Co-amplification of mitochondrial pseudogenes in Calomys musculinus (Rodentia, Cricetidae): a source of error in phylogeographic studies

73

NOTE / NOTE

Co-amplification of mitochondrial pseudogenes in Calomys musculinus (Rodentia, Cricetidae): a source of error in phylogeographic studies Rau´l E. Gonza´lez-Ittig and Cristina N. Gardenal

Abstract: In a previous phylogeographic study of the rodent Calomys musculinus, 24 haplotypes of the mitochondrial DNA D-loop region were detected using the restriction fragment length polymorphism technique (PCR-RFLP). Seven percent of the individuals showed patterns in which the sum of the sizes of the restriction fragments exceeded the size of the original PCR product. In the present paper we analyze possible causes of these atypical haplotypes. PCR products were cloned, and two or three different clones from a single individual were detected by their RFLP patterns. Nine clones with different restriction patterns were selected for sequence analyses. A maximum parsimony phylogenetic analysis revealed two well-supported paraphyletic groups. One group comprised sequences showing low nucleotide divergence compared with the most common haplotypes detected in the phylogeographic study. The other group was basal to the three species of Calomys other than C. musculinus included in the study; the mutations in the short portion of the cytochrome b gene amplified corresponded to 12 amino acid substitutions. The results suggest that two independent insertions of mtDNA sequences into the nucleus occurred; these sequences would co-amplify in the PCR procedure. Identification of pseudogenes is crucial to obtain reliable reconstruction of the intraspecific genealogy in phylogeographic studies. Key words: Calomys musculinus, phylogeography, phylogeny, mtDNA D-loop region, pseudogenes. Re´sume´ : Au cours d’une e´tude phyloge´ographique ante´rieure chez le rongeur Calomys musculinus, 24 haplotypes pour la re´gion en boucle D de l’ADN mitochondrial ont e´te´ de´tecte´s par PCR-RFLP (polymorphisme de taille des fragments de restriction sur produits PCR). Sept pour cent des individus affichaient des profils chez lesquels la somme des fragments exce´dait la taille de l’amplicon original. Dans le pre´sent travail, les auteurs ont examine´ les causes possibles de ces haplotypes atypiques. Les amplicons ont e´te´ clone´s et deux ou trois clones diffe´rents provenant d’un meˆme individu ont e´te´ de´tecte´s sur la base de leur profil RFLP. Neuf clones montrant des profils diffe´rents ont e´te´ se´quence´s. Une analyse phyloge´ne´tique base´e sur la parcimonie a re´ve´le´ deux groupes paraphyle´tiques bien supporte´s. L’un de ceux-ci comprenait des se´quences montrant une faible divergence nucle´otidique par rapport aux haplotypes les plus communs de´tecte´s au cours de l’analyse phyloge´ne´tique. L’autre groupe e´tait ancestral aux trois espe`ces de Calomys e´tudie´es autres que le C. musculinus. Les mutations situe´es dans la petite portion amplifie´e du ge`ne codant pour le cytochrome b correspondaient a` 12 substitutions d’acides amine´s. Ces re´sultats sugge`rent que deux insertions inde´pendantes de l’ADNmt se seraient produites dans le noyau et que ces se´quences amplifient toutes deux lors de la PCR. L’identification de pseudoge`nes est cruciale en vue de la reconstruction fide`le de la ge´ne´alogie intraspe´cifique lors d’e´tudes phyloge´ographiques. Mots-cle´s : Calomys musculinus, phyloge´ographie, phyloge´nie, ADNmt, re´gion en boucle D, pseudoge`nes. [Traduit par la Re´daction]

Phylogeographic approaches based on the analysis of spatial distribution of intraspecific gene phylogenies have been Received 21 May 2007. Accepted 30 October 2007. Published on the NRC Research Press Web site at genome.nrc.ca on 21 December 2007. Corresponding Editor: L. Bonen. R.E. Gonza´lez-Ittig and C.N. Gardenal.1 Ca´tedra de Gene´tica de Poblaciones y Evolucio´n, Facultad de Ciencias Exactas, Fı´sicas y Naturales, Universidad Nacional de Co´rdoba, Ve´lez Sarsfield 299, 5000 Co´rdoba, Argentina. 1Corresponding

author (e-mail: [email protected]).

