Mitochondrial DNA diversity and origin of indigenous pigs in South China and their contribution to western modern pig breeds

WANG Chen, CHEN Yao-sheng, HAN Jian-lin , MO De-lin, LI Xiu-jin, LIU Xiao-hong

1 State Key Laboratory of Biocontrol, School of Life Sciences, Sun Yat-sen University, Guangzhou 510006, P.R.China

2 CAAS-ILRI Joint Laboratory on Livestock and Forage Genetic Resources, Institute of Animal Science, Chinese Academy of Agricultural Sciences, Beijing 100193, P.R.China

3 International Livestock Research Institute (ILRI), Nairobi 00100, Kenya

Abstract Indigenous pigs in South China are valuable genetic resources with many specific and unique characters, which have played an important role in the establishment of some western modern pig breeds. However, the origin and genetic diversity of indigenous pigs in South China have not been fully understood. In the present study, we sequenced 534 novel mitochondrial DNA (mtDNA) D-loop and assembled 54 complete mitogenome sequences for all 17 indigenous pig breeds from Fujian,Guangdong, Guangxi and Hainan in South China. These data were analyzed together with previously published homologous sequences relevant to this study. We found that all 13 coding genes of the mitogenomes were under purifying selection,but ND1 had the most variable sites and CYTB contained the most non-synonymous SNPs. Phylogenetic analysis showed that all indigenous pigs in South China were clustered into the D haplogroup with D1a1, D1b, D1c and D1e sub-haplogroups found to be dominant. Haplotype and nucleotide diversities of D-loop sequences ranged from 0.427 to 0.899 and from 0.00342 to 0.00695, respectively, among which all pigs in Guangdong had the lowest diversity. The estimates of pairwise FST, gene flow (Nm) and genetic distance (Da) indicated that most of these indigenous pig breeds differentiated from each other significantly (P＜0.05). Among the western modern breeds, Berkshire and Yorkshire had significant Asian matrilineal footprints from indigenous pigs in South China, especially the Spotted pigs distributed in Guangdong and Guangxi. The neutrality test (Fu’s FS) indicated that indigenous pigs from Fujian and Guangxi had gone through recent population expansion events (P＜0.05). It is concluded that indigenous pigs in South China were most likely derived from the Mekong region and the middle and downstream regions of Yangtze River through Guangxi and Fujian. Our findings provide a complete and in-depth insight on the origin and distribution pattern of maternal genetic diversity of indigenous pigs in South China.

Keywords: pig, South China, mitogenome, D-loop, genetic diversity

1. Introduction

China is the country that owns the largest number of pigs and pig breeds in the world. Chinese indigenous pigs had undergone natural and human-driven selection for more than 8 000 years (Yuan and Flad 2002), and they harbor a higher level of diversity than European pigs (Li et al. 2017).According to their geographical distributions, phenotypes and production performances, Chinese indigenous pig breeds have been divided into six types of North China,Southwest, Lower-Yangtze River Basin, Central China,South China and Plateau (Zhang et al. 1986; Megens et al. 2008; Yang et al. 2017). To adapt to cold or hot environments, indigenous pigs from northern or southern China have evolved many adaptive genome variants (Ai et al. 2015). Pigs of the South China type are mainly distributed in Fujian, Guangdong, Guangxi and Hainan. The appearances of most pig breeds developed in this region are black and white. Moreover, these pigs have many unique characteristics, such as desired meat quality, early sexual maturity, small body size, strong heat tolerance and excellent adaptability to crude fiber and low nutrient feeds. According to historical records, indigenous pigs in South China have played an important role in the improvement of modern pig breeds worldwide (Wang et al. 2011).

