Variations in mitochondrial cytochrome b region among Ethiopian indigenous cattle populations assert Bos taurus maternal origin and historical dynamics

Objective This study was carried out to assess the haplotype diversity and population dynamics in cattle populations of Ethiopia. Methods We sequenced the complete mitochondrial cytochrome b gene of 76 animals from five indigenous and one Holstein Friesian×Barka cross bred cattle populations. Results In the sequence analysis, 18 haplotypes were generated from 18 segregating sites and the average haplotype and nucleotide diversities were 0.7540±0.043 and 0.0010±0.000, respectively. The population differentiation analysis shows a weak population structure (4.55%) among the populations studied. Majority of the variation (95.45%) is observed by within populations. The overall average pair-wise distance (F ST) was 0.049539 with the highest (F ST = 0.1245) and the lowest (F ST = 0.011) F ST distances observed between Boran and Abigar, and Sheko and Abigar from the indigenous cattle, respectively. The phylogenetic network analysis revealed that all the haplotypes detected clustered together with the Bos taurus cattle and converged to a haplogroup. No haplotype in Ethiopian cattle was observed clustered with the reference Bos indicus group. The mismatch distribution analysis indicates a single population expansion event among the cattle populations. Conclusion Overall, high haplotype variability was observed among Ethiopian cattle populations and they share a common ancestor with Bos taurus.


INTRODUCTION
Cattle are believed to have originated from Bos primigenius in the southwest Asia (Bos taurus) between 8,00010,000 YA and south Asia (Bos indicus) [1,2], and spread throughout the old world following the human trade and migration [3].The Horn and North of Africa are con sidered to be the ancient gateways for the dispersal of domesticates into the African continent [2,4].The Bos taurus group arrived in these regions, in Ethiopia in particular, around 7,000 years BC [5].Those regions are the cradle of both NearEast Bos taurus, and Arabian and Indian Bos indicus cattle migration corridor and sometime considered the secondary hybridi zation zone [6,7].The Bos indicus cattle descended from the putative cattle domestication center in the northern part of the Indian subcontinent, the Indus Valley [8] and arrived to East Africa between 2,0003,000 BC [9].Conversely, the worldwide genomewide analysis of ancestry, divergence and admixture revealed that African taurine cattle were first domes ticated in the Middle East and later hybridized with African aurochs [10].
Being the major entry point of cattle to Africa, the genetic landscape of the current cattle populations in Ethiopia could have been shaped by several introductions of zebu cattle and introgression of the taurine from the NearEast [7].As a result, Ethiopia harbors diversified cattle populations [11].There are 33 morphologically recognized cattle populations in the coun try (http://dad.fao.org/:accessed January 04, 2017) which are classified into humpless Shorthorn, zebu (large EastAfrican zebu, small EastAfrican zebu), zenga and sanga types [12].Both the mitochondrial dloop region and nuclear genomic information concurred that there has been extensive hybrid ization among the indigenous cattle populations in Ethiopia and led to have high level of admixture [6,11,13,14].
Cytochrome b (cyt b) and dloop regions of mtDNA play significant role in unraveling the population history of live stock species.Using the dloop region of mtDNA, history of genetic diversity and maternal origin of 10 Ethiopian indige nous cattle populations have been reported [11].However, there is no any report conducted on Ethiopian indigenous cattle populations yet using cyt b region to further uncover the maternal origin and population dynamics despite the fact that Ethiopia in particular and the East African region as whole are considered the hybridization zone for both Bos-taurus and Bos indicus cattle groups [6].Therefore, in the current study, we sequenced the mitochondrial cyt b region aiming to further unveil the genetic diversity, phyogenetic relationship, mater nal origin and population expansion of the indigenous cattle in Ethiopia.It also complements information to the findings reported on the diversity of the cattle populations using the dloop and autosomal DNA information [6,11,13,14].More over, there is limited information on the cytochrome region of cattle mtDNA at global level hence this work could provide additional insight about the historical dynamics of the cattle in Ethiopia and in the Horn of Africa at large.
Genomic DNA was extracted using Promega Genomic DNA extraction kit [15].A complete cyt b gene (1,140 bp) was amplified using forward primer: 5'CCATAAATAGGTGAA GGTTTGG3' and reverse primer: 5'TTGATGGTGAGAC TGCAGTT3' [16].The amplification was performed with the polymerase chain reaction (PCR) premix reaction volume of 20 μL that includes 10 μL ExTaq, 1 μL forward primer, 1 μL reverse primer, 1 μL template and 7 μL of RNA free H 2 O.The PCR cycling profile involved an initial denaturation step at 95°C for 2 min, followed by the first stage of amplification of 30 cycles involving a denaturation step at 94°C for 40 s, an nealing at 56°C for 40 s, and extension at 72°C for 90 s.The PCR reaction was completed with a final extension step at 72°C for eight min.The PCR products were analyzed by 1% agarose gel electrophoresis and visualized by ultraviolet illumi nation after staining with gel rednucleic acid gel stain (Biotium, Hayward, CA, USA).The PCR products with good quality were sent to Sangon Biotech in China (http://www.sangon.com) for sequencing.
In addition, we examined the starlike clustering patterns of the population expansion by mismatch distribution anal ysis in ARLEQUIN.With the same package, the history of the indigenous cattle was inferred using Fu's F S and Tajima's D neutrality tests which were implemented in 1,000 simula tions.