Genome 51: 73–78 (2008)

extensively used in studies of population genetic structure. In animals, the molecule most widely used for phylogenetic inferences is the mitochondrial DNA (mtDNA) because it usually does not recombine, is maternally inherited, and has a high mutation rate, which produces useful levels of intraspecific polymorphism (Avise 2000). Regardless of the advantages of mtDNA in these studies, some caution must be taken when it is used for phylogenetic purposes. In some species the presence of two or more types of mtDNA in one individual, a phenomenon called ‘‘heteroplasmy’’, has been described; it can be caused by point mutations in a matrilineal line resulting in different genetic variants, by occasional biparental transmission (i.e., Zhao et

doi:10.1139/G07-104

#

2007 NRC Canada

74

al. 2004), or by the systematic mechanism of doubly uniparental inheritance of mtDNA, as observed in several species of Mytilus (Sutherland et al. 1998). Heteroplasmy is often not reported in phylogeographic studies, but it could have serious consequences for the reconstruction of the evolutionary history (Hagelberg 2003). Another source of error is the potential co-amplification of mtDNA with mitochondrial pseudogenes integrated into the nucleus (Numts). Zullo et al. (1991) detected, in the domestic rat, the presence of 2 nuclear fragments of 500 bp showing 80% identity with the mtDNA D-loop region. DeWoody et al. (1999) and Triant and DeWoody (2007a) reported at least 3 independent translocations and duplications of Numts in Microtus rossiaemeridionalis (Rodentia); one of them was present in several species of the genus Microtus (Triant and DeWoody 2007b). Mirol et al. (2000) described the presence of 4 different pseudogenes in rodents of the genus Ctenomys (Rodentia), whose divergence with normal mtDNA varied between 10% and 32%. In a previous study, we performed a phylogeographic analysis of the vesper mice Calomys musculinus (Sigmodontinae, Cricetidae) (Gonza´lez-Ittig and Gardenal 2004). On the basis of restriction site data of the non-coding mtDNA control region (D-loop), we detected 24 haplotypes in specimens from 16 localities; restriction sites were inferred from patterns of restriction fragment length polymorphisms (RFLP–PCR). Some individuals (7%) presented haplotypes in which the sum of the sizes of the restriction fragments exceeded the size of the PCR product; these cases were not included in that study because they hindered the phylogeographic inference. The purpose of the present work is to identify possible causes of those atypical haplotypes and to highlight the consequences of their inclusion in intraspecific phylogenetic studies. From a total of 256 individuals of C. musculinus of Argentina used in the phylogeographic study (Gonza´lez-Ittig 2004), we selected the following individuals because they presented restriction patterns in which the sum of the sizes of the fragments was larger than the size of the PCR product: Mo50 from Molinari (Cordoba Province), LL983 from Laguna Larga (Cordoba Province), U21712 from Uranga (Santa Fe Province), A19297 and A19299 from Alcorta (Santa Fe Province), SL A13 from Donovan (San Luis Province), and SN19630 from San Nicola´s (Buenos Aires Province). DNA extraction and PCR conditions for amplification of the mtDNA D-loop region were as described by Gonza´lez-Ittig and Gardenal (2002). A segment of approximately 1.3 kb was obtained, containing the complete control region, a short portion of the cytochrome b (Cytb) and 12S rRNA genes, and the flanking proline and phenylalanine tRNAs. To eliminate sample contamination, we made at least 3 independent DNA extractions from liver and kidney of each individual, using different stocks of tubes for each procedure. Identical RFLP patterns of the amplified PCR product were obtained in all replicates of the same sample. The PCR product from single individuals was inserted in pCR1II-TOPO1 plasmids (Invitrogen, Carlsbad, California) and transformed into competent Escherichia coli DH5a cells. The presence of inserts was confirmed by a new amplification with the same primers used for the entire D-loop region. After obtaining the restriction pattern of each clone,