Mitochondrial DNA (mtDNA) has been commonly used in evaluating the relationships of wild and domestic animals,such as sheep (Lv et al. 2015), yak (Yue et al. 2016), dog(Ardalan et al. 2011) and camel (Ming et al. 2017). The D-loop in mtDNA tends to be used more widely because of its higher variation than the remaining regions of the mitogenome (Cann et al. 1984). Evidence of mtDNA variation from wild boars and domestic pigs suggested that the origin of wild boars took place in the islands of Southeast Asia and they migrated across the Kra Isthmus on the Malay Peninsula into Mainland Asia without human assistance, then the pigs were independently domesticated in multiple centers throughout the world (Larson et al. 2005,2010). East Asia was divided into nine geographic regions referred to in previous studies on pig genetic diversity:the Mekong region, the upstream region of Yangtze River(URYZ), the middle and downstream region of the Yangtze River (MDYZ), South China (SC), upstream and middle region of the Yellow River (UMYR), the downstream region of the Yellow River (DRYR), Northeast Asia (NEA), South Pacific Islands (SPI), and Australia and New Zealand (AN)(Wu et al. 2007; Ji et al. 2011). Pig domestication in East Asia was initiated mainly in the Mekong region and MDYZ(Wu et al. 2007). Additionally, pigs in URYZ (Jin et al. 2012),Tibetan highland (Yang et al. 2011), the middle Yellow River basin (Larson et al. 2010) and Northeast China (Xiang et al.2017) were also considered as local origins.

To date, a large number of mtDNA sequences of indigenous pigs in China have been published, however,previous samples were collected mainly from the northwest region (Tibet, Qinghai and Gansu) (Fang and Andersson 2006; Wu et al. 2007; Larson et al. 2010; Yang et al.2011; Jin et al. 2012; Zhang et al. 2016), southwest region(Sichuan, Yunnan and Guizhou) (Fang and Andersson 2006;Wu et al. 2007; Larson et al. 2010; Yang et al. 2011; Jin et al. 2012; Zhang et al. 2016), northern region (northeast China and Shandong) (Fang and Andersson 2006; Wu et al. 2007; Larson et al. 2010; Zhang et al. 2016) and MDYZ region (Jiangsu, Zhejiang and Jiangxi) (Fang and Andersson 2006; Wu et al. 2007; Larson et al. 2010; Jin et al. 2012; Huo et al. 2016). In recent years, the complete mitogenomes, which were amplified and sequenced through conventional PCR following a primer walking strategy, have also been used to analyze the genetic relationships among pig breeds (Li et al. 2016). These previous analyses help us to understand the genetic diversity of indigenous pigs in most geographic areas of China, however, the numbers of pig samples and breeds collected from South China were limited to only a few studies (Wu et al. 2007; Larson et al.2010). To gain more complete insights on the origin and distribution pattern of genetic diversity in indigenous pigs in South China, it is essential to conduct a comprehensive study involving additional sampling towards a complete list of indigenous pig breeds from the region.

In the present study, we assembled 54 complete mitogenomes from whole genome re-sequencing data and sequenced 534 novel D-loop sequences of indigenous pigs in Fujian, Guangdong, Guangxi and Hainan. Relevant sequences of indigenous pigs in South China and some modern pigs in Europe were retrieved from the GenBank database for meta-analysis and comparison. Since pigs have evolved under isolated terrains of sub-tropical and tropical landscapes as well as different human selection preferences, we hypothesized that the genetic differentiation among indigenous pig breeds is more significant in South China than other parts of China. Furthermore, Dahuabai pig, the only spotted breed distributed in the Pearl River Delta, may have been involved in the development of some western modern pig breeds. The objectives of this study were: (1) to identify the maternal origin of indigenous pig breeds in South China; (2) to examine the distribution pattern of genetic diversity of indigenous pig breeds in South China for rational conservation and utilization; and(3) to verify which indigenous pig breed(s) from South China have contributed to the establishment of western modern pig breeds.

2. Materials and methods

2.1. Experimental design and treatments

All experiments were carried out according to the protocols approved by the Animal Care and Use Committee of Guangdong Province, China. The approval ID was SYXK(Guangdong) 2011-0112. A total of 550 individuals from the 17 indigenous pig breeds were collected in South China,including Minbei Spotted pig (MBH, n=30), Putian pig (PT,n=31), Guanzhuang Spotted pig (MGZH, n=36), Huai pig(HZ, n=33) from Fujian; Yuedong Black pig (YDH, n=33),Dahuabai pig (DHB, n=24), Meihua pig (MH, n=35), Lantang pig (LT, n=26) and Guangdong Small-ear Spotted pig(GDXE, n=33) from Guangdong; Luchuan pig (LC, n=30),Guizhong Spotted pig (GZH, n=39), Bama Xiang pig (BMX,n=48), Longlin pig (LL, n=36) and Debao pig (DB, n=16)from Guangxi; and Wuzhishan pig (WZS, n=40), Ding’an pig (DA, n=30) and Tunchang pig (TC, n=30) from Hainan.Nearly all the individuals were active breeding boars and sows of these breeds, and shared no immediate parents.

