Hystrix, (as.) 10 (2) (1998): 29-36

MITOCHONDRIAL CYTOCHROME B SEQUENCE DIVERGENCE AMONG SPANISH, ALPINE AND ABRUZZO CHAMOIS (GENUS RUPZCAPRA)

NADIA MUCCI(*), ETTORERANDI(*), LEONARDOGENTILE (* *),

FRANCO MAR1 (* *O) AND MAURIZIOLOCATI (**)

(*) Istituto Nazionale per la Fauna Selvatica, via Ca Fornacetta 9, 40064 Ozzano dell'Emilia (BO),Italy. (**) Centro Studi Ecologici Appenninici, Parco Nazionale d'Abruzzo, viale Santa Lucia, 67032 Pescasser-oli (AQ),Italy. (") via Lanino 6, 20144 Milano, Italy

ABSTRACT - We have studied genetic divergence and phylogenetic relationships of Alpine, Spanish and Abruzzo chamois (genus Rupicapra) by sequencing a region of 330 nucleotides within the mito- chondrial DNA cytochrome b gene (mtDNA cyt 6).These sequences were aligned with additional ho- mologous sequences of Caprinae: Japanese serow, Chinese goral, Canadian mountain goat, Mishmi takin, muskox, Sardinian mouflon and domestic goat. Results suggest that, using representatives of the Bovini as outgroups, the Caprinae constitute a monophyletic clade. However, inferred phyloge- netic relationships among and within tribes of Caprinae were poorly defined and did not reflect cur- rent evolutionary and taxonomical views. In fact, the Asian Rupicaprini goral and serow constituted a strongly supported clade, which included the muskox, while the takin grouped with Uvis. Therefore, the monophyly of Ovibovini was not supported by cyt b sequences. Species of Rupicapra joined a strongly supported monophyletic clade, which was distantly related to the Asian rupicaprins and Ore- amnos. Therefore, the monophyly of the Rupicaprini was not supported by these cyt h sequences. There were sister species relationships within Rupicapra, Spanish and Alpine chamois and the Abruz- zo chamois (Rupicupru pyrenaica ornata) was strictly related to the Spanish chamois (Rupicapra pyre- naica parva), as previously suggested by allozyme data and hiogeographic reconstructions. Key words: Rupicapra pyrcnaica parva, Rupicupra pyrenuica or-nata, Rupicupra rupicapi-U,Caprinae, genetic divergence, rnitochondrial DNA, cytochrome h.