Genome Vol. 51, 2008

the restriction sites were inferred. Plasmids showing different restriction patterns in the D-loop insertion were selected and used as templates in the DNA sequencing reactions, performed at Macrogen Inc. (Seoul, Korea) with an ABI PRISM1 3700 DNA analyzer (PE Applied Biosystems, Foster City, California, USA). Sequences were aligned with ClustalX (Thompson et al. 1997) and corrected by eye. Phylogenetic analysis was performed using the PHYLIP version 3.66 package (Felsenstein 2006). An heuristic search, carried out with 1000 bootstrap replicates, was used to construct a consensus phylogenetic tree. We included homologous sequences of 3 other species of the genus Calomys (tribe Phyllotini): C. callidus, C. venustus, and C. laucha (accession Nos. DQ926660, DQ926661, and DQ926662, respectively). We used as outgroups sequences of Oligoryzomys longicaudatus and Pseudoryzomys simplex (acc. Nos. AY863416 and AY863422), two species of the tribe Oryzomyini, which is basal to the tribe Phyllotini (Smith and Patton 1999). Genetic distances between the sequences were calculated according to the Kimura 2-parameter model (Kimura 1980). Of the 37 clones recovered from individual LL983, 36 showed pattern A for the enzymes HaeIII, AseI, and AluI and pattern B for NlaIII. The remaining clone, identified as LL983 A17, presented different restriction patterns for the 4 enzymes. For each enzyme, when pattern A or B was superimposed with the pattern of clone LL983 A17, the original pattern was reconstituted (Figs. 1a, 1b). From the cloned PCR product of individual U21712, a total of 245 clones presented pattern B for RsaI and pattern A for AseI, AluI, Tsp509I, and MseI (Fig. 1); one clone had pattern B for MseI, pattern C for RsaI, and pattern C for Tsp509I. Three clones (U21712 C151, U21712 C198, and U21712 C247) showed the same patterns as LL983 A17 for AseI and AluI and one pattern not detected before for Tsp509I. For this last enzyme, the combination of patterns A and C and the new one reconstituted the original pattern B (Fig. 1c). Also, the combination of patterns A and B of the enzyme MseI reconstituted the original pattern D (Fig. 1d). From individual SN19630, 60 clones were recovered with the restriction pattern A for NlaIII and a new one for RsaI. The clone SN19630 B45 presented a new restriction pattern for both enzymes. However, it was not possible to isolate a third clone that, combined with the two clones found, would reconstitute the original restriction pattern. Thirty-three of the 58 clones isolated from individual SL A13 presented pattern C and 25 showed pattern B for the enzyme RsaI (Fig. 1e). In summary, using this cloning procedure we could obtain different restriction patterns from PCR products of the same individual and were able to determine that from some of those original patterns, new ones or combinations of others already described could be originated (Fig. 1). To confirm the inferred restriction sites for most clones, and because of the inability to do so in others, the following clones were selected for sequencing: SN19630 B11, SN19630 B45, LL983 A17, LL983 A21, U21712 C10, U21712 C151, A19299 E22, A19297 F178, and Mo50 D3 (acc. Nos. DQ029205 to DQ029213). For the majority of the clones, sequence analysis confirmed the restriction sites inferred from electrophoretic patterns; the exceptions were 2 clones showing very atypical restriction patterns (clones #

2007 NRC Canada

Gonza´lez-Ittig and Gardenal

75

Fig. 1. Restriction patterns of different clones reconstitute the original pattern when superimposed. (a) Lanes 1 to 3: clone presenting pattern A, the original pattern B, and pattern of clone LL983 A17 for enzyme AseI. Lanes 4 to 6: clone with pattern B, the original pattern C, and pattern of clone LL983 A17 for enzyme NlaIII. (b) Lanes 1 to 3: clone with pattern A, the original pattern B, and pattern of clone LL983 A17 for enzyme HaeIII. Lanes 4 to 6: clone with pattern A, the original pattern B, and pattern of clone LL983 A17 for enzyme AluI. (c) Lanes 1 to 4: pattern of clone U21712 C151, clone with pattern C, clone with pattern A, and the original pattern B for enzyme Tsp509I. (d) Lanes 1 to 3: the original pattern D, clone with pattern A, and clone with pattern B for enzyme MseI. (e) Lanes 1 to 3: clone with pattern B, clone with pattern C, and the original pattern D for enzyme RsaI. M, 100 bp ladder.