RESULTS mtDNA sequence variation and genetic diversity
The full length of the mitochondrial cyt b gene, 1,140 bp, was sequenced for 76 animals representing six cattle populations in Ethiopia.Eighteen (18) segregating/polymorphic sites were detected in the entire cyt b region.The transition to transver sion ratio of polymorphic sites is 5:1 and the average codon bias index was 0.566.From the polymorphic sites, seven of them are parsimony informative sites and the remaining 11 are two variant singleton variable sites.
From the 18 segregating sites, we observed 18 haplotypes (Table 1), and only five of the haplotypes (H1, H3, H5, H8, and H10) are shared by more than one population.The first hap lotype (H1) is the most frequent haplotype observed in 37 sequences and the third haplotype (H3) is the second fre quently detected in 16 sequences.Both H1 and H3 are shared by animals from the five indigenous cattle populations.The number of haplotypes in each population ranged from four (Boran and HF×Barka cross) to seven (Horro).Among the thirteen unique haplotypes, H15 and H16 belong to the cross bred population (Figure 1).No median vector is observed among the cattle populations studied indicating absence of unsampled or extinct haplotypes.The average haplotype and nucleotide diversities were 0.7540±0.043and 0.0010±0.000,respectively (Table 1).The highest and lowest haplotype di versities were observed in Abigar (H d = 0.8333±0.127)and Sheko (H d = 0.6282±0.143)cattle, respectively.

Population differentiation and phylogenetic analysis
The AMOVA revealed that there is only 4.55% of variation among the cattle populations studied, and the remaining 94.45% explains variation within the population (Table 2).The overall average pairwise and Reynold' s distance estimates were 0.0496 and 0.0540, respectively (Table 3).The pairwise distance es timation indicated that Boran showed highest deviation from Abigar (F ST = 0.12446), Sheko (F ST = 0.11360), and F 1 (F ST = 0.19437).Similar trends were observed using the Reynold's distance.However, we observed little or no differentiation between Sheko and Abigar, and Boran and Horro cattle pop ulations with both distance measures.On the other hand, both the phylogenetic network and NJ tree analysis show absence of clear phylogeographic structure among the cattle popula tions (Figures 1, 2).Interestingly, five individuals of the reference Bos taurus population clustered together with Ethiopian cattle populations, however, no Ethiopian cattle haplotype matched to Bos indicus reference population (Figure 1).

Population and historical demographic dynamics
In this study, the unimodal peak (Figure 3) indicates a single population expansion event held among the cattle populations.
There was no significant variation on the overall neutrality and mismatch distribution tests (Tajima's D = -1.0312±0.484[p>0.05];F S = -1.9374±1.043[p>0.05])(Table 4).Similar ob servation was reported for Chinese cattle [21].However, there were negative and significant values of Tajima's D test for Abigar and Fu's F S test for Abigar, Horro, and Sheko cattle populations.This indicates natural selection pressure in the populations resulted from excess of rare alleles [22].obtained in this study is much lower than those reported for Chinese cattle in the same mtDNA region (S = 105) [21], but higher than the segregating sites (S = 3 from 18 animals) re ported for Leiqiong cattle (Bos indicus type) of China [16].The transition to transversion ratio (5:1) was slightly lower than the value obtained for Chinese cattle populations (5.8:1) [21].

