RESEARCH ARTICLE

Diversity and Molecular Phylogeny of Mitochondrial DNA of Rhesus Macaques(Macaca mulatta) in Bangladesh

M. KAMRUL HASAN1*, M. MOSTAFA FEEROZ2, LISA JONES‐ENGEL3, GREGORY A. ENGEL3,4,SREE KANTHASWAMY1, AND DAVID GLENN SMITH1

1Molecular Anthropology Laboratory, Department of Anthropology, University of California, Davis (UC Davis), California2Department of Zoology, Jahangirnagar University, Savar, Dhaka, Bangladesh3University of Washington, National Primate Research Center, Seattle, Washington4Swedish Medical Center, Seattle, Washington

While studies of rhesusmacaques (Macacamulatta) in the eastern (e.g., China) and western (e.g., India)parts of their geographic range have revealed major genetic differences that warrant the recognition oftwo different subspecies, little is known about genetic characteristics of rhesus macaques in thetransitional zone extending from eastern India and Bangladesh through the northern part of Indo‐China, the probable original homeland of the species. We analyzed genetic variation of 762 base pairs ofmitochondrial DNA from 86 fecal swab samples and 19 blood samples from25 local populations of rhesusmacaque in Bangladesh collected from January 2010 to August 2012. These sequences were comparedwith those of rhesus macaques from India, China, and Myanmar. Forty‐six haplotypes defined by 200(26%) polymorphic nucleotide sites were detected. Estimates of gene diversity, expected heterozygosity,and nucleotide diversity for the total population were 0.9599� 0.0097, 0.0193� 0.0582, and0.0196� 0.0098, respectively. A mismatch distribution of paired nucleotide differences yielded astatistically significantly negative value of Tajima’s D, reflecting a population that rapidly expandedafter the terminal Pleistocene. Most haplotypes throughout regions of Bangladesh, including an isolatedregion in the southwestern area (Sundarbans), clustered with haplotypes assigned to the minorhaplogroup Ind‐2 from India reflecting an east towest dispersal of rhesusmacaques to India. Haplotypesfrom the southeast region of Bangladesh formed a cluster with those fromMyanmar, and represent theoldest rhesus macaque haplotypes of Bangladesh. These results are consistent with the hypothesis thatrhesus macaques first entered Bangladesh from the southeast, probably from Indo‐China, thendispersed westward throughout eastern and central India. Am. J. Primatol. 76:1094–1104, 2014.© 2014 Wiley Periodicals, Inc.

Key words: rhesus macaque; Macaca mulatta; mtDNA; haplotypes; HVR 1; Bangladesh

INTRODUCTION

The genus Macaca is believed to have originatedin northern Africa as early as 7 million years ago(Mya) [Eudey, 1980] and migrated through theMiddle East and northern India approximately 3Mya. Macaques had passed through most ofChina and reached the Indonesian archipelago asearly as 2.5 Mya where the common ancestor of thefascicularis species group originated [Abegg &Thierry, 2002; Delson, 1980; Fooden, 1976]. It hasbeen hypothesized that rhesus macaques divergedfrom a fascicularis‐like ancestor that reached themainland of Indo‐China from Indonesia by approxi-mately 1 Mya [Delson, 1980; Fooden, 2006; Stevison& Kohn, 2009; Tosi et al., 2003]. They subsequentlydispersed throughout most of Asia but the origin,route, and timing of this dispersal is not wellunderstood.

The marked genetic homogeneity of Indianrhesus macaques [Smith & McDonough, 2005] mayresult from their location at the terminus of a serialfounder effect from the westward dispersal of the

Contract grant sponsor: National Institutes of Health;contract grant numbers: RR000169‐48, RR005090, RR025781;contract grant sponsor: National Institute of Allergy andInfectious Diseases; contract grant numbers: R01 AI078229,R01AI078229‐03S1, R03 AI064865.

�Correspondence to: M. Kamrul Hasan, Molecular AnthropologyLaboratory, Department of Anthropology, University of Califor-nia, Davis (UC Davis), CA 95616. E‐mail: mkhasan@ucdavis.edu

Received 25 November 2013; revised 26 March 2014; revisionaccepted 1 April 2014

DOI: 10.1002/ajp.22296Published online 8 May 2014 in Wiley Online Library(wileyonlinelibrary.com).

American Journal of Primatology 76:1094–1104 (2014)

© 2014 Wiley Periodicals, Inc.

species, as reported for modern human populationsexpanding eastward from Africa [Ramachandranet al., 2005]. Although a genetic study was conductedon a small population of free ranging Nepali rhesusmacaques [Kyes et al., 2006], studies of geneticvariation of rhesus macaques has primarily focusedon animals in captive breeding facilities, limitingsuch studies to rhesus macaques originating in Indiaand China [Smith & McDonough, 2005]. Unfortu-nately, virtually no molecular research on rhesusmacaques of Bangladesh has been reported. Apreliminary survey of the mtDNA diversity in 39individuals from five different localities ofBangladesh [Feeroz et al., 2008] detected a total ofonly seven haplotypes. Recently, we conducted astudy of rhesusmacaques and associated foamy virusin Bangladesh, and used microsatellites to charac-terize population structure in macaques from sixurban populations [Engel et al., 2013; Feerozet al., 2013]. This study found the monkeys in acentral population (Madhupur) to reflect admixturewith the Sundarbans population and suggested thatcontemporary human agency such as monkey per-formers and monkey pet owners are playing a role inthe genetic and viral composition of some of thesepopulations.