LL983 A17 and U21712 C151). According to genetic distances based on the Kimura 2-parameter model, the sequences formed two different groups. Group 1 consisted of 7 sequences diverging by 0.5% to 1.5%. The sizes of the PCR products ranged between 1314 and 1319 bp. These sequences represent the two most common restriction patterns detected in the phylogeographic study (haplotypes 1 and 2) or differ from them by 1 or 2 restriction sites. Group 2 was formed by sequences of the two very atypical clones LL983 A17 and U21712 C151, which diverged by 0.8%. The sizes of the PCR products were 1284 and 1285 bp, respectively, the difference being caused by several short deletions. Genetic distances among clones of the two groups ranged from

31% to 33%. The enzyme AluI is ‘‘diagnostic’’ of the two types of sequences: pattern A characterizes all sequences of group 1 and pattern B characterizes those of group 2 (Fig. 1b). The phylogenetic tree based on the complete sequences of PCR products (Fig. 2) shows a similar picture: sequences are distributed in two paraphyletic clusters, the main one being formed by the 7 sequences of the group 1 defined by Kimura 2-parameter distances; this group is more closely related to the cluster formed by C. laucha, C. callidus, and C. venustus than to group 2, which is situated in a basal position. These relationships, estimated on the basis of the D-loop region, show a general correspondence with the #

2007 NRC Canada

76

Genome Vol. 51, 2008

Fig. 2. Maximum parsimony consensus tree based on the complete mtDNA D-loop sequences of nine clones of Calomys musculinus. An heuristic search was used for the analysis; 1000 bootstrap replicates and 10 random sequence additions for each replicate were performed. Only values higher than 50% are indicated. Sequences of Oligoryzomys longicaudatus and Pseudoryzomys simplex were used as outgroups.

Table 1. Alignment of amino acids corresponding to 110 bp of the Cytb gene of Calomys musculinus mtDNA. Clone A19299 E22 A19297 F178 Mo50 D3 SN19630 B11 SN19630 B45 LL983 A21 U21712 C10 LL983 A17 U21712 C151

Amino acid sequence PFTMIGQISSILYFSIIVIFMPVASMIENNILKLHY PFTMIGQISSILYFSIIVIFMPVASMIENNILKLHY PFTMIGQISSILYFSIIVIFMPVASMIENNILKLHY PFTMIGQISSILYFSIIVIFMPIASMIENNILKLHY PFTMIGQISSILYFSIIVIFMPIASMIENNILKLHY PFTMIGQISSILYFSIIVIFMPVASMIENNILKLHY PFTMIGQISSILYFSIIVIFMPVASMIENNILKLHY TFMMTGQIVSLSYFTIIVIFMPITSMMKNDILKLHA TFMMTGQIVSLSYFTIIVIFMPITSMMKNDILKLHA

Note: The last two entries correspond to group 2 sequences. Differences are indicated in boldface.

phylogenies of the genus obtained by Salazar-Bravo et al. (2001), Almeida et al. (2007), and Haag et al. (2007) using the Cytb gene. The complete sequence of the PCR product includes approximately 110 bp of the Cytb gene. In this short region we translated the codons of the different clones into amino acids (Table 1); 13 differences can be observed when groups 1 and 2 are compared. Most of the mutations in group 2 produced amino acid substitutions, which could generate nonfunctional proteins. The results presented here could be explained by assuming at least two independent insertions of mtDNA sequences (Numts) into the nucleus of C. musculinus: one translocation event would be relatively recent (group 1), comprising sequences that would not have had enough time to accumulate significant amounts of mutations. The other group of sequences would represent ‘‘old’’ translocations. In rodents of

the genus Microtus, Triant and DeWoody (2007a) observed at least three independent insertions of Numts; two of them were ‘‘old’’ but the third was very recent and led the authors to indicate that integration of mitochondrial genes into the nucleus would be an ongoing evolutionary process. A similar phenomenon was described by Mirol et al. (2000) in the genus Ctenomys (Rodentia), where 4 independent translocations were inferred. The group of less divergent sequences detected in our study could also be interpreted as the result of heteroplasmy. In one individual (SL A13) we recovered patterns B and C (Fig. 1d), which represent the two most common haplotypes (1 and 2); matings among individuals with haplotypes 1 and 2 would be highly probable and, if the egg’s mechanism of mitochondrial selection failed, individuals with both haplotypes could be generated. Although this observation suggests the possibility of biparental inheritance, the methodology #