DISCUSSION
No transversion mutation was observed on the dloop region of mtDNA in Ethiopian cattle populations [11].Whereas, 62  polymorphic sites (52 transitions, five transversions, five indels) were detected in the dloop analysis of Kenana and Butan cattle populations in the neighboring Sudan [23].In the cur rent study, we observed 18 haplotypes (H = 18) which is lower than the haplotypes reported for Chinese cattle (H = 47) in same target region [21], however, higher than the haplotypes (H = 3) reported for Leiqiong cattle [16].In 117 dloop sequen ces, 81 haplotypes were identified in 10 Ethiopian indigenous cattle populations [11].The relative higher number of haplo types observed in the dloop could be because of high mutation rate of the dloop region compared to cyt b [24].For north Ethiopian cattle populations, 11 indicine and one taurine haplotypes were detected from five Ychromosome micro satellites analyzed.The current study revealed that the average haplotype diver sity (H d = 0.7540) as well as nucleotide diversity (π = 0.00100) (Table 1) are slightly lower than the diversities observed for Chinese cattle (H d = 0.848; π = 0.00923) [22].This could be because of presence of wider gene pool of maternal origins in Chinese cattle compared to Ethiopian cattle.However, lowest average values of haplotype (H d = 0.0741) and nucleotide (π = 0.0012) diversities were reported for Leiqiong cattle [16].This could be because of sampling bias, in which most of the individuals included in the study (n = 18) in the latter cattle population could be collected from similar haplotype origin and as a consequence very few number of haplotypes (H = 3) detected in the cyt b region.Based on the analysis of Ychro mosome microsatellites of north Ethiopian cattle populations, the average haplotype diversity was 0.617167±0.02617[7].

Population differentiation
The population differentiation analysis revealed very weak  populations structure (4.55%) among the cattle populations studied and the highest variation (95.45%) is explained by within population variation (Table 2).However, no (zero) percentage of variation was reported among Ethiopia cattle populations using the dloop [11].With same target region (dloop), only 2.4% of variation was observed between Kenana and Butan cattle in Sudan [23].Very low percentage of varia tion was reported for North Ethiopian cattle based on the analysis of Ychromosome simple sequence repeats (SSR) mar kers [7].Very low population differentiation estimates were also reported based on the nuclear DNA analyses in Ethio pian indigenous cattle populations (SSR markers: 1.3% [13] and 1.1% [25]; single nucleotide polymorphism CHIP panel: 2% [6]).The very weak phylogeographic structure and low genetic differentiation could be a result of a recent common ancestral origin, multiple introgressions and strong genetic exchange among the indigenous cattle populations [6,13,25].
On the other hand, the highest withinindividual genetic vari ability observed in Ethiopian cattle provides an untapped opportunity for adaptation to changing environments and for implementation of withinbreed genetic improvement schemes [6] in which genetic variability enables adaptation of natural populations to changing environments.Overall, the relation ships among Ethiopian cattle populations, which represent a mosaic of the humped zebu and taurine, reflect their evolu tionary history of origin and admixture rather than their phenotype based categorization as zebu, sanga and zenga [6], as we concluded in this study.Among Ethiopian indigenous cattle populations, the pair wise (F ST ) and Reynold' s distances vary from -0.01113 (between Sheko and Abigar) to 0.12446 (between Abigar and Boran) and zero (between Boran and Horro, Abigar, and Sheko) to 0.13292 (between Boran and Abigar), respectively (Table 3).However, we observed no differentiation between Sheko and Abigar despite the fact that Abigar cattle have the highest in dicine allele frequencies compared to other sanga cattle in Ethiopia [26].This could be due to two reasons: i) Sheko is an African short horn taurine and Abigar, which is a sanga type, is produced from introgression of Bos taurus and zebu (Bos indicus) groups [26]; and hence, Abigar partly shares same ancestral origin of Sheko; ii) it could be because of high gene flow of Abigar cattle population towards Sheko breeding tract.A recent extensive genomewide study revealed that there is very high allelic gene flow between Sheko and other local cattle populations as a consequence Sheko cattle population is get ting assimilated [6].The close geographical proximity between the two cattle populations (Sheko and Abigar) could facilitate the gene flow easily since isolationbydistance plays funda mental role for differentiation of populations.
On the other side, we observed highest differentiation be tween Boran and Abigar (F ST = 0.12446), and Boran and Sheko (F ST = 0.1136).This could be because of low gene flow, geo graphical isolation, ecological factors and morphological adaptation to local conditions.The Boran cattle, a humped breed reflecting indicine introgression, carry genes adaptive to harsh environment and heat tolerance developed by a long term natural selection, display high resistance to heat and ticks and good productivity on poor forage and low amounts of water, whereas, Sheko and Abigar are found in humid and forest areas where burden of tsetse flies is high [27].

