Conserved Primers for Rapid Sequencing of the Complete Mitochondrial Genome from Carnivores, Applied to Three Species of Bears

Phylogenetic relationships of the order Carnivora have been extensively studied. However, depending on the type of data, species sampling, and method of analysis used, carnivores have been placed in nearly every possible position throughout the eutherian phylogenetic tree (for a review, see Novacek 1992 ). The integrity of the order itself has, however, remained intact. Undisputedly monophyletic, the order Carnivora nonetheless constitutes a very adaptable and heterogeneous group (Wayne et al. 1989 ) whose evolution has been marked by several events of parallel or convergent evolution and rapid radiation. Consequently, phylogenetic relationships between and within many carnivore families are also still largely unresolved. For example, the families Procyonidae, Ailuridae, Mustelidae, Pinnidedia, and Ursidae are joined at a polytomy. Within Pinnipedia, the closer affinity of Otariidae to either Phocidae or Obodenidae has long been debated. Additionally, the monophyly of the family Mustelidae has been challenged, with true mustelids and mephitids thought to belong to different branches within the superfamily Musteloidea, also including Procyonidae and Ailuridae. The phylogeny of the family Ursidae is another typical example. Six species of bears, grouped under the subfamily Ursinae, are thought to have diverged at the beginning of the Pliocene through such a rapid radiation event (Thenius 1990 ) that the order of species divergence is difficult to determine. Hence, the speciation event that led to extant ursine bears (American black bear, Ursus americanus; brown bear, U. arctos; polar bear, U. maritimus; Asiatic black bear, U. thibetanus; sun bear, U. malayanus; and sloth bear, U. ursinus) is usually represented as a polytomy. Rapid radiations remain unresolved largely because available data contain very few informative changes essential to infer the correct phylogenetic tree. More resolving power can be reached by using longer DNA sequences to provide a sufficient number of informative characters. Moreover, it has been shown that the combination of multiple genes in a single large data set has the potential to raise a weak phylogenetic signal above the noise level (Bull et al. 1993 ). Combining these two advantages, complete mitochondrial (mtDNA) sequences (approximately 17 kbp and 13 protein-coding genes) offer great potential for recovering phylogenies. With the specific goal of resolving the ursine polytomy, and keeping in mind the variety of questions related to carnivore phylogeny that are yet to be resolved, a new approach to complete mitochondrial genome sequencing was developed to obtain complete mitochondrial sequences from closely related carnivore species. A series of primers were designed, based on conserved regions identified from an alignment of published complete mitochondrial genomes from carnivores (harbor seal, Phoca vitulina, X63768, Arnason and Johnsson 1992 ; gray seal, Halichoerus grypus, X72004, Arnason and Gullberg 1993 ; domestic cat, Felis catus, U20753, Lopez, Cevario, and O'Brien 1996 ; and dog, Canis familiaris, U96639, Kim et al. 1998 ). Eleven primer pairs were designed for the amplification of fragments covering the entire mitochondrial genome. Each fragment is longer than 1.4 kbp and covers more than one gene, therefore, limiting the risk of amplifying nuclear copies. The primer design also ensures that there is sufficient overlap of the fragments, in order to obtain the sequence of the primer sites and their flanking nucleotides. Also, 3–12 internal primers were designed to complete the sequence of each fragment in both the directions. A list of all the primers used to sequence the complete mitochondrial genome from a bear species is presented in table 1 . This approach was used to sequence the mitochondrial genome of the three North American ursine species: U. americanus, U. arctos, and U. maritimus. The sequences obtained show that the organization of the bear mitochondrial genome conforms with that of other mammalian species. The length of the mtDNA of the black, brown, and polar bear individuals used in this study is 16,841, 17,020, and 17,017 bp, respectively. However, these numbers are approximate because the multiple 10-bp tandem repeat region located in the D-loop is too long to be sequenced directly. This region of the mitochondrial molecule, reported to show high levels of heteroplasmy by Hoelzel et al. (1994) , is approximately 350 bp long in the black bear individual sequenced and 600 bp in brown and polar bears. These complete mitochondrial sequences have been deposited into GenBank under the accession numbers AF303109 for black bear, AF303110 for brown bear, and AF303111 for polar bear. To confirm the authenticity and the position of our new sequences, the cytochrome b gene from our complete mtDNAs was analyzed along with previously published cytochrome b sequences from the eight extant species of bears. To facilitate comparisons with previous studies, the weighted parsimony method used by Talbot and Shields (1996) as well as