Bangladesh is located at the transitional zonebetween the Indo‐Himalayas and Indo‐China sub‐regions [Stanford, 1991]. Macaque migrations fromwest to east and vice versawere physically impossiblewithout traversing Bangladesh. However,Bangladesh is a small and densely populated countrythat is crisscrossed by hundreds of rivers that restrictthe distribution of primates. Its total area of147,570 km2 is home to more than 160 million people(1,084 people/km2).Many primate habitats have beendestroyed by increasing human population pressure,industrialization, and rapid urbanization [Feerozet al., 2011], and many rhesus macaque populationshave been confined inside human settlements [Hasanet al., 2013].

Bangladesh currently supports five species ofmacaques: rhesus macaques (Macaca mulatta), pig‐tailed macaques (Macaca leonina), cynomolgousmacaques (Macaca fascicularis), Assamese maca-ques (Macaca assamensis), and stump‐tailed mac-aques (Macaca arctoides) [Feeroz, 2001; Feerozet al., 2011; Hasan, 2003; Khan, 1982]. The latterfour macaque species are distributed only in thenortheastern and southeastern hill areas of thecountry and their population density is very low[Islam et al., 2000]. Rhesus macaques are foundthroughout the country except for the northwesternpart where no primate species lives [Hasanet al., 2013; Khan, 1982]. This species is found inall types of natural forest, tea gardens, plantedforests, and human settlement areas and is the onlyprimate species found in the Sundarbans mangroveforest in the southwest of the country [Hasan

et al., 2013]. Rhesus macaque populations in naturalhabitats of the country can be divided into the threemajor sub‐populations defined by geographic andanthropogenic barriers shown in Figure 1: (i) central,(ii) eastern, and (iii) southwestern. The eastern sub‐population is connected by forest that traverses thestate of Tripura in Northeast India (see Map), whichborders Bangladesh to the north, south, and west,and the Indian state of Mizoram to the east, dividingeastern Bangladesh into a northeast and southeastsector.

Detailed molecular research on rhesus macaquepopulations of Bangladesh could address some

Fig. 1. Map showing four sub‐populations and sampling sites.Color dots indicate sampling sites in each sub‐population (pink,sky blue, and yellow dots indicate the central, eastern, andsouthwestern sub‐populations, respectively). The deep greencolor in the map indicates existing forests and the light greencolor indicates continuous forest connecting the northeast andsoutheast sectors of the eastern sub‐population through India.Numbers indicate corresponding local populations of rhesusmacaques sampled as follows: 1. Sadhana, 2. Dhamrai, 3.Narayanganj, 4. Bormi, 5. Madhupur, 6. Satchari, 7. Kalenga,8. Fenchuganj, 9. Chashnipeer, 10. Syed Jahan, 11. Malnichara,12. Khadimnagar, 13. Haripur, 14. Jaintapur, 15. Chandpur, 16.Chunati, 17. Fashiakhali, 18. Dulahazara, 19. Charmuguria, 20.Kolargaon, 21. Naria, 22. Nandanshar, 23. Kartikpur, 24.Wajirpur, and 25. Karamjal, Sundarbans. Areas bounded by ared line represent the three sub‐populations of rhesus macaquesin Bangladesh separated by river barriers. Sampling site 5 liesjust on the western side of the main channel of the BrahmaputraRiver before it course was shifted westward by the 1,762earthquake.

Am. J. Primatol.

Diversity of Rhesus Macaques in Bangladesh / 1095

questions about the origin and distribution of rhesusmacaques in their geographic range. The presentstudy focuses exclusively on genetic variation in themitochondrial DNA of rhesus macaque populationswithin Bangladesh and comparison of this variationwith other populations of the species’ geographicrange. We characterized the (a) population structureand genetic diversity of rhesus macaques ofBangladesh; (b) geographical distribution of geneticvariation; (c) genetic diversity within and betweenthe local populations of rhesus macaques; and (d)phylogenetic relationship of Bangladeshi rhesusmacaques to the other rhesus macaque populationsin their geographic range.

We hypothesize that (i) the rhesus macaquepopulation of Bangladesh comprises three sub‐populations isolated from one another by riverbarriers and (ii) rhesus macaques expanded toBangladesh from the east, expanded rapidly duringHolocene times and dispersed westward along theIndo‐Ganghetic Plain and Indus River valley to theirpresent geographic range through what are nowIndia, Pakistan, and Afghanistan. We predict that (i)rhesus macaque sub‐populations in Bangladesh willbe genetically distinct from each other and thatgenetic distances among them will not be highly

correlated with corresponding geographic distancesdue to river barriers to gene flow; (ii) Tajima’s test ofselective neutrality will exhibit a negative value of Dcharacteristic of a rapidly expanding Bangladeshirhesus population; and (iii) the southeastern portionof the eastern sub‐population will exhibit geneticsimilarity to Myanmar rhesus macaque populationsand contain the greatest genetic diversity while thecentral and western population of Bangladeshirhesus macaques will be more genetically similar toIndian rhesus macaques.