Phylogenetic analysis
We constructed median joining network to reveal the phylo genetic relationships among the cattle populations studied (Figure 1).The network illustrates a starlike pattern indicat ing a population expansion event and this strengthened by unimodal population expansion event detected (Figure 3).This observation is consistent with the dloop region of the mtDNA reported on Ethiopian indigenous cattle populations [11].In addition, among the five globally identified maternal origins (T, T 1 , T 2 , T 3 , and T 4 ) of Bos taurus and two lineages (I 1 and I 2 ) of Bos indicus [2,28,29], all the 81 haplotypes detected in ten Ethiopian cattle populations clustered with Haplogroup T1 [11].However, double peaks were reported for North Ethiopia cattle populations using Y chromosome microsat ellite haplotype mismatch distribution analysis [7].All the detected haplotypes (H = 50) in Kenana and Butan cattle of Sudan belong to haplogroups T1 [24].In the dloop analysis of Ethiopian cattle, a major haplotype occurred at the H1 with the highest frequency which showed a broad geographic dis tribution and represents the possible ancestral haplotype of T1 lineage [11].In the current study, we also observed the highest frequency at the H1 and the second highest haplotype occurrence at the H3 (Figure 1).
Interestingly, in the current study, nine of haplotypes of the Bos taurus reference sequences clustered with Ethiopian in digenous cattle populations.All these haplotypes are from USA, European and Asian Bos-taurus cattle.In contrast, we did not observe any haplotype from Bos indicus reference sequences being shared or clustered with Ethiopian indige nous cattle despite the fact that four of the indigenous cattle populations included in this study are from Bos indicus group.Despite the proportion of variation of influence, influence of European taurine on North Ethiopian cattle was reported but exhibited a weaker influence compared to Asian taurine origin [28].According to Bradley et al [30], African Bos taurus and African Bos indicus share the same African set of taurine mitochondrial DNA haplotypes suggesting the pattern of zebu influence on the African continent was a process of introgres sion rather than replacement of African taurine cattle with unmixed Asian zebu as we also observed in this study.In line with this, the genomewide analysis unveiled presence of sub stantial taurine introgression in Ethiopia zebu, sanga, and zenga cattle [6].As a result of subsequent introgression be Tarekegn et al (2018) Asian-Australas J Anim Sci 31:1393-1400 tween the African taurine type and SouthAsian zebu (Bos indicus) the existing populations, for instance Ethiopian cattle, have been created with different proportions of taurine and indicine backgrounds [1,9].The north Ethiopian cattle breeds have been heavily (>90%) influenced by Zebu, followed by African, European and the NearEastern tourine, respectively [25].A recent worldwide admixture and divergence analysis uncovered that the indicine ancestry in African cattle is higher in East Africa (74%) [10], which confirms the former premise that explains the East African region could be considered as the cradle of African zebu [3].However, Edea et al [6] still argues that the stronger influence of zebu ancestry in Ethio pian cattle can be attributed to the replacement of taurine by Bos indicus than introgression due to adaptation of zebu to harsh environments like tolerance for heat, ticks, drought and poor forage.
Over all, the cluster of the haplotypes detected in Ethiopian indigenous and the cross bred cattle populations to the Bos taurus origin of the reference populations asserts the former finding that shows all African cattle possess taurine humpless longhorns (Bos taurus) [11,26].In line with this, no haplotype which clustered to Bos indicus type was observed from 10 in digenous cattle populations analyzed using the dloop region and attributed to the challenge of recurrent drought and rinderpest epidemics [11].This could lead to mtDNA more sensitive to demographic expansion like population fragmen tation and bottleneck, the east African region suffer [11].In contrast, the very limited number of taurine alleles of the Ychromosome reported for the north Ethiopian cattle popul ation could be a result of recent crossbreeding or incomplete introgression of zebu patrilines [7].Overall, the different data sets reported on Ethiopia cattle populations indicate the distri bution of taurus cattle vary in the different parts of the country.