the maximum likelihood method using the HKY + gamma model used by Waits et al. (1999) were applied to a data set comprising cytochrome b sequences from Talbot and Shields (1996) , an additional Asiatic black bear sequence (Matsuhashi et al. 1999 ), and our three new sequences. Two different trees that could, however, not be distinguished statistically were obtained. This is not surprising because previous studies have shown that such short sequences are insufficient to resolve the ursine polytomy. At this point in our study, our purpose is not to thoroughly revisit bear phylogeny but mainly to validate our approach. A few findings are relevant to this objective. The cytochrome b sequence from the black bear used in this study is characteristic of the western haplotype previously identified in several independent studies (Cronin et al. 1991 ; Paetkau and Strobeck 1996 ; Byun, Koop, and Reimchen 1997 ; Wooding and Ward 1997 ). With a “western” North American black bear and an Asiatic black bear from Japan (Matsuhashi et al. 1999 ) being added to the original study of Talbot and Shields (1996) , who used an “eastern” black bear and an Asiatic black bear, unfortunately of an unknown origin, the bootstrap support for a clade comprising the two black bear species was improved from 40 to 98. This analysis suggests that the use of highly divergent conspecifics, as we did in this case for the two black bear species, can substantially improve confidence in the tree topology. A few representatives from each ursid species, corresponding to subspecific clades, will likely be helpful in resolving all of the nodes of the ursine polytomy, thereby emphasizing the need for an efficient method to obtain sequence data from very closely related taxa. The complete mitochondrial sequences obtained here were used in a preliminary phylogenetic analysis, along with previously published sequences from other carnivores (as listed previously). Phylogenetic analyses were performed using various data subsets from the complete genome nucleotide sequence. Artiodactyl and perissodactyl sequences (Bos taurus, V00654, Anderson et al. 1982 ; and Equus caballus, X79547, Xu and Arnason 1994 ) were used to root the trees. Maximum parsimony, maximum likelihood, and neighbor-joining methods for estimating trees from complete sequence data gave the same tree topology. This corroborates the conclusions of Russo, Takezaki, and Nei (1996) showing that when many mitochondrial genes are used, all methods appear to converge toward a single tree. The tree topology obtained corresponds to what is currently accepted as the putative true tree (see Cao et al. 1994 ; Kuma and Miyata 1994 ; Janke, Xu, and Arnason 1997 ; Cao et al. 1998 ). This analysis verifies that extensive sequence from the mitochondrial genome provides reliable data and sufficient resolution for building mammalian phylogenies (Honeycutt and Adkins 1993 ) and distinguishing nodes resulting from rapid radiation episodes such as the ursine speciation events. Our strategy offers many advantages when the ultimate goal is to sequence numerous individuals of the same or closely related species. A complete mitochondrial sequence can be obtained in just 73 sequencing reactions, that is, at least 10 times less than procedures that involve cloning of random fragments. Moreover, the sequencing procedure is completely PCR-based, thus bypassing purification of mitochondria, DNA restriction digestion, and cloning steps common to other methods. Finally, the 11 primer pairs used for the amplification of mitochondrial fragments are located in highly conserved regions so as to transfer easily to other carnivores. They have indeed been used to amplify mtDNA from three other ursids, nine phocids, one odobenid, two otariids, five true mustelids, a mephitid, one procyonid, three canids, and two felids. Although we expect the majority of the primers to be widely useful, some of the internal primers presented here will show unavoidable mismatches that may lead to poorer PCR performance. Consequently, only a few new primers may be required to complete the sequence of any carnivore mitochondrial genome. Moreover, as more species are sequenced, a list of alternative primers will become available to facilitate further sequencing. This opens the door to various complete mtDNA phylogenetic analyses at the order, family, subfamily, and species levels among carnivores and should certainly lead to a better understanding of carnivore phylogeny. Dan Graur, Reviewing Editor Keywords: direct sequencing complete mitochondrial genome mtDNA phylogeny Carnivora Ursidae Address for correspondence and reprints: Isabelle Delisle, Department of Biological Sciences, CW 405, Biological Sciences Building, University of Alberta, Edmonton, Alberta, Canada T6G 2E9. idelisle@ualberta.ca . Table 1 Primer List Table 1 Primer List Table 1 Continued Table 1 Continued I.D. holds a scholarship from the NSERC. C.S. is supported by grants from the NSERC and from Parks Canada. We would also like to thank two anonymous reviewers of an early version of this manuscript for their helpful comments.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.