Molecular Ecology (2001) 10 , 2241–2247© 2001 Blackwell Science Ltd Blackwell Science, Ltd Mitochondrial DNA variation and biogeography of eastern gorillas M. I. JENSEN-SEAMAN*† and K. K. KIDD‡* Department of Anthropology Yale University, P.O. Box 208277, New Haven, CT 06520-8277, USA, ‡ Department of Genetics, Yale University, P.O. Box 208005, New Haven, CT 06520-8005, USA AbstractMitochondrial DNA variation in 109 individuals from four populations of wild livinggorillas in East Africa was ascertained by sequencing the first hypervariable segment of thecontrol region, or ‘ D -loop’, amplified from noninvasively collected hair and faeces. D -loophaplotypes from eastern gorillas fell into two distinct clades, each with low levels of geneticdiversity; most observed haplotypes within each clade differing by only one or two muta-tions. Both clades show evidence of population bottlenecks in the recent past, perhapsconcomitant with the tropical forest reduction and fragmentation brought on by globalcooling and drying associated with the last glacial maximum. Keywords : beringei , D -loop, graueri , hominid, hominoid, phylogeography Received 10 January 2001; revision received 7 June 2001; accepted 7 June 2001 Introduction The climatic history of Africa during the last 1 Myr hasbeen one of change, with warm wet times alternating withcool dry periods (Hamilton 1975; Livingstone 1975; Jolly et al . 1997). The lowered temperatures and increasedaridity during the Pleistocene glaciations reduced andfragmented tropical forests into island refugia (Livingstone1967, 1982; Grubb 1982; Haffer 1982; Kingdon 1989). Aslarge-bodied, forest-dwelling primates, the history ofgorillas ( Gorilla gorilla ) is expected to be closely tied to thehistory of the African forests, with the fragmentation offorests potentially fragmenting gorilla populations. Also,as gorillas live at reasonably uniform densities across sitesand across habitats (Yamagiwa 1999), population sizes arepredicted to decrease in proportion to the decrease in thesize of the forests, which may have been reduced to only5–25% of the present during the last glacial maximum at ≈ 18 000 years ago (Livingstone 1967, 1975; van ZinderenBakker & Coetzee 1972; Moeyersons & Roche 1982;Chapman 1983).Gorilla populations are found in equatorial Africa in twowidely separated regions: West Africa and East/CentralAfrica, separated by nearly 1000 km. The eastern gorillapopulations (Fig. 1) are usually assigned to the subspecies G. g. beringei and G. g. graueri , although increasingly theyare being referred to as the separate species G. beringei (Groves 1996, 2001; Sarmiento & Butynski 1996). Themountain gorillas of the Virungas are the type populationof G. g. beringei , and today number around 300 individuals(Harcourt 1995). The gorillas of the Bwindi ImpenetrableNational Park, located ≈ 40 km to the north of the Virungasalso number ≈ 300 individuals (McNeilage et al . 1998), andhave traditionally been placed in beringei , althoughSarmiento et al . (1996) suggest that these gorillas are not thesame subspecies as the Virunga gorillas based on morpho-logical data. The isolated population of gorillas found onMt. Tshiaberimu, numbering only ≈ 16 – 18 individuals(Butynski & Sarmiento 1995), has been placed by Groves(1970) in graueri , though later noting they ‘tend towards beringei in some respects’ (p. 168, Groves & Stott 1979).Recently Sarmiento & Butynski (1996) suggested revivingthe subspecies rex-pygmaeorum , originally proposed bySchwarz (1927), for the Tshiaberimu gorillas. The Kahuzi-Biega populations, usually considered graueri , currentlyaccount for well over half of all gorillas in East Africa (Hall et al . 1998). Schouteden (1947, referenced in Groves & Stott1979) favoured including the Kahuzi-Biega populations in rex-pygmaeorum , restricting the subspecies graueri to thetype population from the Itombwe Massif west of LakeTanganyika. Correspondence: K. K. Kidd. Fax: 1 203 785 6568; E-mail:[email protected]†Present address Human and Molecular Genetics Center, MedicalCollege of Wisconsin, Milwaukee, WI 53226, USA. MEC1365.fm Page 2241 Tuesday, August 21, 2001 4:44 PM2242 M. I. JENSEN-SEAMAN and K. K. KIDD © 2001 Blackwell Science Ltd, Molecular Ecology , 10, 2241–2247 The maternally inherited mitochondrial DNA (mtDNA)control region, or ‘ d- loop’, is one of the most rapidly evolv-ing, highly polymorphic genetic markers in most mam-mals examined to date (Avise 2000). Here we examine thepatterns of d- loop DNA sequence diversity in several popu-lations of eastern gorillas in order to infer relationshipsbetween populations, genetic structure within popula-tions, and past changes in population size. These resultsare examined in the light of climatic and vegetationchanges in Pleistocene Africa. Materials and methods Hair samples from vacated night nests were collected from(Fig. 1): Mt. Tshiaberimu, D. R. Congo ( n = 15); VirungaVolcanoes National Park, Rwanda ( n = 17); Bwindi Impenet-rable National Park, Uganda ( n = 48). Additional hair sampleswere obtained from captive gorillas (Houston ZoologicalGardens, n = 1; Royal Zoological Gardens at Antwerp, n = 2).Two of these captive gorillas, ‘M’kubwa’ (Houston) and‘Mukisi’ (Antwerp), were captured in the 1950s in theWalikale region, D. R. Congo, probably from in or near thenorthern part of the lowland sector of the Kahuzi-BiegaNational Park. Fecal samples were collected from nests fromthe highland sector of the Kahuzi-Biega National Park, D. R.Congo ( n = 29), and stored in 100% ethanol until extraction.The Tshiaberimu samples are from two different identifiedsocial groups, and the samples from highland Kahuzi-Biegaand the Virungas are from habituated social groups, makingit unlikely that the same individuals were repeatedly sampled.Using a slight modification of a protocol described byWalsh et al . (1991), ≈ 5 mm of the root portion of individualhairs was cut directly into 0.5 mL screw-top microcentri-fuge tubes, using a fresh disposable scalpel blade and dis-posable tweezers. To each tube 250 µ L of 5% Chelex resin(Bio-Rad, Hercules, CA, USA), 0.039 m DTT, and 0.2 µ g/ µ Lproteinase K was added. Tubes were then vortexed briefly,incubated 4–12 h at 55 ° C while shaking, vortexed, boiledfor 8 min, vortexed, centrifuged for 1 min, and stored at4 ° C. Fecal material with a wet volume of ≈ 300 µ