- Email: [email protected]
Mitochondrial DNA reveals multiple introductions of domestic chicken in East Africa
Molecular Phylogenetics and Evolution 58 (2011) 374–382
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Mitochondrial DNA reveals multiple introductions of domestic chicken in East Africa J.M. Mwacharo a,⇑, G. Bjørnstad b,c, V. Mobegi c, K. Nomura d, H. Hanada d, T. Amano e, H. Jianlin c,f, O. Hanotte a,c a
School of Biology, Centre for Genetics and Genomics, The University of Nottingham, University Park, Nottingham NG7 2RD, UK Norwegian School of Veterinary Science, Department of Basic Sciences & Aquatic Medicine, P.O. Box 8146, N-0033 Oslo, Norway c International Livestock Research Institute, P.O. Box 30709, Nairobi 00100, Kenya d Laboratory of Animal Genetics and Breeding, Department of Animal Science, Tokyo University of Agriculture, 1737 Funako Atsugi-Shi, Kanagawa 243-0034, Japan e Nakagawa 5-21-22, Tsuzuki-Ku, Yokohama-Shi, Kanagawa-Ken 224-0001, Japan f CAAS-ILRI Joint Laboratory on Livestock and Forage Genetic Resources, Institute of Animal Science, Chinese Academy of Agricultural Sciences (CAAS), Beijing 100094, China b
a r t i c l e
i n f o
Article history: Received 1 November 2010 Accepted 29 November 2010 Available online 5 December 2010 Keywords: Gallus gallus Village chicken Control region Indian Ocean
a b s t r a c t Chicken were possibly domesticated in South and Southeast Asia. They occur ubiquitously in East Africa where they show extensive phenotypic diversity. They appeared in the region relatively late, with the first undisputed evidence of domestic chicken in Sudan, around 700 BC. We reveal through a detailed analysis of mitochondrial DNA D-loop sequence diversity of 512 domestic village chickens, from four East African countries (Kenya, Ethiopia, Sudan, Uganda), the presence of at least five distinct mitochondrial DNA haplogroups. Phylogeographic analyses and inclusion of reference sequences from Asia allow us to address the origin, ways of introduction and dispersion of each haplogroup. The results indicate a likely Indian subcontinent origin for the commonest haplogroup (D) and a maritime introduction for the next commonest one (A) from Southeast and/or East Asia. Recent introgression of commercial haplotypes into the gene pool of village chickens might explain the rare presence of two haplogroups (B and C) while the origin of the last haplogroup (E) remains unclear being currently observed only outside the African continent in the inland Yunnan Province of China. Our findings not only support ancient historical maritime and terrestrial contacts between Asia and East Africa, but also indicate the presence of large maternal genetic diversity in the region which could potentially support genetic improvement programmes. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Village chickens (Gallus gallus) are ubiquitous in East Africa with a total population of 100.8 million birds (FAOSTAT, 2007). These birds fulfil various roles ranging from socio-cultural to sustaining livelihoods, suggesting a long historical presence in the region. Chicken are not a native species to Africa where no wild Gallus sp. is found (Delacour, 1977). A previous phylogenetic study of chicken mitochondrial DNA (mtDNA) control region suggested that the Indochinese red junglefowl subspecies Gallus gallus gallus was the primary maternal ancestor of all domestic fowls G. g. domesticus and that Southeast Asia (Thailand) was the likely center of domestication (Fumihito et al., 1996). A recent study, which analysed a larger fragment of the mtDNA control region encompassing the first hypervariable segment, in a comparatively large and diverse gene pool of domestic chickens from a wide geographic area (Europe and Asia), was the first to suggest multiple maternal geographic centers of origin for the domestic species (Liu et al., 2006). They identified nine divergent clades, (called here ⇑ Corresponding author. Fax: +44 (0) 115 951 3251. E-mail address: [email protected] (J.M. Mwacharo). 1055-7903/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2010.11.027
haplogroups), seven of which included both domestic chickens and wild red junglefowl subspecies haplotypes. Three of the nine clades involved mtDNA control region haplotypes found across Europe and Asia, whereas the other six clades contained haplotypes found exclusively in South and Southeast Asia. Both Fumihito et al. (1996) and Liu et al. (2006) studies lacked samples from the African continent. More recently, three studies have made an attempt to address the origin of African village chickens through the analysis of partial mtDNA D-loop sequences. Muchadeyi et al. (2008) observed two distinct haplogroups in Zimbabwe village chickens which they postulated came from Southeast Asia and the Indian subcontinent. Similarly, Razafindraibe et al. (2008) observed two haplogroups in Madagascar village chicken and speculated that one was of Indonesian and the other of African continental origin or an introgression from commercial lines. At the opposite a single haplogroup thought to be of Indian origin was observed in Nigeria village chickens by Adebambo et al. (2010), while no information is yet available for the East African region. Since historical times, Africa and Asia have been linked via maritime and terrestrial corridors. The Indian Ocean was one of the world’s earliest arenas that provided a maritime corridor linking
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Africa and Asia. Following the monsoon wind patterns of the Indian Ocean, Asia and East Africa witnessed important seafaring and maritime exchanges across historical times. Asia and Africa were also connected via a terrestrial route that traversed the Arabian Peninsula, a region which occupies a key geographic junction, with the African landmass to the West and the Asian continent to the East (Boivin and Fuller, 2009). These terrestrial and maritime routes witnessed the long distance movement of people, livestock and crops within and between continents. Domesticates that dispersed from Asia to Africa included cattle, pigs, chicken, broomcorn millet, bananas, water yam and taro and those that moved from Africa to Asia are donkey’s, sorghum, pearl millet, finger millet, cowpeas and hyacinth beans (Fuller and Boivin, 2009). Village chickens in East Africa are found across the entire region and in all agro-ecological zones; they show large within population phenotypic diversity in plumage colour, feather morphology and pattern, skin colour, comb type etc. (Msoffe et al., 2001; Dana et al., 2010a). Raised under free-range scavenging system, village chickens contribute substantially to egg and meat production under almost a zero input system under small holder subsistence economies (Sonaiya, 1997; Kitalyi, 1998). The possible geographic origin of East African village chickens remains unknown, with the Indian subcontinent, Southeast Asia and North Africa as possible primary centers of origin and entry in agreement with our knowledge of the history of trading of the region. To investigate the origin of East African village chicken, we generated mtDNA control region sequences from 512 village chickens and analysed their phylogenetic relationships with representative Asian haplotypes from the nine haplogroups described by Liu et al. (2006). In addition, we evaluated the phylogeographic structure of the village chickens to assess their pattern of diffusion across East Africa. Our results reveal an unexpected complex pattern of introduction of domestic chicken in the region and provide new information concerning the history of trading and human contacts between East Africa and Asia.
375
Triton X-100), 2.5 mM of each dNTP, 10 pM of each primer and 1 unit of Taq DNA polymerase (Promega, Madison WI, USA). Thermo-cycling conditions were: 94 °C (3 min), 35 cycles of 94 °C (1 min), 58 °C (1 min) and 72 °C (2 min), and a final extension step at 72 °C (10 min). PCR products were purified using the WizardÒ SV Gel and PCR Clean-Up Kit (Promega, Madison WI, USA). Purified products were sequenced directly using the BigDyeÒ Terminator v3.1 (Applied Biosystems, USA) on an ABI prism 3100 Avant DNA analyzer. The two PCR primers and an internal primer f1-3m (50 -TGGTTCCTCGGTCAGGCACATCC-30 ) were used for the sequencing reactions.
2.3. Sequence and phylogenetic analysis For each sample, three fragments were generated. These were edited manually using BioEdit 7.0 (Hall, 1999) and subsequently joined to reconstruct a fragment of 700 bp. The fragments were aligned using Clustal X 1.83 (Thompson et al., 1997) against a reference sequence (GenBank accession number X52392; Desjardins and Morais, 1990). Subsequent analyses were restricted to the first 397 bp incorporating the first hypervariable segment (HVS1). We set to determine the number of haplogroups present in East Africa village chickens by constructing a MedianJoining (MJ) network (Bandelt et al., 1999) using NETWORK 4.5 (fluxus-engineering.com). This analysis was augmented by constructing a phylogenetic tree involving the haplotypes observed in East Africa using the Neighbour-Joining (NJ) algorithm as implemented in MEGA 4.0 (Tamura et al., 2007) following 1000 bootstrap replications. To portray the affinity of East Africa haplotypes to those observed in Asia and Africa, a MJ network incorporating the 30 haplotypes downloaded from the GenBank (see Table S1) was also constructed. The nomenclature of the haplogroups observed in this study compared to the study of Liu et al. (2006), Silva et al. (2008) and Muchadeyi et al. (2008) are shown in Table S2.
2. Materials and methods 2.1. Sample collection and DNA extraction Blood samples from 512 genetically unrelated village chickens were collected from 23 populations in four countries in East Africa (Table 1). All samples were from unimproved village chickens raised under free-range scavenging. Two mature birds were sampled per flock and the sampling strategy and characteristics of the sampling locations were described previously (Mwacharo et al., 2007). Genomic DNA was extracted from either whole blood using phenol–chloroform or from air dried blood preserved on FTA classic cards (Whatman Biosciences) using the manufacturers protocol. To ascertain the genetic affinities of the study populations to Asiatic and other African chickens, 30 haplotypes were downloaded from the GenBank and were included in the analysis (see Table S1). For this purpose, the central and most common haplotypes for the nine clades observed in Liu et al. (2006) were selected for the study to determine the possible origins of chicken found in East Africa within the geographic range of the wild ancestor, the red junglefowl. 2.2. PCR amplification and sequencing The first 700 bp of the mtDNA D-loop region were amplified via PCR using primers AV1F2 (50 -AGGACTACGGCTTGAAAAGC-30 ) and CR1b (50 -CCATACACGCAAACCGTCTC-30 ). PCR amplifications were carried out in 25 ll reaction volumes containing 20 ng genomic DNA, 1 X PCR buffer (10 mM Tris–HCl pH 8.3, 50 mM KCl, 0.1%
2.4. Population genetic variability and structure Genetic variation (nucleotide diversity, haplotype diversity and nucleotide differences) for each population were calculated using DnaSP 5.10 (Librado and Rozas, 2009). Population genetic structure was assessed by nested analysis of molecular variance (AMOVA). The groupings used for AMOVA were as follows: (i) the overall dataset assuming no groups; (ii) between the haplogroups observed on the MJ network; (iii) between chicken populations found in different countries (country specific groups); (iv) between populations for haplogroup D only and (v) between populations for haplogroup A only. Phi (U) statistics representing haplotype correlations at various levels of hierarchical clusters UCT, USC, and UST (Excoffier et al., 2005) were calculated. Significance testing was performed using 10,100 coalescent simulations in Arlequin 3.1 (Excoffier et al., 2005). A Mantel test was used to assess the non-random association between genetic differentiation (FST) and geographic distances (km) between populations using the IBDWS v3.05 software (http://ibdws.sdsu.edu). Population pairwise FST values were calculated using Arlequin 3.1 (Excoffier et al., 2005). Geographic distances between populations were calculated using the MapCrow Travel Distance Calculator between central towns within the sampling locations based on geographical coordinates (http://www. mapcrow.info/). For the analysis all negative FST values were set to zero. This analysis was limited to haplogroup D with the widest geographic range in the region. Several comparative analyses were performed across all countries and between all countries.