METHODSSample Collection

Swabs from the surface of 86 fecal samples and 19blood samples were collected from 25 local freeranging rhesus macaque populations belonging tothe three sub‐populations of the country fromJanuary 2010 to August 2012 (Table I). Each of thethree sub‐populations is geographically isolated fromeach other by physical barriers to natural dispersion(e.g., major river systems) and/or subsequent anthro-pogenic effects (e.g., habitat destruction resulting in alack of natural corridors for movement). Fecal swabs

TABLE I. Sampling Sites With the Number of Recorded Haplotypes

Sampling sites GPS coordinationNo. of

samplesNo. of

haplotypes Haplotype names

Central sub‐population1. Sadhana N 23°42.1920 E 90°25.4770 8 2 BR36, BR372. Dhamrai N 23°55.070 E 90°12.420 5 2 BR34, BR353. Narayanganj N 23°36.9030 E 90°30.7160 8 4 BR1, BR5, BR38, BR394. Bormi N 24°14.5880 E 90°31.3910 5 3 BR26, BR27, BR285. Madhupur N 24°41.193, E90°08.569 7 1 BR23

Eastern sub‐population6. Satchari N 24°07.414 E 91°26.56.660 7 5 BR9, BR10, BR11, BR43, BR457. Kalenga N 24°09.540 E 91°37.320 4 2 BR31, BR338. Fenchuganj N 24°39.5430 E 91°58.5530 1 1 BR319. Chashnipeer N 24°54.5350 E 91°52.65630 5 2 BR16, BR1710. Syed Jahan N 24°55.7450 E 91°52.5720 7 4 BR17, BR20, BR21, BR2211. Malnichara N 24°56.2510 E 91°52.0740 2 2 BR18, BR3112. Khadimnagar N 24°57.4380 E 91°55.9790 1 1 BR1713. Haripur N 24°58.0540 E 92°01.5800 1 1 BR1914. Jaintapur N 25°05.9300 E 92°07.7680 1 1 BR1415. Chandpur N 23°13.6960 E 90°38.5430 6 2 BR1, BR716. Chunati N 21°55.4970 E 92°3.4960 4 3 BR3, BR41, BR4217. Fashiakhali N 21°40.450 E 92°04.490 2 2 BR44, BR4618. Dulahazara N 21°40.0840 E 92°4.7880 1 1 BR40

Southwest sub‐population19. Charmuguria N 23°10.2490 E 90°10.0360 5 3 BR8, BR12, BR1320. Kolargaon N 23°16.2760 E 90°28.6080 2 1 BR121. Naria N 23°18.3170 E 90°24.7130 2 1 BR122. Nandanshar N 23°17.8680 E 90°28.6330 2 1 BR123. Kartikpur N 23°17.7320 E 90°28.7370 6 3 BR1, BR2, BR424. Wajirpur N 22°49.2610 E 90°15.0370 4 2 BR29, BR3025. Karamjal, SB N 22°25.4570 E 89°35.6470 9 5 BR6, BR15, BR24, BR25, BR32

Am. J. Primatol.

1096 / Hasan et al.

were fixed in lysis buffer (1M Tris–HCl, 0.5M EDTA,5MNaCl and SDS). The animals in some populationswere trapped for blood samples and released to thesame population. Sampling was conducted under theUniversity of Washington Institutional Animal Careand Use Committee and Bangladesh Ministry ofEnvironment and Forest Permit. Detailed trappingand sampling techniques were previously reported inJones‐Engel et al. [2006] and Feeroz et al. [2013].Biological samples were transported to the UnitedStates under the permission of Convention onInternational Trade in Endangered Species.

Laboratory Analyses

The Promega PCR Clean‐up System (Promega,Madison,WI, USA)was used to extract DNA from thefecal swabs using the manufacturer’s protocol. TheQIAGEN (QI Aamp DNA Blood Mini Kit, Qiagen,MD, USA) kit was used for blood DNA extraction.Extracted DNAs were quantified using the Qu-bitdsDNA HS Assay kit and diluted to 10ng/ml forPCR amplification. We analyzed 762 base pairs (bp)of the non‐coding D‐loop of mtDNA (nps 15,777–16,539), including hypervariable region I (HVR I) anda portion of cytochrome b gene. The mtDNAsequences were amplified using primers 15167F(50–30) ATGCAAGGCGCCACGATTT and 16050R(50–30) CCGAGCGAATGCCACC. This primer pairwas designed to replace a pair of previously usedprimers that readily amplified a pseudogene(numtDNA) [Smith & McDonough, 2005]. Thispseudogene [Bensasson et al., 2001; Collura &Stewart, 1995] was easily identifiable by its extremedivergence from all rhesus sequences amplified withthe redesigned pair of primers, an abundance oftransversions and indels, a paucity of polymorphic,but consistently heteroplasmic, sites and the failureof its HVS I to readily align with the referenceMacaca sylvanus sequence. The redesigned primersyielded sequences with specific mutations thatprovide the haplogroup structure characteristic oftrue mtDNA sequences, with the majority of thevariation in the HVR I region [Smith & McDonough,2005]. This conspicuous difference between thepseudogene and the mtDNA fragments amplifiedprovided assurance that pseudogene sequenceswere not included in the sequences analyzed in thisstudy.

For amplification, 2ml of DNA template was usedin a 25ml of PCR reaction (nuclease free water, 10mMof dNTP, 10� PCR buffer, 50mM of MgCl2, 10mM offorward and reverse primers and Platinum TaqDNApolymerase). Amplification was carried out under thefollowing thermocycler conditions: 95°C for 3min, 36cycles of 95°C for 20 sec denaturing, 55°C for 10 secannealing, and 72°C for a 40 sec extension). Success-ful amplification was confirmed by the identificationof a fragment of the appropriate size after 6%

polyacrylamide Minigel electrophoresis. PCR prod-ucts were cleaned‐up using 2ml of ExoSAP‐IT to 5mlof PCR product at 37°C for 15min followed by 15minat 80°C.