A total of 36 genomic DNA samples from 12 breeds,representing three pigs per breed, were re-sequenced for the whole genomes used to assemble their mitogenomes,including MBH, PT, MGZH, HZ, YDH, DHB, LT, GDXE,TC, GZH, LL and DB. The other 534 samples, including 20 out of the 36 samples used for whole genome resequencing, were Sanger sequenced for D-loop. To obtain a more comprehensive insight on the origin and significance of indigenous pigs in South China, we retrieved publicly available whole genome re-sequencing data of 18 indigenous pigs from South China (BMX, LC and WZS:GenBank accession no. SRA096093). Moreover, we downloaded additional complete mitogenomes of 134 wild boars and indigenous pigs in Asia and Europe as well as D-loop sequences of 136 wild boars and indigenous pigs in South China and 70 western modern pigs (Berkshire, n=8;Duroc, n=24; Landrace, n=20; and Yorkshire, n=18) from the GenBank. All the information of these sequences is shown in Appendices A and B.

2.2. Animals and feeding

Among the 17 indigenous pig breeds sampled in our study,nine were collected from the national conservation farms,including PT, HZ, YDH, LT, DHB, LC, BMX, TC and WZS,while other breeds were sampled from either local pig farms or smallholder pig keepers. All individuals were identified by the standard characteristics of each breed (Zhang et al. 1986; Wang et al. 2011). According to available feed resources in the distribution ranges of these indigenous breeds, various diets were formulated to meet their nutrient requirements and fresh water was available ad libitum.

2.3. Sample collection and DNA extraction

Only ear tissues were collected, and then animals were released immediately following a treatment of the wounds with an antiseptic preparation. All the tissues were preserved in 75% ethanol and stored at -80℃ prior to DNA extraction. Total genomic DNA was isolated using a standard phenol-chloroform method.

2.4. Re-sequencing of the whole genomes and Sanger sequencing of the mtDNA D-loop

A total of 36 genomic DNA samples were used to construct individual paired-end sequencing libraries with an insert size of 350 bp, following the Illumina DNA sample preparation protocol. Sequencing was performed to generate an average at 105.18 Gb of 150 bp long paired-end reads for each sample on an Illumina HiSeq X Ten System.

Approximately 680 bp of the mtDNA D-loop of 534 samples were successfully PCR amplified using a pair of previously published primers (Larson et al. 2005). The primers were synthesized by Shanghai Sangon Biotech Co., Ltd. (China). PCR amplification was set up using 50 μL volume, including 25 μL 2×Taq PCR StarMix (GenStar,Beijing, China), 2 μL of each primer (10 pmol μL-1), 1 μL genomic DNA (200 ng μL-1) and 20 μL ddH2O. The PCR reaction was performed under the following themocycling conditions: 94°C for 2 min, followed by 40 cycles of 94°C for 30 s, 57°C for 30 s and 72°C for 1 min, and a final extension at 72°C for 5 min. The PCR products were purified and sequenced using the PCR forward primer by Guangzhou IGE Biotechnology Ltd. (Guangzhou, China).