376
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Table 1 Sampling locations, population abbreviations, sample sizes and genetic diversity measures and their standard deviations (SD) for each population and the five lineages observed from the Median-Joining network analysis. Country/population
Sample size
Haplogroup (number of individuals observed)
Number of haplotypes
Haplotype diversity (SD)
Nucleotide diversity (SD)
Mean number of nucleotide differences
Kenya Kilifi (KF) Taita (TT) Muranga (MG) Meru (MR) Marsabit (MT) Kitui (KT) Nairobi (NB) Kisii (KS) Homa Bay (HB) Kakamega (KK) Nandi (ND) Naivasha (NV) All
14 13 15 14 60 14 4 15 14 14 14 20 211
D D D D D D D D D D D D D
(11); A (3) (5): A (8) (5); A (10) (5); A (9) (9); A (51) (5); A (9) (3); A (1) (15) (14) (13); A (1) (14) (15); A (4); B (1) (114); A (96); B (1)
6 (D = 4; A = 2) 8 (D = 3; A = 5) 4 (D = 3; A = 1) 4 (D = 2; A = 2) 7 (D = 43; A = 3) 6 (D = 3; A = 3) 3 (D = 2; A = 1) 4 (D = 4) 6 (D = 6) 7 (D = 6; A = 1) 5 (D = 5) 5 (D = 3; A = 1; B = 1) 25 (D = 16; A = 8; B = 1)
0.747 0.923 0.543 0.648 0.637 0.747 0.833 0.638 0.736 0.912 0.593 0.647 0.857
(0.111) (0.050) (0.133) (0.116) (0.053) (0.111) (0.222) (0.093) (0.109) (0.059) (0.144) (0.088) (0.012)
0.00883 0.01318 0.01199 0.01138 0.00623 0.01262 0.01469 0.00249 0.00302 0.00747 0.00246 0.00919 0.01192
(0.00223) (0.00147) (0.00252) (0.00208) (0.00126) (0.00208) (0.00558) (0.00064) (0.00078) (0.00169) (0.00084) (0.00233) (0.00023)
3.5055 5.2308 4.7619 4.5165 2.4718 5.0109 5.8333 0.9905 1.1978 2.9670 0.9780 3.6474 4.734
Ethiopia Debre Birhan (DB) Nekemte (NE) Jimma (JM) All
15 15 12 43
D D D D
(14); E (1) (14); C (1) (12) (40); E (1); C (1)
3 3 2 6
0.257 0.257 0.167 0.374
(0.141) (0.141) (0.134) (0.093)
0.00302 0.00302 (0.00226) 0.00042 (0.00034) 0.00319 (0.00129)
1.2000 0.1667 1.265
Sudan Abu Naama (AN) Delleng Nuba (DN) Bhari Khartoum (BK) Shilluk (SH) All
27 21 17 70 135
D D D D D
(27) (21) (17) (68); E (2) (133); E (2)
5 (D = 5) 4 (D = 4) 5 (D = 5) 7 (D = 6; E = 1) 13 (D = 12; E = 1)
0.279 0.724 0.581 0.265 0.413
(0.112) (0.059) (0.131) (0.069) (0.054)
0.00056 0.00259 0.00170 0.00163 0.00177
(0.00028) (0.00037) (0.00048) (0.00068) (0.00040)
0.2222 1.0286 0.6765 0.6459 0.701
Uganda Teso (TS) Langi (LG) Nganda NG) Nkonjo (NK) All
33 26 31 33 123
D D D D D
(33) (26) (31) (33) (123)
4 3 4 4 9
0.278 0.341 0.295 0.375 0.322
(0.098) (0.110) (0.102) (0.102) (0.054)
0.00073 0.00091 0.00078 0.00132 0.00096
(0.00027) (0.00031) (0.00029) (0.00043) (0.00019)
0.2917 0.3600 0.3097 0.5227 0.380
Haplogroups Haplogroup D Haplogroup A Haplogroup B Haplogroup E Haplogroup C All East Africa
410 96 1 4 1 512
410 96 1 4 1 –
0.456 (0.031) 0.622 (0.031)
0.00169 (0.00015) 0.00223 (0.00025)
0.6687 0.8873
0.638 (0.024)
0.00745 (0.00042)
2.9508
(D = 2; E = 1)) (D = 2; C = 1) (D = 2) (D = 4; E = 1; C = 1)
(D = 4) (D = 3) (D = 4) (D = 4) (D = 9)
30 8 1 1 1 41
2.5. Population demographic structure Population dynamics were inferred on the basis of mismatch distribution patterns (Rogers and Harpending, 1992) for all East African chickens and for the different clusters observed from the MJ network and the NJ tree. Departures of the observed mismatch distributions from the simulated model of expansion were tested with v2 test of goodness of fit and Harpending’s raggedness index ‘‘r’’ (Harpending, 1994) following 1000 coalescent simulations. The two tests were augmented with the Fu’s FS statistic (Fu, 1997), a coalescent based estimator of neutrality, whose significance was also tested with 1000 coalescent simulations in Arlequin 3.1 (Excoffier et al., 2005). 3. Results 3.1. mtDNA D-loop sequence variability and haplotype distribution pattern Sequences spanning the first 397 bp of the mtDNA D-loop and which included the first hypervariable segment were used for analysis. From 512 sequences, 41 haplotypes defined by 37 polymorphic sites were generated. For this study, the individual haplotypes were abbreviated EA (East Africa) A/B/C/D/E followed by a number (see Fig. 1 and Table S3). We observed 35 transitions and four transversions and a mean number of nucleotide
differences (k) between haplotypes of 2.951 ± 1.548. Five haplotypes (EAD1, EAD3, EAD8, EAA1 and EAA8) were the most frequently observed with a total of 421 out of 512 individuals (82.22%). Haplotype EAD3 was the only haplotype observed in all the 23 study populations (Table S3 and Fig. 2). The next commonest haplotype was EAA1 observed in seven populations all from Kenya. The rest of the haplotypes occurred at low frequencies and some of them were specific to particular countries and populations while others were shared between populations within and between countries (Table S3). For instance, haplotype EAB1 was specific to Kenya, haplotype EAC1 to Ethiopia and haplotype EAE1 was observed in Ethiopia and Sudan. Excluding haplotype EAD3, Kenya shared one haplotype each with Ethiopia (EAD6) and Sudan (EAD4) and two with Uganda (EAD4 and EAD13). Sudan shared two haplotypes with Uganda (EAD4 and EAD25), while Ethiopia shared one haplotype (EAD17) with Uganda. The study populations showed a wide range of haplotype (0.167 ± 0.134–0.923 ± 0.05) and nucleotide (0.00042 ± 0.00034– 0.01469 ± 0.00558) diversities (Table 1). These values however are within the range of those observed in other chicken populations from Africa (see Muchadeyi et al., 2007) and Asia (see Liu et al., 2006; Oka et al., 2007; Silva et al., 2008). Generally, both haplotype (0.923 ± 0.050 for Taita) and nucleotide diversities (0.01469 ± 0.00558 for Nairobi) were higher for populations in Kenya (Table 1). The overall correlations between sample size
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
377
haplotype (EAB1, EAC1, EAE1). Haplogroup B was observed in one population from Kenya (Naivasha), E in one population in Ethiopia (Debre Berhan) and another from Sudan (Shilluk), and C was found in one population from Ethiopia (Nekemte) (Fig. 2). The links between haplogroups D, A and E were well resolved (Fig. 3). Haplogroup D was separated from A by three mutations and from E by seven mutations. The link between B and C and between these two and D was however not well resolved. Haplogroup D was connected to these two by 11 median vectors (mv) and B was connected to C via two median vectors. The median vectors may represent either un-sampled haplotypes, haplotypes never introduced into Eastern Africa or introduced into Eastern Africa but becoming extinct shortly upon arrival or later. A star-like pattern radiating from haplotypes EAD3 and EAA2, which are central to haplogroups D and A and thus could represent ancestral haplotypes, is clearly visible. 3.3. Population dynamics and maternal genetic structure
Fig. 1. Sequence variation of 41 haplotypes derived from 512 village chickens observed in the mtDNA D-loop region. The number of individuals sharing the same haplotypes is indicated in the right column by ‘‘N’’. Mutations are scored relative to the White Leghorn reference sequence (GenBank accession No. X52392; Desjardins and Morais, 1990). Dots () denote identity with the reference sequence.
and these two measures of genetic diversity were negative and significant (Spearmans rho: sample size versus haplotype diversity = 0.49803, P = 0.0172; sample size versus nucleotide diversity = 0.52929; P = 0.0098).