Sequencing was performed on an ABI PRISM3130xlGenetic Analyzer (Applied Biosystems, FosterCity, CA) using BigDye Terminator v3.1 (AppliedBiosystems) reaction mix, 5� BigDye sequencingbuffer, 1mM of forward and reverse primers, Exo-SAP‐IT PCR product and nuclease free water. Thethermocycler conditions were 96°C for 1min dena-turing, 25 cycles of 96°C for 10 sec, 50°C for 5 sec, and60°C for 4min. BigDyeXTerminator solution andSAM solution were added to the PCR product andrigorously vortexed for 30min for the purification ofthe PCR products. The electropherograms werealigned with Sequencher v5.0 (Gene Codes Corp.,Ann Arbor, MI).

Data AnalysisThe 105 Bangladeshi rhesus macaque sequences

were compared with 109 Indian, 21 Nepali, 13Myanmar, and 88 Chinese rhesus macaque sequen-ces that have already been described [Kyeset al., 2006; Smith & McDonough, 2005] (Fig. 2).Phylogenetic analyseswere carried out inMEGAv5.2using the maximum composite likelihood distancemodel, and a neighbor‐joining (NJ) tree was con-structed from a bootstrap analysis with 1,000replicates. Other distance models employed inMEGA v5.2 generated trees with identical topology

Fig. 2. Map of the geographic range ofMacacamulatta; pink dotsare the sampling locations in Bangladesh and origin of mtDNAsequences used for comparison. Locations 1–4 in India (Kashmir,New Delhi, Uttar Pradesh, and Lucknow, respectively), 5 inNepal (Kathmunda), 6–9 in Bangladesh (northeast, central,southwest, and southeast, respectively), 10 in Myanmar(Yangun), 11–15 in China (China West: 11 and 12, SichuanandKunming, respectively; ChinaSouth: 13 and 14, Guangxi andGuangdong, respectively; ChinaEast: 15, Suzhou), 16 in Vietnam[Smith & McDonough, 2005].

Am. J. Primatol.

Diversity of Rhesus Macaques in Bangladesh / 1097

and near‐identical bootstrap values. TheARLEQUINsoftware package (version 3.5.1.2) was used toestimate haplotype frequencies and molecular diver-sity indices and construct a mismatch distribution.Network 4.611was used to construct a median‐joining (MJ) haplotype network rooted with theM. sylvanus reference sequence cited above.

This research adhered to the American Society ofPrimatologists principles for the ethical treatment ofnon‐human primates. The captured animals em-ployed have been managed in compliance withInstitutional Animal Care and Use Committee(IACUC) regulations.

RESULTSGenetic Variation Within Bangladesh

Overall genetic variationA total of 46 mtDNA haplotypes was detected in

25 local rhesus populations in Bangladesh (Table I).GenBank accession numbers for these sequencesare KJ767675–KJ767715. A total of 200 (26%) of the762 nucleotide sites studied were polymorphic andno indels (insertion/deletion) were observed. Most ofthe substitutions (78.2%) were transitions resultingin a transition: transversion ratio of approximately3.6:1. A gene diversity of 0.9599� 0.0097 andexpected heterozygosity of 0.0193� 0.0582 wereestimated for the total population. Themean numberof pair wise differences and average nucleotidediversity among the 762 bp were 14.0766� 6.3639and 0.0196� 0.0098, respectively. Nucleotide compo-sition was C (28.21%), T (27.12%), A (32.41%), and G(12.27%). The negative value of Tajima’s D (D¼�2.1104) for haplotypes of rhesus macaques ofBangladesh was statistically significant at the 0.01level of probability with an excess of haplotypes of lowfrequency, reflecting a geologically recent history ofrapid population expansion.

Distribution of Haplotypes

MtDNA haplotypes of the three rhesus sub‐populations of Bangladesh formed the three distincthaplogroups in the haplotype network shown inFigure 3.

Haplogroup 1This haplogroup includes haplotypes from all

three sub‐populations. These include seven haplo-types from the eastern sub‐population, five haplo-types from central sub‐population and twohaplotypes from the southwest sub‐population. Thisis the oldest of the three haplogroups; haplotypesBR43, BR44, BR45, and BR46 are the oldesthaplotypes when the haplotype network is rootedwith theM. sylvanus reference sequence (NC002764)as shown in Figure 3.

Haplogroup 2This haplogroup predominantly comprises hap-

lotypes from the southwest sub‐population but alsoincluded haplotypes from each of the two other sub‐populations. Most of the haplotypes of the southwestsub‐population diverged from a single haplotype ofthe central sub‐population (BR23) in the haplotypenetwork (Fig. 3) that is also the root of somehaplotypes from the eastern sub‐population (BR14,BR31). The Sundarbans sub‐population (Southwest),which is 256 km from the central sub‐population andcrisscrossed bymany large rivers, is isolated bymanyanthropogenic factors, but three of its haplotypes(BR24, BR25, and BR32) are derived from a singlecentral haplotype, BR23 (Fig. 3). Haplotype BR1 ofthis haplogroup was shared by all three sub‐populations including the central sub‐population,which is about 50 km away from the other sub‐populations, and located on both sides of a large river(the Meghna).