2.5. Reference-guided assembly of complete mitogenomes and cleaning of D-loop sequences

All the whole genome re-sequencing reads of 54 individuals were aligned to the reference mitogenome (GenBank accession no. EF545571.1) of a wild boar from Fujian using BWA-MEM (Li and Durbin 2009) with default parameters.SAMtools (Li et al. 2009) was used to convert mapped reads into BAM format, and then to sort, merge and extract the paired reads that were mapped to the reference mitogenome. Duplicate reads were removed with Picard(http://broadinstitute.github.io/picard/). Variants including SNPs and InDels were called and filtered, and eventually the complete mitogenomes were generated using GATK 3.7(McKenna et al. 2010). The D-loop, two rRNAs, 13 proteincoding genes and 22 tRNAs of the mitogenomes were determined following the annotations in the reference mitogenome. The 534 novel D-loop sequences obtained in this study were manually cleaned using Chromas 2.6 (https://chromas.software.informer.com/) and restricted to 643 bp between the 1st and 643rd nucleotides of the reference mitogenome for following the analyses. All new complete mitogenome and D-loop sequences were deposited in the GenBank database (MG250514-MG250567 and KY673797-KY674330).

2.6. Statistical analyses

Diversity and natural selectionIn total, we analyzed 188 complete mitogenomes (including 70 indigenous pigs in South China and another 118 pigs in Asia and Europe,see Appendix A) and 774 D-loop sequences (including 40 wild boars and 664 indigenous pigs in South China and 70 western modern pigs, see Appendix B). All the sequences were aligned using MEGA 7.0 (Kumar et al. 2016) and all positions containing gaps were excluded. The percentage of nucleotide compositions, and numbers of observed transitions (Ti) and transversions (Tv) were examined using ARLEQUIN version 3.5 (Excoffier and Lischer 2010). The polymorphic sites (S), nucleotide diversity (π), number of haplotype (H) and haplotype diversity (Hd) were estimated using DnaSP 5.10 (Librado and Rozas 2009). Selective pressures on the 13 protein-coding genes in mitogenomes of indigenous pigs in South China were estimated by the relative rates of nonsynonymous and synonymous substitutions (ω=dN/dS) using the CodeML function implemented in PAMLX (Xu and Yang 2013) with options of runmode=0, model=0 and fix α=0.

PhylogenyBoth mitogenome and D-loop sequences were classified into specific haplogroups using MitoToolPy(Peng et al. 2015). For phylogenomic analysis of 188 mitogenomes, the best fitting model GTR+I+G was selected using the Modeltest 3.7 implemented in PAUP 4.0 (Posada 2003) for the construction of the phylogenetic tree using MrBayes 3.2 (Ronquist and Huelsenbeck 2003), with the following parameters: ngen=6 000 000, samplefreq=1 000,nchains=4 and sumt burnin=1 500. An African warthog sequence (Phacochoerus africanus; GenBank accession no. DQ409327) was added as an outgroup. Network 5.0(Bandelt et al. 1995) was used to construct a phylogenetic network for all D-loop sequences of D haplogroup.

Genetic differentiation and demographyOnly D-loop data were included for these analyses. The pairwise genetic differentiations between the 17 pig breeds in South China were estimated by the pairwise fixation index (FST) using ARLEQUIN 3.5, the gene flow (Nm) using DnaSP 5.10 and the net genetic distance (Da) using MEGA 7.0. The Analysis of Molecular Variance (AMOVA) was performed using ARLEQUIN 3.5. Inferences of the demographic history were based on the Tajima’s D (Tajima 1989) and Fu’s FS(Fu 1997) implemented in the ARLEQUIN 3.5, as well as the mismatch distribution using DnaSP 5.10.

3. Results

3.1. Characteristics of the complete mitogenomes

The whole genome re-sequencing data for each sample reached an average depth of 3 028-fold of the reference mitogenome (EF545571.1). The lengths of 54 complete mitogenomes of indigenous pigs in South China varied between 16 690 and 16 702 bp. The PCR amplified D-loop sequences of the 20 samples included in the whole genome re-sequencing verified their sequencing reliability. For all 70 mitogenomes of indigenous pigs in South China,the variations of each of 13 protein-coding genes were analyzed (Table 1). The average nucleotide compositions of all genes exhibited an AT-bias. The mean Ti/Tv ratio(ratio of transition and transversion) was 10.44. Among the 13 coding genes, ND1 had the most variable sites while CYTB carried the most non-synonymous SNPs.The values of dN/dS (relative rates of nonsynonymous and synonymous substitutions) of all genes were lower than 1, suggesting that purifying selection had acted on the functions of coding sequences in the mitogenomes of indigenous pigs in South China.