~NTRODUCTION et al., 1990; Randi et al., 1991) and molec- Some aspects of phylogenetic relationships ular (Chikuni et al., 1995; Groves and and systematics of the subfamily Caprinae Shields, 1996) findings. are controversial. Three living tribes (Rupi- The chamois (genus Rupicapra) is an evo- caprini, Ovibovini and Caprini; Grubb, lutionary advanced rupicaprin taxa (Geist, 1993) are universally recognized, but their 1987), only distantly related to the other monophyly have been repeatedly ques- Rupicaprini both from a morphologic (Gen- tioned, after re-evaluation of fossil records try, 1990) and genetical (Hartl et al., 1990; (Gentry, 1990; 1992), quantitative morpho- Randi et al., 1991) point of view. Based on metrics (Gentry, 1992), biochemical (Hartl the admittedly poor fossil records, Masini 30 N. Mucci et al. and Lovari (1988) suggested that the Rupi- MATERIAL AND METHODS caprini originated during the Miocene in DNA Sequencing Asia. Rupicapra is supposed to have origi- Tissue and blood samples were collected nated recently in the Villafranchian (2.5 mil- from two samples of Spanish (Riipicupra lion years ago) and evolved in west Eurasia pyrenaica parva), Abruzzo (R. p. ornata) during middle and late Pleistocene (Geist, and Alpine (R. rupicapra) chamois. Total 1987; Masini and Lovari, 1988). DNA was extracted from about 20 mg of Present west European populations of Rup- minced tissues in 550 ml of lysis buffer icapru have probably relictual distribu- (0.05M Tris/HCl. pH 8.0, 0.02M EDTA, I% tions, restricted to the Cantabrian moun- SDS, 0.2M NaCl, 1% b-mercaptoethanol), tains and the Pyrenees, the Alpine range, and digested overnight with 0.2 mg/ml pro- and central Italian Apennines. Recently all teinase-k. After RNAse treatment, 350 ml of the existing populations of chamois were 5M NaCl were added and proteins were pel- classified in the single polytypic species leted by centrifugation (15,000 rpm, 30 Rupicapra rupicapra, with nine (Dolan, min). Surnatants were cleaned by 1963) or ten (Knaus and Schroeder, 1983) pheno1:chloroform:isoamyl alchool extrac- subspecies. A taxonomical revision of Rup- tions, and DNA was ethanol precipitated icapra by Lovari and colleagues (Lovari and resuspendend inTE. DNA was collected and Scala, 1980; Scala and Lovari, 1984; from 100 ml of whole blood, lysed by Nascetti et al., 1985; Masini and Lovari, adding 1 ml of chilled sterile water. Nuclei 1988), recommended the inclusion of the and white cells were pelleted (15,000 rpm, Iberian and Apennine populations in a new 5 min) and resuspended in 550 ml of lysis species R. pyrenuica, with three sub- buffer, then DNA was extracted following species: R. 17. pyrenaica and R. p. purw the same procedure described for tissues. (the Spanish chamois), and R. p. ornata The entire mtDNA cyt b was amplified by (the Abruzzo chamois). A close relation- polymerase chain reaction (PCR) using ship between Spanish and Abruzzo primers L1 (14464) = 5’ - CGA AGC TTG chamois were suggested by Lovari (1985) ATA TGA AAA ACC ATC GTT G - 3’, and based on behavioral observations support- H1 (15707) = 5’ - GTC TTC ACT TTT ed by genetic analyses of allozyme elec- TGG TTT ACA AGA C - 3’, of which 5’ trophoretic variability (Nascetti et al., ends bind to nucleotides nos. 14,464 and 198.5). 15,707 of the bovine tRNA-Glu and tRNA- Although most studies at population level Thr genes (Anderson et al., 1982). Amplifi- have used mtDNA restriction fragment cations were performed using a 9600 Perkin analyses (Hammer et al., 1995), more de- Elmer thermocycler, with the following pro- tailed information about the genetic conse- tocol: 94°C for 2 min; 94°C for 15 sec, 50 quences of recent speciation events and ge- - 55 “C for 15 sec, 72°C for 1 min (30 cy- ographic subdivision of genetic variability cles), 72°C for 10 min. PCR products were among conspecific populations can be ob- purified by 1.5% low-melting agarose gel tained by nucleotide sequencing (Wenink et and agarase (Boehringer) digestion. All se- al., 1996). In this paper we aim to compare quences were obtained by double-strand nucleotide sequences of a portion of the mi- DNA cycle sequencing using DTaq Seque- tochondrial DNA cytochrome h gene of R. nase (Amersham) and the following ruppicapra, R. p. yur~~uand R. p. ortzatu. primers: L1, L2 (14520) = 5’- AAC ATC DNA sequence divergence among Kzq?i- CGA AAA TCA CAC CC - 3’, L3 (14583) cupm will be evaluated within the frame- = 5’- CCA TCC AAC ATC TCA GCA TGA work of molccular evolution of the Capri- TGA AA - 3’. Mitochondrial DNA cyt h se- me.

32 N. Mucci et al.

Table I. Pairwise estimated % DNA distances (Tamura and Nei, 1993) among 13 taxa of the Bovini.

Taxon 123 4 5 6 7 8 9 10111213

I - B. bubalis00 13 13 15 13 12 12 13 11 16 13 16 16 2 - B. taurus 00 10 15 16 16 15 16 15 16 17 19 17 3 - B. javanicus 00 15 17 15 17 18 1.5 20 1.5 19 16 4 - C. crispus 00 8 11 11 12 12 13 11 13 7 5 - N. caudatus 00 11 11 12 12 14 1.5 14 9

6 - 0. americanus 00 9 11 9 11 9 12 12 7 - R. rupicapra 00 4310 9 11 11 8 - R. p. ornata 00 3 13 9 12 13 9 - R.p.parva 00 11 8 10 11 I0 - C. hircus 00 12 14 14 II - 0. musirnon 00 11 15 12 - B. t. taxicolor 00 13