To understand the historical dynamics of the study populations, mismatch distribution patterns were evaluated at two levels (see Fig. 4): (i) across all the populations analysed and (ii) separately for haplogroups D and A. Analysis for haplogroups B, E and C was not performed because they were each represented by a single haplotype. The analysis was augmented with Fu’s FS (Fu, 1997) statistic. For all the tests, the simulated sum of squares differences did not differ significantly from the observed (P > 0.05), while Harpending’s raggedness index ‘‘r’’ was marginally significant for haplogroup A (P = 0.048). Fu’s FS statistic was significant for the overall dataset (P = 0.001) and for haplogroup D (P = 0.000) but not for haplogroup A (P = 0.100). These results support a model of demographic expansion over all East African chicken populations and for haplogroup D. However, the raggedness index and Fu’s FS statistics do not support an expansion for haplogroup A. To reveal the maternal genetic structure across the East African region, we performed an analysis of molecular variance (AMOVA) at five hierarchical levels (Table 2). Overall, 73.81% of the genetic variation is observed within populations. However this number increases to 87.95% if only the commonest haplogroup D is considered and declines to 42.82% considering the five haplogroups as hierarchical clusters. Interestingly, comparing the results for haplogroups D and A, we observe that the among population distribution of variation is more than six times higher for haplogroup A (39.18% versus 6.08%). To test whether genetic differentiation was directly proportional to geographic proximity, we performed a Mantel test involving pairwise FST values against geographic distance between populations. This analysis was limited to haplogroup D, the only one found across Eastern Africa. All the correlations were not significant (P < 0.05), except the ones that included populations from Kenya (Fig. S2a–h). This may be an ‘‘artefact’’ of the presence of haplogroup A limiting the number of observations for the D haplogroup in Kenya. 3.4. Phylogeographic relationships of East African mitochondrial DNA haplotypes with Asia
3.2. Network and phylogenetic relationships The Median-Joining network (MJ) and phylogenetic tree (NJ) of East African haplotypes is shown in Fig. 3 and Fig. S1 respectively. Both revealed the grouping of haplotypes in two major haplogroups and three isolated haplotypes. We refer to these five groups of sequences (haplogroups) or individual sequences as haplogroups A–E (see Table S3). Haplogroup D is the commonest followed by haplogroup A (Table S3 and Fig. 2). The remaining three haplogroups (B, C and E) are each represented by a single
Table S2 provides the nomenclature of the different haplogroups across studies while the MJ network (Fig. 5) reveal the affinity of East African haplotypes to those from Asia and other regions in Africa. All haplotypes of haplogroup D in East Africa belong to a group including haplotypes observed in Europe, Middle East, India (Liu et al., 2006), Sri Lanka (Silva et al., 2008) and Zimbabwe (Muchadeyi et al., 2008). Haplogroup A was also observed in Southeast and East Asia, including Japan (Liu et al., 2006) and in the South African region (Muchadeyi et al., 2008). Worldwide, this
378
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Fig. 2. Geographic distribution of the indigenous chickens sampled in East Africa and used in this study. The shaded area in each pie is proportional to the number of individuals in each population observed for each haplogroup. (Population abbreviations: KF = Kilifi; TT = Taita; MG = Muranga; MR = Meru; MT = Marsabit; KT = Kitui; NB = Nairobi; KS = Kisii; HB = Homa Bay; KK = Kakamega; ND = Nandi; DB = Debre Berhan; NE = Nekemte; JM = Jimma; AN = Abu Naama; DN = Delleng Nuba; BK = Bhari Khartoum; SH = Shilluk; TS = Teso; LG = Langi; NG = Nganda; NK = Nkonjo). Colour codes: Yellow = Haplogroup D; Blue = Haplogroup A; Light green = Haplogroup B; Red = Haplogroup E; Dark green = Haplogroup C. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
is the haplogroup with the largest geographic distribution being observed from Japan to Africa. Haplotype EAA2 for this haplogroup is identical to haplotype A1 from clade A in Muchadeyi et al. (2008). Haplogroup B, represented here with a single haplotype and individual found in Kenya was observed previously in South China and Japan (Liu et al., 2006), Sri Lanka (Silva et al., 2008) and in commercial egg layers (Muchadeyi et al., 2008). Haplogroup C, also represented by a single haplotype in our study, was previously observed in Western China (Liu et al., 2006), Sri Lanka (Silva et al., 2008) and in commercial broilers and populations from Northwest Europe (Muchadeyi et al., 2008). Finally, the last haplogroup in East Africa, haplogroup E, groups together with a set of haplotypes found within the Yunnan province of China (Liu et al., 2006).
4. Discussion One conspicuous characteristic of village chickens in Africa is their extensive phenotypic variation in colours, feather types; morphology and body size (Msoffe et al., 2001; Halima et al., 2007; Dana et al., 2010a). Microsatellite analysis has also revealed extensive genetic diversity in African village chickens (Wimmers et al., 2000; Muchadeyi et al., 2007). In this study, we analysed partial
mtDNA D-loop sequences from 512 village chickens sampled in East Africa to determine their ancestry, origin and dispersal patterns. We reveal the existence of at least five genetically distinct mtDNA D-loop haplogroups, two of which (haplogroups A and D) are shared with other African chicken populations. Specifically, haplogroups D and A have been observed previously in village chickens from the South African region (Muchadeyi et al., 2008; Razafindraibe et al., 2008), haplogroup D in West Africa (Adebambo et al., 2010) while haplogroups B, C and E are reported for the first time for the African continent. Our results indicate a different origin and history of the two major haplogroups (A and D) in East Africa (Fig. 5 and Table S2). Indeed, they are related to geographically distinct Asian mitochondrial DNA haplogroups, primarily South Asia for haplogroup D and East and South Asia for haplogroup A (Liu et al., 2006). We do observe a distinct geographic distribution of the two haplogroups in East Africa with haplogroup A exclusively found in Kenya and haplogroup D found across all four countries studied. Last but not least AMOVA analyses indicate that the diversity of haplogroup D is found mainly within populations, while 40% of the diversity of haplogroup A is found among populations. It supports an independent and possibly more recent arrival and history of haplogroup A compared to haplogroup D and more particularly a coastal arrival of haplogroup A in Kenya. Mantel test on the other hand reveals
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
379
Fig. 3. Median-Joining network of 41 haplotypes observed in 512 East African village chickens based on the polymorphic sites of the mtDNA D-loop region. Circled areas are proportional to the haplotype frequencies. Median vectors are represented by ‘‘mv’’.
the absence of geographic structure within the East African region (Uganda, Sudan, Ethiopia) for haplogroup D indicating that the populations found within and between these countries are panmictic with exchange of genetic material not being uncommon. The point of entry and subsequent dispersal of chickens on the African continent remain unsolved. It has been suggested that chicken arrived in Africa first through Egypt, where the most ancient undisputed presence of domestic fowls are found (Coltherd, 1966; Houlihan and Goodman, 1986) and later through the Indian Ocean coastline (MacDonald, 1992; MacDonald and Edwards, 1993). Haplogroup D is predominant in East (this study) and West Africa (Adebambo et al., 2010) and also occurs in the South African region (Muchadeyi et al., 2008; Razafindraibe et al., 2008). Three routes by which this haplogroup could have arrived in East Africa from its supposedly Indian subcontinent origin are possible: (i) following terrestrial routes through the Middle East (Iraq, Syria, Jordan) into Egypt from where it diffused southwards into Sudan, Ethiopia and Uganda; (ii) through the Horn of Africa via the Arabian Peninsula/Gulf of Aden with subsequent dispersion westwards to Sudan and southwards to Kenya and Uganda; or (iii) directly to Coastal East Africa via the Indian ocean trading and subsequently movement inland towards Uganda, Sudan and Ethiopia. Our genetic data cannot yet distinguish between these three possibilities in the absence of information on mitochondrial DNA diversity from Northern Africa (e.g. Egypt) and the Arabian Peninsula. However, archaeological evidences indicate that chicken arrived in Egypt and Sudan much earlier than in Kenya, Uganda or Ethiopia. For instance, a silver bowl with an image of a domestic
fowl was excavated at Tell Basta dating to the late XIX (BC 1307–1196) or early XX (BC 1196–1070) Egyptian dynasties (Houlihan and Goodman, 1986). Also, chicken were already known in Sudan by the mid 7th century BC being depicted on four ivory plaques from the tomb of Queen Yeturow at Nuri (Dunham, 1955). Earliest finds from East Africa date only to between 800 and 1400 AD (MacDonald and Edwards, 1993). So archaeological information favours a terrestrial arrival of domestic chicken in sub-Saharan Africa from the North East of the African continent rather than an arrival through the Horn of Africa. The second major haplogroup in East Africa, haplogroup A, is found only in Kenya. This haplogroup has also been observed in Zimbabwe (Muchadeyi et al., 2007) and Madagascar (Razafindraibe et al., 2008), but it is absent in West Africa (Adebambo et al., 2010). It occurs in the same group with representative haplotypes from clades C and D in Liu et al. (2006) (see Fig. 5) and the authors speculated that it could be derived from South and Southwest China and/or surrounding regions. Muchadeyi et al. (2008) speculates that it originates from Southeast Asia, while Razafindraibe et al. (2008) speculated without corroborating evidences that the haplogroup may originate from Indonesia following the Austronesian arrivals on the island. Whether haplogroup A originates from South/Southwest China and/or surrounding areas or Southeast Asia remains unclear. However, our data indicate that this haplogroup is present both in the South and East African regions suggesting either an early trading contact all along the coast of East Africa with East Asia or an arrival in a single African coastal region and subsequent dispersal along the
380
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Overall
Observed Simulated
Relative frequencies (%)
40
30
20
10
0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
Number of mismatches Observed Simulated
60
40
20
0
1
2
3
4
5
6
Observed
Lineage A
Simulated
50
Relative frequencies (%)
Relative frequencies (%)
Lineage D
40 30 20 10 0
7
1
2
Number of mismatches
3
4
5
6
7
8
Number of mismatches
Fig. 4. Mismatch distribution patterns for the overall, haplogroup D and haplogroup A for mtDNA D-loop region sequence haplotypes generated using 512 village chickens sampled in East Africa.
Table 2 Population genetic structure estimated from the hierarchical analysis of molecular variance (AMOVA) based on the frequencies of the 41 haplotypes observed from 512 mtDNA Dloop sequences of East African village chicken. Hierarchical clusters
Hierarchy
Variance components
Percent of variation
F
P-value
1. Overall
1
Within populations Among populations
73.81 26.19
– 0.26191
– 0.0000
2. Among countries
4
Within populations Among populations within countries Among countries
70.85 13.62 15.52
0.15520 0.16128 0.29145
0.01271 0.00000 0.00000
3. Among countries (haplogroup D)
4
Within populations Among populations within countries Among countries
87.95 6.08 5.97
0.05972 0.12054 0.06468
0.0000 0.0000 0.0000
4. Haplogroup A only (Kenya)
1
Within populations Among populations
60.82 39.18
0.39176
0.0000
5. Between the 5 haplogroups
5
Within populations Among populations within haplogroups Among haplogroups
42.82 9.07 48.11
0.48110 0.17478 0.57179
0.0000 0.0000 0.0000
coastal area of the continent. When such an event took place is unknown. Our data seem to suggest that it most likely happened after the arrival of the D haplogroup in East Africa given the more restricted geographic distribution of A compared to D in the region. The biological history of haplogroups B and C appear to be similar. This is the first time that they are reported in African village chickens. The origins of these two haplogroups are uncertain but Liu et al. (2006) proposed that they originate from Yunnan province, and/or surrounding areas (Myanmar, Thailand). Of interest is that haplogroup B is identical to haplotypes found in commercial brown and white egg layers and haplogroup C has close genetic affinity to haplotypes found in commercial broilers and chicken from Northwest Europe (Muchadeyi et al., 2008). In the past two or more decades, crossbreeding programmes involving village chickens and commercial layers and broilers to improve produc-
tivity have been undertaken by various development agencies and governments in East Africa (Pedersen et al., 2000). It is possible that these two haplogroups have been introduced to East Africa in recent times via commercial flocks through such crossbreeding activities. Similar introgression has also been shown to have occurred in Dutch fancy breeds of chicken based on mtDNA analysis (Dana et al., 2010b). These haplotypes therefore could represent signatures of recent introgression of commercial broiler and layer mtDNA haplotypes into village chickens. A recent introgression theory is supported by a lack of diversity of the haplogroups in East Africa and their presence in a single population. From a conservation and utilisation of indigenous chicken genetic diversity point of view, it indicates that exotic commercial genotypes might have successfully introgressed some indigenous village chicken populations in East Africa. The extent of such introgression will benefit to be assessed with different sets of
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
381
Fig. 5. Median-Joining network showing the relationship between 41 haplotypes from 512 East African village chickens and 30 haplotypes downloaded from the GenBank based on the polymorphic sites of the mtDNA D-loop region. Circled areas are proportional to the haplotype frequencies. Median vectors are represented by ‘‘mv’’. Colour codes are as follows: Black = Liu et al. (2006) haplotypes; Grey = Muchadeyi et al. (2008) haplotypes; Brown = Silva et al. (2008) haplotypes; Light Blue = Adebambo et al. (2010) haplotypes; Yellow = Haplogroup D; Blue = Haplogroup A; Light Green = Haplogroup B; Red = Haplogroup E; Dark Green = Haplogroup C (this study). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
genetic markers (e.g. autosomal single nucleotide polymorphisms, microsatellites). The fifth haplogroup, haplogroup E, was observed only in Ethiopia and Sudan with the same haplotype observed in both countries. Like haplogroup B and C, it is reported for the first time in African village chickens. Haplogroup E has not been reported in commercial chicken (Muchadeyi et al., 2008) and a presence through crossbreeding with commercial lines is therefore not supported. Its close relationship with chickens mitochondrial DNA sequences from Yunnan province in China (Liu et al., 2006) support an origin of the haplotype in Yunnan province and/or adjacent regions. How it reached the African continent remains unknown. It can be argued that this haplogroup reached East Africa either independently or accompanying one of the major haplogroups, A or D. However, from the geographic distribution of haplogroup A in East Africa it is unlikely that haplogroup E arrived in the region with haplogroup A. Indeed, haplogroup A occurs only in Kenya while haplogroup E is found in Ethiopia and Sudan. Haplogroup D is the commonest in Ethiopia and Sudan and therefore E could have arrived as a companion to D. However, the data of Liu et al. (2006), indicates that haplogroups E and D are present in separate regions within the geographic range of the wild ancestor; supporting a presence in East Africa following separate introductions. Haplogroup E’s presence within the Yunnan Province of
China raises the interesting hypothesis of the arrival of haplogroup E into Africa following the early AD 15th Chinese maritime expeditions to Southeast Asia, South Asia, the Middle East and East Africa (Duyvendak, 1939; Beaujard, 2005). For example, the haplogroup might have entered the continent through the Horn of Africa possibly Mogadishu in modern day Somali republic where Chinese emissaries are reported to have received a Giraffe and Rhino as presents for their rulers (Duyvendak, 1939; The Ming Dynasty, AD 1368–1644). Alternatively, this haplogroup might have reached the African continent from other geographic areas bordering the Indian Ocean for which the genetic diversity of native chickens has not been studied yet (e.g. Myanmar). In conclusion, this paper presents the likely origin, point of entry and subsequent dispersal of chickens in East Africa. We show for the first time the existence of five haplogroups in village chickens in East Africa. Two of these haplogroups are shared with chickens from West and Southern Africa while three are reported for the first time for Africa. Using chicken as the model organism and mtDNA as a marker, we demonstrate ancient trading contacts between Asia and East Africa. The study provides a basis for further understanding of the dynamics of human interactions and their interrelationships with livestock not only as a source of nourishment but also as a commodity of trade and a potential genetic marker to trace signatures of migration and trading across and within continents.