3.2. Phylogenetic analyses

To investigate the relationships of indigenous pigs in South China with other pigs around the world, 188 mitogenomes were included in their phylogenetic reconstruction (Fig. 1).It was clear that these mitogenomes were classified into two major clades of Asian and European origins. In the Asian clade, Asian wild boars were mainly dispersed in the basal haplogroups of A and D, while domestic pigs were present in the secondary D sub-haplogroups. The indigenous pigs in South China were dominant in sub-haplogroups of D1b, D1c and D1e. Furthermore, sub-haplogroup D1a1,which was derived from pigs all over China, had about half of the samples which originated from South China.Seven mitogenomes of western modern pigs were also clustered into the Asian clade, including one Berkshire and one Yorkshire present in D1a1, two Yorkshires in D1b,two Berkshires and one Yorkshire in D3, confirming the maternal contribution of indigenous Asian pigs to western modern breeds.

3.3. Genetic diversity of D-loop of indigenous pigs in South China

To achieve a complete and in-depth understanding of the genetic diversity of indigenous pigs in South China, we analyzed the D-loop sequences of 664 indigenous pigs and 40 wild boars from the four regions. There were 40 variable sites including nine singleton variable sites and 31 parsimony informative sites, which defined 67 haplotypes:57 from indigenous pigs and 15 from wild boars. There were five shared haplotypes between indigenous pigs and wild boars in the region. Compared to 136 previously publishedhomologous D-loop sequences, we identified 31 novel haplotypes in the newly sampled indigenous pigs in the four provinces (Appendix C). The genetic diversity in terms of Hd and π at the breed level are shown in Table 2. For

breeds with more than 16 samples, Hd ranged from 0.427 in DHB to 0.899 in LL and the π ranged from 0.00342 in MH to 0.00695 in MBH. Additionally, all five indigenous pig breeds in Guangdong and MGZH distributed in southwestern Fujian bordering northeastern Guangdong had relatively low Hd ranging from 0.427 in DHB to 0.716 in GDXE. Twenty of the 67 haplotypes were private to individual breeds, each represented by a single sample. There were six private haplotypes in PT, four in Hainan wild boars, two each in LL and HZ, and one each in LT, LC, LG (Lingao pig), GZH,MGZH and Fujian wild boars.

Table 1 Summary of the 13 genes in complete mitogenomes of 70 indigenous pigs from South China in this study1)

Fig. 1 Phylogenetic tree of 188 complete mitogenomes of wild boars and domestic pigs from around the world. Indigenous pigs from the four regions in South China are labled with black dots while the western modern pigs are labeled with blue dots. All the samples in each haplogroup are shown in different colors.

3.4. Population differentiation and demographic history of indigenous pigs in South China

Based on the AMOVA of D-loop data, only 1.27% (P＞0.05)of the total variation was found among the four regions,however, up to 15.95% (P＜0.01) and 82.78% (P＜0.01)of the total variation were present among specific breeds within provinces and within all breeds, respectively (Table 3),suggesting that most of these indigenous pig genetic resources were unique within the four regions. The pairwise FSTestimates ranged from -0.0196 to 0.4347, the Nm from -141.28 to 509.29 and the Da from -0.00009 to 0.00294 (Tables 4 and 5). Significant differentiations(FSTvalues, P＜0.05) were observed between most indigenous pig breeds, of which the three breeds of MGZH,DHB and MH were significantly differentiated (P＜0.01) from the remaining 14 breeds. All the FST, Nm and Da estimates suggested that DB from Guangxi was close to DA from Hainan and LC from Guangxi.

When the effects of natural selection and past demographic changes were examined in each of the 17 breeds using the Tajima’s D and Fu’s FSestimates (Appendix D), only LL from Guangxi had a significantly negative Fu’s FSvalue (-4.372,P＜0.05). In addition, significantly negative Fu’s FSestimates were also observed in indigenous pigs from Fujian (-7.960,P＜0.05) and Guangxi (-14.996, P＜0.01) (Table 6). However,the mismatch distribution analysis did not reveal any signal of population expansion in any indigenous pig breeds(Appendix E). Furthermore, the mismatch distribution patterns of these indigenous pig breeds were rather similar across the four regions (Fig. 2).

