 30% + 15%. els .2% were also observed, but only in the The mucosal heteroplasmy of the second wom- muscle of three subjects within the MT-HV2; an was 27% and the mean of her embryos was all other heteroplasmy levels were low (Payne 32% + 23%. Six of the embryos had no detect- et al. 2013).

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 13 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

Patients with mtDNA polymerase g muta- 2012). Given a mutant density of 3.5 1025, tions, which decreased the fidelity of mtDNA this tissue had about one mtDNA mutation per replication, had elevated MT-HV2 mutations every two mtDNA molecules. Therefore, there is and these increased in frequency with age. considerable genetic heterogeneity within the When first-degree relatives of patients with mu- thousands of mtDNAs within a somatic cell. tations in the nDNA genes required for mtDNA The frequency and heteroplasmy levels of po- maintenance were analyzed, 40% of the variants tentially maternally transmitted low-hetero- in a given individual were shared with their ma- plasmy mutations merit further examination. ternal relatives, as compared to only 12% of the Bovine. Proof that low-heteroplasmy vari- variants shared with unrelated individuals. Sev- ants can be transmitted through the maternal enty-one percent of the shared variants among germline would be if mtDNAs harboring one relatives were primarily found in skeletal muscle of two variant alleles were to alternately appear as opposed to 13% shared muscle variants in in successive generations. Such mtDNA allelic unrelated individuals. Therefore, according to switching across maternal generations has been this study, low-level heteroplasmies (0.2%– reported for bovine lineages. A synonymous 2%) are present in all individuals and a signifi- sequence variant in the URF-5 (now ND5) cant proportion of these can be transmitted gene at nt 12792C.T, detected as an HaeIII through the maternal lineage (Payne et al. restriction fragment length polymorphism, 2013). was observed to switch from one allele to the Using the Illumina platform, quality control other within two maternal generations in a criteria for mtDNA sequence validation and a 1982 study (Hauswirth and Laipis 1982). This heteroplasmy cutoff of 2%, a much lower fre- variant was then linked to four mtDNA control quency of heteroplasmic mutations was report- region variants at nts 16074, 16079, 16231, and ed for blood and mucosal mtDNAs of nine 16250, generating four different haplotypes. individualsfromthreefamilies.Fourheteroplas- These also switched among generations over mic mtDNA variants were reported, with one an eight-generation bovine pedigree (Olivo apparent germline mutation. Of the remain- et al. 1983). While the observed haplotypes ap- ing variants one was prominent and two were peared to be homoplasmic in the animals stud- low-heteroplasmy variants (Goto et al. 2011). ied, three offspring from one cow were found While these studies suggest that low-hetero- to be heteroplasmic, suggesting that genotyp- plasmy variants are ubiquitous and can be ma- ic switching was occurring by the germline ternally inherited, the rise of NGS to detect very transmission of low-heteroplasmic genotypes low-level heteroplasmic mutant mtDNAs may (Ashley et al. 1989). be subject to artifacts. For example, most cur- A more extensive bovine survey of mtDNA rent protocols for making NGS libraries use variation revealed that a control region variant PCR and PCR polymerases are error prone that changed a G to a C at the end of a homo- and thus could generate spurious mutations. polymeric run of Gs switched between the G and One effort to overcome the potential of PCR the C allele in 13 different mother–daughter artifacts is “duplex sequencing.” This method pairs (Koehler et al. 1991). These early bovine requires that all sequence variants be confirmed observations were the first to lead to the hypoth- by identification of the complementary nucle- esis that there was an mtDNA copy-number otide change on both DNA strands. The esti- bottleneck in the mammalian female germline. mated error rate of the duplex sequencing Mouse. The Shoubridge laboratory has method was estimated to be 1/109. When this studied maternal germline segregation of approach was applied to a human brain sample, mtDNAs in heteroplasmic mice. In their system, the mtDNA mutation rate was found to be the mtDNAs of two different mouse lineages, 3.5 1025, lower than that expected for the NZB/BinJ (NZB) and BALB/cByJ (BALB), above studies. Still this is much higher than were combined by removing a bleb of cytoplasm reported nDNA mutation rates (Schmitt et al. from a one-cell embryo and fusing it to the one-

14 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

cell embryo of the other mtDNA strain. The het- contained 200 mtDNAs. Based on the as- eroplasmic embryos were then implanted into sumption that the average number of mtDNAs foster mothers at the two-cell stage. In the initial per PGC and oogonium remained relatively study, five founder females carrying 3.1% to constant throughout the PGC replication phase 7.1% of the donor mtDNA were studied. These and on the observed distribution of oocyte and were crossed with BALB males and the progeny offspring NZB/BALB mtDNA heteroplasmy analyzed. The mean mtDNA heteroplasmy lev- levels, the Shoubridge laboratory concluded els of the offspring were found to be similar to that there must be 185 (range 76–867) that of the founder mother, but the hetero- mtDNAs in an oogonium. The Shoubridge lab- plasmy levels of the individual offspring varied oratory concluded that the rapid segregation of with the highest heteroplasmy levels being the mtDNA heteroplasmy in mammals could be 29.6% in one pup (Jenuth et al. 1996). explained by a drastic reduction in the number In the mouse, the primary oocytes are of mtDNA segregation units, a bottleneck, oc- thought to be derived from 50 primordial curring in the PGCs of the female germline. germ cells (PGCs) located at the base of the Such a reduction in mtDNA segregating units allantois of a 7.5-d postcoitum (dpc) mouse would greatly increase the rate of genetic drift, embryo. These cells are alkaline phosphatase leading to rapid segregation of different mtDNA (ALP) positive, permitting them to be identified types in different female germline cells. They and isolated. The PGCs migrate to the germinal estimated that the number of segregating ridge where they grow and differentiate into oo- mtDNA units in PGCs was 200 and that the gonia. The oogonia then proliferate by mitosis multiple cell divisions of the oogonia required during embryonic development. In the mouse, to generate that large a number of primary oo- the oogonia undergo 15 divisions to generate cytes were sufficient to account for the observed 25,000 primary oocytes. In humans, about variance in heteroplasmy frequency of oocytes six to seven million primary oocytes are pro- and offspring. They summarized: “Our study duced through roughly 24 cell divisions. The suggests that the probability of inheriting one oogonia will either degenerate or differentiate of two mtDNA genotypes can be modeled as a into primary oocytes through asymmetric divi- binomial sampling process ...” Thus, “It seems sion, generating a primary oocyte and a daugh- unlikely that strong positive or negative selec- ter oogonium. By birth, most oogonia have ei- tion for pathogenic mtDNAs occurs in the oo- ther differentiated into primary oocytes or cyte in early embryogenesis ...” (Jenuth et al. degenerated. The primary oocytes undergo oo- 1996). In short, germline heteroplasmy segrega- genesis in which they enter meiosis and become tion is the product of the stochastic sampling arrested in prophase I where they remain until process known as genetic drift. puberty begins and individual proto-oocytes Eleven years after the initial Shoubridge complete differentiation, form follicles, and study reported an mtDNA PGC mtDNA bottle- can be ovulated. Within the follicle, the oocyte neck, Cao and associates published a paper re- completes the first meiotic division generating porting that the number of mtDNAs in PGCs the first polar body and enters the second mei- was not as low as surmised by Shoubridge. Us- otic division where it becomes arrested at meta- ing ALP staining to identify PGCs in embryos phase II. At fertilization, meiosis II is complet- between 7.5 and 13.5 dpc, Cao and associates ed, the second polar body is extruded and the determined by quantitative real-time polymer- female and male pronuclei approach each other ase chain reaction (PCR)(qRT-PCR) that the and fuse. average mtDNA copy number was 1561 + 161 Based on ultrastructural analysis of mouse (range 1350–1732), and that the minimum oogonia, the Shoubridge laboratory estimated mtDNA copy number of the smallest PGCs that there were 40 mitochondria per oogo- was 953. They also estimated that there were nium and assuming five mtDNAs per mito- 100 mitochondria in a single PGC. Additional chondrion they concluded that an oogonium germ cell mtDNA copy-number estimates in-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 15 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

cluded PGCs at 13.5 dpc at 3.66 103, primary observations, these authors built a mathemati- oocytes at 1.16 103, and mature oocytes at cal model that encompassed both a severe 1.57 105 mtDNAs per cell. By contrast, quan- mtDNAcopy-number bottleneck in early mam- tification of the mtDNAs in the somatic cells of malian PGCs as well as a subsequent mtDNA 7.5-dpc embryos was reported as low, ranging amplification phase, the combination of the from 57 to 3345 mtDNAs percell. One complex- two being able to account for the rapid germline ity of these assessments discovered by this re- segregation of mtDNA heteroplasmy (Cree et al. search group was that staining embryo cells 2008). with ALP partially inhibited mtDNA quantifi- At the end of 2008, the Shoubridge labora- cation. As an alternative approach to identifying tory published another paper in which they germline cells, the authors used mice that ex- quantified the mtDNA copy number in the pressed the enhanced green fluorescent protein germ cells of NZB/BALB heteroplasmic mice. (EGFP) driven by the 18-kb Oct-4 promoter In addition, they determined the proportion of (GOF-18/GFP). Oct-4 is transcribed in germ the NZB and BALB mtDNAs in the germ cells at cells at d 9.5 to 13.5 dpc. Cao and colleagues different stages. Based on the concern that ALP then reported that the average mtDNA copy staining might result in spuriously low mtDNA number for PCGs isolated using GOT-18/GFF copy numbers, this team used EGFP transcribed was 1408 and 1294 in two experiments versus from the Oct4 (Pou5fl) promoter and then man- 673 and 736 for ALP-stained cells. Hence, they ually isolated the germ cells. They then quanti- corrected their ALP-stained PGC mtDNA esti- fied the mtDNA copy number in 8.5-dpc cells, mates by multiplying by 1.92. Based on these observing a mean mtDNA copy number of results, this group concluded that there was no 280 with a median of 145. By 10.5 dpc, they constriction of mtDNA content in PGCs, and found that the mtDNA copy number had in- thus that the rapid mtDNA germline hetero- creased to a mean of 2800 and median of plasmy segregation was not the product of a 2200. At 14.5 dpc when the PGCs have colo- physical bottleneck in the number of mtDNA nized the gonad, the mtDNA copy number within the PGCs (Cao et al. 2007). had risen to 6000 per cell. Thus, these data This report was followed a year later by a indicate that the mtDNA copy number decreas- report from the Chinnery laboratory reaffirm- es 700-fold from oocyte to PGC, but then in- ing that the mtDNA copy number in PGCs was creases 10- to 20-fold during expansion of the on the order of 203 at 7.5 dpc, but that the PCG population and the colonization of the mtDNA copy number in older PGCs increased gonad (Wai et al. 2008). While this was consis- to 1529 mtDNAs/cell by 14.5 dpc (Cree et al. tent with the observations of Cree et al. (2008), 2008). To isolate PCGs without ALP staining, the Shoubridge laboratory then analyzed the this group identified PGCs by the fluorescence proportion of NZB and BALB mtDNAs during of GFP transcribed from the Stella (Dppa3) pro- embryonic development. This led to the sur- moter, which is specific for PGCs. Their studies prising conclusion that the proportion of NZB revealed that the mouse oocyte contains 2.28 and BALB mtDNAs did not change markedly 105 mtDNAs, that the mtDNAs do not replicate during the mtDNA constriction and prolifera- until the PGCs are formed, that the median tion cycle of the PGCs and the oogonia. Hence, mtDNA copy number in 5.5-dpc PGCs is 203, the Shoubridge laboratory concluded that “de- and the mean is 451. However, by 13.5 dpc when spite the severe reduction in mtDNA copy num- the number of primary oocytes is 25,791, ber, the mitochondrial genetic bottleneck does the median mtDNA copy number is 1529. not occur during embryonic oogenesis.” How- Hence, the mtDNA copy number per cell in- ever, when they examined the mtDNA hetero- creases from d 5.5 to 13.5. They also observed plasmy variance in PGCs, oogonia, and primary that by 14.5 dpc, female germline cells had a oocytes in primordial follicles with ,10,000 lower mtDNA content then male germline cells: mtDNA per cell to that of postnatal mature ovu- 1376 + 601 versus 2152 + 951. From these lated oocytes and primary oocytes in secondary

16 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

follicles that harbor .10,000 mtDNAs per cell, the female germline. In a follow-up report, the they found that the mtDNA heteroplasmy var- Chinnery laboratory argued that the discrep- iance had increased significantly. They then hy- ancy in the conclusions of the two Shoubridge pothesized that the “...genetic bottleneck must laboratory studies was the result of the frequent- be the result of the selective replication of a ran- ly extreme bias in the relative levels of the NZB dom subset of mtDNA templates during the versus the BALB mtDNAs, one mtDNA type growth and maturation of the ovarian follicles,” always being present at a low percentage heter- which must start in the primordial follicles. oplasmy. This is the product of the hetero- To identify this differential replication phase, plasmy being derived from a small bleb of cyto- the researchers pulse-labeled female pups with plasm from one oocyte fused to a much larger bromo-deoxyuridine (BrdU) injected at d 1 cytoplasm of the recipient oocyte. Because a (P1) and (P4) postpartum. They then analyzed greater variance and thus range of heteroplasmy the mtDNA labeling of oocyte mtDNAs within levels can be generated from a starting mtDNA mtDNA nucleoids, the nucleoids identified by heteroplasmic ratio of 50:50 than can be gener- their association with mtDNA-binding protein ated from a starting ratio of 5:95, Chinnery ar- (Tfam), polymerase g (PLOG), and single- gued that the variance levels observed by Shou- strand binding protein (mt-SSB). This revealed bridge did not reflect the true extent of the that only a limited number of mtDNAs were heteroplasmy segregation (Samuels et al. 2010). replicating, replicating mtDNAs did not always The question of whether or not there was a correlate with Tfam reactivity, and the replicat- severe reduction in mtDNA copy number in ing mtDNAs were not consistently associated PGCs was again raised by Cao and collabora- with the Balbiani body. The Balbiani body is tors. They questioned the effectiveness of previ- an aggregate of Golgi elements surrounded by ous studies to identify true PGCs at 7.4 dpc. To mitochondrial and endoplasmic reticulum. The increase the reliability of studying only PGCs, Shoubridge group then concluded that the pur- Cao and associates identified the PGCs using pose of the reduced mtDNA copy number in three different protein markers and also distin- early PGCs to 200 mtDNAs is to permit selec- guished among PGCs isolated from early-bud tion against cells with high percentages of dele- (EB) and late-bud (LB) 7.5-d embryos. The first terious mtDNA mutations. However, “The ge- marker used was Blimp1, which is expressed in netic bottleneck for neutral (and less deleterious nascent PGCs and is PGCs specific. By intro- mtDNA sequence variants that are associated ducing into mice a bacterial artificial chro- with most human disease) occurs during folli- mosome harboring a monomeric red fluores- culogensis in early postnatal life ...” (Wai et al. cent protein gene (mRFP) expressed from the 2008). Blimp1 promoter (Blimp1-mRFP), they were Although this Wai-Shoubridge study did able to isolate positive cells manually. The ap- not correlate the proposed differential mtDNA propriate cells were further validated by immu- replication bottleneck with the Balbiani-body- nohistochenical staining for Stella (PGC7) and associated mitochondrial cloud, comparative by ALP staining. Quantification of the mtDNA interspecific studies have been used to argue copy number of the Blimp1-mRFP-positive EB that the Balbiani body is important for regional cells at 7.5 dpc gave a mean value of 1396 mtDNA proliferation (Zhou et al. 2010). mtDNAs/cell and LB cells at 7.5 dpc of 1479. The conclusion of the Wai-Shoubridge At 13.5 dpc, female germ cells gave a mean copy study that segregation did not occur during em- number of 1747 and male germ cells of 2039. bryonic oogenesis contradicted their previous Thus, Cao again concluded that mtDNA segre- genetic-based conclusions (Jenuth et al. 1996) gation does not occur because of a low mtDNA as well as the conclusions of the Chinnery group copy number in the early PGCs. Rather, segre- that a low mtDNA copy number in the early gation must occur later in female germ cell de- PGCs was one of two major factors in the rapid velopment without reduction in mtDNA copy segregation of the mtDNA heteroplasmy along number, presumably because of differential

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 17 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

replication of mtDNAs during oocyte matura- Primate. Although the progeny of a hetero- tion. Consistent with this conclusion, they plasmic female mice can have an array of heter- point out that mouse models heteroplasmic oplasmic ratios, the segregation rate does not for highly deleterious mtDNA mutations in- seem as extreme as has been observed in human cluding an mtDNA 4696-nt deletion (Sato pedigrees harboring the tRNALeu(UUR) nt 3243 et al. 2007) and an ND6 frameshift mutation A.G mutation or the ATP6 nt 8993T.G (Fan et al. 2008) produced pups with lower per- L126R mutations. To determine whether there centages of mutant mtDNAs in successive lit- aresignificant differences in mtDNA heteroplas- ters. If the mtDNAvariance were entirely deter- mic segregation between rodents and primates, mined by segregation and/or selection early mtDNA segregation was studied in cytoplasmic in female germline development, then all litters mixing experiments of rhesus macaque oocytes. should be generated from oocytes with the same The mtDNAs of macaque oocytes were distribution of heteroplasmy levels. Thus, the mixed by karyoplast-cytoplast fusion. A karyo- distribution of mtDNA heteroplasmy across fe- plast is a portion of the cytoplasm of a cell that male litters with advancing maternal age should contains the nucleus surrounded by the cellu- be the same. If the mutant mtDNAs were seg- lar plasma membrane. A cytoplast is a fragment regating later in oocyte maturation, then the of the cell that lacks the nucleus but contains sicker oocytes would be preferentially lost most of the cytoplasm, mitochondria, and from the ovary during the female reproductive mtDNAs. In this macaque experiment, a micro- lifespan, resulting in the successive decline in pipette was inserted under the zona pellucida of mutant mtDNAs over sequential litters. Why a metaphase stage-II embryo and 50% of the then might the PGCs retain high mtDNA copy cytoplasm plus the nucleus was extracted. This numbers while adjacent somatic cells have lower karyoplast was then inserted under the zona pel- mtDNA levels? These authors speculate that the lucida of anotheroocyte from which the nucleus high PGC mtDNA copy number permits the and half of the cytoplasm had been removed. retention of heteroplasmic mutations that can The two cell fragments were then fused to gen- be transmitted through the female germline and erate a “reconstituted cell,” which was fertilized permit subsequent adaptation to changing en- by intracytoplasmic sperm injection (ICSI). vironments (Cao et al. 2009). The karyoplast and cytoplast donors were Clearly, there is no consensus as to the cellu- derived from Indian and Chinese origin ma- lar and molecular mechanism of rapid germline caques, which differed in their mtDNA control mtDNA heteroplasmic segregation. It could re- region (D-loop) sequence, permitting the fate sult from the rapid segregation of heteroplasmic of the two mtDNA haplotypes to be monitored mtDNAs because of genetic drift resulting from through development. The mean heteroplasmy a physical bottleneck in the number of mtDNA in 15 analyzed reconstituted oocytes was 54.9% segregating units in the PGCs (Jenuth et al. 1996; + 10%. In two cell embryos, half of the embry- Cree et al. 2008), a combination of an initial os had 50% of each haplotype in each blasto- severe reduction in the mtDNA population mere, but the other half of the embryos exhibited size followed by further segregation during sub- significant differences among the blastomeres, sequent replication (Cree et al. 2008; Khrapko the most divergent case being 36% and 70%. 2008), an aggregation of a larger number of The divergence in percentage heteroplasmy in- mtDNAs into homogeneous segregating units creased in four- and eight-cell embryos, the co- such as multiple mtDNA containing nucleoids efficient of variance increasing from 17.7% to (Carling et al. 2011), the replication of only a 25.0% to 30.9% and the range increasing from small proportion of the mtDNAs in the primor- 13.3% to 25.3% to 43.2% in two-, four-, and dial follicle cells leading to biased transmission eight-cell embryos, respectively. By the eight- of a few mtDNAs (Wai et al. 2008; Carling et al. cell stage, some embryo blastomeres differed 2011; Jokinen and Battersby 2013), or to some by as much as 10% and 80% of the Indian and as-yet unidentified factors. Chinese mtDNAs (Lee et al. 2012).