I3 - 0.moschatus 00 quences of other taxa of Ruminantia were heuristic searches with TBR and MULPARS obtained from the GenBank and used for options in use, and with 10 random-addi- phylogenetic analyses: cattle (Bos taurus), tions of sample sequences. Support of clus- banteng (B. javanicus), and water buffalo ters and clades was evaluated by bootstrap (Buhalus bubalis) (Kikkawa et al., 1995 and (Felsenstein, 1985), as a percentage of unpubl.; GenBank accession nos. D34635, group recurrence based on 1,000 boot- D34636, D34637); Mishmi takin (Budovcas strapped replications, with both MEGA and taxicolor taxicolor), muskox (Ovibos PAUP. moschatus), Chinese goral (Nemorhaedus caudatus), Japanese serow (Capricornis crisps), (Groves and Shields, 1996; RESULTS U17868, U17862, U17861, U17865). More- over, we have added new cyt b sequences of We sequenced a region of 330 nucleotides the domestic goat (Capra hircus), and Sar- of mtDNA cyt b gene of Spanish, Abruzzo dinian muflon (Ovis gmelini). and Alpine chamois. Theye sequences were aligned with corresponding sequences of Data Analysis three species of Bovini and other Caprinae Phylogenetic analyses of the aligned se- (Fig. 1). The Tamura and Nei (1993) pair- quences were performed with neighbor-join- wise distances among the sequences are ing (NJ; Saitou and Nei, 1987), and maxi- shown in Table 1. This distance matrix was mum parsimony (MP; Swofford, 1993) pro- used to build the NJ tree (Fig. 2). The MP cedures using MEGA (Kumar et al., 1993) tree, obtained by PAUP with heuristic and PAUP (Swofford, 1993) computer pro- searches randomly repeated 10 times, was grams. NJ trees were computed using Tamu- identical to the NJ tree (not shown). In both ra and Nei’s (1993) DNA distances. Parsi- distance and parsimony analyses representa- mony analyses, excluding uninformative tives of the Bovini (Bos taurus, B. jai,anicus substitutions, were performed through and Bubalus huhalis) were considered as 33