Haplogroup 3This haplogroup primarily comprises haplotypes

from the eastern sub‐population (11 haplotypes) butalso included two haplotypes (BR6 and BR15) fromthe southwest sub‐population (Sundarbans) and fourhaplotypes from central sub‐population as illustratedby the haplotype network (Fig. 3). Three haplotypes(BR34, BR35, and BR37) of the central sub‐popula-tion and two haplotypes (BR11 and BR33) of thenortheast sub‐population derive from a single haplo-type (BR15) of the Sundarbans (southwestern) sub‐population, though these sites are 200–260km apartand isolated from each other by many rivers andother barriers.

The phylogenetic (NJ) tree of the haplotypes fromthese three sub‐populations formed five distinct sub‐clusters (Fig. 4). Consistent with the haplotypenetwork (Fig. 3), the haplotypes from the eastern(BR45 and BR46) sub‐population formed the root ofthis tree. All the remaining haplotypes except thosefrom the southeast and two from the northeast (BR43and BR45) formed a single sub‐cluster with lowbootstrap value.Most of the haplotypes of the studiedpopulations exhibited locality‐specific distributions.Except for BR1, BR17, and BR31, each of thehaplotypes is unique to a local population of rhesusmacaques in Bangladesh. BR1 was shared by fourdifferent local populations (Kolargaon, Kartikpur,Nandanshar, and Naria) of the southwest sub‐population, one local population (Chandpur) of theeastern sub‐population, and one local population(Narayanganj) of the central sub‐population of thecountry. BR17 was shared by two local populations(Khadimnagar and Syed Jahan shrine) and BR31was shared by three local populations (Kalenga,Fenchuganj, and Malnichara) of the eastern sub‐population. Approximately 64% of the 25 local

Am. J. Primatol.

1098 / Hasan et al.

populations studied exhibited more than a singlehaplotype.

Relationship Between Geographic Distanceand Genetic Distance

The sampling locations of rhesus macaques inBangladesh are separated from each other bygeographic distances by between 3 and 393km. Insome locations where local populations of rhesusmacaques were separated from each other by a shortgeographic distance, human settlements, and/orriver systems of the country precluded gene flowbetween them. The correlation between the geo-graphic and genetic distances among these localpopulations was positive and statistically significant

(r¼ 0.1726,P< 0.001) as shown in Figure 5. However,when the southeastern sub‐population is omittedfrom theanalysis, the correlation between genetic andgeographic distance is not statistically significant,suggesting that the southeastern sector of the easternsub‐population is the population least effectivelyisolated by geographic barriers.

Relationship With Rhesus PopulationsOutside Bangladesh

When the haplotypes of Bangladeshi rhesusmacaques were compared with the available sequen-ces for rhesus macaques in other populations [Kyeset al., 2006; Smith & McDonough, 2005], threehaplotypes (BR46, BR43, and BR45 from east)

Fig. 3. Median‐joining haplotype network rooted with aMacaca sylvanus sequence illustrating three different haplogroups. Colors of thecircles represent haplotypes from specific sub‐populations of rhesus macaques in Bangladesh using the color codes cited in Figure 1 andthe dark blue represents the M. sylvanus sequence, the root of the network. Sizes of the circles indicate the number of samples of thatparticular haplotype.

Am. J. Primatol.

Diversity of Rhesus Macaques in Bangladesh / 1099

formed a separate sub‐cluster representing the root ofthe NJ tree in Figure 6 with a 95% bootstrap value.Most of the remaining Bangladeshi haplotypesformed a separate sub‐cluster with haplotypes ofthe Ind‐2 haplogroup of India defined by Smith andMcDonough [2005]. Three haplotypes (BR40, BR41,and BR42) from the southeastern part of Bangladeshclustered with all the Myanmar haplotypes. One

haplotype (BR44) from the southeast formed the rootof all Chinese samples (though with a very lowbootstrap value). All the haplotypes from haplogroupInd‐1 from India [Smith &McDonough, 2005] formeda separate sub‐cluster that included none of theBangladeshi haplotypes. All the rhesus populationsof Bangladesh (except one northeast local populationand all southeastern populations) formed a single

Fig. 4. Neighbor‐joining (NJ) tree of rhesusmacaquehaplotypes showingfive sub‐clusters. Color bars represent the three sub‐populationsof rhesus macaques in Bangladesh using the color codes cited in Figures 1 and 2. Capital letters on the nodes correspond to nodes inFigure 5.

Am. J. Primatol.

1100 / Hasan et al.

monophyletic sub‐cluster, suggesting the monophy-letic origin of most rhesus macaques in Bangladesh.

DISCUSSIONHaplotype diversity among the rhesus macaques

of Bangladesh exhibited a locality‐specific distribu-tion. All but three of the haplotypes (BR1, BR17, andBR31), each shared by the adjoining local popula-tions, except Narayanganj, represented a specificlocal population. Historically, no forest existed inNarayanganj and its vicinity, and the area has beenused as a river port and business hub for at least1,000 years [Islam et al., 2003]. Among the fourhaplotypes from Narayanganj, two haplotypes (BR1and BR5) resembled those from the sub‐populationsof other riverine areas (Chandpur, Kolargaon, Nan-danshar, Naria, and Kartikpur), but two others(BR38, BR39) resembled those from the easternsub‐population. Human‐aided dispersal of monkeysto Narayanganj may have been facilitated by thearea’s ancient status as a river port and business hub.Significantly, none of the many rhesus macaque localpopulations between Narayanganj and the easternregion of the country resembled the Narayanganjpopulation. This result is consistent with our recentreport [Feeroz et al., 2013] in which we discussNarayanganj as Bangladesh’s “melting pot” forreleased and/or escaped pet macaques.