18 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

Three embryonic stem cell (ESC) lines were Selection in mtDNA Germline Segregation derived from the inner cell mass cells of heteroplasmic blastocytes and found to harbor All of the above studies demonstrate that heteroplasmy levels of the cytoplast mtDNA of an intrinsic feature of heteroplasmic mtDNA 97.9%, 93%, and 5%. From nine clones of the mutations is their “imperfect transmission” 93% cell line, six were homoplasmic for the (Chapman et al. 1982), which can be greatly cytoplast mtDNA and three were heteroplasmic accentuated along the maternal lineage by bot- in the range of 90.7% to 92.9%, with the direc- tlenecks in the mtDNA copy number that fos- tion of segregation being independent of the ter intracellular heteroplasmic mtDNA genetic nuclear origin (Lee et al. 2012). drift (Hauswirth and Laipis 1982; Charlesworth Reconstituted oocyte-derived embryos were and Charlesworth 2010). Genetic drift is driven implanted in females and two fetuses were an- by ﬂuctuations in gene pool size and, as the alyzed for their heteroplasmy levels. The male constriction of the gene pool size increases, the fetus harbored 26.3% of the cytoplast mtDNA, segregation rate of a heterogeneous pool of ge- while the female fetus harbored 93.8% of the netic variants toward homogeneity also increas- cytoplast mtDNA, with her tissue levels ranges (Takahata and Slatkin 1983). The rapid seg- ing from 91.1% in blood to 98.4% in kid- regation of heteroplasmic mtDNA genotypes neys. Recovery of the ovaries from the female along the female germline by an extreme “bot- fetus and analysis of 51 primordial oocytes tleneck” counters the accumulation of large revealed a continuous range of cytoplast donor numbers of mtDNA mutations over many gen- mtDNAs from 3.7% to 99.2%. Hence, the erations, which would progressively erode en- somatic cell heteroplasmy level of the female fe- ergetic function and ultimately compromise tus was essentially independent of the level of the viability of the species, a process known as heteroplasmy in her germline cells (Lee et al. “Muller’s ratchet.” In a sense, rapid mtDNA al- 2012). lelic drift acts like nuclear gene recombination A difference in heteroplasmy segregation in in that it increases diversity while limiting the somatic versus germline cells was already appar- accumulation of deleterious mutations. By rap- ent in the epiblast cell lineages, resulting in idly sorting out heteroplasmic alleles along the marked asymmetric segregation of the mtDNAs maternal lineage, mtDNA genetic drift permits into the somatic tissues, even though the female the introduction of new mtDNA mutations germline remained capable of generating the full into the population in the near homoplasmic range of possible heteroplasmic levels. Hence, state. These genotypes can then be acted on by there appears to be two mtDNA segregation sys- selection with deleterious alleles being removed tems, one for somatic tissues that tends to move by purifying selection thus maintaining the in- rapidly toward homoplasmyand is already func- tegrity of the maternal germline. Nonpathogen- tioning in the early embryo and the other that ic variants would accumulate along radiating is acting in the female germline and generates a maternal lineages at random, and rare advan- diverse array of oocyte heteroplasmy levels (Lee tageous alleles could become enriched within et al. 2012). regional bioenergetic environments to found This difference between primate germline regional haplogroups. While severe mtDNA and somatic cell lineage heteroplasmy levels is bottlenecks will increase the probability that a reminiscent of the observation of Cao et al. deleterious allele will be presented, resulting in (2007, 2009) who reported that the mtDNA genetic disease and in the rapid demise of that copy number in the somatic cells of mouse em- lineage (Bergstrom and Pritchard 1998), a severe bryos was markedly lower than that of the fe- mtDNA germline bottleneck would also in- male germline cells. Such a somatic lineage crease the presentation of rare beneﬁcial alleles, mtDNA bottleneck would be conducive to the thus increasing species diversity and adaptabil- rapid segregation of mtDNA heteroplasmy in ity to changing environments. Thus, in contrast somatic cells. to nuclear genetics, in mtDNA genetics, genetic

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 19 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

drift and selection are not two mutually exclu- ate deletions by intramolecular recombination sive or competing forces. Rather, they work to- within postmitotic tissues (Wallace and Fan gether to inhibit the accumulation of deleterious 2009). Hence, it is possible that the maternal mutations and facilitate the fixation of advanta- transmission of the 4696-nt mtDNA deletion geous ones. is the product of maternal transmission of a If we are to have a complete understanding duplication, although this has not been deter- of origin and biological importance of mtDNA mined. genetic diversity, we need to understand the role Regardless of the molecular nature of the that selection plays in shaping mtDNA genetic maternally transmitted rearranged mtDNA, diversity. Todate, only a few mouse models have analysis of the level of the deleted mtDNA in been generated and studied that provide insight pups derived from heteroplasmic female mice into the importance of selection in sorting out revealed a striking directional loss of the rear- mtDNA heteroplasmic mtDNA mutations. ranged mtDNA in successive litters. Among Mouse Germline Segregation of an mtDNA five females with tail deletion levels between Deletion. In one mouse model system, an 20% and 60%, the mean deletion heteroplasmy mtDNA deletion was introduced into the mouse of all of the pups declined precipitously. By the female germline via somatic cell cytoplasmic third litter, all of the pups had less than 10% hybrid (cybrid) fusions (Bunn et al. 1974; Wal- deleted mtDNA, with many having no detect- lace et al. 1975). A rearranged mtDNA was re- able deleted mtDNA in their tail tissue. Interest- covered from mouse brain by fusion of synap- ingly, in two of the females, the mean mtDNA tosomes, neuronal synaptic bouton fragments heteroplasmy of the pups of the first litter was containing mitochondria and mtDNAs, to cul- higher than that of the tail of the mother. The tured cells lacking mtDNA (r0) (King and At- same striking decline in the percentage of delet- tardi 1989; Chomyn 1996). Because r0 cells can- ed mtDNA was seen in the oocytes of two fe- not grow without uridine and pyruvate, r0 cells males that were superovulated at d 44 and 117. that acquire partially functional mitochondria In both cases, the oocyte heteroplasmy level in can be selected by growth in media lacking the first superovulation oocytes was less than either pyruvate or uridine. The resulting synap- that of the females and the reduction of het- tosome cybrids were screened for mtDNA mu- eroplasmy in the oocytes in the subsequent su- tations and one clone was identified that har- perovulation was even more marked. The same bored a 4696-nt mtDNA deletion that removed reduction in mtDNA deletion levels was ob- six tRNAs and seven structural genes. served in the litters of mice in which the ovar- Next, enucleated cell cytoplasts from this ies of heteroplasmic deletion mice were grafted mtDNA 4696-nt deletion cell line were fused into ovarectimized C57BL/6 mice harboring to single-cell mouse embryos, the embryos im- Mus spretus mtDNAs (Sato et al. 2007). Appar- planted into foster mothers, and the resulting ently, then, oocytes with high mtDNA deletion mice found to be heteroplasmic for the deleted levels can be ovulated while the mouse is young, mtDNA. Unlike human mtDNA deletion mu- but, as the female mouse ages, the high deletion tations, which are not generally maternally in- oocytes are progressively lost. Hence, there must herited, this deletion was transmitted through be selection against the formation or ovulation the female germline (Inoue et al. 2000). The of the oocytes with highest deletion levels heteroplasmic mice have COX-negative fibers throughout the female mouse’s lifespan. in their heart and muscle, renal dysfunction, Mouse Germline Segregation of a Frameshift and male mice with 70% rearranged mtDNAs mtDNA Mutation. An analogous finding has are infertile (Inoue et al. 2000; Nakada et al. been reported for a mouse line that harbored a 2006). heteroplasmic mtDNA ND6 frameshift mtDNA In humans, it has been observed that mutation. This mutation was the result of an mtDNA duplications can be maternally inher- insertion of a C at nt 13885 within the mtDNA ited and that these duplicated molecules gener- (ND6 13885insC). The original mtDNA frame-

20 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

shift mutation was isolated in mouse L cells duced by the mouse polymerase g (D257A linked to an mtDNA homoplasmic COI PolgA exo2) mutator gene. This revealed that T6589C missense mutation. When this cell while deleterious tRNA mutations were in- line was enucleated and fused to a female mouse troduced into the germline, there was a dearth embryonic stem cell (mfESC), one of the cybrid of deleterious polypeptide gene mutations clones was found to harbor mtDNAs that had (Stewart et al. 2008). reversed the ND6 13885insC mutation by de- Germline Segregation of Single-Base tRNA letion of an adjacent T, thus restoring the read- Deletion Mutations. Using the mtDNA poly- ing frame. Hence, this mfESC cybrid was homo- merase g (D257A PolgA exo2) mutator gene plasmic for the COI T6589C missense mutation to generate germline mtDNA mutations, a but heteroplasmic for the ND6 13885insC mouse was isolated that was heteroplasmic for frameshift mutation. a tRNAMet nt 3875 C deletion (tRNAMet 3875 Injection of this mfESC cybrid into mouse delC) mutation. The inheritance of the mtDNA blastocysts resulted in the generation of one fe- harboring this mutation was monitored and male mouse that was heteroplasmic for the ND6 found to be biased toward loss of the tRNA 13885insC frameshift mutation at 44% + 3% mutation. The tRNAMet 3875delC mutation in- in all of her tissues, with the percentage frame- hibits aminoacylation, but the biochemical de- shift mutation being 47% in her tail. When this fect was partially ameliorated by the up-regula- female was mated, all of the pups of her first and tion of the other mtDNA tRNAs and all of the second litter had tail mtDNAs that were 14% mtDNA mRNAs except for that of ND6 (Freyer frameshift, in the third litter all of the pups et al. 2012). had 6% frameshift mtDNA, and in the fourth, Mice harboring the tRNAMet 3875delC mu- fifth, and sixth litters the frameshift mtDNA tation had relatively similar heteroplasmy levels was lost. When females that harbored 14% across tissues and these did not change with frameshift mtDNAs were mated, they produced age. Yet, even with selective breeding, the levels pups with either 6% or 0% frameshift mtDNAs. of tRNAMet 3875delC mutation heteroplasmy When a female with 14% frameshift mtDNA could not be increased above 86%. Hence, se- was superovulated and 11 oocytes were geno- lection is clearly acting against this tRNA muta- typed, two oocytes were found to harbor 16% tion at high levels of heteroplasmy. The mtDNA frameshift mtDNAs, one harbored 14%, one genotypes of 44 mothers and their 533 off- 12%, two 6%, and five 0% frameshift mtDNAs spring were analyzed. When the heteroplasmy (Fan et al. 2008). Hence, there was a directional of the mothers was binned into groups of 40%– and concerted loss of the frameshift mtDNA 60%, 61%–70%, and .70%, the heteroplasmy over successive generations. Furthermore, the levels of the pups of the 40%–60% and .70% distribution of oocyte heteroplasmy levels was mothers were found to have undergone nonran- not Gaussian but was truncated such that there dom segregation. This conclusion was based on were oocytes with substantially less mutant comparison of the offspring distribution of ge- mtDNA than the mother’s tail genotype, but notypes observed versus prediction made from none had significantly more mutant mtDNAs the Kimura distribution that would be expected than the mother. Therefore, germ cells or oo- for a random genetic drift model. In contrast to cytes with a higher percentage of frameshift the ND6 13885insC frameshift mutation and mtDNAs must have been selectively removed the nt 4696 deletion mtDNA mice in which prior to ovulation. Thus, mammalian females the heteroplasmy levels of the pups of successive have a prefertilization mtDNA selection that litters declined with age, the heteroplasmy dis- eliminates those oocytes harboring the most tribution of the tRNAMet nt 3875delC mutation deleterious mtDNA mutations. mtDNAs did not significantly differ among Independent evidence that a selection pro- generations (Freyer et al. 2012). cess exists in the female germline was obtained Analysis of female germ cell mtDNA by analyzing the fate of mutant mtDNAs pro- tRNAMet 3875delC mutation levels in 819 Stel-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 21 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

la-GFP-isolated PGCs from 18 embryos at tained in the female germline by two opposing 13.5 d revealed awide variation in heteroplasmy selective forces, one enriching for the mutant levels, with some PGCs cells being 100% mu- mtDNA at intermediate heteroplasmy levels tant. Thus, germline cells could be homoplas- and the other strongly selecting against the mu- mic for the mutant even though no homoplas- tant at high levels of heteroplasmy. mic animals were generated. This indicated that Haplogroup Incompatibility in Heteroplas- early in female germ cell development there was mic Mice. To determine whether mtDNA hap- a stochastic distribution of heteroplasmic levels, logroups might also show nonrandom segrega- but that the highest levels of mutant mtDNAs tion along the mouse female germline, NZB were lost as the oogonia oroocytes matured. The and 129S6 mtDNAs were combined within the authors conclude that “...the variance in het- female mouse germline by fusing cytoplasts eroplasmy among offspring is determined em- from a culture mouse cell line harboring NZB bryonically, consistent with a prenatal germline mtDNAs with a mfESCs derived from a 129S6 bottleneck governing the segregation of mutat- mouse. The NZB-129 heteroplasmic mfESC ed mtDNA.” Analysis of 340 oocytes from the cybrids were injected into blastocysts to generate ovaries of ﬁve 3.5-d-old neonate female mice a heteroplasmic mouse line. NZB and 129 revealed a relatively random distribution of mtDNAs differ in 91 nucleotide positions in- mtDNA genotypes, consistent with a Kimura cluding 15 nonsynonymous amino acid substi- distribution prediction of random genetic drift. tutions, ﬁve tRNAvariants, seven rRNAvariants, From this, the authors concluded that the dis- and 11 control region variants. The heteroplas- tribution of mtDNA tRNAMet 3875delC muta- mic female mice were backcrossed for 20 gen- tion genotypes of the oocytes was not truncated erations with C57BL6/J2 mice originally ob- by selection. However, when the mean mutation tained from the Jackson Laboratory, and all level of the pups of a heteroplasmic female were subsequent females were mated with males plotted against the mother’s tRNAMet 3875delC from the same C57BL/6 J line maintained by heteroplasmy levels, the offspring of females brother/sister mating. with 45%–60% heteroplasmy had pups with The C57BL6/J2 NZB-129 heteroplasmic an up to 14% increased mutation load, while mice were then allowed to segregate the two females with 65%–83% heteroplasmy levels ex- mtDNAs. This revealed a strong bias in segrega- hibited a progressive decline in the average tion toward the loss of the NZB mtDNAs. Over tRNAMet nt 3875delC mutation levels of 24 generations, 171 mice segregated to ,3% 10%. The authors concluded “That selection NZB mtDNAs within one to two generations. is likely to have occurred after the oocyte stage Three out of seven mice that lost the NZB ...,” (Freyer et al. 2012), but it is noteworthy mtDNAs bred true as pure 129 mtDNAs. By that none of the oocytes genotyped contained contrast, more than 10 generations of selective the tRNAMet 3875delC mutation heteroplasmy breeding were required to generate 12 females at levels greater than 85%, even though such with .97% NZB mtDNAs, and of three of these genotypes were present in the PGCs. Therefore, only one bred true for the NZB mtDNA. the descendants of the PGCs with 100% mutant In unrestricted segregation experiments, of mtDNA must have been eliminated prior to the 864 pups born to heteroplasmic mothers, 67% formation of the mature oocytes. had a mean increase in 129 heteroplasmy, 28% For the tRNAMet 3875 delC mutation, two had an increase in NZB heteroplasmy, and only factors appear to be acting on mutant hetero- 5% had a mean heteroplasmy similar to their plasmy levels. At 45%–60% mutant in the mother. Moreover, the tendency to segregate mother, there is a tendency for the mutant NZB mtDNAs was highest when the mother’s mtDNAs to increase in the offspring, while at heteroplasmy level was in the 60%–80% range, heteroplasmy levels of 65%–83%, there was a with the tendency to lose the NZB mtDNAs de- tendency for their loss. Hence, the tRNAMet clining when the percentage of heteroplasmy of 3875delC mutation levels seem to be main- the mother was strongly biased toward either

22 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

129 or NZB mtDNAs. “Therefore, for mice with oplasmy variance among samples and experi- approximately equal proportions of NZB and ments and to calculate the statistical signifi- 129 mtDNAs, there was a strong bias to decrease cance of these differences, (2) development of NZB mtDNA among generations. The enrich- appropriate mathematical models to describe ment of the 129 mtDNAs was seen in the ovaries the origin and transmission of mtDNA hetero- of the heteroplasmic females and to a lesser ex- plasmy, and (3) development of theory and tent in the oocytes of the heteroplasmic females. methods to determine the role of selection in The tendency to segregate the NZB mtDNAs was mtDNA heteroplasmy transmission. not simply the result of incompatibility between Statistical Analysis of mtDNA Heteroplasmy the C57BL/6 J nucleus and the NZB mtDNAs Variance. A major limitation for defining the because the fertility and fecundity of the homo- principles covering heteroplasmy transmission plasmic NZB mice was equal to or better than has been difficulty in comparing the hetero- that of the 129 mice with the same nucleus” plasmy variance levels observed across different (Sharpley et al. 2012). Thus, the segregation pat- experimental systems. Difficulties arise at two tern of the NZB-129 heteroplasmic mice was levels, correcting for the effects of parental allele similar to that of the tRNAMet nt 3875delC mu- frequencies on the potential range of heter- tation mice,althoughless directional orconcert- oplasmy variances of their descendants and dif- ed than was seen for the mtDNA 4696-nt dele- ficulties in calculating the standard error of tion and ND6 13885insC frameshift mutations. heteroplasmy variance estimates for testing the This series of experiments unequivocally statistical significance of observed differences. demonstrates that selection acts on mtDNA ge- The heteroplasmy variance of offspring or notypes within the female germline, acting to derived cells is a function of the heteroplasmy weed out the most severe mtDNA mutations. level of the mothers for offspring or of the pa- Furthermore, the bias toward the elimination rental cells for derived cells, the parental het- of a deleterious mtDNA is correlated with the eroplasmy designated as p0. This fact becomes severity of the mitochondrial defect associated particularly important when the heteroplasmic with the mtDNA mutation. level of one of the mtDNA types in very low, Mathematical Models to Detect Selection in ,10%. Under these conditions, the range of mtDNA Segregation. The striking inconsisten- heteroplasmy values of the offspring or derived cies observed among different estimates of the cells will be constrained by the frequency of the mtDNA copy numbers in mouse female germ- low parental allele, making the range of hetero- line cells, the variation in estimates of heteroplasmic variance levels aberrantly low. Hetero- plasmy level variances in progeny cells and off- plasmic variance will then be maximum when spring of different studies and model systems, p0 ¼ 0.5 and minimal when p0, 0.1 (Samuels and the role of selection in defining the trans- et al. 2010). To correct for this systematic dis- mission of heteroplasmic mtDNA genotypes tortion in variance capacity, Samuels and col- mean that it is currently impossible to define laborators have used Mendelian population ge- the factors that determine the origin and inher- netics theory (Wright 1969) to normalize the itance of pathogenic mtDNA mutations. In an heteroplasmy variances in proportion to p0. effort to more accurately compare different ex- This was accomplished by dividing the observed periments and ultimately to better understand heteroplasmy variance by the term p0(1-p0). mtDNA mutant transmission, significant ef- This term is at maximum (0.25) when p0 is forts have been made to model the transmission 0.5, and becomes progressively smaller as p0 of heteroplasmic mtDNAs along the female gets smaller. Hence, dividing the observed var- germline. iance by p0(1-p0) proportionately increases Unfortunately, three factors have limited population variance estimates for low values the investigator’s ability to generalize the re- of p0. With this correction, Samuels and associ- sults from different studies: (1) development ates discovered that a number of conclusions of methods for comparing differences in heter- about changes in heteroplasmy variance drawn