1_

Ovrs m,ls,m"w - - Budorrus I mxicolor

I Oreamnos amerrranits

A Nemorhoedus raudalus

Capricornis crispus

Oluhor niosrhatus r Birhalur huhnlis 98 nos taunrs

Bos juvanicus

0.00 0.05 0.10 Tamura and Nei DNA distance

Figure 2. Neighbor-joining tree computed using Tamura and Nei (1993) DNA distances and the cyt sequences of' Caprinae showed in Fig. 1. The tree was rooted using three \pecks of Bovini as an out- group. Branch lenglhs are indicated and correspond to the estimated percentage nucleotide divergence among taxa. Bootstrap value5 of the groups which occurred with frequencies greater than 50% after 1,000 bootstrap runs are indicaled. outgroups of thc Caprinae, according to Genetic distance among the Caprinae range morphological (Gentry, 1992) and molecu- from 15% sequence divergence among Oiis, lar findings (Allard et al.. 1992; Cronin et Oiihos and Nemoi.haedus, to 3% sequence al., 1996). divergence among R. p. parva and R. p. ot-- Our results indicate the existence of two izatu. The Spanish and Alpine chamois have clearly defined lineages: a) the Asian Rup- about 4% sequence divergence. It should be icaprini Cupricornis and Nemorhaedus, observed that most of the sequence diver- which group with Ovibos; b) a cluster gence among oniuto and pai~ais attributed which groups the three Rupicupra taxa. to the long branch which leads to oriiata in Ovis is weakly linked to Biidorcas; Capra the NJ tree (Fig. 2). This branch is four and Orcaninns are intermediate between times longer than its sister branch leading to the Asian Rupicaprini and Kupicupra. panw and is the consequence of apparent Bootstrapping supports the clade of the heterogeneous rate of molecular evolution Asian Rupicaprini with Ovibos (88%) and in the ortiata lineage. the Rz4pic.api.a clade (95%), but topological relationships of Ovis,Budoi-cus and Capi.u are weak (Fig. 2). The Spanish and Alpine Discussio~ chamois are sister lineages, and the Abruz- Partial nucleotide sequences of the mtD- zo chamois is strictly related to the pyre- NA cyt b gene suggest that the Asian Rup- tiaicu lineage. icaprini Neniorhaedus and Cupricornis 34 N. Mucci et al. represent a basal evolutionary lineage Molecular evolution of the cyt h of Capri- among the Caprinae, in accordance with nae exhibits heterogeneous substitution current evolutionary reconstructions and rates at different nucleotides, as shown by taxonomy (Geist, 1987). The gcnus Rupi- branches of different lengths in the NJ tree cupr-a is monophyletic. It is weakly relat- (Fig. 2). Heterogenous rates can be due to ed to Ovis, Cupl-a, and Or-enmnos, but is different substitution rates among evolu- only distantly related to the Asian Rupi- tionary lineages or to the genetic conse- caprini. The deeper genetic gap in Rupi- quences of population bottlenecks and ex- cupr-U is between the two species R. mpi- tended periods of low effective population capl-u and R. yyrenaicn. The Abruzzo size. Heterogeneous rates can bias the esti- chamois K. p. or-nnta is more strictly relat- mation of genetic distances and the recon- ed to the Spanish than to the Alpine struction of phylogenetic trees. In particu- chamois, as previously suggested by al- lar, the long branch associated with the or- lozyme analyses and biogeographic recon- nata lineage could be a consequence of the structions (Nascetti et al., 1985; Masini segregation of an ancestral mtDNA haplo- and Lovari. 1988). Therefore, this mtDNA type in the Abruzzo population, which was cyt b phylogeny suggests that the Rupi- isolated for centuries and underwent a re- caprini may not be treated as a mono- cent dramatic bottleneck (Lovari, 19x9). phyletic group and, consequently, it cannot The genetic effects of population bottle- be assumed that all living Rupicaprini are necks were described by Nascetti et al. ancestors of the Ovibovini and Caprini (I 995), which found the Abruzzo chamois (Randi et al., 1991). The cluster of ances- population to be monomorphic at 25 al- tral Asian Rupicaprini excludes Or-eanzizos lozyme loci. and Riipic,upru. The combined effects of isolation and bot- Morphometrical analyses (Gentry, 1992) tleneck could have fixcd an ancestral showed that Rupirap-u has a unique com- mtDNA haplotype in the Abruzzo chamois bination of skeletal traits and does not population, a haplotype which could prc- group with the other Rupicaprini. but is date the time of reproductive isolation of linked to the Caprini. The monophyly of the the population. Mitochondrial DNA poly- Ovibovini tribe is not supported, because morphisms were shown by Hammer et al. Ovibos is linked to the Asian Rupicaprini (1995) in some Alpine chamois popula- and Birdorcas to Ovis (Groves and Shields, tions. Biased estimation of inter- and in- 1996). From a skeletal point of view, the tra-populalion sequence variability, and Ovibovini appears as a monophyletic group the possible retention of an ancestral mtD- (Gentry, 1992), probably as a consequence NA haplotype in the Abruzzo chamois of the evolutionary convergence of mor- population, make any estimate of diver- phological traits (Groves and Shields, gence time and any taxonomical conclu- 1996). The great genetic divergence among sion from this preliminar data set inadvis- Cupl-a and Ovis could potentially disrupt able. the monophyly of the Caprini tribe. These results are in accordance with allozyme and ACKNOWLEDGEMENTS other mtDNA sequence data (Hart1 et al.. We would like to thank the Direction and btaff ol' 1990: Randi et al., 1991: Groves and Abruzzo National Park. G. Tosi (Milan Universi- Shields. 1996). Moreover, mitochondrial ty) and C. Nores (INDUROT - Oviedo - Spain) phylogenies of Caprinae are substantially for their help in collecting samples. We are also supported by sequences of autosomal k-ca- grateful to S. Imvari. M. Festa-Bianchet and sein genes (Chikuni et al., 1995: Cronin et some anonymous reviewers who greatly con- al., 1996). tributcd to improving this paper. REFERENCES tochrome b sequences. Mol. Phylogenet. Evol., 5: 467-476. Allard, M.W., Miyamoto, M.M., Jareki, L., Grubb, P., 1993. Order Artiodactyla. In: Kraus, E and Tennant, M.R., 1992. DNA D.E. Wilson and D.M. Reeder (eds.), systematics and evolution of the artio- Mammal species of the world. A taxo- dactyl family Bovidae. Proc. Natl. Acad. nomic and geographic reference. Smith- Sci. USA, 89: 3972-3976. sonian Institution Press, Washington and Anderson, S., De Bruijn, H.L., Coulson, London, pp.: 377-414 A.R., Eperon, I.C., Sanger, E and Young, Hammer, S., Nadlinger, K. and Hartl, G.B., Y.G., 1982. Complete sequence of bovine 1995. Mitochondrial DNA differentiation mitochondrial DNA: Conserved features in chamois (genus Rupicupra): implica- of the mammalian mitochondrial tions for taxonomy, conservation and genome. J. Mol. Biol., 156: 683-717. management. Acta Theriologica Suppl., Chikuni. K., Mon. Y.. Tabata, T., Saito, M., 3: 145-155. Monma, M. and Kosugiyama, M.. 1995. Hartl, G.B., Burger, H., Willing, R. and Molecular phylogeny based on the kap- Suchentrunk, E, 1990. On the biochemical pa-casein and cyt h sequences in the systematics of the Caprini and Rupicapri- mammalian suborder Ruminantia. J. Mol. ni. Biochem. Syst. Ecol., 18: 175-182. Evol., 41: 859-866. Kikkawa, Y., Amano, T. and Suzuki, H.. Cronin, M.A., Stuart, R., Pierson, B.J. and 1995. Analysis of genetic diversity of do- Patton, J.C., 1996. K-casein gene phy- mestic cattle in East and Southeast Asia logeny of higher ruminants (Pecora, Ar- in terms of variations in restriction sites tiodactyla). Mol. Phylogenet. Evol.. 6: and sequences of mitochondrial DNA. 295-3 11. Biochem. Genet., 33: 51-60. Dolan, J.M., 1963. Beitrag zur systematis- Knaus, W. and Schriider, W., 1983. Das chen Gliederung des Tribus Kupicupini Gamswild. Paul Parey, Hamburg und Simpson 1945. Zeitschrift fur zoologis- Berlin. che Systematik und Evolutionsforschung, Kumar. S., Tamura, K. and Nei, M., 1993. 1: 31 1-407. MEGA: Molecular Evolutionary Genetic Felsenstein, J., 1985. Confidence limits on Analysis, Version 1.O. The Pennsylvania phylogenies: An approach using the boot- State University, University Park, PA strap. Evolution. 39: 83-791. 16802. Geist. V.. 1987. On the evolution of the Lovari, S., 1985. Behavioural repertoire of Caprinae. In: H. Soma (ed.), The biology the Abruzzo Chamois Rupicupr-a pye- and management of Cupricor-nis and re- nuicu or-nuta (Ncumann, 1899) (Artio- lated mountain antelopes. Croom Helm, dactyla: Bovidae). Saugetierkundliche London, pp. : 3-40. Mitteilungen, 32: 113-136. Gentry, A.W., 1990. Evolution and dispersal Lovari, S.. 1989. L'evoluzione del camoscio of the African Bovidae. In: G.A. Bubenik appenninico. Le Scienze, 247: 46-55. and AB. Bubenik (eds.), Horns, prong- Lovari. S. and Scala, C., 1980. Revision of horns and antlers: Evolution, morphology, Rupicupr-u genus. 1. A statistical re-eval- physiology and social significance. uation of Couturier's data on the mor- Springer-Verlag. New York, pp.: 195-230. phometry of six subspecies of chamois. Gentry, A.W., 1992. The subfamilies and Boll. Zool., 47: 113-124. tribes of the family Bovidae. Mammal. Masini, E and Lovari. S., 1988. Systemat- Rev. 22: 1-32. ics, phylogenetic relationships, and dis- Groves. P. and Shields, G.F., 1996. Phylo- persal of the chamois (Rup17icq~i.uspp.). genetics of the Caprinae based on cy- Quaternary Research; 30: 339-349. 36 N. Mucci et a1