2007 NRC Canada

Gonza´lez-Ittig and Gardenal

used in the present study does not allow discrimination between heteroplasmy and very recent translocations of Numts. The high number of amino acid substitutions in the sequences of clones LL983 A17 and U21712 C151 supports the idea that they represent a basal group of pseudogenes, since non-synonymous substitutions would accumulate when a DNA segment has lost functional restrictions. These sequences appear even more ancestral than the homologous sequences in 3 other species of the genus Calomys. As Bensasson et al. (2001) stated, some Numts can resemble ancestral mitochondrial sequences or molecular ‘‘fossils’’, given that the substitution rate within the nuclear genome is less than that of mtDNA. Because pseudogenes generally have different evolutionary histories, their inclusion in haplotype networks in phylogeographic studies can produce distortions in the genealogy of genes (Bensasson et al. 2001; Triant and DeWoody 2007c). In our first analysis of C. musculinus, based on all the restriction patterns obtained without taking into account whether some of them represented nuclear pseudogenes, the haplotype network showed important genetic gaps, with the main lineages spread over a wide area. This phylogeographic pattern would correspond to a species in which genes evolved separately during a long period owing to an ancestral geographic fragmentation, and then a recent secondary admixture occurred (model II; Avise 2000). However, when the most divergent pseudogenes (group 2) were identified and eliminated from the analysis, clones of group 1 were recognized, and each clone was treated as a separate sample, a very different scenario emerged (model V according to Avise 2000). A recent range expansion with low to moderate contemporary gene flow was inferred, which is in line with evidence from data using other molecular markers regarding the genetic structure of C. musculinus populations (Chiappero et al. 2002; Chiappero and Gardenal 2003). The identification of these divergent sequences will prevent the inclusion of Numts in phylogenetic networks, favoring a reliable reconstruction of the haplotype genealogy.

Acknowledgements This work was supported by the Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas (CONICET), Argentina, and the Secretarı´a de Ciencia y Tecnologı´a (SECYT), Universidad Nacional de Co´rdoba, Argentina. We appreciate advice on the analysis of some of the data given by Erika Hagelberg, Nestor Bianchi, and Luis Conci. The critical revision of the manuscript by Antonio Blanco and by Peter Larsen is warmly appreciated. R.E.G.I. is a Postdoctoral Fellow and C.N.G. is a Career Investigator of CONICET.

References Almeida, F.C., Bonvicino, C.R., and Cordeiro-Estrela, P. 2007. Phylogeny and temporal diversification of Calomys (Rodentia, Sigmodontinae): implications for the biogeography of an endemic genus of the open/dry biomes of South America. Mol. Phylogenet. Evol. 42: 449–466. doi:10.1016/j.ympev.2006.07.005. PMID:16971145. Avise, J.C. 2000. Phylogeography. The history and formation of species. Harvard University Press, Cambridge, Mass. Bensasson, D., Zhang, D.X., Hartl, D.L., and Hewitt, G.M. 2001.