Historical dynamics of the cattle populations
The bellshaped curve of the population dynamics obtained from the mismatch distribution analysis reveals presence of population expansion (Figure 3).However, the pvalues of the sum of square deviation (SSD = 0.01901±0.009;p = 0.375) and raggedness index (r = 0.1564±0.039;p = 0.318) tests were not significant.On the other hand, negative values for all the populations studied and negative and significant values of the coalescentbased neutrality tests (Tajima' s D for Abigar, F S value for Abigar, Horro, and Sheko) obtained also suggest existence of a population expansion event.
The absence or little deviation of Horro from Boran could mean that Horro's zebu line could be Boran.Horro is from zenga group: cross of zebu and sanga cattle [12].We exhibited similar observation between Sheko and F 1 populations which could be because of their taurus background.On the other hand, moderate population differentiation was observed be tween tsetse susceptible Bos indicus representative (Boran) and resistant Bos taurus group (Sheko) and F 1 populations (Table 3).This observation is strengthened by highest (0.229) Nei's corrected distance observed between Boran and Sheko using 10 microsatellite markers [14].Solomon [14], reported very narrow among population variation (2.2%) for the same study populations (Boran, Abigar, Guraghe, Horro, and Sheko) with the same SSR markers.
In conclusion this study revealed highest mtDNA varia tions among Ethiopian cattle populations and the haplotypes clustered into a haplogroup which goes in line with previous reports on East African cattle populations [11,24].Moreover, our finding asserted that the cattle populations studied are from the Bos taurus ancestral origin given the fact that they are highly influenced by the Bos indicus group.The dynamics of population history showed presence of a onetime but rapid and recent population expansion in all the cattle populations studied.

Figure 1 .
Figure 1.Phylogenetic network graph of Ethiopian cattle populations.The graph was constructed to obtain insights on the genetic relationships between the haplotypes and determine the origin of the maternal lineages in the context of Bos taurus and Bos indicus reference haplotypes retrieved from the Gene bank.All the mutations and character states were weighted equally.All haplotypes detected in Ethiopian cattle are clustered with the Bos taurus reference haplotypes indicating Ethiopian cattle populations are from Bos taurus origin.

Mitochondrial DNA sequence variation and genetic diversity
In this study, we analyzed the complete mitochondrial cyt b gene in five indigenous and one cross bred cattle populations in Ethiopian together with published sequences of Bos taurus and Bos indicus cattle.The number of segregating sites (S = 18)

Figure 2 .
Figure 2. Neighbor joining (NJ) tree of Ethiopian cattle populations.The NJ tree was constructed based on the Kimura two-parameter model algorithm in MEGA6 [19] to demonstrate the relationships among the haplotypes found in the study cattle populations.The NJ tree shows absence of clear phylogeographic structure among Ethiopian cattle populations.Five haplotypes of the Bos taurus reference individuals clustered together with Ethiopian cattle; however, no haplotype from the reference Bos indicus animals clustered together with Ethiopian cattle.

Figure 3 .
Figure 3. Demographic dynamics among the indigenous cattle populations in Ethiopia.The demographic dynamics of each population was inferred from mismatch distribution patterns following 1,000 coalescent simulations.The figure shows a uni-modal peak indicating a onetime population expansion in each population.

Table 1 .
Genetic diversity indices of Ethiopian cattle populations using cyt b region of mtDNA Cyt, cytochrome; N, number of samples; S, segregating sites; H, number of haplotypes; K, nucleotide differences; H d , haplotype diversity; π, nucleotide diversity; F 1 , Holstein Friesian (HF) × Barka cross.

Table 2 .
Analysis of molecular variance (AMOVA) of Ethiopian cattle populations