Nascetti, G., Lovari. S., Lanfranchi, P., Scala, C. and Lovari. S., 1984. Revision of Berducou, C., Mattiucci, S., Rossil L. and Kz4pic~pl-ugenus. 11. A skull and horn sta- Bullini, L., 1985. Revision of Rupicqwa tistical comparison of Ri~piccipurupi- genus. 111. Elcctrophoretic studies rqro oi-nata and R. rirpicopra pyre- demonstrating species distinction of l7aiCa. Boll. ZOO^.. 5 1: 285-294. chamois populations of the Alps from Swofford, D.L., 1993. PAUP: Phylogenetic those of the Apennines and Pyrenees. In: Ananlysis Using Parsimony, Version 3.1. S. Lovari (ed.), The biology and man- Illinois Nat. Hist. Survey, Champaign, agement of mountain ungulates. Croon Illinois. Helm, London, pp.: 57-62. Tamura, K. and Nei, M., 1993. Stimation of Randi, E., Fusco, G.. Lorenzini, R., Toso, S. the number of nucleotidc substitutions in and Tosi, G., 1991. Allozyine divergence the control region of mitochondrial DNA and phylogenetic relationships among in humans and chimpanzees. Mol. Biol. Capa, Ovis and Rupicapra (Artiodacty- Evol.. 10: 512-526. la, Bovidae). Heredily, 67: 281-286. Wenink, P.W., Baker, A.J., Rosner, H.-U. Saitou, N. and Nei, M., 1987. The neighbor- and Tilanus, M.G.J., 1996. Global mito- joining method: A new method for re- chondrial DNA phylogeography of hol- constructing phylogenetic trees. Mol. Bi- arctic breeding dunlins (Calidl-is alpina). 01. Evol., 4: 406-425. Evolution, 50: 3 18-330.