382
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Acknowledgements The authors wish to thank all farmers who provided the village chickens for sampling. The following members of the chicken diversity consortium also contributed to the study by assisting in sample collection: D. Tadelle, G. Abebe (Ethiopia); J. Hirbo, W.N. Mnene (Kenya); Veterinaire Sans Frontieres (Sudan); E. Ssewannyana and L. Serungoji (Uganda). ILRI research is principally funded by program grants from the United Kingdom, Japan, The European Union, Ireland, and France, as well as unrestricted funding from other donors to the Consultative Group on International Agricultural Research (CGIAR). This work was funded in large part through a Monbukagakusho scholarship to the first author from the Japanese Government. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ympev.2010.11.027. References Adebambo, A.O., Mobegi, V.A., Mwacharo, J.M., Oladejo, B.M., Adewale, R.A., Ilori, L.O., Makanjuola, B.O., Afolayan, O., Bjørnstad, G., Jianlin, H., Hanotte, O., 2010. Lack of phylogeographic structure in Nigerian village chickens revealed by mitochondrial DNA D-loop sequence analysis. Int. J. Poult. Sci. 9, 503–507. Bandelt, H.-J., Forster, P., Röhl, A., 1999. Median-joining networks for inferring intraspecific phylogenies. Mol. Biol. Evol. 16, 37–48. Beaujard, P., 2005. The Indian Ocean in Eurasia and African world systems before the sixteenth century. J. World Hist. 16, 411–465. Boivin, N., Fuller, D.Q., 2009. Shell middens, ships and seeds: exploring coastal subsistence, maritime trade and the dispersal of domesticates in and around the ancient Arabian Peninsula. J. World Prehist. 22, 113–180. Coltherd, J.B., 1966. The domestic fowl in ancient Egypt. Ibis 108, 217–223. Dana, N., Dessie, T., van der Waaij, L.H., van Arendonk, J.A.M., 2010a. Morphological features of indigenous chicken populations of Ethiopia. AGRI 46, 11–23. Dana, N., Megens, H.-J., Crooijmans, R.P.M.A., Hanotte, O., Mwacharo, J., Groenen, M.A.M., van Arendonk, J.A.M., 2010b. East Asian contributions to Dutch traditional and western commercial chickens inferred from mtDNA analysis. Anim. Genet. doi: 10.1111/j.1365-2052.2010.02134.x. Delacour, J., 1977. The Pheasants of the World, second ed. Spur Publications, Liss, Hants, England. Desjardins, P., Morais, R., 1990. Sequence and gene organisation of the chicken mitochondrial genome: a novel gene order in higher vertebrates. J. Mol. Biol. 212, 599–634. Dunham, D., 1955. The Royal Cemeteries of Kush II, Nuri. Boston, Massachusets museum of Fine Arts. Duyvendak, J.J.L., 1939. The True Dates of the Chinese Maritime Expeditions in the Early Fifteenth Century. T’oung Pao, Second Series, vol. 34, Livr. 5, pp. 341–413. Excoffier, L.G., Laval, G., Schneider, S., 2005. Arlequin ver 3.0: an integrated software package for population genetics data analysis. Evol. Bioinform 1, 47–50. FAOSTAT, 2007. Food and Agriculture Organization statistical databases. CDROM. Fu, Y.-X., 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147, 915–925. Fuller, D.Q., Boivin, N., 2009. Crops, Cattle and Commensals Across the Indian Ocean: Current and potential archaeobiological evidence: in ‘‘Etudes Océan Indien no 42/43: Plantes et Sociétés’’, Paris: Centre d’Études et de Recherche sur L’ Océan Indien Occidental (CEROI), pp 13–46. Fumihito, A., Miyake, T., Takada, M., Shingu, R., Endo, T., Gojobori, T., Kondo, N., Ohno, S., 1996. Monophyletic origin and unique dispersal patterns of domestic fowls. Proc. Natl. Acad. Sci. USA 93, 6792–6795.
Halima, H., Neser, F.W.C., van Marle-Koster, E., de Kock, A., 2007. Phenotypic variation of native chicken populations in northwest Ethiopia. Trop. Anim. Health Prod. 39, 507–513. Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Symp. Ser. 41, 95–98. Harpending, H., 1994. Signature of ancient population growth in a low-resolution mitochondrial DNA mismatch distribution. Hum. Biol. 66, 591–600. Houlihan, P.F., Goodman, S.M., 1986. The Birds of Ancient Egypt. Aris and Phillips Ltd., Teddington House, Warminster, England. Kitalyi, A.J., 1998. Village Chicken Production Systems in Rural Africa. Household Food Security and Gender Issues. FAO Animal Production and Health Paper No. 142. FAO, Rome Italy, 81pp. Librado, P., Rozas, J., 2009. DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25, 1451–1452. Liu, Y.P., Wu, G.S., Yao, Y.-G., Miao, Y.-W., Luikart, G., Baig, M., Beja-Pereira, A., Ding, Z.-L., Palanichamy, M.G., Zhang, Y.-P., 2006. Multiple maternal origins of chickens: out of the Asian jungles. Mol. Phylogenet. Evol. 38, 12–19. MacDonald, K.C., 1992. The domestic chicken (Gallus gallus) in sub-Saharan Africa: A background to its introduction and its osteological differentiation from indigenous fowls (Numidinae and Francolinus sp). J. Archaeol. Sci. 19, 303–318. MacDonald, K.C., Edwards, D.N., 1993. Chickens in Africa: the importance of Qasr Ibrim. Antiquity 67, 584–590. Msoffe, P.L., Minga, U.M., Olsen, J.E., Yongolo, M.G., Jull-Madsen, H.R., Gwakisa, P.S., Mtambo, M.M., 2001. Phenotypes including immunocompetence in scavenging local chicken ecotypes in Tanzania. Trop. Anim. Health Prod. 33, 341–354. Muchadeyi, F.C., Eding, H., Wollny, C.B.A., Groeneveld, E., Makuza, S.M., Shamseldin, R., Simianer, H., Weigend, S., 2007. Absence of population substructuring in Zimbabwe chicken ecotypes inferred using microsatellite analysis. Anim. Genet. 38, 332–339. Muchadeyi, F.C., Eding, H., Simianer, H., Wollny, C.B.A., Groeneveld, E., Weigend, S., 2008. Mitochondrial DNA D-loop sequences suggest a Southeast Asian and Indian origin of Zimbabwean village chickens. Anim. Genet. 39, 615–622. Mwacharo, J.M., Nomura, K., Hanada, H., Jianlin, H., Hanotte, O., Amano, T., 2007. Genetic relationships among Kenyan and other East African indigenous chickens. Anim. Genet. 38, 485–490. Oka, T., Ino, Y., Nomura, K., Kawashima, S., Kuwayama, T., Hanada, H., Amano, T., Takada, M., Takahata, N., Hayashi, Y., Akishinonomiya, F., 2007. Analysis of mtDNA sequences shows Japanese native chickens have multiple origins. Anim. Genet. 38, 287–293. Pedersen, G., Permin, A., Minga, U.M., 2000. Possibilities for smallholder poultry projects in Eastern and Southern Africa. In: Proceeding of a Workshop Morogoro, Tanzania 22–25 May 2000. Network for Smallholder Poultry Development. The Royal Veterinary and Agricultural University. Copenhagen, Denmark, 61pp. Razafindraibe, H., Mobegi, V.A., Ommeh, S.C., Rakotondravao, J., Bjørnstad, G., Hanotte, O., Jianlin, H., 2008. Mitochondrial DNA origin of indigenous Malagasy chicken: implications for a functional polymorphism at the Mx gene. Ann. NY Acad. Sci. 1149, 77–79. Rogers, A.R., Harpending, H., 1992. Population growth makes waves in the distribution of pairwise genetic differences. Mol. Biol. Evol. 9, 552–569. Silva, P., Guan, X., Ho-Shing, O., Jones, J., Xu, J., Hui, D., Notter, D., Smith, E., 2008. Mitochondrial DNA-based analysis of genetic variation and relatedness among Sri Lankan indigenous chickens and the Ceylon Junglefowl (Gallus lafayetti). Anim. Genet. 40, 1–9. Sonaiya, E.B., 1997. African Network on Rural Poultry Development. In: Proceedings ANRPD Workshop, Addis Ababa, Ethiopia, pp. 134–143. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596– 1599. The Ming Dynasty (AD 1368–1644).. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The ClustalX windows interface. flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl. Acids Res. 24, 4876–4882. Wimmers, K., Ponsuksili, S., Hardge, T., Valle-Zarate, A., Mathur, P.K., Horst, P., 2000. Genetic distinctness of African, Asian and South American local chickens. Anim. Genet. 31, 159–165.