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 23 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

from various studies involving animals with low occurs in the maternal germline between the maternal p0 could not be substantiated (Sa- PGCs and the oocytes. muels et al. 2010). Application of these analytical approaches Once the effects of extreme values of p0 on to mouse versus human mtDNA heteroplasmic variance estimates were corrected, it became variance data revealed that human heteroplasmy necessary to calculate the standard error (SE) variance was greater than mouse. Comparison of the variance so that the statistical significance of the 3243A.G mutation heteroplasmy levels of differences in variance among different sam- of the 82 oocytes derived from the hysterectomy ple determinations could be calculated. At low of the carrier mother (quadriceps ¼ 18.11%; levels of p0, the required sample size to obtain leukocytes ¼ 7.25%) with the heteroplasmy statistical significance rapidly increases. Won- variance of oocytes from the NZB-BALB mouse napinij and colleagues (2010) pointed out that mtDNA studies revealed that the human nor- many experimental estimates of mtDNA heter- malized oocyte A3243G, mutation level vari- oplasmy variance have been based on sample ance was 0.13, while the mouse normalized oo- sizes 20. Their simulations indicate that these cyte mutation level variance was 0.02–0.04. variance estimates have very large SEs. When Similarly, using an aggregate of human pedi- Wonnapinij and colleagues estimated the SE of grees harboring the NARP 8993T.G, LHON the variance for the original 1996 NZB-BALB 11778G.A and 3460G.A, and MERRF 8344 mtDNA segregation of Jenuth and collaborators A.G mutations encompassing 72 mother– (1996), they concluded that once variance nor- child pairs with heteroplasmy levels in the range malization was applied and variance SEs were of 40%–60%, Wonnapinij and colleagues con- calculated, the reported differences in hetero- cluded that the normalized heteroplasmy mu- plasmy levels between primary and mature oo- tation level variance of human offspring was cytes were not statistically significant. However, 0.37, while that for mouse offspring was the heteroplasmy variance of the PGCs was sig- 0.04–0.12 (Wonnapinij et al. 2010). This com- nificantly lower than that of primary and ma- parison assumes that the nature of the genetic ture oocytes. This analysis supports the hypoth- differences between human mtDNA 3243A.G esis that an mtDNA bottleneck occurs during mutant and between NZB and BALB mtDNAs the transition of PGCs to primary oocytes. does not influence the heteroplasmy segrega- Similarly, Wonnapinij and colleagues rean- tion rate. alyzed the mouse mtDNA heteroplasmy varia- A Neutralist Mathematical Model to Describe tion of the Wai and associates (2008) study. mtDNA Heteroplasmy Segregations. To deter- Once the heteroplasmy variances were normal- mine whether mtDNA heteroplasmy variance ized and SE was estimated, Wonnapinij and col- has been affected by nonrandom factors such leagues concluded that the heteroplasmy vari- as selection, it was necessary to define the ex- ance values determined on postnatal d 11 and pectation for heteroplasmy variance if only sto- after relative to those of d 8 and earlier were chastic factors were contributing to mtDNA not significantly different. Therefore, the con- segregation. The calculation of the expectations clusion of Wai and associates (2008) that an of random heteroplasmy transmission required mtDNA heteroplasmy bottleneck occurs during a theoretical model. Generally, researchers have postnatal folliculogenesis could not be substan- used the Sewall Wright variance formula, V ¼ 2t/Ne tiated (Wonnapinij et al. 2010). Wonnapinijand p0(1-p0)(1-e ) for determining the effect of colleagues also pointed out that the mtDNA random genetic drift on mtDNA heteroplasmy heteroplasmy variance levels in the early post- inheritance. Despite its simplicity, this approach natal period differed between the Jenuth study ignores much of the information that is present (Jenuth et al. 1996) and Wai study (Wai et al. in the total heteroplasmy distribution, especial- 2008). From these analyses, Wonnapinij and ly when this is not symmetric. Additionally, a colleagues concluded that the major bottleneck common assumption when comparing hetero- responsible for mtDNA heteroplasmy variance plasmy distributions between cells or individu-

24 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

als is that the population follows a normal dis- parametercalled b, and treats b as one parameter tribution. However, Wonnapinij and colleagues (b ¼ e2t/Ne). Parameter b is determined from have argued that the normal distribution is not the heteroplasmy variance either of a number the optimal choice for representing the poten- of offspring from a single mother or by pooling tial distribution of mtDNA heteroplasmic geno- the heteroplasmyvaluesofoffspring frommoth- types for two main reasons. First, the normal (or ers with similar heteroplasmy levels (b ¼ 1-[V/ Gaussian) distribution is defined over a range of p0 (1-p0)] or b ¼ 1-Vnormalized). Wonnapinij and infinite minus and plus tails, yet genetic varia- colleagues refer to parameter b as the bottleneck tion is constrained to allele frequencies ( p0) be- parameter, which then determines the width of tween zero and one. Second, the normal distri- the heteroplasmy distribution in the offspring bution is always symmetric, whereas mtDNA (Wonnapinijet al. 2008; Samuels et al. 2013). To heteroplasmy distribution is often skewed. determine the statistical significance of the ob- Therefore, Wonnapinij and colleagues (2008, served differences in the fit of experimental data 2010) argued that while the normal distribution with the Kimura distribution, the WCS-K tool can provide a reasonable model for the stochas- uses the Kolmogorov-Smirmov (KS) test (Won- tic dispersion of heteroplasmy allele frequencies napinij et al. 2008). with r0 values close to 0.5, it is less accurate at Using the Kimura model, Wonnapinij and extreme allele frequencies. colleagues reassessed the variance estimates and To predict the entire heteroplasmy proba- conclusions of a number of previous experi- bility distribution including the probabilities mental studies on heteroplasmic segregation. of a single allele becoming fixed or lost, Won- When the authors applied the WCS-K tool to napinij and colleagues developed a tool based the heteroplasmy distribution of 82 oocytes re- on Kimura’s distribution, henceforth designat- covered from the hysterectomy of the 3243A.G ed the Wonnapinij/Chinnery/Samuels-Kimura mutation woman (Brown et al. 2001), they con- (WCS-K) tool. In the 1950s, Kimura (1955) de- cluded that the oocyte mtDNA distribution fit veloped a model of the possible allele frequency the expectations of the Kimura distribution and distribution for a finite population resulting thus was by inference stochastic. Application of from the random sampling of gametes. His con- the tool to the Jenuth and collaborators data on tinuous model includes a set of probability dis- the transmission to oocytes of the mouse NZB tribution functions that describe gene frequency versus BALB mtDNAs from heteroplasmic distributionsofpopulationsunderpurerandom mothers (Jenuth et al. 1996), the authors con- genetic drift. These functions are represented by cluded that six of the eight mouse line datasets three equations: a probability f(0, t)for losing an were consistent with the expectations of the allele, a probability f(1, t) for fixing that allele, Kimura distribution for random genetic drift, and a probability distribution function w(x,t) but that two datasets at the extreme ends of that the allele is present at frequency x in the the heteroplasmic distribution failed to fit the population. This model has three parameters: Kimura distribution because of unexpectedly the initial gene frequency, p0; the effective pop- low numbers of oocytes that lacked one or the ulation size, Ne; and the number of generations, other mtDNA. Of course, the determination of t. In the case of mtDNA heteroplasmy, p0 is in- whether an mtDNA is present or absent de- terpreted as the founder’s heteroplasmy propor- pends on the sensitivity of the method used to tion or as a surrogate of the parental allele fre- detect the minor allele. quency calculated as the mean heteroplasmy in Analysis of two datasets on heteroplasmic the derived cells or offspring of the progenitor. Drosophila eggs (Solignac et al. 1984; de Stor- Ne is interpreted as a parameter related to the deur et al. 1989) resulted in five of six datasets number of segregating mtDNA units. t is inter- fitting the Kimura distribution, the inference preted as the number of generations. The WCS- being that their distribution arose by random K tool avoids complications related to the defi- segregation. In one D. simulans dataset in which nition of Ne by combining it with t into a single the mtDNA heteroplasmy was generated by

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 25 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

cytoplasmic injection of siIII and siII mtDNA number of mitochondria or mtDNA molecules genomes (two naturally occurring Drosophila may be drastically reduced (Jansen and de Boer mtDNA sequences), the heteroplasmy distribu- 1998; Krakauerand Mira 1999). If the numberof tion was found to fit the Kimura distribution segregating units remains 200 (Jenuth et al. and thus was random. However, in this case the 1996; Cree et al. 2008; Wai et al. 2008), then founder female harbored 38.5% heteroplasmy this assumption is still valid. If, however, the and by the third generation the mean hetero- numberof mtDNA segregating units were to ap- plasmy of the offspring was p0 ¼12.7% (V ¼ proach N ¼ 8, as suggested in one scenario for 0.0096). Therefore, although this striking direc- the 3243A.G genotype distribution of the 82 tional shift could not have occurred by chance, oocytes removed from a hysterectomy (Brown the WCS-K tool was unable to detect deviation et al. 2001), then this term could introduce from randomness. Finally, in a Drosophila data- error into the conclusions of the Kimura model. set derived from eggs 30 generations away from Another concern is whether the KS test for the founder, the results did not fit the expecta- significance is appropriate and sufficiently sen- tions of the Kimura distribution, suggesting sitive to evaluate differences between the ex- that factor(s) in addition to genetic drift had perimentally observed heteroplasmy distribu- acted on this mtDNA segregation pattern tion and the predictions of the Kimura model, (Wonnapinij et al. 2008). especially when the heteroplasmy levels are Thus, by applying the WCS-K tool to het- 0.5. By implementing the KS nonparametric eroplasmy distributions, Wonnapinij and col- test, the WCS-K tool avoids assuming that the leagues concluded that many were consistent data were sampled from Gaussian distributions, with stochastic processes. Still, even this tool but there are drawbacks to using a nonparamet- showed that some cases deviated significantly ric test. If the populations are approximately from the expectations of randomness. Gaussian (as in the case of 0.5 heteroplasmy In contrast to the seemingly modest role level), the nonparametric tests have less power played by selection in determining transmission (are less likely to give you a small p value), es- of heteroplasmy when analyzed using the WCS- pecially with small sample sizes (Stephens K tool, the mouse experiments that analyzed 1974). Additionally, the KS test is most sensitive the transmission of deleterious mutations in- when the distribution functions differ in a glob- cluding the mtDNA 4696 deletion (Sato et al. al fashion near the center of the distribution. 2007), the ND6 13885insC mutation (Fan et al. But if there are repeated deviations between 2008), the tRNAMet 3875delC mutation (Freyer the distribution functions or the distribution et al. 2012), and the NZB mtDNA in 129-NZB functions have (or are adjusted to have) the haplogroup heteroplasmy (Sharpley et al. 2012) same mean values, then the distribution func- all unequivocally show evidence of mtDNA tions cross each other multiple times and the germline purifying selection. One set of poten- maximum deviation between the distributions tial limitations of using the Kimura distribution is reduced (Babu and Feigelson 2006). to model a stochastic distribution of mtDNA Because the predictive ability of the WCS-K genotypes resides in the three parameters that tool is based on the assumption of random ge- relate to the neutralist basis of the distribution. netic drift in a finite large population, the These include: (1) a large population of N dip- strength of the predictions of the WCS-K tool loid parents, (2) no mutation, selection, or mi- is strongly influenced by sample size. This gration, and (3) no overlapping generations makes it possible to change the statistical signi- (Kimura 1955). One of the mathematical as- ficance of a comparison simply by increasing sumptions of the continuous Kimura model is the range of mtDNA heteroplasmy genotypes that the effective size N must be sufficiently large binned together. When the Kimura distribution so that terms of order 1/N2 and higher can be was applied to analysis of heteroplasmy trans- omitted without serious error. Studies of mam- mission of the tRNAMet 3875delC mutation, malian embryonic development suggest that the even for more central p0 values where the effects

26 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

on “b” are minimized, quite large sample sizes be located in a heterozygous individual with a were required to show the directionality associ- significant probability of being lost in subse- ated with selection (Freyer et al. 2012). While quent generations of diploid individuals by statistical evidence of selection was obtained drift. A cell with 1% mutant mtDNAs within a when binning the data from mothers in the cell of 5000 mtDNAs would harbor 50 mutant range of 40%–60% heteroplasmy (Freyer et al. mtDNAs. Because cytokinesis divides the cell in 2012), it was not when binning females with half, low-frequency mtDNA alleles are much heteroplasmy levels of 50%–60% (data not less likely to be lost in successive generations shown). This limitation is compounded be- than nDNA alleles. Finally, the mitochondria cause the currently available program can only and mtDNAs within the cell are not autono- analyze transmissions through one generation, mous units. Rather, they function as a collective thus limiting detection of cumulative effects of population through repeated fission and fusion selection. Moreover, the effects of population and sharing of all their gene products. Subtle size fluctuations on selection are also not con- alterations in the mtDNA sequence and het- sidered and therefore not modeled (Wonnapinij eroplasmy level can have significant effects on et al. 2008, 2010). cellular energy production, REDOX regulation, Whilethese technical considerationsareper- ROS production, Ca2þ regulation, mtDNA tinent, perhaps a more important reason why copy number, mitochondrial fission and fusion, the WCS-K tool is relatively insensitive in iden- mtDNA replication and turnover, etc. For ex- tifying the effects of selection on mtDNA mu- ample, 20% of normal mtDNAs within a cell tations is that the Kimura model was developed can mask the respiratory-deficient phenotypic for changes in the allele frequency of Mendelian of 80% 3243A.G mutant mtDNAs (Yoneda genes within finite populations. The Mendelian et al. 1994). Also, low-frequency deleterious mode of inheritance was the only genetic model mtDNAs can be retained in a cell lineage for known in the 1950s when Kimura developed many generations without adversely affecting his model. However, the Mendelian rules of in- the cellular phenotype. An ND5 gene frameshift heritance derive from the behavior of chro- mutation, which was found to be homoplasmic mosomal genes and the biological unit upon in a human oncocytoma tumor, was present which selection acts on Mendelian genes is the at low-heteroplasmy levels in the normal tissues diploid organism. This unit of selection is fun- of the patient and his two sisters. Thus, this damentally different from the situation with the deleterious mutation was silently transmitted mtDNA in which selection acts on cells and through the maternal lineage (Gasparre et al. organisms with several thousand-fold ploidy. 2008). Hence, classical concepts of genotype– This does not adverselyaffect the mathemat- phenotype associations and their interaction ics of allelic transmission if all alleles are neutral. with selection are violated by mtDNA genetics. However, it has a profound effect onwhether the In conclusion, the formulations of Wonna- Kimura model can detect the effects of purifying pinij and associates (Wonnapinij et al. 2008, selection during the transmission of a deleteri- 2010) have provided a significant improvement ous heteroplasmy mtDNA mutation. In diploid for the mathematical analyses of mtDNA heter- animals, selection acts on units of two gene cop- oplasmy variance in cells and offspring. How- ies for which there are only three allelic combi- ever, the WCS-K tool is less powerful at detect- nations: a/a, a/b, and b/b and two to three ing the role of selection on mtDNA mutation quantized phenoptypic states. For mtDNA ge- inheritance. netics, selection acts on the cellular phenotype, which is a composite of the several thousand Heteroplasmic mtDNA Mutation Segregation mtDNAs with different percentages of two or in Somatic Cells and Tissues more allele systems. For an nDNA allele at a population fre- Heteroplasmic mtDNAs also segregate in so- quency of 1%, virtually every minor allele will matic cells and tissues and this greatly compli-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 27 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

cates phenotypic predictions in patients. Al- frequency. Only cell lines that were predomi- though it is standard practice to test for mutant nantly homoplasmic were stable, with the ex- nDNA genes in patient blood, the mtDNA het- ception of one cell line, which maintained a eroplasmy level in blood cells can be quite stable 30% 3243A.G mutant heteroplasmy different from that of brain, heart, muscle, or (Yoneda et al. 1992). kidney, the organs most susceptible to life- threatening mitochondrial disease. Segregation of Human Somatic Cell The ﬁrst reports of directional segregation mtDNA Mutations of mtDNA heteroplasmy in mouse and human somatic cells appeared in the 1970s. In these Shortly after it was discovered that mtDNA de- experiments, the fate of the mtDNAs was letions accumulate in muscle of patients with monitored in replicating somatic cell hybrids mitochondrial myopathy (Holt et al. 1988), it and cybrids via the mtDNA chlorampheni- was found that mtDNA deletions were progres- col (CAP)-resistance (CAPR) or -sensitivity sively lost from blood cells in CPEO and Kearn (CAPS) genetic markers. CAPR is the product Sayre patients, though they become progressive- of mutations in the mtDNA 16S rRNA gene ly enriched in muscle (Shoffner et al. 1989). (Blanc et al. 1981a,b). When CAPR and CAPS Moreover, mtDNA deletions in humans were cells of the same species and same cell lineage rarely transmitted through the female germline (mouse L cell to mouse L cell, human HeLa cells (Shoffner et al. 1989), though mtDNA duplica- to human HeLa cells) were fused, the mixture of tions can be maternally transmitted (Ballinger CAPR and CAPS mtDNAs was retained for ex- et al. 1992, 1994). tended periods and seemed to segregate sto- The absence of mtDNA deletions in blood chastically. By contrast, when cell lines of the cells when they are prevalent in postmitotic tis- same species but different origin (mouse L cell sues led to the hypothesis that severely deleteri- to mouse RAG cells, human HeLa cells to hu- ous mtDNA mutations might be inhibiting man WiL2 [WAL2]or HT1080 cells) were fused, bone marrow stem cell replication. Those stem then one of the two mtDNAs was directionally cells that spontaneously segregated the deleted and rapidly lost. In the case of cybrids prepared mtDNA would then have a proliferative advan- between enucleated CAPR HeLa cells and tage and repopulate the bone marrow. WAL2A cells, the heteroplasmy could only be The MELAS tRNALeu(UUR) 3243A.G mu- retained by continuous CAP selection (Wallace tation, like mtDNA deletions, commonly pre- et al. 1976, 1977; Bunn et al. 1977; Wallace 1981, sents as CPEO and is also progressively lost from 1982, 1986). The directionality of mtDNA seg- the blood cells, though not as rapidly as mtDNA regation was even more extreme when cells of deletions. Comparison of the 3243A.G heter- different species were fused. When CAPR hu- oplasmy levels for 13 patients tested in white man cells were fused to CAPS mouse cells and blood cells, oral mucosa cells, and urinary tract the hybrids selected in CAP, the surviving hy- epithelial cells consistently showed that the brids retained the human mtDNA and all of the blood cell mutant level was lowest and the uri- human chromosomes, and lost all of the mouse nary epithelial cells heteroplasmy was highest. mtDNA and segregated the mouse chromo- For one patient the levels were 20%, 30%, and somes. This is the opposite of the well-estab- 65% and for another were 30%, 50%, and 80%. lished mouse–human hybrid segregation pat- Because of their consistently high mutation het- tern (Wallace et al. 1976; Giles et al. 1980). eroplasmy level, urinary epithelial cells are con- When clinically relevant mtDNA mutations sidered the best surrogate for postmitotic tissues became available, these mutants were also when testing for 3243.G heteroplasmy levels introduced into cultured cells. In six different of the body (Monnot et al. 2011). cybrids lines heteroplasmic for the MELAS The level of the 3243A.G mutation pro- tRNALeu(UUR) 3243A.G mutation, the mutant gressively declines in blood cells over the life of 3243A.G mtDNA progressively increased in the individual. In a longitudinal analysis of 34