Table 2 Number of samples, haplotype and nucleotide diversity of 664 indigenous pigs and 40 wild boars in South China1)

Table 3 AMOVA of the 17 indigenous pig breeds based on mtDNA D-loop sequences

3.5. Phylogeographic analysis

A phylogenetic network was drawn for all D-loop haplotypes of indigenous pigs and wild boars in South China along with western modern pigs carrying the Asian haplogroup D, to better visualize their phylogeographic patterns (Fig. 3). In total, 26 of the 70 western modern pigs were distributed in D haplogroup, eight in D1a, 12 in D1b, one in D1e and five in D3 sub-haplogroups. Among the four western breeds,seven out of the eight Berkshires and 15 out of the 18 Yorkshires carried Asian mtDNA haplotypes, suggesting their significant Asian matrilineal footprints. All the Asian haplotypes of Berkshire were of D1a (3/7) and D3 subhaplogroups (4/7), while most of the Asian haplotypes of Yorkshire were of D1b sub-haplogroup (11/15), followed by two in D1a, one in D1e and one in D3 sub-haplogroups.Duroc and Landrace carried limited Asian haplotypes (3/24 and 1/20, respectively).

The five haplotypes shared between wild boars and indigenous pigs in South China suggested their genetic admixture. In Fujian, 60% of the wild boars were of haplogroup A while the remaining ones were of D1 subhaplogroups. In Hainan, 84% of the wild boars were in haplogroup D (8% in D1, 52% in D2 and 24% in D3 subhaplogroups) and the rest were in haplogroup A (Table 7).Indigenous pigs in South China were dominant in the following four sub-haplogroups: D1a (216 individuals,32.53%), D1b (141 individuals, 21.23%), D1c (71 individuals,10.69%) and D1e (147 individuals, 22.14%) (Table 7).There were four core haplotypes of H2, H3, H13 and H15,which were in the centers of D1b, D1c, D1e and D1a subhaplogroups, respectively. H13 had the highest frequency,including two wild boars and 25 indigenous pigs from Fujian, 60 indigenous pigs from Guangdong, 31 indigenous pigs from Guangxi and 16 indigenous pigs from Hainan. H2 was the second most frequent haplotype, observed in 112 pigs from the four regions, of which Guangxi had the largest share with 42 pigs. Moreover, several other haplotypes surrounding the H2 were also from Guangxi. Compared with other sub-haplogroups, D1a had unique haplotype compositions with several relatively frequent haplotypes surrounding H15. Its star-like networking pattern among major haplotypes indicated a recent population expansion.

Table 6 Results of neutralisty test of the 17 indigenous pig breeds in the four regions

Fig. 2 The mismatch distribution of D-loop sequences of indigenous pigs from the four regions in South China. A, Fujian. B,Guangdong. C, Guangxi. D, Hainan.

Fig. 3 A reduced median network of mtDNA D-loop haplotypes of all the indigenous pigs and wild boars from the four regions in South China, together with western moden pigs carrying the Asian haplogroup D. These indigenous pigs were from Fujian(FJD), Guangdong (GDD), Guangxi (GXD) and Hainan (HND), while the wild boars were from Fujian (FJW) and Hainan (HNW).Sequences of western modern pigs (EUD) were downloaded from the GenBank database. Samples from each different regions are indicated by different colors. Each haplotype is represented by a circle with the size proportional to its frequency. The length of each branch is proportional to the number of mutaions on the respective branch.