77 Mitochondrial pseudogenes: evolution’s misplaced witnesses. Trends Ecol. Evol. 16: 314–321. doi:10.1016/S0169-5347(01) 02151-6. PMID:11369110. Chiappero, M.B., and Gardenal, C.N. 2003. Restricted gene flow in Calomys musculinus (Rodentia, Muridae), the natural reservoir of Junin virus. J. Hered. 94: 490–495. doi:10.1093/jhered/ esg095. PMID:14691315. Chiappero, M.B., Sabattini, M.S., Blanco, A., Caldero´n, G.E., and Gardenal, C.N. 2002. Gene flow among Calomys musculinus (Rodentia, Muridae) populations in Argentina. Genetica, 114: 63–72. doi:10.1023/A:1014620707407. PMID:11990761. DeWoody, J.A., Chesser, R.K., and Baker, R.J. 1999. A translocated mitochondrial cytochrome b pseudogene in voles (Rodentia: Microtus). J. Mol. Evol. 48: 380–382. doi:10.1007/ PL00013154. PMID:10093228. Felsenstein, J. 2006. PHYLIP (Phylogeny Inference Package). Version 3.66 [computer program]. Department of Genetics, University of Washington, Seattle, Wash. Available from http:// evolution.genetics.washington.edu/phylip.html. Gonza´lez-Ittig, R.E. 2004. Estructura poblacional de Calomys musculinus (Rodentia, Muridae) estimada mediante polimorfismos en el ADN mitocondrial. Doctoral thesis, Universidad Nacional de Co´rdoba. Gonza´lez-Ittig, R.E., and Gardenal, C.N. 2002. Haplotype diversity of the mitochondrial DNA D-loop region in Calomys musculinus (Rodentia, Muridae) detected by PCR-RFLP. Biochem. Genet. 40: 293–302. doi:10.1023/A:1020250701882. PMID:12392167. Gonza´lez-Ittig, R.E., and Gardenal, C.N. 2004. Recent range expansion and low levels of contemporary gene flow in Calomys musculinus: its relationship with the emergence and spread of Argentine haemorrhagic fever. Heredity, 93: 535–541. doi:10. 1038/sj.hdy.6800546. PMID:15316554. Haag, T., Muschner, V.C., Freitas, L.B., Oliveira, L.F.B., Langguth, A.R., and Mattevi, M.S. 2007. Phylogenetic relationships among species of the genus Calomys with emphasis on South American lowland taxa. J. Mammal. 88: 769–776. doi:10. 1644/05-MAMM-A-319R1.1. Hagelberg, E. 2003. Recombination or mutation rate heterogeneity? Implications for mitochondrial Eve. Trends Genet. 19: 84–90. doi:10.1016/S0168-9525(02)00048-3. PMID:12547517. Kimura, M. 1980. A simple model for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16: 111–120. doi:10.1007/BF01731581. PMID:7463489. Mirol, P.M., Mascheretti, S., and Searle, J.B. 2000. Multiple nuclear pseudogenes of mitochondrial cytochrome b in Ctenomys (Caviomorpha, Rodentia) with either great similarity to or high divergence from the true mitochondrial sequence. Heredity, 84: 538–547. doi:10.1046/j.1365-2540.2000.00689.x. PMID:10849078. Salazar-Bravo, J., Dragoo, J.W., Tinnin, D.S., and Yates, T.L. 2001. Phylogeny and evolution of the neotropical rodent genus Calomys: inferences from mitochondrial DNA sequence data. Mol. Phylogenet. Evol. 20: 173–184. doi:10.1006/mpev.2001. 0965. PMID:11476627. Smith, M.F., and Patton, J.L. 1999. Phylogenetic relationships and the radiation of sigmodontine rodents in South America: evidence from cytochrome b. J. Mamm. Evol. 6: 89–128. Sutherland, B., Stewart, D., Kenchington, E.R., and Zouros, E. 1998. The fate of paternal mitochondrial DNA in developing female mussels Mytilus edulis: implications for the mechanism of doubly uniparental inheritance of mitochondrial DNA. Genetics, 148: 341–347. PMID:9475744. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., and #

2007 NRC Canada

78 Higgins, D.G. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24: 4876–4882. Triant, D.A., and DeWoody, J.A. 2007a. Extensive mitochondrial DNA transfer in a rapidly evolving rodent has been mediated by independent insertion events and by duplications. Gene, 401: 61–70. doi:10.1016/j.gene.2007.07.003. PMID:17714890. Triant, D.A., and DeWoody, J.A. 2007b. Molecular analyses of mitochondrial pseudogenes within the nuclear genome of arvicoline rodents. Genetica, 132: 21–33. doi:10.1007/s10709-0079145-6. PMID:17333478.

Genome Vol. 51, 2008 Triant, D.A., and DeWoody, J.A. 2007c. The occurrence, detection, and avoidance of mitochondrial DNA translocations in mammalian systematics and phylogeography. J. Mammal. 88: 908–920. Zhao, X., Li, N., Guo, W., Hu, X., Liu, Z., Gong, F., et al. 2004. Further evidence for paternal inheritance of mitochondrial DNA in the sheep (Ovis aries). Heredity, 93: 399–403. doi:10.1038/sj. hdy.6800516. PMID:15266295. Zullo, S., Sieu, L.C., Slightom, J.L., Hadler, H.I., and Eisenstadt, J.M. 1991. Mitochondrial D-loop sequences are integrated in the rat nuclear genome. J. Mol. Biol. 221: 1223–1235. PMID:1942048.

#

2007 NRC Canada