28 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

individuals carrying the 3243A.G mutation, tional segregation of liver and kidney toward 17 symptomatic and 17 carriers all showed an NZB mtDNAs and of spleen toward 129 age-related decline in blood mutant hetero- mtDNAs was confirmed. In addition, it was plasmy levels. For the symptomatic subjects found that seminal vesicles, ovaries, and pancre- the heteroplasmy level declined 0.534% per as also segregated toward 129 mtDNAs. How- year, while for the asymptomatic carriers the ever, tail, brain, lung, and heart maintained rel- heteroplasmy level declined 0.215% per year. atively constant heteroplasmy (Sharpley et al. At the same time, the symptomatic individual’s 2012). Hence, striking differences exist between phenotypes became worse while the asymptom- germline and somatic cell mtDNA hetero- atic individuals remained asymptomatic (Meh- plasmy segregation. Because the somatic tissue razin et al. 2009). A mathematical model was cells contain thousands of mtDNAs, this segre- developed for the decline of 3243A.G hetero- gation is unlikely to be caused by a bottleneck plasmy in blood and compared to longitudinal involving a dramatic reduction in mtDNA copy data that has been collected on the blood cell number. Rather, this must involve a process by heteroplasmy levels. The model and the data which either some mtDNAs are selectively rep- indicate that the loss of heteroplasmy is an ex- licated or eliminated or certain mitochondria ponential function of the form: Mage-corrected ¼ are differentially propagated. m(t)eSt, where m(t) is the average mutational In contrast to the NZB/BALB mouse nuclei level as a function of time and S ¼ 0.020 + (Jenuth et al. 1997), it was discovered that the 0.003 (1/year) (Rajasimha et al. 2008). Mus musculus castaneous (CAST/Ei) mouse nu- While the percentage of the mtDNA cleus does not cause the tissue-specific segrega- tRNALeu(UUR) 3243A.G mutation hetero- tion of the NZB and BALB mtDNAs. This dis- plasmy declines with age in blood, the hetero- covery provided the opportunity to map genetic plasmy level of the MERRF tRNALys 8344A.G loci involved in the mtDNA sorting process. mutation does not (Rajasimha et al. 2008). Heteroplasmic female Mus musculus mice were Hence, different mtDNA mutations show differ- crossed with male Mus musculus castaneous ent tissue-specific stabilities indicating that a (CAST/Ei) mice and the tissue-specific segrega- variety of physiological processes must modu- tion of the NZB-BALB mtDNAs was found to late the fate of heteroplasmic variants in tissues. be inherited as a simple Mendelian recessive trait. Using F2 intercross animals, Battersby, Shoubridge, and associates mapped three quan- Segregation of Mouse Somatic Cell titative trait loci (QTLs). One locus on chromo- mtDNA Mutations some 5 at D5mit25, designated Smdq1 (segrega- The original Jenuth et al. 1996 manuscript re- tion of mitochondrial DNA QTL#1) accounted ported that germline segregation of mouse for 35% of the variance in the heteroplasmy NZB-BALB heteroplasmy was random. Howev- segregation phenotype of liver. A second locus er, in a subsequent paper these authors reported found on chromosome 2 (Smdg2) accounted that NZB-BALB mtDNAs segregated direction- for 16% of the heteroplasmy variance in the ally with age in different tissues (Jenuth et al. kidney. A third locus on chromosome 6 1997). Over the life of an NZB-BALB hetero- (Smdq3) accounted for 20% of the variance plasmic animal, the kidney and liver progres- for the spleen (Battersby et al. 2003). Analysis sively increased the NZB mtDNA, while the of the maximal oxygen consumption rates of the blood and spleen cells accumulated the BALB liver mitochondria of NZB-BALB heteroplas- mtDNA (Jenuth et al. 1997). The heteroplasmy mic animals with NZB mtDNAs levels between of the tail, cerebral cortex, gastrocnemius mus- 0% and 91%–97% on the BALB background cle, heart ventricle, and lung were reported to revealed no significant difference. In the liver, remain relatively stable (Jenuth et al. 1997). the rate of enrichment of the NZB mtDNAs In a similar experiment involving mouse was constant with age with a selective advantage NZB and 129-mtDNA heteroplasmy, the direc- over BALB of 14% per replication cycle. No

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 29 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

difference was found in the relative incorpora- the two Gimap3 alleles, the overexpression of tion rates of BrdU of the two mtDNAs. Hence, the CAST/Ei allele in heteroplasmic mice the advantage of the NZB mtDNA did not ap- slowed the mtDNA segregation rate (Jokinen pear to be caused by increased replication rate. et al. 2010). Also, partial hepatectomy followed by liver lobe regeneration did not affect the NZB-BALB mtDNA ratio. Surprisingly, however, when he- Mitochondrial Physiological Control of Somatic and Germ Cell mtDNA patocytes were explanted from heteroplasmic Segregation mouse livers, the majority of the in vitro cul- tures reversed the segregation direction, losing It has been suggested that in hematopoietic tis- the NZB mtDNA while enriching the BALB sues the selection is age-dependent and propor- mtDNAs. Therefore, the authors surmised that tional to the initial heteroplasmy level. This direction and mechanism of segregation was not implies that selection occurs at the organelle the result of differential mitochondrial physiol- level. Because Gimap3 is a mitochondrial outer ogy or mtDNA replication rate and thus might membrane protein, it might participate in mi- depend on mitochondrial maintenance or mitochondrial partitioning or sorting. Gimap3 is tochondrial turnover (Battersby and Shou- part of a vertebrate-specific gene family and the bridge 2001). protein is anchored to the mitochondrial outer To determine whether the directional segre- membrane by its carboxyl terminus. It has been gation of the NZB mtDNAs in hematopoietic reported to be important in T-cell survival and tissues was the result of immune surveillance, development. One speculation is that “Gimap3 NZB-BALB heteroplasmic females were crossed could act as a scaffold on mitochondrial mem- onto nuclear backgrounds that lacked the Tap1, branes to remodel mitochondria in response to b2m, or Rag1 genes. Absence of these genes specific stimuli” (Jokinen et al. 2011; Jokinen should impair the presentation and recognition and Battersby 2013). of mtDNA-coded peptides. However, the kinet- ics of selection for the BALB mtDNA was un- Regulation of mtDNA Heteroplasmy altered indicating that segregation was not the by tRNAArg 9821 insA-Generated ROS result of presentation of mitochondrially encoded peptides by the major histocompatibility The discovery that Gimap3 regulates mtDNA locus (MHC) (Battersby et al. 2005). segregation in hematopoietic cells does not Fine mapping of the chromosome 6 seem to be the basis for the enrichment of the (Smdq3) locus led to the identification of the NZB mtDNAs in either liver or kidney, each of Gimap3 locus (GTPase of immunity-associated which is modulated by a different genetic locus. protein 3), which proved to be mutant in It has been proposed that because the rate of CAST/Ei relative to BALB/c, with the CAST/ mtDNA selection remains constant in the liver Ei gene being differentially spliced. The Gimap3 with age and is independent of the initial heter- gene has two AUG start codons in exons 3 and 4 oplasmy level selection for the NZB mtDNA with a stop codon in exon 4 upstream of AUG may act at the mtDNA level (Jokinen et al. start codon. When all five exons are spliced to- 2011; Jokinen and Battersby 2013). This pro- gether as in BALB the second AUG is used to posal corresponds to the discovery of an impor- initiate the polypeptide. In CAST/Ei, a G to A tant mtDNA sequence difference between NZB transition in the splice acceptor site of exon 4 versus BALB or 129 mtDNAs. NZB and results in the deletion of exon 4. This results in NIH3T3 mtDNAs differ from BALB and 129 the first AUG being used resulting in the addi- mtDNAs by having an additional A in the length tion of 58 amino acids to the amino terminus of of a homopolymer run of As in the DHU loop the protein without affecting the carboxy-ter- of tRNAArg gene at nt 9821 (tRNAArg 9821insA). minal transmembrane domain. Although the This same variant has been found to be enriched mtDNA copy number was not different between in heteroplasmic cultured mouse L cells (Fan

30 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

et al. 2012). Extensive biochemical analysis of germline segregation of the NZB mtDNAs in mouse L929 r0 cell cybrids that harbor either NZB-129 heteroplasmic mice. Assuming that NIH3T3 or NZB mtDNAs revealed that they the female germline cells have mitochondria have the same mitochondrial respiration rate with a limited number of mtDNAs and fusion as cells with fewer As. However, the similarity and fission are inhibited, then each mitochon- in respiration rate proved to be the result of a drion would act as a semi-autonomous unit. compensatory 1.6- to 1.9-fold increase in Those PGCs and proto-oocytes with higher per- mtDNA copy number in the tRNAArg 9821 centages of NZB mtDNAs would generate in- insA cells. This increased mtDNA copy number creased ROS that would damage the mitochon- was induced by increased mitochondrial ROS dria and mtDNAs resulting in preferential loss production associated with the tRNAArg 9821 of the germline cells with the higher NZB insA allele. When the increased ROS production mtDNAs. This effect would increase until the was neutralized by antioxidants, the mtDNA percentage of NZB mtDNAs reached a suffi- copy number declined and mitochondrial res- ciently high level that the increased ROS sig- piratory chain function was reduced. The in- naled to the nucleus to increase the mtDNA creased ROS also partially inactivated the mito- copy number. The biochemical difference be- chondrial a-ketoglutarate dehydrogenase and tween the predominant 129 and NZB cells aconitase enzymes thus reducing the efficiency would then be negated permitting the very of the tricarboxylic acid cycle and the tRNAArg high NZB lineages to segregate toward homo- 8921 insA cells had a reduced capacity to grow in plasmic NZB. galactose (Moreno-Loshuertos et al. 2006). In contrast to germline cells, somatic tissue Because an alteration in a mitochondrial cells and cultured cells have mitochondria that tRNA would be expected to inhibit mitochon- contain multiple nucleoids and mitochondria drial protein synthesis, this could impede the that actively fuse within a cell mixing their con- synthesis of mitochondrial OXPHOS enzymes, tents and gene products. In this case, the mito- inhibit electron transport, and result in in- chondrial polysomes of a heteroplasmic somatic creased ROS production. Hence, a major factor cell can translate both mtDNAs and mRNAs in the enrichment of the NZB mtDNAs in liver and can complement each other in trans. This could be increased ROS production as a result of trans complementation of mtDNAs in hetero- tRNAArg 9821 insA variant-induced partial mi- plasmic cells was first demonstrated in somatic tochondrial protein synthesis defect (Fan et al. cell hybrids and cybrids harboring CAPR and 2012). The increased ROS would increase the CAPS mtDNAs linked to different polypeptide mtDNA copy number and this effect would be- electrophoretic forms of the ND3 gene, MVI come increasingly predominant as the propor- versus MVII. In heteroplasmic cells in which tion of NZB mtDNAs increased in the tissues. the mtDNA translation products were differen- Battersby and Shoubridge have argued that tially labeled by growth in 35S-methionine in the the tRNAArg 9821 insA could not be the basis of presence of cytosolic ribosome inhibitor, eme- the NZB segregation because they were able to tine, both MVI and MVII were equally labeled. map two different chromosomal loci in the When CAP was added, both MVI and MVII BALB/c CAST/Ei crosses that abolished NZB continued to be labeled. This means that the mtDNA directional segregation (Battersby and CAPR ribosomes had to be translating the MV Shoubridge 2007). However, the gene(s) re- mRNA from the CAPS mtDNA (Oliver and Wal- sponsible for the liver and kidney replicative lace 1982; Oliver et al. 1983). advantage of the NZB mtDNAs could be In the case of somatic tissues harboring NZB nDNA genes involved in regulating mitochon- mtDNAs with the tRNAArg 9821insA allele drial ROS production and/or REDOX regula- mixed with either BALB or 129 mtDNAs, the tion. ribosomes would use both mutant and normal A ROS-driven mtDNA copy-number hy- tRNAs in proportion to the mtDNA hetero- pothesis could also explain the directional plasmy. Because inhibition of polysome elonga-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 31 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

tion would be limited by the “slowest” ribo- multiple enzymes and transcription factors via some, the presence of even a small proportion oxidation-reduction of thiol-disulfides (Wal- of the inhibitory NZB mtDNA tRNAArg 9821 lace 2012). Because the Jackson Laboratory insA would impede protein synthesis, reduce strain of C57BL/6 J, which lacks the Nnt gene mitochondrial OXPHOS, and increase ROS was used in the NZB-129 heteroplasmy studies, production. The increased ROS production it is possible that mitochondrial ROS produc- would then drive an increase in mtDNA copy tion caused by the NZB tRNAArg 9821 insA al- number. The resulting repeated cycles of ampli- lele was enhanced on this strain background re- fication of the mtDNAs within NZB mtDNA- sulting in enhanced selection against PGCs or containing cells could result in the progressive primary or mature oocytes with significant lev- enrichment of the NZB mtDNAs. Hence, the els of the NZB mtDNA and increased ROS pro- tRNAArg 9821 insA allele could account for the duction. progressive increase in NZB mtDNAs in liver and kidney of NZB and BALB or 129 hetero- Regulation of mtDNA Heteroplasmy plasmic mice. by Mitochondrial Fusion and Fission This hypothesis would also be consistent with the dramatic reversal of NZB segregation Other mitochondrial quality assurance mecha- seen for NZB-BALB heteroplasmic hepatocytes nisms such as mitochondrial dynamics and mi- when grown in culture in high glucose medium tochondrial autophagy (mitophagy) might also (Battersby and Shoubridge 2001). The high glu- be important in mtDNA heteroplasmy segrega- cose would fully reduce the electron transport tion (Youleand van der Bliek 2012; Jokinen and chain of the cells resulting in increased ROS pro- Battersby 2013). The mitochondria in somatic duction regardless of the haplotype combina- tissues and cultured cells are highly dynamic tion. The increased mtDNA copy-number effect undergoing a frequent fusion and fission cycle. would be neutralized permitting the more ener- Perturbations of mitochondrial fission and fu- getically efficient BALB mtDNA to take over. sion have been shown to affect the segregation of If mitochondrial ROS production is impor- heteroplasmic mtDNAs. Mitochondrial fission tant in the replicative segregation of NZB versus is initiated by Drp1, a member of the dynamin BALB or 129 mtDNA heteroplasmy animals, protein family. Mitochondrial fusion is initi- then one might anticipate that antioxidant de- ated by the Mfn1 and Mfn2 proteins that medi- fenses might also be relevant in regulating their ate the fusion of the outer mitochondrial mem- segregation. The degree of enrichment for the branes and OpaI, which mediates the fusion NZB mtDNA in liver and kidney and against of the mitochondrial inner membranes (Youle NZB mtDNA in the germline could then be and van der Bliek 2012). In a muscle-derived influenced by mouse strain-specific genetic var- rhabdomyosarcoma cultured cell line carrying iation. One potentially relevant nDNA genetic 80% tRNALeu(UUR) 3243A.G mtDNAs, RNAi variant is the loss of the nicotinamide nucleo- knockdown of the mitochondrial fusion Drp1 tide transhydrogenase gene (Nnt) on chromo- and hFis1 gene mRNAs resulted in an average some 13 in the Jackson Laboratory strain 13% and 11% increase in the mutant mtDNA C57BL/6 J (Toye et al. 2005; Huang et al. over the wild type, respectively. By contrast, 2006; Fielder et al. 2012). NNT is a mitochon- RNAi knockdown of Opa1 mRNA did not affect drial inner membrane protein, which uses the the mutant heteroplasmy level. Hence, in this potential energy of the mitochondrial electro- replicating cell line, increased mitochondrial fu- chemical gradient to transfer reducing equiva- sion was permissive for retention of mutant lents from NADH to NADPH. The higher re- mtDNAs (Malena et al. 2009). ducing potential of NADPH is required for the In mice in which the mitochondrial fusion reduction of mitochondrial lipid peroxides and genes, Mfn1 and Mfn2, were knocked down in of H2O2 to H2O via the glutathione peroxidases. skeletal muscle via floxed Mfn1 and Mfn2 genes NADPH is also essential for the regulation of and the muscle-specific MLC1f promoter-driv-

32 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

en Cre, profound mtDNA depletion ensued. At 9748 and CWD nt 10155-15945) were fused 7 to 8 weeks, the double mutant mice exhibited into somatic cell hybrids. The homoplasmic hyperproliferation of muscle mitochondria, mtDNA deletion cells each had fragmented mi- mtDNA depletion to 250 mtDNAs/nucleus tochondria and were deficient in mitochondrial versus 3500 mtDNA/nucleus in normal protein synthesis and respiration. However, in mouse muscle, a 5-fold increase in mtDNA the hybrids containing the two deleted mtDNAs point mutations, and a 14-fold increase in the mitochondrial were elongated and func- mtDNA deletions. Comparison of the mtDNA tional in both protein synthesis and respiration. deletion levels at 8–13 weeks revealed an 80- By using hybridization probes that were internal fold increase in mtDNA deletions in Mfn2/2, to the two deletions and differentially fluores- Mfn2þ/2 mice (2.3 1025/genome versus cently labeled, the presence of the two deleted 2.8 1027/genome). When the homozy- mtDNAs were monitored. This revealed that the gous mtDNA polymerase g mutator locus, respiring hybrids harbored both mtDNAs in PolgAD257A, was combined with the Mfn2/2 elongated mitochondria and most of the nucle- mutation, the animals manifested a profound oids hybridized to either one or the other probe, OXPHOS defect with complex I activity though in some cases the two hybridization (Chen et al. 2010). Hence, both mtDNA fission probes seemed to overlap. When the hybrid cells and fusion are important in maintaining were released from selection of respiratory suf- mtDNA integrity and the segregation of ficiency (uridine) within 12 d, 55% of the cells mtDNAs within cells and tissues. had segregated to pure CWD mtDNA and 42% had segregated to pure FLPD, with only 3% re- taining both mtDNAs. As the mtDNAs segre- Regulation of mtDNA Heteroplasmy gated the nucleoids resolved into either one by Inter-mtDNA Complementation mtDNA or the other. This indicated that nucle- Mitochodnrial fission and fusion permit oids do not exchange mtDNAs, and that the mtDNA mutant complementation within cells. instances where the deleted mtDNAs colocal- The mtDNAs are contained in nucleoids and ized in the hybrids was the result of two nucle- the distribution of mtDNAs within nucleoids oids being adjacent to each other and thus not and the partitioning of nucleoids into daughter resolved by light microscopy (Gilkerson and mitochondria could be important in hetero- Schon 2008; Gilkerson et al. 2008). plasmy segregation. In nucleoids, the mtDNAs Further evidence that the nucleoids are are complexed with numerous copies of the clones of a single mtDNA genotype have come mtDNA-packaging protein and transcription from the analysis of cultured cells that were het- factor, TFAM. Knockdown of TFAM in HeLa eroplasmic for a 7522 nt deletion, 80% dele- cells perturbs the distribution of mtDNAs into tion. By fragmenting the mitochondria, sorting daughter nucleoids and mitochondria (Kasa- the pico-green-stained mitochondria in a cell shima et al. 2011). The number of mtDNAs sorter, and performing differential PCR ampli- per nucleoid has been estimated to be as low fication for the deleted and normal mtDNA, as one (1.4) (Kukat et al. 2011) and between it was again concluded that each nucleoid con- two and 10 (Gilkerson et al. 2008; Poe et al. tained only one type of mtDNA (Poe et al. 2010; Rebelo et al. 2011). If mtDNA nucleoids 2010). Therefore, the current evidence favors contain more than one mtDNA and nucleoids the conclusion that individual nucleoids harbor could exchange mtDNAs, then nucleoids could only one type of mtDNA. be heteroplasmic, which would significantly affect the dynamics of mtDNA segregation. Regulation of mtDNA Heteroplasmy To determine whether nucleoids are hetero- by Mitophagy plasmic, two cell lines each harboring a homoplasmic, nonoverlapping mtDNA deletion that The mtDNAs continually replicate within so- deleted one or more tRNAs (FLPD nt 7846- matic cells including postmitotic cells, with the