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Mitochondrial DNA reveals multiple introductions of domestic chicken in East Africa J.M. Mwacharo a,⇑, G. Bjørnstad b,c, V. Mobegi c, K. Nomura d, H. Hanada d, T. Amano e, H. Jianlin c,f, O. Hanotte a,c a
School of Biology, Centre for Genetics and Genomics, The University of Nottingham, University Park, Nottingham NG7 2RD, UK Norwegian School of Veterinary Science, Department of Basic Sciences & Aquatic Medicine, P.O. Box 8146, N-0033 Oslo, Norway c International Livestock Research Institute, P.O. Box 30709, Nairobi 00100, Kenya d Laboratory of Animal Genetics and Breeding, Department of Animal Science, Tokyo University of Agriculture, 1737 Funako Atsugi-Shi, Kanagawa 243-0034, Japan e Nakagawa 5-21-22, Tsuzuki-Ku, Yokohama-Shi, Kanagawa-Ken 224-0001, Japan f CAAS-ILRI Joint Laboratory on Livestock and Forage Genetic Resources, Institute of Animal Science, Chinese Academy of Agricultural Sciences (CAAS), Beijing 100094, China b
a r t i c l e
i n f o
Article history: Received 1 November 2010 Accepted 29 November 2010 Available online 5 December 2010 Keywords: Gallus gallus Village chicken Control region Indian Ocean
a b s t r a c t Chicken were possibly domesticated in South and Southeast Asia. They occur ubiquitously in East Africa where they show extensive phenotypic diversity. They appeared in the region relatively late, with the first undisputed evidence of domestic chicken in Sudan, around 700 BC. We reveal through a detailed analysis of mitochondrial DNA D-loop sequence diversity of 512 domestic village chickens, from four East African countries (Kenya, Ethiopia, Sudan, Uganda), the presence of at least five distinct mitochondrial DNA haplogroups. Phylogeographic analyses and inclusion of reference sequences from Asia allow us to address the origin, ways of introduction and dispersion of each haplogroup. The results indicate a likely Indian subcontinent origin for the commonest haplogroup (D) and a maritime introduction for the next commonest one (A) from Southeast and/or East Asia. Recent introgression of commercial haplotypes into the gene pool of village chickens might explain the rare presence of two haplogroups (B and C) while the origin of the last haplogroup (E) remains unclear being currently observed only outside the African continent in the inland Yunnan Province of China. Our findings not only support ancient historical maritime and terrestrial contacts between Asia and East Africa, but also indicate the presence of large maternal genetic diversity in the region which could potentially support genetic improvement programmes. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Village chickens (Gallus gallus) are ubiquitous in East Africa with a total population of 100.8 million birds (FAOSTAT, 2007). These birds fulfil various roles ranging from socio-cultural to sustaining livelihoods, suggesting a long historical presence in the region. Chicken are not a native species to Africa where no wild Gallus sp. is found (Delacour, 1977). A previous phylogenetic study of chicken mitochondrial DNA (mtDNA) control region suggested that the Indochinese red junglefowl subspecies Gallus gallus gallus was the primary maternal ancestor of all domestic fowls G. g. domesticus and that Southeast Asia (Thailand) was the likely center of domestication (Fumihito et al., 1996). A recent study, which analysed a larger fragment of the mtDNA control region encompassing the first hypervariable segment, in a comparatively large and diverse gene pool of domestic chickens from a wide geographic area (Europe and Asia), was the first to suggest multiple maternal geographic centers of origin for the domestic species (Liu et al., 2006). They identified nine divergent clades, (called here ⇑ Corresponding author. Fax: +44 (0) 115 951 3251. E-mail address: [email protected] (J.M. Mwacharo). 1055-7903/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2010.11.027
haplogroups), seven of which included both domestic chickens and wild red junglefowl subspecies haplotypes. Three of the nine clades involved mtDNA control region haplotypes found across Europe and Asia, whereas the other six clades contained haplotypes found exclusively in South and Southeast Asia. Both Fumihito et al. (1996) and Liu et al. (2006) studies lacked samples from the African continent. More recently, three studies have made an attempt to address the origin of African village chickens through the analysis of partial mtDNA D-loop sequences. Muchadeyi et al. (2008) observed two distinct haplogroups in Zimbabwe village chickens which they postulated came from Southeast Asia and the Indian subcontinent. Similarly, Razafindraibe et al. (2008) observed two haplogroups in Madagascar village chicken and speculated that one was of Indonesian and the other of African continental origin or an introgression from commercial lines. At the opposite a single haplogroup thought to be of Indian origin was observed in Nigeria village chickens by Adebambo et al. (2010), while no information is yet available for the East African region. Since historical times, Africa and Asia have been linked via maritime and terrestrial corridors. The Indian Ocean was one of the world’s earliest arenas that provided a maritime corridor linking
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Africa and Asia. Following the monsoon wind patterns of the Indian Ocean, Asia and East Africa witnessed important seafaring and maritime exchanges across historical times. Asia and Africa were also connected via a terrestrial route that traversed the Arabian Peninsula, a region which occupies a key geographic junction, with the African landmass to the West and the Asian continent to the East (Boivin and Fuller, 2009). These terrestrial and maritime routes witnessed the long distance movement of people, livestock and crops within and between continents. Domesticates that dispersed from Asia to Africa included cattle, pigs, chicken, broomcorn millet, bananas, water yam and taro and those that moved from Africa to Asia are donkey’s, sorghum, pearl millet, finger millet, cowpeas and hyacinth beans (Fuller and Boivin, 2009). Village chickens in East Africa are found across the entire region and in all agro-ecological zones; they show large within population phenotypic diversity in plumage colour, feather morphology and pattern, skin colour, comb type etc. (Msoffe et al., 2001; Dana et al., 2010a). Raised under free-range scavenging system, village chickens contribute substantially to egg and meat production under almost a zero input system under small holder subsistence economies (Sonaiya, 1997; Kitalyi, 1998). The possible geographic origin of East African village chickens remains unknown, with the Indian subcontinent, Southeast Asia and North Africa as possible primary centers of origin and entry in agreement with our knowledge of the history of trading of the region. To investigate the origin of East African village chicken, we generated mtDNA control region sequences from 512 village chickens and analysed their phylogenetic relationships with representative Asian haplotypes from the nine haplogroups described by Liu et al. (2006). In addition, we evaluated the phylogeographic structure of the village chickens to assess their pattern of diffusion across East Africa. Our results reveal an unexpected complex pattern of introduction of domestic chicken in the region and provide new information concerning the history of trading and human contacts between East Africa and Asia.
375
Triton X-100), 2.5 mM of each dNTP, 10 pM of each primer and 1 unit of Taq DNA polymerase (Promega, Madison WI, USA). Thermo-cycling conditions were: 94 °C (3 min), 35 cycles of 94 °C (1 min), 58 °C (1 min) and 72 °C (2 min), and a final extension step at 72 °C (10 min). PCR products were purified using the WizardÒ SV Gel and PCR Clean-Up Kit (Promega, Madison WI, USA). Purified products were sequenced directly using the BigDyeÒ Terminator v3.1 (Applied Biosystems, USA) on an ABI prism 3100 Avant DNA analyzer. The two PCR primers and an internal primer f1-3m (50 -TGGTTCCTCGGTCAGGCACATCC-30 ) were used for the sequencing reactions.
2.3. Sequence and phylogenetic analysis For each sample, three fragments were generated. These were edited manually using BioEdit 7.0 (Hall, 1999) and subsequently joined to reconstruct a fragment of 700 bp. The fragments were aligned using Clustal X 1.83 (Thompson et al., 1997) against a reference sequence (GenBank accession number X52392; Desjardins and Morais, 1990). Subsequent analyses were restricted to the first 397 bp incorporating the first hypervariable segment (HVS1). We set to determine the number of haplogroups present in East Africa village chickens by constructing a MedianJoining (MJ) network (Bandelt et al., 1999) using NETWORK 4.5 (fluxus-engineering.com). This analysis was augmented by constructing a phylogenetic tree involving the haplotypes observed in East Africa using the Neighbour-Joining (NJ) algorithm as implemented in MEGA 4.0 (Tamura et al., 2007) following 1000 bootstrap replications. To portray the affinity of East Africa haplotypes to those observed in Asia and Africa, a MJ network incorporating the 30 haplotypes downloaded from the GenBank (see Table S1) was also constructed. The nomenclature of the haplogroups observed in this study compared to the study of Liu et al. (2006), Silva et al. (2008) and Muchadeyi et al. (2008) are shown in Table S2.
2. Materials and methods 2.1. Sample collection and DNA extraction Blood samples from 512 genetically unrelated village chickens were collected from 23 populations in four countries in East Africa (Table 1). All samples were from unimproved village chickens raised under free-range scavenging. Two mature birds were sampled per flock and the sampling strategy and characteristics of the sampling locations were described previously (Mwacharo et al., 2007). Genomic DNA was extracted from either whole blood using phenol–chloroform or from air dried blood preserved on FTA classic cards (Whatman Biosciences) using the manufacturers protocol. To ascertain the genetic affinities of the study populations to Asiatic and other African chickens, 30 haplotypes were downloaded from the GenBank and were included in the analysis (see Table S1). For this purpose, the central and most common haplotypes for the nine clades observed in Liu et al. (2006) were selected for the study to determine the possible origins of chicken found in East Africa within the geographic range of the wild ancestor, the red junglefowl. 2.2. PCR amplification and sequencing The first 700 bp of the mtDNA D-loop region were amplified via PCR using primers AV1F2 (50 -AGGACTACGGCTTGAAAAGC-30 ) and CR1b (50 -CCATACACGCAAACCGTCTC-30 ). PCR amplifications were carried out in 25 ll reaction volumes containing 20 ng genomic DNA, 1 X PCR buffer (10 mM Tris–HCl pH 8.3, 50 mM KCl, 0.1%
2.4. Population genetic variability and structure Genetic variation (nucleotide diversity, haplotype diversity and nucleotide differences) for each population were calculated using DnaSP 5.10 (Librado and Rozas, 2009). Population genetic structure was assessed by nested analysis of molecular variance (AMOVA). The groupings used for AMOVA were as follows: (i) the overall dataset assuming no groups; (ii) between the haplogroups observed on the MJ network; (iii) between chicken populations found in different countries (country specific groups); (iv) between populations for haplogroup D only and (v) between populations for haplogroup A only. Phi (U) statistics representing haplotype correlations at various levels of hierarchical clusters UCT, USC, and UST (Excoffier et al., 2005) were calculated. Significance testing was performed using 10,100 coalescent simulations in Arlequin 3.1 (Excoffier et al., 2005). A Mantel test was used to assess the non-random association between genetic differentiation (FST) and geographic distances (km) between populations using the IBDWS v3.05 software (http://ibdws.sdsu.edu). Population pairwise FST values were calculated using Arlequin 3.1 (Excoffier et al., 2005). Geographic distances between populations were calculated using the MapCrow Travel Distance Calculator between central towns within the sampling locations based on geographical coordinates (http://www. mapcrow.info/). For the analysis all negative FST values were set to zero. This analysis was limited to haplogroup D with the widest geographic range in the region. Several comparative analyses were performed across all countries and between all countries.
376
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Table 1 Sampling locations, population abbreviations, sample sizes and genetic diversity measures and their standard deviations (SD) for each population and the five lineages observed from the Median-Joining network analysis. Country/population
Sample size
Haplogroup (number of individuals observed)
Number of haplotypes
Haplotype diversity (SD)
Nucleotide diversity (SD)
Mean number of nucleotide differences
Kenya Kilifi (KF) Taita (TT) Muranga (MG) Meru (MR) Marsabit (MT) Kitui (KT) Nairobi (NB) Kisii (KS) Homa Bay (HB) Kakamega (KK) Nandi (ND) Naivasha (NV) All
14 13 15 14 60 14 4 15 14 14 14 20 211
D D D D D D D D D D D D D
(11); A (3) (5): A (8) (5); A (10) (5); A (9) (9); A (51) (5); A (9) (3); A (1) (15) (14) (13); A (1) (14) (15); A (4); B (1) (114); A (96); B (1)
6 (D = 4; A = 2) 8 (D = 3; A = 5) 4 (D = 3; A = 1) 4 (D = 2; A = 2) 7 (D = 43; A = 3) 6 (D = 3; A = 3) 3 (D = 2; A = 1) 4 (D = 4) 6 (D = 6) 7 (D = 6; A = 1) 5 (D = 5) 5 (D = 3; A = 1; B = 1) 25 (D = 16; A = 8; B = 1)
0.747 0.923 0.543 0.648 0.637 0.747 0.833 0.638 0.736 0.912 0.593 0.647 0.857
(0.111) (0.050) (0.133) (0.116) (0.053) (0.111) (0.222) (0.093) (0.109) (0.059) (0.144) (0.088) (0.012)
0.00883 0.01318 0.01199 0.01138 0.00623 0.01262 0.01469 0.00249 0.00302 0.00747 0.00246 0.00919 0.01192
(0.00223) (0.00147) (0.00252) (0.00208) (0.00126) (0.00208) (0.00558) (0.00064) (0.00078) (0.00169) (0.00084) (0.00233) (0.00023)
3.5055 5.2308 4.7619 4.5165 2.4718 5.0109 5.8333 0.9905 1.1978 2.9670 0.9780 3.6474 4.734
Ethiopia Debre Birhan (DB) Nekemte (NE) Jimma (JM) All
15 15 12 43
D D D D
(14); E (1) (14); C (1) (12) (40); E (1); C (1)
3 3 2 6
0.257 0.257 0.167 0.374
(0.141) (0.141) (0.134) (0.093)
0.00302 0.00302 (0.00226) 0.00042 (0.00034) 0.00319 (0.00129)
1.2000 0.1667 1.265
Sudan Abu Naama (AN) Delleng Nuba (DN) Bhari Khartoum (BK) Shilluk (SH) All
27 21 17 70 135
D D D D D
(27) (21) (17) (68); E (2) (133); E (2)
5 (D = 5) 4 (D = 4) 5 (D = 5) 7 (D = 6; E = 1) 13 (D = 12; E = 1)
0.279 0.724 0.581 0.265 0.413
(0.112) (0.059) (0.131) (0.069) (0.054)
0.00056 0.00259 0.00170 0.00163 0.00177
(0.00028) (0.00037) (0.00048) (0.00068) (0.00040)
0.2222 1.0286 0.6765 0.6459 0.701
Uganda Teso (TS) Langi (LG) Nganda NG) Nkonjo (NK) All
33 26 31 33 123
D D D D D
(33) (26) (31) (33) (123)
4 3 4 4 9
0.278 0.341 0.295 0.375 0.322
(0.098) (0.110) (0.102) (0.102) (0.054)
0.00073 0.00091 0.00078 0.00132 0.00096
(0.00027) (0.00031) (0.00029) (0.00043) (0.00019)
0.2917 0.3600 0.3097 0.5227 0.380
Haplogroups Haplogroup D Haplogroup A Haplogroup B Haplogroup E Haplogroup C All East Africa
410 96 1 4 1 512
410 96 1 4 1 –
0.456 (0.031) 0.622 (0.031)
0.00169 (0.00015) 0.00223 (0.00025)
0.6687 0.8873
0.638 (0.024)
0.00745 (0.00042)
2.9508
(D = 2; E = 1)) (D = 2; C = 1) (D = 2) (D = 4; E = 1; C = 1)
(D = 4) (D = 3) (D = 4) (D = 4) (D = 9)
30 8 1 1 1 41
2.5. Population demographic structure Population dynamics were inferred on the basis of mismatch distribution patterns (Rogers and Harpending, 1992) for all East African chickens and for the different clusters observed from the MJ network and the NJ tree. Departures of the observed mismatch distributions from the simulated model of expansion were tested with v2 test of goodness of fit and Harpending’s raggedness index ‘‘r’’ (Harpending, 1994) following 1000 coalescent simulations. The two tests were augmented with the Fu’s FS statistic (Fu, 1997), a coalescent based estimator of neutrality, whose significance was also tested with 1000 coalescent simulations in Arlequin 3.1 (Excoffier et al., 2005). 3. Results 3.1. mtDNA D-loop sequence variability and haplotype distribution pattern Sequences spanning the first 397 bp of the mtDNA D-loop and which included the first hypervariable segment were used for analysis. From 512 sequences, 41 haplotypes defined by 37 polymorphic sites were generated. For this study, the individual haplotypes were abbreviated EA (East Africa) A/B/C/D/E followed by a number (see Fig. 1 and Table S3). We observed 35 transitions and four transversions and a mean number of nucleotide
differences (k) between haplotypes of 2.951 ± 1.548. Five haplotypes (EAD1, EAD3, EAD8, EAA1 and EAA8) were the most frequently observed with a total of 421 out of 512 individuals (82.22%). Haplotype EAD3 was the only haplotype observed in all the 23 study populations (Table S3 and Fig. 2). The next commonest haplotype was EAA1 observed in seven populations all from Kenya. The rest of the haplotypes occurred at low frequencies and some of them were specific to particular countries and populations while others were shared between populations within and between countries (Table S3). For instance, haplotype EAB1 was specific to Kenya, haplotype EAC1 to Ethiopia and haplotype EAE1 was observed in Ethiopia and Sudan. Excluding haplotype EAD3, Kenya shared one haplotype each with Ethiopia (EAD6) and Sudan (EAD4) and two with Uganda (EAD4 and EAD13). Sudan shared two haplotypes with Uganda (EAD4 and EAD25), while Ethiopia shared one haplotype (EAD17) with Uganda. The study populations showed a wide range of haplotype (0.167 ± 0.134–0.923 ± 0.05) and nucleotide (0.00042 ± 0.00034– 0.01469 ± 0.00558) diversities (Table 1). These values however are within the range of those observed in other chicken populations from Africa (see Muchadeyi et al., 2007) and Asia (see Liu et al., 2006; Oka et al., 2007; Silva et al., 2008). Generally, both haplotype (0.923 ± 0.050 for Taita) and nucleotide diversities (0.01469 ± 0.00558 for Nairobi) were higher for populations in Kenya (Table 1). The overall correlations between sample size
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
377
haplotype (EAB1, EAC1, EAE1). Haplogroup B was observed in one population from Kenya (Naivasha), E in one population in Ethiopia (Debre Berhan) and another from Sudan (Shilluk), and C was found in one population from Ethiopia (Nekemte) (Fig. 2). The links between haplogroups D, A and E were well resolved (Fig. 3). Haplogroup D was separated from A by three mutations and from E by seven mutations. The link between B and C and between these two and D was however not well resolved. Haplogroup D was connected to these two by 11 median vectors (mv) and B was connected to C via two median vectors. The median vectors may represent either un-sampled haplotypes, haplotypes never introduced into Eastern Africa or introduced into Eastern Africa but becoming extinct shortly upon arrival or later. A star-like pattern radiating from haplotypes EAD3 and EAA2, which are central to haplogroups D and A and thus could represent ancestral haplotypes, is clearly visible. 3.3. Population dynamics and maternal genetic structure
Fig. 1. Sequence variation of 41 haplotypes derived from 512 village chickens observed in the mtDNA D-loop region. The number of individuals sharing the same haplotypes is indicated in the right column by ‘‘N’’. Mutations are scored relative to the White Leghorn reference sequence (GenBank accession No. X52392; Desjardins and Morais, 1990). Dots () denote identity with the reference sequence.
and these two measures of genetic diversity were negative and significant (Spearmans rho: sample size versus haplotype diversity = 0.49803, P = 0.0172; sample size versus nucleotide diversity = 0.52929; P = 0.0098).