Sundarbans is the largest continuous mangroveforest in the world, covering an area of 6,017 km2 andcrisscrossed by fourmajor rivers and their tributaries[Barlow et al., 2011; Khan, 2011]. In spite ofgeographic and anthropogenic barriers, the rhesus

macaque population from Sundarbans (Karamjal)resembled those of central (Madhupur) and northeast(Satchari, Kalenga, Fenchuganj, and Malnichara)Bangladesh. These three sites are geographicallyisolated fromeach otherwith several rhesusmacaquelocal populations intervening among them whosehaplotypes lack resemblance to those from Sundar-bans.Madhupur forest was located on thewest side ofthemain channel of Brahmaputra. Themain channelof the Brahmaputra was shifted southward, openingthe present main channel (Jamuna in Bangladesh),due to the tectonic uplift of the Madhupur Tractduring the 1,762 earthquake episode, placing Mad-hupur forest in between old and present channels ofBrahmaputra [Bhuiyan et al., 2010; Sussanet al., 1904]. The genetic resemblance of rhesusmacaques from Madhupur region to Sundarbans inthe southwest is reasonable because the Brahmapu-tra was not a barrier between these two regionsbefore 1,762 earthquake. The intervening localpopulations between Sundarbans and Madhupurmay have expanded following a different dispersalroute. Large areas of Bangladesh have been defor-ested over the last two decades. The moist deciduousforests of Bhawal and Madhupur once covered muchof the areas between the Padma, Jamuna, andMeghna rivers [Feeroz et al., 2013]. This extensiveforest cover throughout the central Bangladeshwould have allowed movement of rhesus macaques.

As in other Asian countries, non‐human primate(NHP) pets are not uncommon in Bangladesh. Own-ers of NHP pets face a severe problem whenconfronted with a monkey acquired as an infantthat has grown into an aggressive adult [Feerozet al., 2013] and, consequently, usually release theiradult pets in the nearest NHPpopulation. In additionto NHP pet ownership, performing monkey owner-ship is a centuries old tradition in Bangladesh.Owners of performing monkeys usually capturemonkeys from the forests and urbanmonkey habitatsand train them for performance, and monkey tradingis common among the monkey performers’ communi-ty [Akhtar, 2014; Feeroz et al., 2013]. Release ofuntrainable monkeys to the nearest monkey popula-tion by this community is also common[Akhtar, 2014]. That Sundarbans (Karamjal) is themajor source of the performing monkeys [Engelet al., 2013; Feeroz et al., 2013] is consistent withour findings. However, Karamjal is situated at thenorthern periphery of Sundarbans and its rhesusmacaques may not be representative of those of theentire Sundarbans. Detailed sampling throughoutthe Sundarbans could clarify the rhesus macaquedistribution in this remotest forest as well as in itsgeographic range.

The haplotypes fromChandpur, another old riverport, and Shariatpur (Nandanshar, Kolargaon, Kar-tikpur, and Naria), geographically isolated from eachother by the huge riverMeghna (about 5.5 kmwidth),

Fig. 5. Relationship between genetic distance and geographicdistance among the rhesus macaques of Bangladesh. The red boxidentifies all samples from the southeastern region ofBangladesh. Note that the correlation between genetic andgeographic distances is not statistically significant when thesamples from the southeast region are not included.

Am. J. Primatol.

Diversity of Rhesus Macaques in Bangladesh / 1101

resembled each other and include a common haplo-type (BR1). The extension of the Meghna into India(the Brahmaputra river) has been argued to provide abarrier to gene flow separating the two majorsubspecies of rhesus [Melnick et al., 1993] andAssamese [Fooden, 1988] macaques. Monkeys might

cross this river as stow‐aways on water vehicles, byriver rafting or with the aid of humans. If, as ouranalysis suggests, the rhesus macaques in these twoareas share a recent common ancestry, the Meghnahas not provided a barrier to gene flow, perhaps dueto meandering of its tributaries and/or their

Fig. 6. Neighbor‐joining tree based on 762bp sequences of mitochondrial DNA of rhesus from regional populations within their naturalrange. The size of the base of each triangle is proportional to sample size. Note that the oldest haplotypes are from Bangladesh and thatmost of the Bangladesh samples cluster with the Ind‐2 haplotypes from India reflecting a westward direction of gene flow.

Am. J. Primatol.

1102 / Hasan et al.

circumnavigation downstream toward the Deltawhere the tributaries narrow.

All the southeast rhesus macaque populations ofBangladesh and a single local population of thenortheast (Satchari) resembled each other, forminga separate sub‐cluster. Though the northeast(Satchari) and the southeastern populations areabout 250 km from each other, these regions areconnected to each other through the hill forest ofTripura of eastern India (Fig. 1). Haplotypes from theeastern region of Bangladesh are the oldest hap-lotypes and are genetically distant from the otherhaplotypes of the country. The southeast region ofBangladesh is close to and continuous with theAracan hill range of Myanmar, and the sequences ofrhesus macaques in the two areas form a cluster,apparently reflecting their recent common ancestry.As the southeastern sector of the eastern sub‐population is the oldest and rhesus macaques inBangladesh rapidly expanding in the recent past,rhesus macaques probably first expanded northwardand westward through the country from southeast-ern Bangladesh.