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 33 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

excess mtDNAs being removed by mitophagy. period resulted in the selective loss of the COI Thus, mitophagy provides another mechanism mutant mtDNAs. Hence, mitophagy is capable by which heteroplasmic mtDNA mutations can of modulating mtDNA heteroplasmy. be segregated. Mitophagy is preceded by mito- However, in a cell line heteroplasmic for an chondrial fission, which is thought to be asym- mtDNA cytochrome b mutation with a 37% metric such that one daughter mitochondrion reduction in mitochondrial membrane poten- is more energetically competent than the other. tial overexpression of Parkin did not result in a Mitophagy is then envisioned to preferentially reduction in the percentage of mutant mtDNAs degrade the more respiration-compromised (Suen et al. 2010). Mitophagy has been found to mitochondrion. This process is mediated by be dependent on both Parkin expression and Parkin and PTEN-induced kinase 1 (PINK1), mTORC1 inhibition in cells homoplasmic for encoded by the Park2 and Park6 loci, respective- deleterious mtDNA mutations (Gilkerson et al. ly. In healthy mitochondria, PINK1 is incor- 2012). porated into the mitochondrion but is rapidly Because of the multiplicity of cellular pro- degraded by the mitochondrial PARL rhom- cesses that can impinge on mtDNA hetero- boid protease. However, if the mitochondrial plasmy,it is not surprising that different mtDNA membrane potential is significantly reduced, mutations show different directionality and ki- PINK1 becomes stabilized in the mitochondrial netics of segregation in different tissues. outer membrane where it phosphorylates ser- ines on target proteins. PINK1 phosphorylation results in the attraction of the cytosolic protein IMPLICATIONS FOR MEDICINE Parkin to the mitochondrial outer membrane. Complexities of Genetic Counseling Parkin is an E3 ubiquitin ligase that ubiquiti- nates mitochondrial outer membrane target The absolute nature of maternal inheritance proteins (Narendra et al. 2008, 2009, 2010; of the mtDNA creates an intractable dilemma Guo 2010; Youle and van der Bliek 2012). Re- for women that harbor high levels of a delete- cent evidence has implied that Mfn1, Mfn2, and rious mtDNA mutation. They have a high prob- Miro1 (a protein involved in mitochondrial ability that most if not all of their children transport along microtubules) are important will develop devastating diseases. Homoplasmic Parkin ubiquitination targets. The Parkin mod- mtDNA mutations such as those commonly ification of mitochondrial proteins is followed associated with LHON will be transmitted to by the attraction of the p62/SQSTM1 protein to all of a woman’s offspring, though fortunate- the mitochondrion, which commits the mito- ly the penetrance of these milder mtDNA dis- chondrion to LC3-II encapsulation, fusion of ease mutations is incomplete. For more delete- the resulting autophagosome with a lysosome, rious mtDNA mutations, the outcome can be and degradation (Narendra et al. 2012). much worse. In one published pedigree where The relevance of mitophagy to mtDNA het- the mother was heteroplasmic for the mtDNA eroplasmy segregation was first demonstrated in tRNAThr 15923A.G mutation she conceived Parkin overexpression experiments. Cybrid cells and lost seven pregnancies. Five of the concep- that were stably heteroplasmic for a high per- tions were lost prenatally. One boy was born and centage of an mtDNA COI mutation (COX- died at 42 h and another girl was born and died ICA65) had a 55% reduction in mitochondrial at 56 h of severe OXPHOS defects (Fig. 6) (Yoon inner membrane potential and showed an in- et al. 1993). crease in Parkin-bound mitochondria. Partial Even more common are women with sub- inhibition of mitochondrial fusion with vMIA clinical levels of a pathogenic mutation such and of the electron transport chain with azide to as the tRNALeu(UUR) 3243A.G or ATP6 reduce mitochondrial electron transport, in- 8993T.G mutations but repeatedly transmit creased the Parkin bound to the mitochondria. a high percentage of the deleterious mutant Overexpressed Parkin in these cells over a 60-d mtDNA to their offspring. Unfortunately, pre-

34 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

burn and Dahl 2001; Poulton and Bredenoord 2010). Preimplantation Diagnosis. The ambiguities associated with chorionic villus sampling and amniocentesis diagnosis of mtDNA disease have led to investigation of the potential value of preimplantation genetic diagnosis (PGD). By determining the mtDNA heteroplasmic geno- 42 56 type of oocytes or embryos before fertilization or implantation, only those embryos with a de- One published example of a germline lethal Figure 6. sirable mtDNA genotype could be retained for mtDNA mutation. Awoman harboring a heteroplasmic mtDNA tRNAThr A15923G mutation lost five implantation into the mother. Hence, many po- pregnancies in a row (triangles). The two surviving tential embryos could be screened with only the term infants died within 2 to 3 d. Postmortem mito- rare embryo with the optimal genotype permit- chondrial analysis of one infant’s skeletal muscle re- ted to go through gestation. vealed that complex II þ III and complex IVactivities It would be particularly attractive if it were reduced to 6% and 5% of control and liver complex possible to remove the first polar body from an IV activity that was reduced 37% and complex II þ oocyte and determine its mtDNA genotype. III activity that was undetectable. The mother’s twin sister had a similar reproductive history, yet both Only those oocytes with the lowest polar body women were seemingly normal. (From Yoon et al. heteroplasmy level would then be utilized. An- 1993; modified, with permission, from the authors.) other possibility could be to fertilize the oocytes in vitro and then collect and genotype blastomeres or trophoblast cells to identify the low- natal diagnostic procedures that have been ef- heteroplasmy embryos. fective in identifying fetuses with deleterious The potential for PGD received significant chromosomal gene mutations have proven to initial support from a study of the mtDNA het- be relatively unreliable for diagnosing mtDNA eroplasmy genotypes of NZB-BALB mouse po- diseases. These difficulties arise from the exclu- lar bodies, blastomeres, and embryos. In these sively maternal inheritance of the mtDNA and studies, it was found that the mtDNA hetero- the unpredictable nature of heteroplasmic plasmy level of the first polar body was within mtDNA mutation segregation. 0.1%–6.1% that of the rest of the embryo (R2 ¼ Obviously, all forms of prenatal diagnosis 0.99), as compared to the lack of correlation are useless for a woman harboring a homoplas- between the mother’s mtDNA genotype and mic mtDNA mutation because all of her fetuses that of her embryos (R2 ¼ 0.32). Similarly, it will inherit her mutation. By contrast, hetero- was found that the mtDNA heteroplasmy of dif- plasmic mtDNA mutations might be amenable ferent blastomeres within the same embryo were to prenatal diagnosis if the woman was able also within 0%–6% of each other (Dean et al. to conceive a fetus with low levels of hetero- 2003). plasmy. However, this can also be problematic The positive results of the NZB-BALB het- because it is not certain that the fetal brain, eroplasmy study were supported by studies on heart, muscle, or kidney will have the same level mice harboring the 4696 nt deleted mtDNA. of mtDNA mutation as found in the chorionic Analysis of the second polar body following fer- villus or amniocentesis sample. Furthermore, tilization revealed that the polar body mtDNA for most heteroplasmic mtDNA diseases, it is genotype correlated with the percentage of de- unclear what a “safe” level of heteroplasmy letion in the embryo with a coefficient of 0.95. might be. Finally, some women might conceive Unfortunately, for mtDNA deletion mutations multiple times and in every case the conceptus the benefits of PGD were compromised by the will have an unacceptably high level of mutant subsequent preferential replication of the delet- mtDNAs resulting in repeated fetal loss (Thor- ed mtDNA during development. As mtDNA

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 35 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

deletion-containing embryos developed, the 11.8% + 7.8% lower than the embryo. Hence, percentage of deleted mtDNAs progressively in- it was concluded that in humans, polar body creased, rising 17% during the 19 d after fertil- biopsy would not provide a reliable preimplan- ization, 8% for the first 30 d after birth and tation test for human mtDNA diseases (Gigarel before weaning, 8% for the first 100 d after et al. 2011). weaning, and 6% for every subsequent 100 d. The second possibility could be sampling The maximum deletion level observed in any one of the blastomeres of an early cleavage stage zygote was 78%, implying that the oogonia embryo. Studies of rhesus macaque oocytes with higher levels of deletion did not yield oo- having a 50:50 heteroplasmy generated by kar- cytes. Ultimately, mtDNA deletion mice died of yoplast–cytoplast fusion, the heteroplasmy lev- kidney failure, with the maximum deletion level els among primate blastomeres might be very observed in the kidneys being 90% (Sato et al. different. In quantifying the divergence in per- 2005). Because human mtDNA deletions have centage heteroplasmy, the coefficient of vari- not been observed to be transmitted through the ance increased from 17.7% to 25.0% to 30.9% maternal lineage, the human female germ cells and the range increased from 13.3% to 25.3% to must have a lower tolerance for mitochondrial 43.2% in two-, four-, and eight-cell embryos, respiratory deficiency than the mouse (Wallace respectively. By the eight-cell stage, some em- et al. 2013). bryo blastomeres differed by as much as 10% to Based on the mouse polar body studies, 80%. Furthermore, the variance in hetero- sampling the mtDNA heteroplasmy levels of plasmy level among oocytes, blastomeres, and human meiosis I polar bodies would appear to offspring was sufficiently great as to render ab- be the most promising approach for PGD of solute predictions on the transmission of het- heteroplasmic mtDNA diseases. Unfortunately, eroplasmy among cells and across generations when human studies were conducted it was unreliable (Lee et al. 2012). found that the mtDNA genotype of the first In contrast to the macaque reconstituted cell polar body might be very different from that studies, analysis of 3243A.G heteroplasmy lev- of the embryo. The preimplantation embryos els in human embryos revealed that hetero- of three couples participating in assisted repro- plasmy levels of sister blastomeres at the two- duction were studied, one each harboring the cell stage were similar as were the heteroplasmy tRNALeu(UUR) 3243A.G MELAS mutation, levels of embryos at 3 versus 5 d. Fetal extraem- the tRNALys 8344A.G MERRF mutation, and bryonic or embryonic tissues of 12 fetuses from the ATP69185T.C cerebellar ataxia and tubul- seven carrier mothers were also found to be opathy mutation. The maternal heteroplasmy relatively consistent from gestational d 120 to levels were 25%, 75%–95%, and 14%–25%, term. The heteroplasmy levels of the tissues of respectively, in tissues. In only half of the em- four fetuses were also found to be relatively uni- bryos (27 of 51) were the polar body genotypes form: 74% + 0.7% (three tissues), 42% + within +10% of that of the embryo. In those 0.8% (five tissues), 71% + 2% (five tissues), cases where the polar body was homoplasmic and 78% + 0.9% (seven tissues). Comparison for the normal mtDNA the embryo was also of placental versus fetal tissue were also reported homoplasmic normal (n ¼ 7). However, for to harbor similar heteroplasmy levels: two half of the embryos the polar body genotype cases encompassed tissue 3243A.G frequen- was significantly different from that of the cies of 74%–75% and 78%, while a third ranged embryo. In cases where the embryo’s hetero- from 57% to 42%. While the pooled average of plasmy levels were very high, the polar body heteroplasmy levels between chorionic villi and genotype was significantly lower. For polar amniotic fluid samples showed ,10% varia- bodies with .60% mutant the polar body ge- tion, the heteroplasmy levels of individual tro- notype was on the average 3.5% + 5% higher phoblast or amniocyte cells was quite variable. than the embryo. For polar body genotypes that Thus, while there is marked variability in differ- were ,60%, the polar body heteroplasmy was ent embryos from the same woman, the tissues

36 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

within an embryo can be relatively uniform even 0% to 80%. Among the fetuses, those harboring though the individual cells within a tissue can the NARP mutation tended to have either vary considerably (Monnot et al. 2011) ,30% or .65% mutant mtDNAs while the Initial studies have suggested that preim- MELAS embryos again had a continuous distri- plantation embryo blastomeres have relatively bution of heteroplasmy levels, though different similar mtDNA heteroplasmy levels, in contrast tissue samples from the same embryo had sim- to the macaque report. A 30-year-old female ilar heteroplasmy levels (Monnot et al. 2009a). carrier with a 35% 3243A.G mtDNA mutation A third cohort of 8993T.G and 3243A.G heteroplasmy requested PGD after having a patients concluded that for 8993T.G muta- daughter diagnosed at 2 years of age with tional loads fetuses with ,60% mutant could MELAs syndrome and 84% 3243A.G mutant have a “favorable” outcome, but for mutational mtDNAs. Eight unfertilized oocytes displayed loads of .90% there was a poor prognosis. By mutation loads ranging from 9% to 90% and contrast the clinical correlation between geno- eight arrested embryos encompassing cleavage type and phenotype for the 3243A.G mutation and morula stages had mutation loads ranging was significantly less clear. In general, it was stat- from 7% to 91%. Comparison of blastomere ed that “assessment of mtDNA mutation load in genotypes from five cleavage-stage embryos chorionic villi, amniotic cells, or blastocytes ... with mean heteroplasmy levels of 24.8% to show that the mutation percentage is stable at 85.5% revealed standard deviations ranging different times or in different cells of one em- from 0.5% to 2.9% (Treff et al. 2012). bryo” (de Die-Smulders and Smeets 2009). Multiple additional studies have been con- PGC was performed on 13 couples harbor- ducted on the potential of PGD for women that ing the 8993T.G heteroplasmy mutation en- are heteroplasmic for the MELAS tRNALeu(UUR) compassing 20 embryos and 123 blastomeres. 3243A.G and the NARP/Leigh syndrome Of the 20 embryos, five had detectable ATP6 8993T.G mutations. In a study of three 8993T.G heteroplasmy. The heteroplasmy lev- 8993T.G NARP and 18 3243A.G MELAS els of the blastomeres of each embryo were quite preimplantation embryos, the heteroplasmy similar with ranges of 0%–8%, 4%–15%, 4%– levels were reported as stable among the differ- 15%, 9%–20%, 9%–20%, 10%–12%, 11%– ent blastomeres. Another study of 10 8993T.G 22%, 18%–21%, 10%–12%, and 18%–21%. NARP and nine 3243A.G MELAS fetuses con- In contrast to the similarity of heteroplasmy cluded that the various tissues had similar het- levels of blastomeres within an embryo, the heteroplasmy levels and that the heteroplasmy lev- eroplasmy levels among embryos varied consid- els did not change with gestational age. From erably. Interestingly, the heteroplasmy levels of this series, eight children carrying less than 30% individual lymphocytes from a subject with a mutant mtDNA in the prenatal period were 44.3% whole blood heteroplasmy proved to be born and appear healthy at 18 months to 9 years highly variable, ranging from 11% to 70% (Ta- of age (Monnot et al. 2009b). jima et al. 2007). In a second series of 3243A.G and For a healthy woman with 35% of the 8993T.G patients, 33 preimplantation embry- 8993T.G mutation in her blood and who os at 3 to 5 d and 25 fetuses at gestational age of had a daughter who died at 2 years with an 10 to 30 wk were analyzed from 16 unrelated mtDNA mutation load of .95%, chorionic vil- carrier females. Twenty percent of the embryos lus sampling of two subsequent pregnancies re- lacked the mtDNA mutation. Of the 80% that vealed .95% 8993T.G mutation and were had the mutation, the percentage of hetero- terminated. The woman then underwent two plasmy did not differ significantly among blas- rounds of in vitro fertilization in association tomeres of the embryos. While the NARP em- with PGD. Fifty-nine cells from 10 embryos bryos tended to have either ,25% or .95% were analyzed and the mutation load found to mutant mtDNAs, the MELAS embryos har- be consistent within the blastomeres of the bored a wide range of heteroplasmy levels from individual embryos. Tenembryos and seven oo-

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 37 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

cytes that were tested, 12 (70%) had very high and a mutation load in the child’s blood and mutation loads. Of the 10 embryos, six had high urine of 42%–52% (MJ Falk, pers. comm.). mutation loads, three had 30%–40% hetero- Hence, there are still important ambiguities plasmy, but one had a heteroplasmy level of in the results on the effectiveness of human 2.4%. This embryo was implanted and the preg- PGD in identifying and avoiding heteroplasmic nancy resulted in a healthy girl with a cord blood mtDNA disease mutations. One recently recog- mutation load of 4% (Thorburn et al. 2009). nized variable that might be relevant is that dif- Therefore, in contrast to the macaque stud- ferent pathogenic mtDNA mutations behave ies, the heteroplasmy levels of different blasto- differently in both tissues and preimplantation meres of human embryos appear to have rela- embryos. Following in vitro fertilization embry- tively similar mtDNA heteroplasmy levels. os with mtDNA mutant genotypes of 70% Hence, blastomere biopsy and mtDNA hetero- 3243A.G, 90% 8344A.G, and 54% 9185T.G plasmy analysis from eight-cell embryos may were not adversely affected in preimplantation provide a reasonable estimate of the hetero- development. However, the mtDNA copy num- plasmy levels in the remaining blastomeres ber of 3243A.G embryos was found to increase and reflect the heteroplasmy levels in the ensu- from the germinal vesicle stage to the blastocyts ing pregnancy amniocytes and cord blood sam- 2.7-fold, with the amount of the increase ples (Poulton and Bredenoord 2010). correlating with the percent heteroplasmy (R2 While mtDNA analysis of blastomere biop- ¼ 0.47). Hence, for the 3243G mutation, sies appears to be a promising approach for the mtDNA copy number appears to partially determining the heteroplasmy level of preim- compensate for the mtDNA defect. By con- plantation embryos, blastomere biopsy removes trast, preimplantation embryos harboring the a significant proportion of the embryo. An al- 8344A.G mutation did not show any increase ternative could be to biopsy extraembryonic in mtDNA copy number, even though the cells in the trophectoderm of the blastocyst. In mtDNA mutation load of the 8344A.G embry- one published case, a 30-year-old woman har- os was higher than that of the 3243A.G embry- boring 35% 3243A.G requested PGD follow- os: oocyte and embryo heteroplasmy levels ing birth of a MELAS syndrome daughter. Six for the 8344A.G were 81% + 15% and 68% developmentally competent blastocysts with + 17%versus37%+29%and36% + 30% for mean heteroplasmy values ranging from 24% the 3243A.G. Thus, the 3243A.G mutation to 91% were sampled. Three to four trophecto- must generate different physiological signals derm cells were obtained and compared to one than the 8344A .G mutation (Monnot et al. to three inner cell mass samples. The standard 2013). How such mutation-specific effects have deviations of the independent samples from the affected the reliability of PGD remains to be de- various trophoblasts ranged from 2.1% to 5.0% termined. and the means of the trophoblast and inner cell mass samples were all within 3% of each other. Germline Gene Therapy Based on this data, one developmentally competent embryo with trophoblast heteroplasmy While the PGD by blastomere sampling seems levels of 12% was chosen for implantation. promising for determining the heteroplasmy This resulted in the successful delivery of a child levels of the 8993T.G NARP/Leigh syndrome for which a buccal cell DNA analysis at 1 mo mutation and possibly the 3243A.G MELAS revealed a heteroplasmy level of 15%. Subse- mutation, there are still major limitations of this quent tests at 5 and 12 mo of buccal swab and method. First, a woman may not produce oo- urine sediment DNA heteroplasmy by a com- cytes that have low levels of mutant mtDNA, mercial laboratory reported ,10% mutation and, without the appropriate oocyte, her in vi- load (Treff et al. 2012). However, subsequent tro fertilization regime could be unproductive. follow-up of this case has revealed significant Second, while the heteroplasmic levels of the problems in the placenta, infant, and child, ATP6 8993T.G L126R mutation may be con-