To understand the historical dynamics of the study populations, mismatch distribution patterns were evaluated at two levels (see Fig. 4): (i) across all the populations analysed and (ii) separately for haplogroups D and A. Analysis for haplogroups B, E and C was not performed because they were each represented by a single haplotype. The analysis was augmented with Fu’s FS (Fu, 1997) statistic. For all the tests, the simulated sum of squares differences did not differ significantly from the observed (P > 0.05), while Harpending’s raggedness index ‘‘r’’ was marginally significant for haplogroup A (P = 0.048). Fu’s FS statistic was significant for the overall dataset (P = 0.001) and for haplogroup D (P = 0.000) but not for haplogroup A (P = 0.100). These results support a model of demographic expansion over all East African chicken populations and for haplogroup D. However, the raggedness index and Fu’s FS statistics do not support an expansion for haplogroup A. To reveal the maternal genetic structure across the East African region, we performed an analysis of molecular variance (AMOVA) at five hierarchical levels (Table 2). Overall, 73.81% of the genetic variation is observed within populations. However this number increases to 87.95% if only the commonest haplogroup D is considered and declines to 42.82% considering the five haplogroups as hierarchical clusters. Interestingly, comparing the results for haplogroups D and A, we observe that the among population distribution of variation is more than six times higher for haplogroup A (39.18% versus 6.08%). To test whether genetic differentiation was directly proportional to geographic proximity, we performed a Mantel test involving pairwise FST values against geographic distance between populations. This analysis was limited to haplogroup D, the only one found across Eastern Africa. All the correlations were not significant (P < 0.05), except the ones that included populations from Kenya (Fig. S2a–h). This may be an ‘‘artefact’’ of the presence of haplogroup A limiting the number of observations for the D haplogroup in Kenya. 3.4. Phylogeographic relationships of East African mitochondrial DNA haplotypes with Asia
3.2. Network and phylogenetic relationships The Median-Joining network (MJ) and phylogenetic tree (NJ) of East African haplotypes is shown in Fig. 3 and Fig. S1 respectively. Both revealed the grouping of haplotypes in two major haplogroups and three isolated haplotypes. We refer to these five groups of sequences (haplogroups) or individual sequences as haplogroups A–E (see Table S3). Haplogroup D is the commonest followed by haplogroup A (Table S3 and Fig. 2). The remaining three haplogroups (B, C and E) are each represented by a single
Table S2 provides the nomenclature of the different haplogroups across studies while the MJ network (Fig. 5) reveal the affinity of East African haplotypes to those from Asia and other regions in Africa. All haplotypes of haplogroup D in East Africa belong to a group including haplotypes observed in Europe, Middle East, India (Liu et al., 2006), Sri Lanka (Silva et al., 2008) and Zimbabwe (Muchadeyi et al., 2008). Haplogroup A was also observed in Southeast and East Asia, including Japan (Liu et al., 2006) and in the South African region (Muchadeyi et al., 2008). Worldwide, this
378
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Fig. 2. Geographic distribution of the indigenous chickens sampled in East Africa and used in this study. The shaded area in each pie is proportional to the number of individuals in each population observed for each haplogroup. (Population abbreviations: KF = Kilifi; TT = Taita; MG = Muranga; MR = Meru; MT = Marsabit; KT = Kitui; NB = Nairobi; KS = Kisii; HB = Homa Bay; KK = Kakamega; ND = Nandi; DB = Debre Berhan; NE = Nekemte; JM = Jimma; AN = Abu Naama; DN = Delleng Nuba; BK = Bhari Khartoum; SH = Shilluk; TS = Teso; LG = Langi; NG = Nganda; NK = Nkonjo). Colour codes: Yellow = Haplogroup D; Blue = Haplogroup A; Light green = Haplogroup B; Red = Haplogroup E; Dark green = Haplogroup C. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
is the haplogroup with the largest geographic distribution being observed from Japan to Africa. Haplotype EAA2 for this haplogroup is identical to haplotype A1 from clade A in Muchadeyi et al. (2008). Haplogroup B, represented here with a single haplotype and individual found in Kenya was observed previously in South China and Japan (Liu et al., 2006), Sri Lanka (Silva et al., 2008) and in commercial egg layers (Muchadeyi et al., 2008). Haplogroup C, also represented by a single haplotype in our study, was previously observed in Western China (Liu et al., 2006), Sri Lanka (Silva et al., 2008) and in commercial broilers and populations from Northwest Europe (Muchadeyi et al., 2008). Finally, the last haplogroup in East Africa, haplogroup E, groups together with a set of haplotypes found within the Yunnan province of China (Liu et al., 2006).
4. Discussion One conspicuous characteristic of village chickens in Africa is their extensive phenotypic variation in colours, feather types; morphology and body size (Msoffe et al., 2001; Halima et al., 2007; Dana et al., 2010a). Microsatellite analysis has also revealed extensive genetic diversity in African village chickens (Wimmers et al., 2000; Muchadeyi et al., 2007). In this study, we analysed partial
mtDNA D-loop sequences from 512 village chickens sampled in East Africa to determine their ancestry, origin and dispersal patterns. We reveal the existence of at least five genetically distinct mtDNA D-loop haplogroups, two of which (haplogroups A and D) are shared with other African chicken populations. Specifically, haplogroups D and A have been observed previously in village chickens from the South African region (Muchadeyi et al., 2008; Razafindraibe et al., 2008), haplogroup D in West Africa (Adebambo et al., 2010) while haplogroups B, C and E are reported for the first time for the African continent. Our results indicate a different origin and history of the two major haplogroups (A and D) in East Africa (Fig. 5 and Table S2). Indeed, they are related to geographically distinct Asian mitochondrial DNA haplogroups, primarily South Asia for haplogroup D and East and South Asia for haplogroup A (Liu et al., 2006). We do observe a distinct geographic distribution of the two haplogroups in East Africa with haplogroup A exclusively found in Kenya and haplogroup D found across all four countries studied. Last but not least AMOVA analyses indicate that the diversity of haplogroup D is found mainly within populations, while 40% of the diversity of haplogroup A is found among populations. It supports an independent and possibly more recent arrival and history of haplogroup A compared to haplogroup D and more particularly a coastal arrival of haplogroup A in Kenya. Mantel test on the other hand reveals
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
379
Fig. 3. Median-Joining network of 41 haplotypes observed in 512 East African village chickens based on the polymorphic sites of the mtDNA D-loop region. Circled areas are proportional to the haplotype frequencies. Median vectors are represented by ‘‘mv’’.
the absence of geographic structure within the East African region (Uganda, Sudan, Ethiopia) for haplogroup D indicating that the populations found within and between these countries are panmictic with exchange of genetic material not being uncommon. The point of entry and subsequent dispersal of chickens on the African continent remain unsolved. It has been suggested that chicken arrived in Africa first through Egypt, where the most ancient undisputed presence of domestic fowls are found (Coltherd, 1966; Houlihan and Goodman, 1986) and later through the Indian Ocean coastline (MacDonald, 1992; MacDonald and Edwards, 1993). Haplogroup D is predominant in East (this study) and West Africa (Adebambo et al., 2010) and also occurs in the South African region (Muchadeyi et al., 2008; Razafindraibe et al., 2008). Three routes by which this haplogroup could have arrived in East Africa from its supposedly Indian subcontinent origin are possible: (i) following terrestrial routes through the Middle East (Iraq, Syria, Jordan) into Egypt from where it diffused southwards into Sudan, Ethiopia and Uganda; (ii) through the Horn of Africa via the Arabian Peninsula/Gulf of Aden with subsequent dispersion westwards to Sudan and southwards to Kenya and Uganda; or (iii) directly to Coastal East Africa via the Indian ocean trading and subsequently movement inland towards Uganda, Sudan and Ethiopia. Our genetic data cannot yet distinguish between these three possibilities in the absence of information on mitochondrial DNA diversity from Northern Africa (e.g. Egypt) and the Arabian Peninsula. However, archaeological evidences indicate that chicken arrived in Egypt and Sudan much earlier than in Kenya, Uganda or Ethiopia. For instance, a silver bowl with an image of a domestic
fowl was excavated at Tell Basta dating to the late XIX (BC 1307–1196) or early XX (BC 1196–1070) Egyptian dynasties (Houlihan and Goodman, 1986). Also, chicken were already known in Sudan by the mid 7th century BC being depicted on four ivory plaques from the tomb of Queen Yeturow at Nuri (Dunham, 1955). Earliest finds from East Africa date only to between 800 and 1400 AD (MacDonald and Edwards, 1993). So archaeological information favours a terrestrial arrival of domestic chicken in sub-Saharan Africa from the North East of the African continent rather than an arrival through the Horn of Africa. The second major haplogroup in East Africa, haplogroup A, is found only in Kenya. This haplogroup has also been observed in Zimbabwe (Muchadeyi et al., 2007) and Madagascar (Razafindraibe et al., 2008), but it is absent in West Africa (Adebambo et al., 2010). It occurs in the same group with representative haplotypes from clades C and D in Liu et al. (2006) (see Fig. 5) and the authors speculated that it could be derived from South and Southwest China and/or surrounding regions. Muchadeyi et al. (2008) speculates that it originates from Southeast Asia, while Razafindraibe et al. (2008) speculated without corroborating evidences that the haplogroup may originate from Indonesia following the Austronesian arrivals on the island. Whether haplogroup A originates from South/Southwest China and/or surrounding areas or Southeast Asia remains unclear. However, our data indicate that this haplogroup is present both in the South and East African regions suggesting either an early trading contact all along the coast of East Africa with East Asia or an arrival in a single African coastal region and subsequent dispersal along the
380
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Overall
Observed Simulated
Relative frequencies (%)
40
30
20
10
0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
Number of mismatches Observed Simulated
60
40
20
0
1
2
3
4
5
6
Observed
Lineage A
Simulated
50
Relative frequencies (%)
Relative frequencies (%)
Lineage D
40 30 20 10 0
7
1
2
Number of mismatches
3
4
5
6
7
8
Number of mismatches
Fig. 4. Mismatch distribution patterns for the overall, haplogroup D and haplogroup A for mtDNA D-loop region sequence haplotypes generated using 512 village chickens sampled in East Africa.
Table 2 Population genetic structure estimated from the hierarchical analysis of molecular variance (AMOVA) based on the frequencies of the 41 haplotypes observed from 512 mtDNA Dloop sequences of East African village chicken. Hierarchical clusters
Hierarchy
Variance components
Percent of variation
F
P-value
1. Overall
1
Within populations Among populations
73.81 26.19
– 0.26191
– 0.0000
2. Among countries
4
Within populations Among populations within countries Among countries
70.85 13.62 15.52
0.15520 0.16128 0.29145
0.01271 0.00000 0.00000
3. Among countries (haplogroup D)
4
Within populations Among populations within countries Among countries
87.95 6.08 5.97
0.05972 0.12054 0.06468
0.0000 0.0000 0.0000
4. Haplogroup A only (Kenya)
1
Within populations Among populations
60.82 39.18
0.39176
0.0000
5. Between the 5 haplogroups
5
Within populations Among populations within haplogroups Among haplogroups
42.82 9.07 48.11
0.48110 0.17478 0.57179
0.0000 0.0000 0.0000
coastal area of the continent. When such an event took place is unknown. Our data seem to suggest that it most likely happened after the arrival of the D haplogroup in East Africa given the more restricted geographic distribution of A compared to D in the region. The biological history of haplogroups B and C appear to be similar. This is the first time that they are reported in African village chickens. The origins of these two haplogroups are uncertain but Liu et al. (2006) proposed that they originate from Yunnan province, and/or surrounding areas (Myanmar, Thailand). Of interest is that haplogroup B is identical to haplotypes found in commercial brown and white egg layers and haplogroup C has close genetic affinity to haplotypes found in commercial broilers and chicken from Northwest Europe (Muchadeyi et al., 2008). In the past two or more decades, crossbreeding programmes involving village chickens and commercial layers and broilers to improve produc-
tivity have been undertaken by various development agencies and governments in East Africa (Pedersen et al., 2000). It is possible that these two haplogroups have been introduced to East Africa in recent times via commercial flocks through such crossbreeding activities. Similar introgression has also been shown to have occurred in Dutch fancy breeds of chicken based on mtDNA analysis (Dana et al., 2010b). These haplotypes therefore could represent signatures of recent introgression of commercial broiler and layer mtDNA haplotypes into village chickens. A recent introgression theory is supported by a lack of diversity of the haplogroups in East Africa and their presence in a single population. From a conservation and utilisation of indigenous chicken genetic diversity point of view, it indicates that exotic commercial genotypes might have successfully introgressed some indigenous village chicken populations in East Africa. The extent of such introgression will benefit to be assessed with different sets of
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
381
Fig. 5. Median-Joining network showing the relationship between 41 haplotypes from 512 East African village chickens and 30 haplotypes downloaded from the GenBank based on the polymorphic sites of the mtDNA D-loop region. Circled areas are proportional to the haplotype frequencies. Median vectors are represented by ‘‘mv’’. Colour codes are as follows: Black = Liu et al. (2006) haplotypes; Grey = Muchadeyi et al. (2008) haplotypes; Brown = Silva et al. (2008) haplotypes; Light Blue = Adebambo et al. (2010) haplotypes; Yellow = Haplogroup D; Blue = Haplogroup A; Light Green = Haplogroup B; Red = Haplogroup E; Dark Green = Haplogroup C (this study). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
genetic markers (e.g. autosomal single nucleotide polymorphisms, microsatellites). The fifth haplogroup, haplogroup E, was observed only in Ethiopia and Sudan with the same haplotype observed in both countries. Like haplogroup B and C, it is reported for the first time in African village chickens. Haplogroup E has not been reported in commercial chicken (Muchadeyi et al., 2008) and a presence through crossbreeding with commercial lines is therefore not supported. Its close relationship with chickens mitochondrial DNA sequences from Yunnan province in China (Liu et al., 2006) support an origin of the haplotype in Yunnan province and/or adjacent regions. How it reached the African continent remains unknown. It can be argued that this haplogroup reached East Africa either independently or accompanying one of the major haplogroups, A or D. However, from the geographic distribution of haplogroup A in East Africa it is unlikely that haplogroup E arrived in the region with haplogroup A. Indeed, haplogroup A occurs only in Kenya while haplogroup E is found in Ethiopia and Sudan. Haplogroup D is the commonest in Ethiopia and Sudan and therefore E could have arrived as a companion to D. However, the data of Liu et al. (2006), indicates that haplogroups E and D are present in separate regions within the geographic range of the wild ancestor; supporting a presence in East Africa following separate introductions. Haplogroup E’s presence within the Yunnan Province of
China raises the interesting hypothesis of the arrival of haplogroup E into Africa following the early AD 15th Chinese maritime expeditions to Southeast Asia, South Asia, the Middle East and East Africa (Duyvendak, 1939; Beaujard, 2005). For example, the haplogroup might have entered the continent through the Horn of Africa possibly Mogadishu in modern day Somali republic where Chinese emissaries are reported to have received a Giraffe and Rhino as presents for their rulers (Duyvendak, 1939; The Ming Dynasty, AD 1368–1644). Alternatively, this haplogroup might have reached the African continent from other geographic areas bordering the Indian Ocean for which the genetic diversity of native chickens has not been studied yet (e.g. Myanmar). In conclusion, this paper presents the likely origin, point of entry and subsequent dispersal of chickens in East Africa. We show for the first time the existence of five haplogroups in village chickens in East Africa. Two of these haplogroups are shared with chickens from West and Southern Africa while three are reported for the first time for Africa. Using chicken as the model organism and mtDNA as a marker, we demonstrate ancient trading contacts between Asia and East Africa. The study provides a basis for further understanding of the dynamics of human interactions and their interrelationships with livestock not only as a source of nourishment but also as a commodity of trade and a potential genetic marker to trace signatures of migration and trading across and within continents.