Table 7 Haplogroup distribution of the indigenous pigs and wild boars in the four regions compared with western modern pigs

4. Discussion

The next generation sequencing (NGS) technologies can produce a large amount whole genome re-sequencing data covering the associated mitogenome of a sample. Two approaches have been proposed to generate complete mitogenomes from the NGS datasets; the first is de novo assembly using MITObim (Hahn et al. 2013) or genome skimming (Male et al. 2014) while the second is NGS reads mapping which is dependent on a reference mitogenome.The latter is very effective at identifying sequence variants.In our study, we used BWA-MEM (Li and Durbin 2009) to map the NGS reads to the reference mitogenome (EF545571.1).GATK HaplotypeCaller function (McKenna et al. 2010) was employed to call SNPs and InDels simultaneously via a local de novo assembly. The identical D-loop sequences of the 20 samples which were both Sanger sequenced and NGS re-sequenced at very high coverage (e.g., 3028 X), proved the high fidelity of our sequencing data. Compared to other haplogroups/sub-haplogroups, haplotypes of D1b had a unique insertion of 11 nucleotides (TTATAAAACAC) in one or two copies in their D-loop after the 1 093rd nucleotide in the reference mitogenome EF545571.1.

In Chinese cooking culture, pork has been the most important meat in daily life and traditional festivals in South China. Along with the rapid economy growth and changes in custom preferences in China, many lean pig breeds/lines have been introduced from western countries into China for genetic improvement of indigenous pigs. The subsequent indiscriminate crossbreeding and breed replacement have led to a reduction of mtDNA haplotype and nucleotide diversity in some Chinese indigenous pig breeds (Huo et al.2016), and thus to a high risk of extinction. Guangdong is well known as one of the most developed provinces in China. As expected, pigs from this province showed the lowest haplotype diversity (0.810±0.016), while pigs from Fujian and Guangxi had relatively high haplotype diversities(0.923±0.010 and 0.912±0.011, respectively) (Appendix F).Following the significant social transition and urbanization in Guangdong, the number of pigs has been rapidly reduced in the province (Wang et al. 2011). Recent local governmental policy in restricting pig farming in the core corridor of the Pearl River Delta has clearly contributed to the fast decrease in the effective population size and genetic diversity of indigenous pigs in Guangdong. A previous study indicated that pigs in URYZ, MDYZ, SEA, Yunnan and Tibetan highland had higher haplotype diversities than those from SC, NEC and DRYR (Jin et al. 2012). Here, we reevaluated all indigenous pig breeds in SC with an increased number of samples and found a higher haplotype diversity (0.906±0.006) than those reported earlier (0.6679±0.00428, Wu et al. 2007;0.861±0.021, Jin et al. 2012), but the nucleotide diversity(0.00540±0.00006) was similar (0.00539±0.00316, Wu et al.2007; 0.00547±0.0002, Jin et al. 2012). Combined with the AMOVA results, it is evident that the genetic diversity of all indigenous pigs in South China was relatively high, while a considerable genetic differentiation was also present among different breeds in the four regions. Therefore,some immediate conservation efforts supported by genetic information should be taken for the sustainable utilization of these unique breeds, most of which are at a high risk of extinction at the moment.

In China, some indigenous pig breeds were raised by native residents in different areas and assigned different names, but they were considered to be of the same breed. In our study, both GDXE and LC pigs were called Liangguang Small-spotted pig, nevertheless, they are kept in the bordering area of western part of Guangdong and the Luchuan County of eastern Guangxi, respectively. DHB and MH are raised in different areas of Guangdong, but they used to be called Large Black-white pigs. Indigenous pigs in the four counties of LG, TC, DA and Wenchang(WC) in northern Hainan, shared an almost overlapping distribution and a very similar morphology, but have been defined as four separate local breeds with the names of these counties. In our study, the FST, Nm and Da all showed that GDXE was close to LC while DA was close to TC,indicating that GDXE and LC had a similar maternal genetic backgrounds, and similarly for DA and TC which also share large portions of their distribution ranges. Furthermore,five out of the six haplotypes found in GDXE were shared by LC which had 11 haplotypes, indicating that GDXE is probably a sub-population of LC. Nevertheless, interesting results were obtained between DHB and MH with significant differentiation, limited gene flow and a long genetic distance.These two local breeds had a very low haplotype diversity and shared only one haplotype. DHB was sampled on the national conservation farm in the south of Guangdong while MH was from another local conservation farm in the north of Guangdong. Qu et al. (2011) showed that the diversities of the native pigs in the nucleus and conservation herds were significantly lower than that in the randomly bred local populations, suggesting the need to apply NGS information to maintain most of the diversity, fitness and selection signatures in the conservation work of some indigenous pig breeds (Bosse et al. 2015).