According to Smith and McDonough [2005],rhesus macaques throughout their geographic rangeformed at least four distinct clusters (i.e., hap-logroups) of mtDNA: Ind‐1, Myanmar‐Ind‐2 in Indiaand ChiE and ChiW in China/Vietnam. They arguedthat the Ind‐2 haplogroup, comprising approximately5% of the Indian haplotypes, derived from a recentwestward dispersal of rhesus macaques from Myan-mar. When we compared our sequences with those ofSmith and McDonough [2005], we found five distinctclusters: (a) a separate Bangladesh cluster; (b) aMyanmar–Bangladesh cluster; (c) a Bangladesh–India‐2 cluster; (d) an India‐1 cluster; and (e) aChina–Vietnam cluster. Ind‐2 haplotypes from wide-ly geographically dispersed locations in India (threefromNewDelhi, one fromKashmir, and another fromUttar Pradesh) clustered with some Bangladeshirhesus macaque sequences (cluster c, above). Thesethree Ind‐2 sampling sites are very far fromBangladesh sampling sites, and none of the Ind‐1haplotypes resembles any of the Bangladeshi or Ind‐1rhesus haplotypes. Because the samples used bySmith and McDonough [2005] were collected fromcaptive breeding centers, not directly from the wild,the validity of their alleged origin locality isproblematical. The captive breeding centers collectedIndian rhesus macaques from dealers before 1978,when regional origin was not regarded as importantandmay have beenmisrepresented. If valid, the widegeographic distribution of both Ind‐1 and Ind‐2 inIndia might reflect two very early westward dis-persals of rhesus macaques at two differenttimes and/or by two different routes. Samples ofrhesus macaques from wild populations of Assam,Meghalaya, Tripura, and other regions adjoiningBangladesh may clarify this hypothesis.

Macaca mulatta and M. fascicularis probablydiverged during middle Pleistocene times from acommon fascicularis‐like ancestor [Deinard & Smith,2001]. Smith and McDonough [2005] argued that thecurrent geographic distribution of rhesus macaquesin India resulted from a westward dispersal, orredispersal, that postdates the species’ dispersalthrough mainland Southeast Asia and China.Melnick and Kidd [1985] suggested that geneticsimilarity between Chinese rhesus macaques andcynomolgus macaques in Thailand results from thedivergence of rhesus from cynomolgus macaques inThailand during a glacial maximum, followed by thedispersal of cynomolgus macaques to the south. Wefound that the oldest known rhesus macaquehaplotypes of the species geographic range are thosefrom Bangladesh, which may support the argumentthat rhesus macaque originated in rainforests ofChittagong Hill Tracts in southeast Bangladesh thatincludeAracanHill Range ofMyanmar and dispersedeastward towards China–Vietnam and westwardtoward western India as hypothesized by Smithand McDonough [2005]. If so, a refuge may haveexisted in Chittagong Hill Tracts‐Aracan Hill Rangeduring either the last or penultimate glacial maxi-mum. The latter is more consistent with the 162,000years BP divergence date between Indian andChinese rhesus macaques estimated by Hernandezet al. [2007] based on 1,467 single nucleotide poly-morphisms. Plans are currently underway to analyzemore samples from the southeast regions ofBangladesh to evaluate this hypothesis.

ACKNOWLEDGMENTS

This work was supported by a base grant to theCalifornia National Primate Research Center fromNational Center for Research Resources at theNational Institutes of Health (RR000169‐48 to D.G.S.), and RR005090 and RR025781 (D.G.S.) and theNational Institute of Allergy and Infectious Diseases(R01 AI078229; R01AI078229‐03S1; R03 AI064865to L.J.‐E.). We are grateful to the Forest Departmentof Bangladesh for their permission and constantsupport during fieldwork and to the authority of theWildlife Rescue Center (WRC) of the Department ofZoology, Jahangirnagar University for their logisticsduring sample collection. The first author thanksDr. Yoshi Kawamoto and the Primate ResearchInstitute (PRI), Kyoto University, Japan for provid-ing valuable training in molecular genetics.

REFERENCES

Abegg C, Thierry B. 2002. Macaque evolution and dispersal ininsular south‐east Asia. Biological Journal of the LinneanSociety 75:555–576.

Akhtar S. 2014. Monkey performers in Bangladesh: monkeypopulation, health care, human–monkey interaction and

Am. J. Primatol.

Diversity of Rhesus Macaques in Bangladesh / 1103

socio‐economic condition of the monkey owners [MPhildissertation]. Savar, Dhaka, Bangladesh: Department ofZoology, Jahangirnagar University.

Barlow AC, Smith JL, Ahmad IU, et al. 2011. Female tigerPanthera tigris home range size in the Bangladesh Sundar-bans: the value of this mangrove ecosystem for the species’conservation. Oryx 45:125–128.

Bensasson D, Zhang D‐X, Hartle DL, et al. 2001. Mitochondrialpseudogenes: evolution’s misplaced witnesses. Trends inEcology & Evolution 16:314–321.

Bhuiyan MA, Rakib M, Takashi K, et al. 2010. Regulation ofBrahmaputra–Jamuna river around Jamuna bridge site,Bangladesh: geoenvironmental impacts. Journal of WaterResource and Protection 2:123–130.

Collura RV, Stewart CB. 1995. Insertions and duplications ofmtDNA in the nuclear genomes of Old World monkeys andhominoids. Nature 378:485–489.

Deinard A, Smith DG. 2001. Phylogenetic relationships amongthe macaques: evidence from the nuclear locus NRAMP1.Journal of Human Evolution 41:45–59.

Delson E. 1980. Fossil macaques, phyletic relationships and ascenario of deployment. New York, NY: Van NostrandReinhold. p 10–30.

Engel GA, Small CT, Soliven K, et al. 2013. Zoonotic simianfoamy virus in Bangladesh reflects diverse patterns oftransmission and co‐infection. Emerging Microbes & In-fections 2:e58.

Eudey AA. 1980. Pleistocene glacial phenomena and theevolution of Asian macaques. In: Lindburg DG, editor. Themacaques: studies in ecology, behavior and evolution. NewYork, USA: Van Nostrand Reinhold. p 52–83.

Feeroz MM. 2001. Species diversity and population density ofnon‐human primates in north‐east and south‐east ofBangladesh. Ecoprint 8:53–57.

Feeroz MM, Hasan K, Hamada Y, Kawamoto Y. 2008. STRpolymorphism of mtDNA D‐loop in rhesus macaques ofBangladesh. Primates 49:69–72.

Feeroz MM, Hasan MK, Khan MMH. 2011. Biodiversity ofprotected areas of Bangladesh. Vol. I: Rema‐Kalengawildlifesanctuary. Bangladesh: BioTrack. Arannayk Foundation.

Feeroz MM, Soliven K, Small CT, et al. 2013. Populationdynamics of rhesus macaques and associated foamy virus inBangladesh. Emerging Microbes & Infections 2:e29.

Ferguson B, Street SL, Wright H, et al. 2007. Single nucleotidepolymorphisms (SNPs) distinguish Indian‐origin and Chi-nese‐origin rhesus macaques (Macaca mulatta). BMCGenomics 8:43.

Fooden J. 1976. Provisional classification and key to livingspecies of macaques (primates:Macaca). Folia Primatologica25:225–236.

Fooden J. 1988. Taxonomy and evolution of the Sinica group ofmacaques. VI. Interspecific comparisons and synthesis.Fieldiana Zoology 45:1–44.

Fooden J. 2006. Comparative review of fascicularis‐groupspecies of macaques (primates: Macaca). Fieldiana Zoology107:1–43.

Hasan MK. 2003. Holock gibbon of Bangladesh: statusdistribution and ecology [MSc dissertation]. Savar, Dhaka,Bangladesh: Department of Zoology, JahangirnagarUniversity.

Hasan MK, Aziz MA, Alam SR, et al. 2013. Distribution ofrhesus macaques (Macaca mulatta) in Bangladesh: inter‐population variation in group size and composition. PrimateConservation 26:125–132.

Hernandez RD, Hubisz MJ, Wheeler D, et al. 2007. Geneticvariation reveals diametric demographic histories andpatterns of linkage disequilibrium for Chinese and Indianrhesus macaques. Science 316:240–243.

Islam M, Ameen M, Nishat A. 2000. Red book of threatenedmammals of Bangladesh. Dhaka: IUCN.

Islam N, Jamal AA, Chowdhury AM. 2003. Banglapedia:national encyclopedia of Bangladesh. Asiatic Society ofBangladesh. 2:415.

Jones‐Engel L, Engel GA, Heidrich J, et al. 2006. Templemonkeys and health implications of commensalism, Kath-mandu, Nepal. Emerging Infectious Diseases 12:900–906.

Khan MAR. 1982. Wildlife of Bangladesh: a checklist. Dhaka:University of Dhaka.

Khan MMH. 2011. Tigers in the mangroves: research andconservation of the tiger in the Sundarbans of Bangladesh.Dhaka, Bangladesh: Arannayk Foundation Bangladesh.

Kyes RC, Jones‐Engel L, Chalise MK, et al. 2006. Geneticcharacterization of rhesus macaques (Macaca mulatta) inNepal. American Journal of Primatology 68:445–455.

Melnick DJ, Kidd KK. 1985. Genetic and evolutionary relation-ships among Asian macaques. International Journal ofPrimatology 6:123–160.

Melnick DJ, Hoelzer GA, Absher R, Ashley MV. 1993. mtDNAdiversity in rhesus monkeys reveals overestimates ofdivergence time and paraphyly with neighboring species.Molecular Biology and Evolution 10:282–295.

Ramachandran S, Deshpande O, Roseman CC, et al. 2005.Support from the relationship of genetic distance andgeographic distance in human populations for a serialfounder effect originating in Africa. Proceedings of theNational Academy of Sciences of the USA 102:15942–15947.

Smith DG, McDonough J. 2005. Mitochondrial DNA variationin Chinese and Indian rhesus macaques (Macaca mulatta).American Journal of Primatology 65:1–25.

Stanford CB. 1991. The capped langur in Bangladesh:behavioral ecology and reproductive tactics. Vol. 26. NewYork, USA: Karger Publishers.

Stevison LS, Kohn MH. 2009. Divergence population geneticsanalysis of hybridization between 2457 rhesus and cyno-molgus macaques. Molecular Ecology 18:2457–2475.

Sussan E, Sollas J, Sollas HBC. 1904. The face of the earth.Oxford, London: Clarendon Press.

Tosi AJ, Morales JC, Melnick DJ. 2003. Paternal, maternal,and biparental molecular markers provide unique windowsonto the evolutionary history of macaque monkeys. Evolu-tion 57:1419–1435.

Am. J. Primatol.

1104 / Hasan et al.