38 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

sistent throughout the embryo and fetus, the pronuclei) fused to an enucleated oocyte by potential for intertissue differences in hetero- electroshock. About 6% of the mtDNAwas car- plasmy levels for the 3243A.G mutation may ried along with the nucleus. Because the aver- be significantly higher. Third, PGD is useless for age mutant mtDNA in the karyoplast donor homoplasmic women. Because of these limita- was 35%, this resulted in zygotes containing tions, there is increasing interest in developing 2% deleted mtDNAs. Eleven mice were born methods to simply replace the mutant mtDNAs and at weaning their tail heteroplasmy levels in preimplantation embryos by pronuclear or were found to be 6%–12% mutant, mean 11%. spindle transfer. Hence, zygote nuclear transfer resulted in a sig- One early attempt at altering the mtDNA nificant reduction in the transmission of the genotype in human oocytes was ooplasmic deleted mtDNAs. Unfortunately, in the case of transfer. In this technique 20% of the cyto- the mtDNA deletion, the deletion levels pro- plasm of an oocyte from a one woman is inject- gressively increased over the next 300 d to ed into an oocyte from another woman along 5%–44% mutant, mean 23% (Sato et al. 2005). with her husband’s sperm (intracytoplasmic Pronuclear transfer technology has also sperm injection, ICSI), and the embryos im- been extended to human preimplantation em- planted into a woman’s uterus. This technique bryos. Using uni-pronuclear or tri-pronuclear was first used in an attempt to rejuvenate the human embryos that otherwise would have oocytes of older women with the cytoplasm been destroyed, the cytoskeleton was disrupted from the oocytes of younger women thus in- with cytochalasin B and nocodazole, the pronu- creasing their fertility. Multiple children have clei pinched off into a karyoplast, the karyoplast been born following application of this proce- placedunder the zona pellucida of an enucleated dure and are alive today. Some of these children recipient zygote, and the two cells fused together have been shown to be heteroplasmic for both with inactivated viral envelope proteins. About the maternal oocyte mtDNA and the cytoplas- 22% of the resulting reconstituted embryos de- mic donor mtDNA (Barritt et al. 2001a,b). veloped to the eight-cell stage and of those that Given that it has been reported that the mix- received two pronuclei, 8.3% developed to the ing of two different mouse mtDNAs within the blastocyst stage, 50% of the developmental same female germline can lead to offspring with potential of unmanipulated embryos. Using neuro-psychiatric defects (Sharpley et al. 2012), mtDNAcontrol region polymorphisms to mon- concern has been raised that randomly mixing itor the fate of the two mtDNAs, the mean mtDNAs of different mtDNA lineages might be mtDNA carryover along with the pronuclear deleterious. Also, the addition of 20% normal karyoplast was 8.1% + 7.6% (n ¼ 8). Varia- mtDNAs would be unlikely to be sufficient to tion in the percentage of mutant mtDNAs in reduce the risk of the child’s developing the the different blastomeres within one eight-cell mtDNA disease. embryo was found to range from 5.1% to 30.2% Rather than the transfer of mitochondria and within another four-cell embryo to range and mtDNAs, a more promising approach from 8% to 39%. Further refinement of the kar- would be to isolate the nucleus from the oocyte yoplast isolation technique resulted in nine em- or zygote of a woman harboring a deleterious bryos with an average carryover of karyoplast mtDNA mutation and to transfer it into an enu- mtDNAs of 1.68% + 1.81% (n ¼ 9). Four of cleated oocyte or zygote from a woman with these embryos lacked detectable karyoplast normal mtDNAs (Wallace 1987). As a model mtDNA. Of those embryos that harbored kar- for pronuclear transfer, in mouse embryos for yoplast mtDNA the maximum donor mtDNA the heteroplasmic mtDNA 4696 nt deletion, was found in a nine-cell embryo in which seven the zygote nuclei were removed by micropipette blastomeres lacked detectable donor mtDNAs, from the mtDNA mutant zygotes following while the remaining two blastomeres harbored disaggregation of the cytoskeleton and the re- 6.1% and 11.4% donor mtDNAs (Craven et al. sulting karyoplast (plasma membrane bound 2010).

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 39 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

Interestingly, the spread in blastomere het- fuses. The resulting reconstituted oocyte can eroplasmy levels in the first round pronuclear then be fertilized by intracytoplasmic sperm in- transfer embryos (Craven et al. 2010) was rem- jection (ICSI). iniscent of the wide range of blastomere mtDNA This technology was first developed and ap- heteroplasmy genotypes reported for the ma- plied to the oocytes of rhesus macaque mon- caque embryo fusion experiments. This was keys. It resulted in preimplantation embryos much greater than the range of heteroplasmy that developed normally to generate pluripotent levels observed in embryos generated from stem cells and when blastocysts of four- to eight- NZB/BALB heteroplasmic mice or in the blas- cell cleavage embryos were transferred into the tomeres of heteroplasmic human preimplanta- reproductive tract of females, one pair of twins tion embryos. One possible explanation for this and two singleton infants were born. These four difference is that the oocyte cytoplasm may be monkeys had the maternal chromosomes of the relatively viscous resulting in the nonrandom spindle donor but the mtDNA of the cytoplast partitioning of the mitochondria in different donor. No spindle donor oocyte mtDNA was regions of the single-cell embryo. Depending detected at a detection sensitivity of 3% (Ta- on the angles of the initial cleavages, the chibana et al. 2009). Three years after birth, the mtDNAs of the karyoplast and the cytoplast four-spindle transfer monkeys were phenotypi- could be asymmetrically distributed into cally like control monkeys and showed no sig- daughter blastomeres resulting in differing het- nificant change in mtDNA carryover in blood eroplasmy levels. By contrast, oocytes and em- and skin samples (Tachibana et al. 2013). Sub- bryos derived from heteroplasmic germline cells sequent analysis revealed that the macaque kar- of the two mtDNAs could be more evenly dis- yoplasts carried 3538 + 2310 mtDNAs while tributed throughout the cytoplasm such that the cytoplast contributed 576,948 + 209,069 cleavages would generate blastomeres with sim- mtDNA (n ¼ 9) resulting in 0.6% hetero- ilar mtDNA heteroplasmy levels. plasmy. Of 102 spindle transfer oocytes 62% One concern with the pronuclear transfer developed into blastocyts. Two female embryos procedure is that it destroys an embryo. This generated from spindle transfer oocytes were concern has been overcome by transfer of permitted to develop in utero for 135 d and the chromosomes from the prefertilization then analyzed for mtDNA heteroplasmy. Of 24 mtDNA mutant oocyte to an enucleated oocyte oocytes isolated from the ovaries of the female with normal mtDNA. In this case, however, the fetuses 22 had no detectable spindle donor chromosomes of the mature oocyte are arrested mtDNA while one from each fetus was hetero- in meiotic metaphase II. Hence, this method plasmic with 16% and 14% heteroplasmy, re- involves the physical transfer of the meiotic II spectively (Lee et al. 2012). spindle. One of the technological limitations of In the spindle transfer technique, the meta- spindle transfer is that both the spindle donor phase II (MII) spindle of the oocyte is visualized karyoplast and the mtDNA donor cytoplast in the oocytes using the polarizing microscope must be ready at the same time, requiring paired and is removed by aspiration with a micro- induction of ovulation of two women, which pipette within a bleb of the ooycte cytoplasm can be quite difficult. Studies on macaque and surrounded by the cell membrane. The oocytes have shown that the spindle donor spindle karyoplasts encompass 1.5% of the oocyte can be frozen using a vitrification proce- volume of an enucleated oocyte cytoplast. dure. These oocytes can then be thawed, the Because the spindle is relatively free of sur- spindle removed, and transferred to a fresh enu- rounding cytoplasm and mitochondria very cleated oocyte without diminution of fertiliza- little mtDNA is transferred. The karyoplast is tion. Blastocyst formation efficiency was 88% treated with a Sendai virus-derived fusigen and 68% of that of controls. Vitrification of the and placed in the perivitelline space of the cy- cytoplasm donor oocytes, by contrast, blocked toplast opposite the first polar body where it development (Tachibana et al. 2013). Hence,

40 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

spindle transfer has the potential of eliminating ferentiating into pancreatic cells, neurons, fi- most of the mutant mtDNAs from the maternal broblasts, and cardiomyocytes. Analysis of the lineage, but low levels of heteroplasmy can still mtDNA heteroplasmy level for all preimplanta- resurface during germline transmission. tion embryos studied was 0.31% + 0.27%. Of To assess the feasibility of moving this tech- the three ESC lines that were generated, two nology into the clinic, spindle transplantation lacked detectable karyoplast mtDNA while one has been applied to human oocytes. Of 64 spin- had 2.79% + 0.27% karyoplast mtDNA. The dle-transfer oocytes, 94% survived ICSI and heteroplasmic ESC line subsequently lost the 75% formed pronuclei. However, in contrast karyoplast mtDNA, the spindle donor mtDNAs to macaque results only about half (48%) becoming undetectable after passage 14 and re- formed two normal pronuclei and two polar mained so in 40 clones. Fibroblasts derived from bodies. The remainders were abnormal. Only one of the spindle transfer stem cell lines were 13% of the nonspindle transfer embryos were converted back to induced pluripotent stem abnormal. Of the normal spindle transfer em- (iPS) cells, yet the mtDNAs from the karyoplast bryos, 76% went on to form blastocysts. From did not reappear. Finally, analysis of the mito- the 16 blastocysts, nine formed ESCs (56%) and chondrial respiratory complexes, respiration, these were indistinguishable from the controls. and media acidification of one of the stem cell The mean karyoplast derived mtDNA levels lines revealed that it was comparable to that of were very low, 0.5% + 04% for spindle transfer embryonic stem cells and iPS cells that had not oocytes and embryos and 0.6% + 0.9% for hu- undergone spindle transfer (Paull et al. 2013). man ESCs (Tachibana et al. 2013). While both the pronuclear and spindle In an independent study using parthenoge- transfer techniques have been refined to transfer netically activated embryos that have undergone a minute amount of nuclear donor mtDNA, endoreduplication to establish a diploid karyo- the possibility still remains that heteroplasmy type, spindle transfer was also analyzed for its can resurface in subsequent ESCs or maternal capacity to replace the mtDNA of the spindle offspring (Paull et al. 2013). While low-level donor mother. As before, the spindle was re- heteroplasmy of a pathogenic mutation would moved from the oocyte in a bleb of cytoplasm, probably be masked by the predominance of the resulting karyoplast placed under the zona normal mtDNAs, minimizing the risk of a clin- pellucida of a comparably enucleated oocyte, ical phenotype, there is a risk of incompatibility and the karyoplast fused to the cytoplast by ei- between the two mtDNA haplogroup lineages ther Sendai viral fusigen or electric shock. Sen- (Sharpley et al. 2012). This concern could be dai fusion proved preferable because electro- mitigated by screening potential oocyte mtDNA shock was prone to induce premature oocyte donor women for their mtDNA haplogroups. activation. This study determined that the num- Then the mtDNA haplogroup of the woman ber of mtDNAs carried by the karyoplast was harboring the deleterious mtDNA mutation 1129 + 785 or 0.36% of the total mtDNA in a could be matched with that of an oocyte donor; metaphase II oocyte, which contains 311,146 only matching mtDNA haplogroups would + 206,521 mtDNAs. An important innovation then be used. of this study was the realization that partial disaggregation of the karyoplast spindle complex Alternatives and Ethical Considerations by a 2-h exposure to room temperature or prior vitrification followed by partial reaggregation of It has been argued that it may be unethical to the spindle at 37ºC significantly increased the manipulate human embryos in an effort to re- efficiency of subsequent second polar body ex- move the risk of mtDNA disease. Certainly, all trusion and embryo development. Several em- appropriate preclinical tests must be performed bryos were permitted to develop into blastocysts in an effort to reduce the risk for adverse out- and from these ESC lines with normal karyo- comes in developing new human therapies. The types were derived. These were capable of dif- question remains is human mtDNA germline

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 41 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

gene therapy “ethical.” To answer this question, Barritt J, Willadsen S, Brenner C, Cohen J. 2001a. Cytoplas- we must factor into our cost-benefit analysis the mic transfer in assisted reproduction. Hum Reprod Up- date 7: 428–435. implications not only for society but also for the Barritt JA, Brenner CA, Malter HE, Cohen J. 2001b. Mito- family and their prospective children. Women chondria in human offspring derived from ooplasmic harboring severely pathogenic mtDNA muta- transplantation. Hum Reprod 16: 513–516. tions want to have healthy children that are Battersby BJ, Shoubridge EA. 2001. Selection of a mtDNA sequence variant in hepatocytes of heteroplasmic mice free of pain and can live productive lives. Cur- is not due to differences in respiratory chain function rently children with high heteroplasmy levels of or efficiency of replication. Hum Mol Genet 10: 2469– severely deleterious mtDNA mutations experi- 2479. ence the progressive loss of mental and physical Battersby BJ, Shoubridge EA. 2007. Reactive oxygen species and the segregation of mtDNA sequence variants. Nat capabilities, often experiencing unremitting Genet 39: 571–572; author reply 572. discomfort and pain and many will ultimately Battersby BJ, Loredo-Osti JC, Shoubridge EA. 2003. Nuclear progress to premature death. For such families, genetic control of mitochondrial DNA segregation. Nat this can mean repeated medical crises, numer- Genet 33: 183–186. ous emergency room admissions, devastating Battersby BJ, Redpath ME, Shoubridge EA. 2005. Mitochon- drial DNA segregation in hematopoietic lineages does medical bills, the loss of time, affection, and not depend on MHC presentation of mitochondrially resources for other family members, ending encoded peptides. Hum Mol Genet 14: 2587–2594. only by the death of the child. Therefore, if the Baudouin SV,Saunders D, Tiangyou W, Elson JL, Poynter J, Pyle A, Keers S, Turnbull DM, Howell N, Chinnery PF. society is to rule on the morality of helping 2005. Mitochondrial DNA and survival after sepsis: A women have healthy children, than the families prospective study. Lancet 366: 2118–2121. who have suffered the ravages of mtDNA disease Bergstrom CT, Pritchard J. 1998. Germline bottlenecks and and personally paid the cost should be given a the evolutionary maintenance of mitochondrial ge- major voice in the decision. nomes. Genetics 149: 2135–2146. Blanc H, Adams CW, Wallace DC. 1981a. Different nucleotide changes in the large rRNA gene of the mitochondrial DNA confer chloramphenicol resistance on two human ACKNOWLEDGMENTS cell lines. Nucleic Acids Res 9: 5785–5795. Blanc H, Wright CT, Bibb MJ, Wallace DC, Clayton DA. This work is supported by National Institutes of 1981b. Mitochondrial DNA of chloramphenicol-resis- Health grants NS21328, NS070298, AG24373, tant mouse cells contains a single nucleotide change in the region encoding the 30 end of the large ribosomal and DK73691 and by Simons Foundation grant RNA. Proc Natl Acad Sci 78: 3789–3793. 205844 awarded to D.C.W. Blok RB, Gook DA, Thorburn DR, Dahl HH. 1997. Skewed segregation of the mtDNA nt 8993 (T!G) mutation in human oocytes. Am J Hum Genet 60: 1495–1501. Brown WM, Prager EM, Wan A, Wilson AC. 1982. Mito- REFERENCES chondrial DNA sequences in primates: Tempoand mode Reference is also in this collection. of evolution. J Mol Evol 18: 225–239. Brown MD, Torroni A, Reckord CL, Wallace DC. 1995. Phy- Ashley MV,Laipis PJ, Hauswirth WW. 1989. Rapid segrega- logenetic analysis of Leber’s hereditary optic neuropathy tion of heteroplasmic bovine mitochondria. Nucleic Ac- mitochondrial DNA’s indicates multiple independent oc- ids Res 17: 7325–7331. currences of the common mutations. Hum Mutat 6: Babu GJ, Feigelson ED. 2006. Astrostatistics: Goodness-of- 311–325. fit and all that! Astronomical Data Analysis Software and Brown MD, Sun F, Wallace DC. 1997. Clustering of Cauca- Systems XVASP Conference Series 351: 127. sian Leber hereditary optic neuropathy patients contain- Ballinger SW,Shoffner JM, Hedaya EV,Trounce I, Polak MA, ing the 11778 or 14484 mutations on an mtDNA lineage. Koontz DA, Wallace DC. 1992. Maternally transmitted Am J Hum Genet 60: 381–387. diabetes and deafness associated with a 10.4 kb mito- Brown DT, Samuels DC, Michael EM, Turnbull DM, Chin- chondrial DNA deletion. Nat Genet 1: 11–15. nery PF. 2001. Random genetic drift determines the level Ballinger SW, Shoffner JM, Gebhart S, Koontz DA, Wallace of mutant mtDNA in human primary oocytes. Am J Hum DC. 1994. Mitochondrial diabetes revisited. Nat Genet 7: Genet 68: 533–536. 458–459. Brown MD, Starikovskaya E, Derbeneva O, Hosseini S, Allen Balloux F, Handley LJ, Jombart T, Liu H, Manica A. 2009. JC, Mikhailovskaya IE, Sukernik RI, Wallace DC. 2002. Climate shaped the worldwide distribution of human The role of mtDNA background in disease expression: A mitochondrial DNA sequence variation. Proc R Soc B new primary LHON mutation associated with Western 276: 3447–3455. Eurasian haplogroup J. Hum Genet 110: 130–138.

42 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

Bunn CL, Wallace DC, Eisenstadt JM. 1974. Cytoplasmic ence, Abstract S09.3, Vienna Austria, May 23–26, 2009. inheritance of chloramphenicol resistance in mouse tis- Eur J Hum Genet 17: 9–10. sue culture cells. Proc Natl Acad Sci 71: 1681–1685. de Stordeur E, Solignac M, Monnerot M, Mounolou JC. Bunn CL, Wallace DC, Eisenstadt JM. 1977. Mitotic segre- 1989. The generation of transplasmic Drosophila simu- gation of cytoplasmic determinants for chloramphenicol lans by cytoplasmic injection: Effects of segregation and resistance in mammalian cells. I: Fusions with mouse cell selection on the perpetuation of mitochondrial DNA lines. Somatic Cell Genet 3: 71–92. heteroplasmy. Mol Gen Genet 220: 127–132. Cao L, Shitara H, Horii T, Nagao Y, Imai H, Abe K, Hara T, Elliott HR, Samuels DC, Eden JA, Relton CL, Chinnery PF. Hayashi J, Yonekawa H. 2007. The mitochondrial bottle- 2008. Pathogenic mitochondrial DNA mutations are neck occurs without reduction of mtDNA content in common in the general population. Am J Hum Genet female mouse germ cells. Nat Genet 39: 386–390. 83: 254–260. Cao L, Shitara H, Sugimoto M, Hayashi J, Abe K, Yonekawa Fan W, Waymire K, Narula N, Li P, Rocher C, Coskun PE, H. 2009. New evidence confirms that the mitochondrial Vannan MA, Narula J, MacGregor GR, Wallace DC. 2008. bottleneck is generated without reduction of mitochon- A mouse model of mitochondrial disease reveals germ- drial DNA content in early primordial germ cells of mice. line selection against severe mtDNA mutations. Science PLoS Genet 5: e1000756. 319: 958–962. Carling PJ, Cree LM, Chinnery PF.2011. The implications of Fan W, Lin CS, Potluri P, Procaccio V, Wallace DC. 2012. mitochondrial DNA copy number regulation during em- MtDNA lineage analysis of mouse L cell lines reveals the bryogenesis. Mitochondrion 11: 686–692. accumulation of multiple mtDNA mutants and intermo- Chapman R, Stephens J, Lansman R, Avise J. 1982. Models of lecular recombination. Genes Dev 26: 384–394. mitochondrial DNA transmission genetics and evolution Fielder TJ, Yi CS, Masumi J, Waymire KG, Chen HW, Wang in higher eucaryotes. Genet Res 40: 41–57. S, Shi KX, Wallace DC, MacGregor GR. 2012. Compari- Charlesworth B, Charlesworth D. 2010. Elements of evolu- son of male chimeric mice generated from microinjection tionary genetics. Roberts and Company, Greenwood Vil- of JM8.N4 embryonic stem cells into C57BL/6 J and lage, CO. C57BL/6NTac blastocysts. Transgenic Res 21: 1149–1158. Chen H, Vermulst M, Wang YE, Chomyn A, Prolla TA, Freyer C, Cree LM, Mourier A, Stewart JB, Koolmeister C, McCaffery JM, Chan DC. 2010. Mitochondrial fusion Milenkovic D, Wai T, Floros VI, Hagstrom E, Chatzidaki is required for mtDNA stability in skeletal muscle and EE, et al. 2012. Variation in germline mtDNA hetero- tolerance of mtDNA mutations. Cell 141: 280–289. plasmy is determined prenatally but modified during Chinnery PF, Elliott HR, Syed A, Rothwell PM. 2010. Mito- subsequent transmission. Nat Genet 44: 1282–1285. chondrial DNA haplogroups and risk of transient ischae- Fuku N, Park KS, Yamada Y, Nishigaki Y, Cho YM, Matsuo mic attack and ischaemic stroke: A genetic association H, Segawa T, Watanabe S, Kato K, Yokoi K, et al. 2007. study. Lancet Neurol 9: 498–503. Mitochondrial haplogroup N9a confers resistance Chinnery PF, Elliott HR, Hudson G, Samuels DC, Relton against type 2 diabetes in Asians. Am J Hum Genet 80: CL. 2012. Epigenetics, epidemiology and mitochondrial 407–415. DNA diseases. Int J Epidemiol 41: 177–187. Gasparre G, Hervouet E, de Laplanche E, Demont J, Pennisi Chomyn A. 1996. Platelet-mediated transformation of hu- LF,Colombel M, Mege-Lechevallier F,Scoazec JY,Bonora man mitochondrial DNA-less cells. Methods Enzymol E, Smeets R, et al. 2008. Clonal expansion of mutated 264: 334–339. mitochondrial DNA is associated with tumor formation Coskun PE, Wyrembak J, Derbereva O, Melkonian G, Doran and complex I deficiency in the benign renal oncocy- E, Lott IT, Head E, Cotman CW, Wallace DC. 2010. Sys- toma. Hum Mol Genet 17: 986–995. temic mitochondrial dysfunction and the etiology of Alz- Gigarel N, Hesters L, Samuels DC, Monnot S, Burlet P, heimer’s disease and Down syndrome dementia. J Alz- Kerbrat V, Lamazou F, Benachi A, Frydman R, Feingold heimers Dis 20: S293–S310. J, et al. 2011. Poor correlations in the levels of pathogenic Craven L, Tuppen HA, Greggains GD, Harbottle SJ, Murphy mitochondrial DNA mutations in polar bodies versus JL, Cree LM, Murdoch AP, Chinnery PF, Taylor RW, oocytes and blastomeres in humans. Am J Hum Genet Lightowlers RN, et al. 2010. Pronuclear transfer in hu- 88: 494–498. man embryos to prevent transmission of mitochondrial Giles RE, Stroynowski I, Wallace DC. 1980. Characterization DNA disease. Nature 465: 82–85. of mitochondrial DNA in chloramphenicol-resistant in- Cree LM, Samuels DC, de Sousa Lopes SC, Rajasimha HK, terspecific hybrids and a cybrid. Somatic Cell Genet 6: Wonnapinij P,Mann JR, Dahl HH, Chinnery PF. 2008. A 543–554. reduction of mitochondrial DNA molecules during em- Gilkerson RW, Schon EA. 2008. Nucleoid autonomy: An bryogenesis explains the rapid segregation of genotypes. underlying mechanism of mitochondrial genetics with Nat Genet 40: 249–254. therapeutic potential. Commun Integr Biol 1: 34–36. Dean NL, Battersby BJ, Ao A, Gosden RG, Tan SL, Shou- Gilkerson RW, Schon EA, Hernandez E, Davidson MM. bridge EA, Molnar MJ. 2003. Prospect of preimplanta- 2008. Mitochondrial nucleoids maintain genetic auton- tion genetic diagnosis for heritable mitochondrial DNA omy but allow for functional complementation. J Cell diseases. Mol Hum Reprod 9: 631–638. Biol 181: 1117–1128. de Die-Smulders C, Smeets H. 2009. The challenge of pre- Gilkerson RW,De Vries RL, Lebot P,Wikstrom JD, Torgyekes natal and preimplantation genetic diagnosis of mito- E, Shirihai OS, Przedborski S, Schon EA. 2012. Mito- chondrial disorders. European Human Genetics Confer- chondrial autophagy in cells with mtDNA mutations

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 43 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

results from synergistic loss of transmembrane potential Jokinen R, Junnila H, Battersby BJ. 2011. Gimap3: A foot- and mTORC1 inhibition. Hum Mol Genet 21: 978–990. in-the-door to tissue-specific regulation of mitochondri- Gomez R, O’Keeffe T, Chang L, Huebinger R, Minei J, Bar- al DNA genetics. Small GTPases 2: 31–35. ber R. 2009. Association of mitochondrial allele 4216C Kasashima K, Sumitani M, Endo H. 2011. Human mito- with increased risk for complicated sepsis and death chondrial transcription factor A is required for the seg- after traumatic injury. J Trauma 66: 850–857; discussion regation of mitochondrial DNA in cultured cells. Exp Cell 857–858. Res 317: 210–220. Gomez-Duran A, Pacheu-Grau D, Lopez-Gallardo E, Diez- Kazuno AA, Munakata K, Nagai T,Shimozono S, Tanaka M, Sanchez C, Montoya J, Lopez-Perez MJ, Ruiz-Pesini E. Yoneda M, Kato N, Miyawaki A, Kato T. 2006. Identifi- 2010. Unmasking the causes of multifactorial disorders: cation of mitochondrial DNA polymorphisms that alter OXPHOS differences between mitochondrial haplo- mitochondrial matrix pH and intracellular calcium dy- groups. Hum Mol Genet 19: 3343–3353. namics. PLoS Genet 2: e128. Goto Y, Nonaka I, Horai S. 1990. A mutation in the Khrapko K. 2008. Twoways to make an mtDNA bottleneck. tRNALeu(UUR) gene associated with the MELAS subgroup Nat Genet 40: 134–135. of mitochondrial encephalomyopathies. Nature 348: Khusnutdinova E, Gilyazova I, Ruiz-Pesini E, Derbeneva O, 651–653. Khusainova R, Khidiyatova I, Magzhanov R, Wallace DC. Goto H, Dickins B, Afgan E, Paul IM, Taylor J, Makova KD, 2008. A mitochondrial etiology of neurodegenerative dis- Nekrutenko A. 2011. Dynamics of mitochondrial heter- eases: Evidence from Parkinson’s disease. Ann NY Acad oplasmy in three families investigated via a repeatable re- Sci 1147: 1–20. sequencing study. Genome Biol 12: R59. Kimura M. 1955. Solution of a process of random genetic Guo M. 2010. What have we learned from Drosophila models drift with a continuous model. Proc Natl Acad Sci 41: of Parkinson’s disease? Prog Brain Res 184: 3–16. 144–150. Hauswirth WW, Laipis PJ. 1982. Mitochondrial DNA poly- King MP, Attardi G. 1989. Human cells lacking mtDNA: morphism in a maternal lineage of Holstein cows. Proc Repopulation with exogenous mitochondria by comple- Natl Acad Sci 79: 4686–4690. mentation. Science 246: 500–503. Holt IJ, Harding AE, Morgan-Hughes JA. 1988. Deletions of Koehler CM, Lindberg GL, Brown DR, Beitz DC, Freeman muscle mitochondrial DNA in patients with mitochon- AE, Mayfield JE, Myers AM. 1991. Replacement of bovine drial myopathies. Nature 331: 717–719. mitochondrial DNA by a sequence variant within one generation. Genetics 129: 247–255. Holt IJ, Harding AE, Petty RK, Morgan-Hughes JA. 1990. A new mitochondrial disease associated with mitochondri- Koopman WJ, Willems PH, Smeitink JA. 2012. Monogenic al DNA heteroplasmy. Am J Hum Genet 46: 428–433. mitochondrial disorders. N Engl J Med 366: 1132–1141. Huang TT, Naeemuddin M, Elchuri S, Yamaguchi M, Kozy Krakauer DC, Mira A. 1999. Mitochondria and germ-cell HM, Carlson EJ, Epstein CJ. 2006. Genetic modifiers of death. Nature 400: 125–126. the phenotype of mice deficient in mitochondrial super- Krysko DV, Agostinis P, Krysko O, Garg AD, Bachert C, oxide dismutase. Hum Mol Genet 15: 1187–1194. Lambrecht BN, Vandenabeele P. 2011. Emerging role of damage-associated molecular patterns derived from mi- Inoue K, Nakada K, Ogura A, Isobe K, Goto Y, Nonaka I, tochondria in inflammation. Trends Immunol 32: 157– Hayashi J-I. 2000. Generation of mice with mitochondri- 164. al dysfunction by introducing mouse mtDNA carrying a deletion into zygotes. Nat Genet 26: 176–181. Kukat C, Wurm CA, Spahr H, Falkenberg M, Larsson NG, Jakobs S. 2011. Super-resolution microscopy reveals that Jansen RP, de Boer K. 1998. The bottleneck: Mitochondrial mammalian mitochondrial nucleoids have a uniform imperatives in oogenesis and ovarian follicular fate. Mol size and frequently contain a single copy of mtDNA. Cell Endocrinol 145: 81–88. Proc Natl Acad Sci 108: 13534–13539. Jenuth JP,Peterson AC, Fu K, Shoubridge EA. 1996. Random Lee HS, Ma H, Juanes RC, Tachibana M, Sparman M, Wood- genetic drift in the female germline explains the rapid ward J, Ramsey C, Xu J, Kang EJ, Amato P, et al. 2012. segregation of mammalian mitochondrial DNA. Nat Ge- Rapid mitochondrial DNA segregation in primate pre- net 14: 146–151. implantation embryos precedes somatic and germline Jenuth JP, Peterson AC, Shoubridge EA. 1997. Tissue-spe- bottleneck. Cell Rep 1: 506–515. cific selection for different mtDNA genotypes in hetero- Leonard WR, Sorensen MV, Galloway VA, Spencer GJ, plasmic mice. Nat Genet 16: 93–95. Mosher MJ, Osipova L, Spitsyn VA.2002. Climatic influ- Ji F, Sharpley MS, Derbeneva O, Alves LS, Qian P, Wang Y, ences on basal metabolic rates among circumpolar pop- Chalkia D, Lvova M, Xu J, Yao W, et al. 2012. Mitochon- ulations. Am J Human Biol 14: 609–620. drial DNAvariant associated with Leber hereditary optic Liang M, Guan M, Zhao F,Zhou X, Yuan M, TongY,YangL, neuropathy and high-altitude Tibetans. Proc Natl Acad WeiQP,Sun YH, Lu F,et al. 2009. Leber’s hereditary optic Sci 109: 7391–7396. neuropathy is associated with mitochondrial ND1 Jokinen R, Battersby BJ. 2013. Insight into mammalian mi- T3394C mutation. Biochem Biophys Res Commun 383: tochondrial DNA segregation. Ann Med 45: 149–155. 286–292. Jokinen R, Marttinen P, Sandell HK, Manninen T, Teeren- Malena A, Loro E, Di Re M, Holt IJ, Vergani L. 2009. Inhi- hovi H, Wai T, Teoli D, Loredo-Osti JC, Shoubridge EA, bition of mitochondrial fission favours mutant over wild- Battersby BJ. 2010. Gimap3 regulates tissue-specific mi- type mitochondrial DNA. Hum Mol Genet 18: 3407– tochondrial DNA segregation. PLoS Genet 6: e1001161. 3416.

44 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

Mehrazin M, Shanske S, Kaufmann P,Wei Y, Coku J, Engel- tochondrial DNA genes. Proc Natl Acad Sci 84: 7580– stad K, Naini A, De Vivo DC, DiMauro S. 2009. Longi- 7584. tudinal changes of mtDNA A3243G mutation load and Oka T, Hikoso S, Yamaguchi O, Taneike M, Takeda T, Tamai level of functioning in MELAS. Am J Med Genet A 149A: T, Oyabu J, Murakawa T, Nakayama H, Nishida K, et al. 584–587. 2012. Mitochondrial DNA that escapes from autophagy Mishmar D, Ruiz-Pesini E, Mondragon-Palomino M, Pro- causes inflammation and heart failure. Nature 485: 251– caccio V,Gaut B, Wallace DC. 2006. Adaptive selection of 255. mitochondrial complex I subunits during primate radi- Oliver NA, Wallace DC. 1982. Assignment of two mito- ation. Gene 378: 11–18. chondrially synthesized polypeptides to human mito- Mitchell P. 1961. Coupling of phosphorylation to electron chondrial DNA and their use in the study of intracellular and hydrogen transfer by a chemi-osmotic type of mech- mitochondrial interaction. Mol Cell Biol 2: 30–41. anism. Nature 191: 144–148. Oliver NA, Greenberg BD, Wallace DC. 1983. Assignment of MITOMAP. 2012. A human mitochondrial genome data- a polymorphic polypeptide to the human mitochondrial base. http://www.mitomap.org. DNA unidentified reading frame 3 gene by a new peptide Monnot S, Gigarel N, Frydman N, Burlet P,Gobin S, Sinico mapping strategy. J Biol Chem 258: 5834–5839. M, Bonniere M, Rio M, Rotig A, Frydman R, et al. Olivo PD, Van de WalleMJ, Laipis PJ, Hauswirth WW.1983. 2009a. NARP and MELAS mutations differentially Nucleotide sequence evidence for rapid genotypic shifts impact mitochondrial DNA segregation throughout hu- in the bovine mitochondrial DNA D-loop. Nature 306: man embryofetal development. http://www.ashg.org/ 400–402. 2009meeting/abstracts/fulltext/f21163htm. Ortiz RG, Newman NJ, Shoffner JM, Kaufman AE, Koontz Monnot S, Gigarel N, Hesters L, Burlet P,Benachi A, Dumez DA, Wallace DC. 1993. Variable retinal and neurologic Y, Tachdjian G, Rotig A, Frydman R, Munnich A, et al. manifestations in patients harboring the mitochondrial 2009b. The challenge of prenatal and preimplantation DNA 8993 mutation. Arch Ophthalmol 111: 1525–1530. genetic diagnosis of mitochondrial DNA disorders. Eu- Paull D, Emmanuele V,WeissKA, Treff N, Stewart L, Hua H, ropean Human Genetics Conference, Abstract C07.4, Vi- Zimmer M, Kahler DJ, Goland RS, Noggle SA, et al. 2013. enna Austria, May 23–26, 2009. Eur J Hum Genet 17: 24. Nuclear genome transfer in human oocytes eliminates Monnot S, Gigarel N, Samuels DC, Burlet P,Hesters L, Fryd- mitochondrial DNA variants. Nature 493: 632–637. man N, Frydman R, Kerbrat V, Funalot B, Martinovic J, et al. 2011. Segregation of mtDNA throughout human Payne BA, Wilson IJ, Yu-Wai-Man P, Coxhead J, Deehan D, embryofetal development: m.3243A.G as a model sys- Horvath R, Taylor RW,Samuels DC, Santibanez-Koref M, tem. Hum Mutat 32: 116–125. Chinnery PF. 2013. Universal heteroplasmy of human mitochondrial DNA. Hum Mol Genet 22: 384–390. Monnot S, Samuels DC, Hesters L, Frydman N, Gigarel N, Burlet P, Kerbrat V, Lamazou F, Frydman R, Benachi A, Poe BG 3rd, Duffy CF, Greminger MA, Nelson BJ, Arriaga et al. 2013. Mutation dependence of the mitochondrial EA. 2010. Detection of heteroplasmy in individual mito- DNA copy number in the first stages of human embryo- chondrial particles. Anal Bioanal Chem 397: 3397–3407. genesis. Hum Mol Genet 22: 1867–1872. Potluri P, Davila A, Ruiz-Pesini E, Mishmar D, O’Hearn S, Moreno-Loshuertos R, Acin-Perez R, Fernandez-Silva P, Hancock S, Simon MC, Scheffler I, Wallace DC, Procac- Movilla N, Perez-Martos A, Rodriguez de Cordoba S, cio V. 2009. A novel NDUFA1 mutation leads to a pro- Gallardo ME, Enriquez JA. 2006. Differences in reactive gressive mitochondrial complex I-specific neurodegener- oxygen species production explain the phenotypes asso- ative disease. Mol Genet Metab 96: 189–195. ciated with common mouse mitochondrial DNA vari- Poulton J, Bredenoord AL. 2010. Applying pre-implanta- ants. Nat Genet 38: 1261–1268. tion genetic diagnosis to mtDNA diseases: Implications Nakada K, Sato A, Yoshida K, Morita T, Tanaka H, Inoue SI, of scientific advances. 174th ENMC International Work- Yonekawa H, Hayashi JI. 2006. Mitochondria-related shop, Naarden, The Netherlands, March 19–21, 2010. male infertility. Proc Natl Acad Sci 103: 15148–15153. Neuromuscul Disord 20: 559–563. Narendra D, Tanaka A, Suen DF, Youle RJ. 2008. Parkin is Rajasimha HK, Chinnery PF, Samuels DC. 2008. Selection recruited selectively to impaired mitochondria and pro- against pathogenic mtDNA mutations in a stem cell pop- motes their autophagy. J Cell Biol 183: 795–803. ulation leads to the loss of the 3243A!G mutation in Narendra D, Tanaka A, Suen DF, Youle RJ. 2009. Parkin- blood. Am J Hum Genet 82: 333–343. induced mitophagy in the pathogenesis of Parkinson dis- Rebelo AP, Dillon LM, Moraes CT. 2011. Mitochondrial ease. Autophagy 5: 706–708. DNA transcription regulation and nucleoid organiza- Narendra DP,Jin SM, Tanaka A, Suen DF, Gautier CA, Shen tion. J Inherit Metab Dis 34: 941–951. J, Cookson MR, Youle RJ. 2010. PINK1 is selectively sta- Rollins B, Martin MV, Sequeira PA, Moon EA, Morgan LZ, bilized on impaired mitochondria to activate Parkin. Watson SJ, Schatzberg A, Akil H, Myers RM, Jones EG, PLoS Biol 8: e1000298. et al. 2009. Mitochondrial variants in schizophrenia, bi- Narendra D, Walker JE, Youle R. 2012. Mitochondrial qual- polar disorder, and major depressive disorder. PLoS ONE ity control mediated by PINK1 and Parkin: Links to par- 4: e4913. kinsonism. Cold Spring Harb Perspect Biol 4: a011338. Ruiz-Pesini E, Wallace DC. 2006. Evidence for adaptive se- Neckelmann N, Li K, WadeRP,Shuster R, Wallace DC. 1987. lection acting on the tRNA and rRNA genes of the human cDNA sequence of a human skeletal muscle ADP/AT P mitochondrial DNA. Hum Mutat 27: 1072–1081. translocator: lack of a leader peptide, divergence from a Ruiz-Pesini E, Mishmar D, Brandon M, Procaccio V,Wallace fibroblast translocator cDNA, and coevolution with mi- DC. 2004. Effects of purifying and adaptive selection on

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 45 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

D.C. Wallace and D. Chalkia

regional variation in human mtDNA. Science 303: 223– Stephens M. 1974. EDF statistics for goodness of fit and 226. some comparisons. J Am Stat Assoc 69: 730–737. Sadun AA, Carelli V,Salomao SR, Berezovsky A, Quiros PA, Stewart JB, Freyer C, Elson JL, Wredenberg A, Cansu Z, Sadun F, DeNegri AM, Andrade R, Moraes M, Passos A, Trifunovic A, Larsson NG. 2008. Strong purifying selec- et al. 2003. Extensive investigation of a large Brazilian tion in transmission of mammalian mitochondrial DNA. pedigree of 11778/haplogroup J Leber hereditary optic PLoS Biol 6: e10. neuropathy. Am J Ophthalmol 136: 231–238. Strauss KA, Dubiner L, Simon M, Zaragoza M, Sengupta PP, Sadun AA, La Morgia C, Carelli V. 2011. Leber’s Hereditary Li P, Narula N, Dreike S, Platt J, Procaccio V, et al. 2013. Optic Neuropathy. Curr Treat Options Neurol 13: 109– Severity of cardiomyopathy associated with adenine nu- 117. cleotide translocator-1 deficiency correlates with mtDNA Samuels DC, Wonnapinij P, Cree LM, Chinnery PF. 2010. haplogroup. Proc Natl Acad Sci 110: 3253–3458. Reassessing evidence for a postnatal mitochondrial ge- Suen DF, Narendra DP, Tanaka A, Manfredi G, Youle RJ. netic bottleneck. Nat Genet 42: 471–473. 2010. Parkin overexpression selects against a deleterious Samuels DC, Wonnapinij P, Chinnery PF. 2013. Preventing mtDNA mutation in heteroplasmic cybrid cells. Proc Natl the transmission of pathogenic mitochondrial DNA mu- Acad Sci 107: 11835–11840. tations: Can we achieve long-term benefits from germ- Tachibana M, Sparman M, Sritanaudomchai H, Ma H, line gene transfer? Hum Reprod 28: 554–559. Clepper L, Woodward J, Li Y, Ramsey C, Kolotushkina Sato A, Kono T, Nakada K, Ishikawa K, Inoue S, Yonekawa O, Mitalipov S. 2009. Mitochondrial gene replacement in H, Hayashi J. 2005. Gene therapy for progeny of mito- primate offspring and embryonic stem cells. Nature 461: mice carrying pathogenic mtDNA by nuclear transplan- 367–372. tation. Proc Natl Acad Sci 102: 16765–16770. Tachibana M, Amato P, Sparman M, Woodward J, Sanchis Sato A, Nakada K, Shitara H, Kasahara A, Yonekawa H, DM, Ma H, Gutierrez NM, Tippner-Hedges R, Kang E, Hayashi J. 2007. Deletion-mutant mtDNA increases in Lee HS, et al. 2013. Towards germline gene therapy of somatic tissues but decreases in female germ cells with inherited mitochondrial diseases. Nature 493: 627–631. age. Genetics 177: 2031–2037. Tajima H, Sueoka K, Moon SY, Nakabayashi A, Sakurai T, Murakoshi Y, Watanabe H, Iwata S, Hashiba T, Kato S, Schaefer AM, Taylor RW, Turnbull DM, Chinnery PF. 2004. et al. 2007. The development of novel quantification assay The epidemiology of mitochondrial disorders–past, for mitochondrial DNA heteroplasmy aimed at preim- present and future. Biochim Biophys Acta 1659: 115–120. plantation genetic diagnosis of Leigh encephalopathy. J Schaefer AM, McFarland R, Blakely EL, He L, Whittaker RG, Assist Reprod Genet 24: 227–232. Taylor RW, Chinnery PF, Turnbull DM. 2008. Prevalence Takahata N, Slatkin M. 1983. Evolutionary dynamics of ex- of mitochondrial DNA disease in adults. Ann Neurol 63: tranuclear genes. Genet Res 42: 257–268. 35–39. Tatuch Y, Christodoulou J, Feigenbaum A, Clarke JTR, Schmitt MW,Kennedy SR, Salk JJ, Fox EJ, Hiatt JB, Loeb LA. Wherret J, Smith C, Rudd N, Petrova-Benedict R, Rob- 2012. Detection of ultra-rare mutations by next-genera- inson BH. 1992. Heteroplasmic mtDNA mutation (T-G) tion sequencing. Proc Natl Acad Sci 109: 14508–14513. at 8993 can cause Leigh disease when the percentage of Sharpley MS, Marciniak C, Eckel-Mahan K, McManus MJ, abnormal mtDNA is high. Am J Hum Genet 50: 852–858. Crimi M, Waymire K, Lin CS, Masubuchi S, Friend N, Thorburn DR, Dahl HH. 2001. Mitochondrial disorders: Koike M, et al. 2012. Heteroplasmy of mouse mtDNA is Genetics, counseling, prenatal diagnosis and reproduc- genetically unstable and results in altered behavior and tive options. Am J Med Genet 106: 102–114. cognition. Cell 151: 333–343. Thorburn D, Wilton L, Stock-Myer S. 2009. Healthy baby Shoffner JM, Lott MT, Voljavec AS, Soueidan SA, Costigan girl born following preimplantation genetic diagnosis for DA, Wallace DC. 1989. Spontaneous Kearns-Sayre/ mitochondrial DNA m.8993T.G mutation. 11th Inter- chronic external ophthalmoplegia plus syndrome associ- national Congress of Inborn Errors of Metabolism, Abstract ated with a mitochondrial DNA deletion: A slip-replica- 117, Platform Presentation 3: Mitochondrial Disorders. tion model and metabolic therapy. Proc Natl Acad Sci 86: Mol Genet Metab 98: 5–6. 7952–7956. Torroni A, Petrozzi M, D’Urbano L, Sellitto D, Zeviani M, Shoffner JM, Lott MT, Lezza AM, Seibel P, Ballinger SW, Carrara F,Carducci C, Leuzzi V,Carelli V,Barboni P,et al. Wallace DC. 1990. Myoclonic epilepsy and ragged-red 1997. Haplotype and phylogenetic analyses suggest that fiber disease (MERRF) is associated with a mitochondrial Lys one European-specific mtDNA background plays a role DNA tRNA mutation. Cell 61: 931–937. in the expression of Leber hereditary optic neuropathy by Snodgrass JJ, Leonard WR, Tarskaia LA, Alekseev VP, Kri- increasing the penetrance of the primary mutations voshapkin VG. 2005. Basal metabolic rate in the Yakut 11778 and 14484. Am J Hum Genet 60: 1107–1121. (Sakha) of Siberia. Am J Human Biol 17: 155–172. Toye AA, Lippiat JD, Proks P, Shimomura K, Bentley L, Snodgrass JJ, Leonard WR, Sorensen MV, Tarskaia LA, Hugill A, Mijat V, Goldsworthy M, Moir L, Haynes A, Mosher MJ. 2008. The influence of basal metabolic rate et al. 2005. A genetic and physiological study of impaired on blood pressure among indigenous Siberians. Am J glucose homeostasis control in C57BL/6 J mice. Diabe- Phys Anthropol 137: 145–155. tologia 48: 675–686. Solignac M, Genermont J, Monnerot M, Mounolou JC. Treff NR, Campos J, Tao X, Levy B, Ferry KM, Scott RT Jr. 1984. Genetics of mitochondria in Drosophila mtDNA 2012. Blastocyst preimplantation genetic diagnosis inheritance in heteroplasmic strains of Drosophila maur- (PGD) of a mitochondrial DNA disorder. Fertil Steril itiana. Mol Gen Genet 197: 183–188. 98: 1236–1240.

46 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics

van den Ouweland JM, Lemkes HHP, Ruitenbeek W, Sand- human and bovine ATP synthase b-subunit: Mitochon- kjujl LA, deVijlder MF, Struyvenberg PAA, van de Kamp drial DNA genes sustain seventeen times more muta- JJP, Maassen JA. 1992. Mutation in mitochondrial tions. Curr Genet 12: 81–90. Leu(UUR) tRNA gene in a large pedigree with maternally Wallace DC, Singh G, Lott MT, Hodge JA, Schurr TG, Lezza transmitted type II diabetes mellitus and deafness. Nat AM, Elsas LJ, Nikoskelainen EK. 1988a. Mitochondrial Genet 1: 368–371. DNA mutation associated with Leber’s hereditary optic Wai T, Teoli D, Shoubridge EA. 2008. The mitochondrial neuropathy. Science 242: 1427–1430. DNA genetic bottleneck results from replication of a sub- Wallace DC, Zheng X, Lott MT, Shoffner JM, Hodge JA, population of genomes. Nat Genet 40: 1484–1488. Kelley RI, Epstein CM, Hopkins LC. 1988b. Familial mi- Wallace DC. 1981. Assignment of the chloramphenicol re- tochondrial encephalomyopathy (MERRF): Genetic, sistance gene to mitochondrial deoxyribonucelic acid pathophysiological, and biochemical characterization of and analysis of its expression in cultured human cells. a mitochondrial DNA disease. Cell 55: 601–610. Mol Cell Biol 1: 697–710. Wallace DC, Brown MD, Lott MT. 1999. Mitochondrial Wallace DC. 1982. Cytoplasmic inheritance of chloram- DNA variation in human evolution and disease. Gene phenicol resistance in mammalian cells. In Techniques 238: 211–230. in somatic cell genetics (ed. Shay JW), Chap. 12. pp. Wallace DC, Fan W, Procaccio V.2010. Mitochondrial ener- 159–187. Plenum, New York. getics and therapeutics. Annu Rev Path 5: 297–348. Wallace DC. 1986. Mitotic segregation of mitochondrial Wallace DC, Lott MT, Procaccio V. 2013. Mitochondrial DNAs in human cell hybrids and expression of chloram- medicine: The mitochondrial biology and genetics of phenicol resistance. Somat Cell Mol Genet 12: 41–49. metabolic and degenerative diseases, cancer, and aging. Wallace DC. 1987. Maternal genes: Mitochondrial diseases. In Emery and Rimoin’s principles and practice of medical In Medical and experimental mammalian genetics: A per- genetics (ed. Rimoin DL, Pyeritz RE, Korf BR). Churchill spective (ed. McKusick VA, Roderick TH, Mori J, Paul Livingstone, Philadelphia. MW), pp. 137–190. A.R. Liss, New York. Wonnapinij P, Chinnery PF, Samuels DC. 2008. The distri- Wallace DC. 2005. A mitochondrial paradigm of metabolic bution of mitochondrial DNA heteroplasmy due to ran- and degenerative diseases, aging, and cancer: A dawn for dom genetic drift. Am J Hum Genet 83: 582–593. evolutionary medicine. Annu Rev Genet 39: 359–407. Wonnapinij P, Chinnery PF, Samuels DC. 2010. Previous Wallace DC. 2007. Why do we have a maternally inherited estimates of mitochondrial DNA mutation level variance mitochondrial DNA? Insights from Evolutionary Medi- did not account for sampling error: comparing the cine. Annu Rev Biochem 76: 781–821. mtDNA genetic bottleneck in mice and humans. Am J Wallace DC. 2008. Mitochondria as chi. Genetics 179: 727– Hum Genet 86: 540–550. 735. Wright S. 1969. Theory of gene frequencies. In Evolution and Wallace DC. 2011. Bioenergetic origins of complexity and the genetics of populations, University of Chicago Press, disease. Cold Spring Harb Symp Quant Biol 76: 1–16. Chicago. Wallace DC. 2012. Mitochondria and cancer. Nat Rev Can- Yoneda M, Miyatake TGA. 1994. Complementation of mu- cer 12: 685–698. tant and wild-type human mitochondrial DNAs coexist- Wallace DC. 2013a. Bioenergetics in human evolution and ing since the mutation event and lack of complementa- disease: Implications for the origins of biological com- tion of DNAs introduced separately into a cell within plexity and the missing genetic variation of common distinct organelles. Mol Cell Biol 14: 2699–2712. diseases. Philos Trans R Soc Lond B Biol Sci doi: Yoneda M, Chomyn A, Martinuzzi A, Hurko O, Attardi G. 10.1098/rstb.2012.0267. 1992. Marked replicative advantage of human mtDNA Wallace DC. 2013b. Mitochondrial bioenergetic etiology of carrying a point mutation that causes the MELAS ence- disease. J Clin Invest 123: 1405–1412. phalomyopathy. Proc Natl Acad Sci 89: 11164–11168. Wallace DC, Fan W. 2009. The pathophysiology of mito- Yoon KL, Ernst SG, Rasmussen C, Dooling EC, Aprille JR. chondrial disease as modeled in the mouse. Genes Dev 1993. Mitochondrial disorder associated with newborn 23: 1714–1736. cardiopulmonary arrest. Pediatr Res 33: 433–440. Wallace DC, Fan W. 2010. Energetics, epigenetics, mito- Youle RJ, van der Bliek AM. 2012. Mitochondrial fission, chondrial genetics. Mitochondrion 10: 12–31. fusion, and stress. Science 337: 1062–1105. Wallace DC, Bunn CL, Eisenstadt JM. 1975. Cytoplasmic Zhang Q, Itagaki K, Hauser CJ. 2010a. Mitochondrial DNA transfer of chloramphenicol resistance in human tissue is released by shock and activates neutrophils via p38 map culture cells. J Cell Biol 67: 174–188. kinase. Shock 34: 55–559. Wallace DC, Pollack Y, Bunn CL, Eisenstadt JM. 1976. Cy- Zhang Q, Raoof M, Chen Y, Sumi Y, Sursal T, Junger W, toplasmic inheritance in mammalian tissue culture cells. Brohi K, Itagaki K, Hauser CJ. 2010b. Circulating mito- In Vitro 12: 758–776. chondrial DAMPs cause inflammatory responses to in- Wallace DC, Bunn CL, Eisenstadt JM. 1977. Mitotic segre- jury. Nature 464: 104–107. gation of cytoplasmic inherited genes for chloramphen- Zhou RR, Wang B, Wang J, Schatten H, Zhang YZ. 2010. Is icol resistance in mammalian cells. II: Fusions with hu- the mitochondrial cloud the selection machinery for man cell lines. Somatic Cell Genet 3: 93–119. preferentially transmitting wild-type mtDNA between Wallace DC, Ye JH, Neckelmann SN, Singh G, Webster KA, generations? Rewinding Muller’s ratchet efficiently. Curr Greenberg BD. 1987. Sequence analysis of cDNAs for the Genet 56: 101–107.

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a021220 47 Downloaded from http://cshperspectives.cshlp.org/ on September 25, 2021 - Published by Cold Spring Harbor Laboratory Press

Mitochondrial DNA Genetics and the Heteroplasmy Conundrum in Evolution and Disease

Douglas C. Wallace and Dimitra Chalkia

Cold Spring Harb Perspect Biol 2013; doi: 10.1101/cshperspect.a021220

Subject Collection Mitochondria

Altered Sulfide (H2S) Metabolism in Ethylmalonic Where Killers Meet−−Permeabilization of the Outer Encephalopathy Mitochondrial Membrane during Apoptosis Valeria Tiranti and Massimo Zeviani Tom Bender and Jean-Claude Martinou Mitochondrial DNA Genetics and the Mitochondrial Biogenesis through Activation of Heteroplasmy Conundrum in Evolution and Nuclear Signaling Proteins Disease John E. Dominy and Pere Puigserver Douglas C. Wallace and Dimitra Chalkia The Role of Mitochondria in Cellular Iron−Sulfur Mitochondrial Trafficking in Neurons Protein Biogenesis: Mechanisms, Connected Thomas L. Schwarz Processes, and Diseases Oliver Stehling and Roland Lill Mechanisms of Mitochondrial Fission and Fusion Mitochondrial Dysfunction and Defective Alexander M. van der Bliek, Qinfang Shen and Autophagy in the Pathogenesis of Collagen VI Sumihiro Kawajiri Muscular Dystrophies Paolo Bernardi and Paolo Bonaldo The Mitochondrial Nucleoid: Integrating Clinical and Molecular Features of POLG-Related Mitochondrial DNA into Cellular Homeostasis Mitochondrial Disease Robert Gilkerson, Liliana Bravo, Iraselia Garcia, et Jeffrey D. Stumpf, Russell P. Saneto and William C. al. Copeland Relevance of Mitochondrial Genetics and Mitochondrial Metabolism, Sirtuins, and Aging Metabolism in Cancer Development Michael N. Sack and Toren Finkel Giuseppe Gasparre, Anna Maria Porcelli, Giorgio Lenaz, et al. Mitochondrial Quality Control Mediated by PINK1 Mechanisms of Protein Sorting in Mitochondria and Parkin: Links to Parkinsonism Diana Stojanovski, Maria Bohnert, Nikolaus Derek Narendra, John E. Walker and Richard Youle Pfanner, et al.

For additional articles in this collection, see http://cshperspectives.cshlp.org/cgi/collection/

Mitochondrial Evolution Michael W. Gray

For additional articles in this collection, see http://cshperspectives.cshlp.org/cgi/collection/