382
J.M. Mwacharo et al. / Molecular Phylogenetics and Evolution 58 (2011) 374–382
Acknowledgements The authors wish to thank all farmers who provided the village chickens for sampling. The following members of the chicken diversity consortium also contributed to the study by assisting in sample collection: D. Tadelle, G. Abebe (Ethiopia); J. Hirbo, W.N. Mnene (Kenya); Veterinaire Sans Frontieres (Sudan); E. Ssewannyana and L. Serungoji (Uganda). ILRI research is principally funded by program grants from the United Kingdom, Japan, The European Union, Ireland, and France, as well as unrestricted funding from other donors to the Consultative Group on International Agricultural Research (CGIAR). This work was funded in large part through a Monbukagakusho scholarship to the first author from the Japanese Government. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ympev.2010.11.027. References Adebambo, A.O., Mobegi, V.A., Mwacharo, J.M., Oladejo, B.M., Adewale, R.A., Ilori, L.O., Makanjuola, B.O., Afolayan, O., Bjørnstad, G., Jianlin, H., Hanotte, O., 2010. Lack of phylogeographic structure in Nigerian village chickens revealed by mitochondrial DNA D-loop sequence analysis. Int. J. Poult. Sci. 9, 503–507. Bandelt, H.-J., Forster, P., Röhl, A., 1999. Median-joining networks for inferring intraspecific phylogenies. Mol. Biol. Evol. 16, 37–48. Beaujard, P., 2005. The Indian Ocean in Eurasia and African world systems before the sixteenth century. J. World Hist. 16, 411–465. Boivin, N., Fuller, D.Q., 2009. Shell middens, ships and seeds: exploring coastal subsistence, maritime trade and the dispersal of domesticates in and around the ancient Arabian Peninsula. J. World Prehist. 22, 113–180. Coltherd, J.B., 1966. The domestic fowl in ancient Egypt. Ibis 108, 217–223. Dana, N., Dessie, T., van der Waaij, L.H., van Arendonk, J.A.M., 2010a. Morphological features of indigenous chicken populations of Ethiopia. AGRI 46, 11–23. Dana, N., Megens, H.-J., Crooijmans, R.P.M.A., Hanotte, O., Mwacharo, J., Groenen, M.A.M., van Arendonk, J.A.M., 2010b. East Asian contributions to Dutch traditional and western commercial chickens inferred from mtDNA analysis. Anim. Genet. doi: 10.1111/j.1365-2052.2010.02134.x. Delacour, J., 1977. The Pheasants of the World, second ed. Spur Publications, Liss, Hants, England. Desjardins, P., Morais, R., 1990. Sequence and gene organisation of the chicken mitochondrial genome: a novel gene order in higher vertebrates. J. Mol. Biol. 212, 599–634. Dunham, D., 1955. The Royal Cemeteries of Kush II, Nuri. Boston, Massachusets museum of Fine Arts. Duyvendak, J.J.L., 1939. The True Dates of the Chinese Maritime Expeditions in the Early Fifteenth Century. T’oung Pao, Second Series, vol. 34, Livr. 5, pp. 341–413. Excoffier, L.G., Laval, G., Schneider, S., 2005. Arlequin ver 3.0: an integrated software package for population genetics data analysis. Evol. Bioinform 1, 47–50. FAOSTAT, 2007. Food and Agriculture Organization statistical databases. CDROM. Fu, Y.-X., 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147, 915–925. Fuller, D.Q., Boivin, N., 2009. Crops, Cattle and Commensals Across the Indian Ocean: Current and potential archaeobiological evidence: in ‘‘Etudes Océan Indien no 42/43: Plantes et Sociétés’’, Paris: Centre d’Études et de Recherche sur L’ Océan Indien Occidental (CEROI), pp 13–46. Fumihito, A., Miyake, T., Takada, M., Shingu, R., Endo, T., Gojobori, T., Kondo, N., Ohno, S., 1996. Monophyletic origin and unique dispersal patterns of domestic fowls. Proc. Natl. Acad. Sci. USA 93, 6792–6795.
Halima, H., Neser, F.W.C., van Marle-Koster, E., de Kock, A., 2007. Phenotypic variation of native chicken populations in northwest Ethiopia. Trop. Anim. Health Prod. 39, 507–513. Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Symp. Ser. 41, 95–98. Harpending, H., 1994. Signature of ancient population growth in a low-resolution mitochondrial DNA mismatch distribution. Hum. Biol. 66, 591–600. Houlihan, P.F., Goodman, S.M., 1986. The Birds of Ancient Egypt. Aris and Phillips Ltd., Teddington House, Warminster, England. Kitalyi, A.J., 1998. Village Chicken Production Systems in Rural Africa. Household Food Security and Gender Issues. FAO Animal Production and Health Paper No. 142. FAO, Rome Italy, 81pp. Librado, P., Rozas, J., 2009. DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25, 1451–1452. Liu, Y.P., Wu, G.S., Yao, Y.-G., Miao, Y.-W., Luikart, G., Baig, M., Beja-Pereira, A., Ding, Z.-L., Palanichamy, M.G., Zhang, Y.-P., 2006. Multiple maternal origins of chickens: out of the Asian jungles. Mol. Phylogenet. Evol. 38, 12–19. MacDonald, K.C., 1992. The domestic chicken (Gallus gallus) in sub-Saharan Africa: A background to its introduction and its osteological differentiation from indigenous fowls (Numidinae and Francolinus sp). J. Archaeol. Sci. 19, 303–318. MacDonald, K.C., Edwards, D.N., 1993. Chickens in Africa: the importance of Qasr Ibrim. Antiquity 67, 584–590. Msoffe, P.L., Minga, U.M., Olsen, J.E., Yongolo, M.G., Jull-Madsen, H.R., Gwakisa, P.S., Mtambo, M.M., 2001. Phenotypes including immunocompetence in scavenging local chicken ecotypes in Tanzania. Trop. Anim. Health Prod. 33, 341–354. Muchadeyi, F.C., Eding, H., Wollny, C.B.A., Groeneveld, E., Makuza, S.M., Shamseldin, R., Simianer, H., Weigend, S., 2007. Absence of population substructuring in Zimbabwe chicken ecotypes inferred using microsatellite analysis. Anim. Genet. 38, 332–339. Muchadeyi, F.C., Eding, H., Simianer, H., Wollny, C.B.A., Groeneveld, E., Weigend, S., 2008. Mitochondrial DNA D-loop sequences suggest a Southeast Asian and Indian origin of Zimbabwean village chickens. Anim. Genet. 39, 615–622. Mwacharo, J.M., Nomura, K., Hanada, H., Jianlin, H., Hanotte, O., Amano, T., 2007. Genetic relationships among Kenyan and other East African indigenous chickens. Anim. Genet. 38, 485–490. Oka, T., Ino, Y., Nomura, K., Kawashima, S., Kuwayama, T., Hanada, H., Amano, T., Takada, M., Takahata, N., Hayashi, Y., Akishinonomiya, F., 2007. Analysis of mtDNA sequences shows Japanese native chickens have multiple origins. Anim. Genet. 38, 287–293. Pedersen, G., Permin, A., Minga, U.M., 2000. Possibilities for smallholder poultry projects in Eastern and Southern Africa. In: Proceeding of a Workshop Morogoro, Tanzania 22–25 May 2000. Network for Smallholder Poultry Development. The Royal Veterinary and Agricultural University. Copenhagen, Denmark, 61pp. Razafindraibe, H., Mobegi, V.A., Ommeh, S.C., Rakotondravao, J., Bjørnstad, G., Hanotte, O., Jianlin, H., 2008. Mitochondrial DNA origin of indigenous Malagasy chicken: implications for a functional polymorphism at the Mx gene. Ann. NY Acad. Sci. 1149, 77–79. Rogers, A.R., Harpending, H., 1992. Population growth makes waves in the distribution of pairwise genetic differences. Mol. Biol. Evol. 9, 552–569. Silva, P., Guan, X., Ho-Shing, O., Jones, J., Xu, J., Hui, D., Notter, D., Smith, E., 2008. Mitochondrial DNA-based analysis of genetic variation and relatedness among Sri Lankan indigenous chickens and the Ceylon Junglefowl (Gallus lafayetti). Anim. Genet. 40, 1–9. Sonaiya, E.B., 1997. African Network on Rural Poultry Development. In: Proceedings ANRPD Workshop, Addis Ababa, Ethiopia, pp. 134–143. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596– 1599. The Ming Dynasty (AD 1368–1644).