The process of pig domestication took place in China earlier than Europe, where pigs were allowed to freely roam in the forests as domesticated herds until the Late Middle Ages, while Chinese pigs had been kept in enclosures at a relatively early stage (White 2011). During the Industrial Revolution, European breeders turned their attention to Asia and imported Chinese pigs to improve commercial traits in their pig breeds (Giuffra et al. 2000; Bosse et al.2014), and previous studies suggested about 20-35%Asian contribution to western modern breeds (Groenen et al. 2012; Bianco et al. 2015). In our study, all four western pig breeds shared Asian mtDNA haplotypes, of which Berkshire and Yorkshire carried significantly frequent Asian mitogenomes, validating previous findings on their footprints of Asian matrilineal origin (Kim et al. 2002; Li et al. 2014). Eighteen Yorkshires had 11 Asian haplotypes of sub-haplogroup D1b, mostly derived from the Spotted pigs in Guangxi. Eight Berkshires carried seven Asian haplotypes of sub-haplogroup D1a and D3 that were mainly shared by DHB and MGZH. MGZH had been improved by the Nankou Spotted pigs from Guangdong since 1924 and are now mainly kept in Shanghang County close to Guangdong (Wang et al. 2011). According to the historical records, Guangdong pig breeds were introduced into ancient Rome and used to breed with primitive Roman pigs(Wang et al. 2011). The shared D-loop haplotypes between Chinese indigenous pigs and western modern pigs were mostly present in pigs from Southeast China, in particular from Guangdong (97.73%), Guangxi (96.00%) and Fujian(86.67%) (Zhang et al. 2016). We therefore postulate that the Spotted pigs in Guangdong and Guangxi might be the main source of the Asian maternal contribution to western modern pigs, e.g., Berkshire and Yorkshire breeds.

Multiple origins of domestication have been confirmed by molecular studies in many domestic animals, such as chickens (Liu et al. 2006), horses (Lei et al. 2009) and goats (Luikart et al. 2001). Pig domestication in East Asia occurred in the Mekong region and MDYZ, harboring the dominant sub-haplogroup D1b and D1a1a, respectively(Wu et al. 2007). From the D-loop phylogenetic network,the major haplotype H2 of sub-haplogroup D1b and most individuals in sub-haplogroup D1b in our study were from Guangxi near the Mekong region. Thirty Late Pleistocene wild boar fossils, which were excavated from three caves in Guangxi and dated back to at least 13 000 years ago,showed a close phylogenetic relationship to Asian pigs (Hou et al. 2014). Sub-haplogroup D1a had a star-like pattern in the phylogenetic network and possessed a high number of individuals of D1a1a. In sub-haplogroup D1e, 60 pigs were from Guangdong followed by 32 pigs and three wild boars from Fujian where the wild boars were defined as MDYZ in other studies (Wu et al. 2007; Ji et al. 2011). Based on the complete coding region of mtDNA, most indigenous pigs in China might be originated from wild boars distributed in the Yangtze River Region and South China (Yu et al. 2013).However, there are no wild boar samples from Guangdong and Guangxi, thus further research should be undertaken with additional wild boar samples from this region, to reveal the complete picture of the origin of indigenous pigs in South China.

5. Conclusion

In the present study, we generated a large number of complete mitogenomes of indigenous pigs in South China based on the whole genome re-sequencing data. A fine regional phylogeographic analysis was achieved involving novel D-loop sequences of indigenous pigs in South China.The major matrilineal components of indigenous pigs in South China suggested their probable migration from the Mekong region and MDYZ through Guangxi and Fujian.The Spotted pigs in Guangdong and Guangxi were the major source of Asian maternal contribution to Berkshire and Yorkshire breeds. These findings improved our understanding of the matrilineal origin and phylogeographic distribution pattern of genetic diversity of indigenous pigs in South China.

Acknowledgements

This work was supported by the Basic Work of Science and Technology Project, China (2014FY120800) and the Science and Technology Project of Guangdong Province,China (2014YT02H042, 2014B020202001).

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm