Asian Elephant Evolutionary Relationships: New Perspectives from Mitochondrial D-Loop Haplotype Diversity

Abstract

Mitochondrial displacement loop (mt D-loop) sequence analyses have greatly improved assessments of genetic diversity, structure, and population dynamics of endangered species threatened by climate change and habitat loss. Tracking population haplotypes of these species using mitochondrial-based markers has opened new avenues for conservation genomics and biodiversity research. Recent studies have used mt D-loop sequences to assess the genetic diversity of the largest land mammal in Asia, the Asian elephant (Elephas maximus), whose populations are rapidly declining. Here, we review haplotype data from mt D-loop sequencing studies and highlight previous population-scale hypotheses pertaining to the origin and diverse genetic profiles of Asian elephants. Retrieving haplotype information from elephant populations can substantially improve estimations of different parameters relevant to their conservation and allow introgression/hybridization dissection of genetic variation to shed light on ongoing evolutionary processes.

1. Introduction

The Asian elephant (Elephas maximus Linnaeus, 1758) [1] is the largest land mammal in Asia and has high ecological importance as an umbrella, keystone, and flagship species [2]. During the Pleistocene and early Holocene epochs, wild populations greatly declined and fragmented from dry to wet forest and grassland habitats. Recent anthropogenic factors, such as poaching, habitat loss, and human conflict, have exacerbated this decline; see also [3,4,5,6,7]. The Asian elephant has three recognized subspecies based primarily on body size and slight differences in coloration, including (1) Elephas maximus maximus (Asian elephant, Linnaeus, 1758) from Sri Lanka, (2) Elephas maximus indicus Cuvier, 1798 (Indian elephant) [8] from the Asian mainland, and (3) Elephas maximus sumatranus Temminck, 1847 (Sumatran elephant) [9] from Sumatra Island, Indonesia [10,11,12,13]. Elephas maximus maximus has the largest body size and the darkest skin color compared with other subspecies, while the skin color of E. maximus indicus, with smaller patches of depigmentation, is darker than that of E. maximus sumatranus [10,11,12,13]. Another tentative unique subspecies based on genetic distance and morphological observations is the Borneo elephant (Elephas maximus borneensis Deraniyagala, 1950), the smallest living elephant in the present day. Their habitat is in the fragmented forests of Borneo Island [12,14,15]. Currently, the most common genetic approach for studying Asian elephant populations or subspecies is the estimation of mitochondrial displacement loop (mt D-loop) sequence variation. Mt D-loops are semi-stable, triple-stranded regions of mitochondrial DNA [16] that can be used to explain phylogenetic topology by tracing evolutionary relationships [12,17,18,19,20,21,22,23]. The Asian elephant is postulated to have two divergent lineages of mt D-loop haplotypes, the ‘α’ and the ‘β’ clades or haplogroups, with approximately 3% sequence divergence [12,17,19,23,24,25,26]. Haplotypes from these two clades coexist within Asian elephant populations across the Asian continent [17,19,25,26]; this is contrary to other mammalian species, such as the crowned shrew (Sorex coronatus), field vole (Microtus agrestis), bank vole (Clethrionomys glarreolus), pygmy shrew (Sorex minutus), and black muntjac (Muntiacus crinifrons), where divergent clades are geographically distributed in Asia and Europe [27,28,29,30,31].

The existence of two clades across the Asian continent has been explained by several hypotheses (discussion in Section 2); however, the origin and evolutionary history of the Asian elephant remains unclear. One critical aspect is the contact zone of Asian elephants of the α and β clades, which do not completely cover Sundaland, including Thailand. Previously, the contact zone between the α and β clades was positioned in Myanmar, which is the ancestral area of the α clade; however, sampling of Asian elephant populations in Thailand to determine the presence of the two haplogroups has been limited [12,25,26,32]. As a bridge between the mainland and the Southeast Asian islands, Thailand is a pivotal area that has not been extensively studied [33]. The Thai Asian elephant population is now hidden under the tip of an iceberg, whereas several haplotypes in other Asian countries have been tentatively identified [25,26,34]. Surprisingly, our recent study on the Asian elephant population in Thailand identified substantial mt D-loop variations within the α and β clades [23], which tend to support the first and second hypotheses for Asian elephant population haplotype distribution. Taking advantage of data sourced from the breakthrough comparative study of mt D-loops of several Asian elephants, we review evidence pertaining to the origin and different genetic profiles of mt D-loop haplogroups and propose a prospective idea for future discussion. Influential factors for the dynamics of Asian elephant genetic diversity are also discussed to build a sustainable approach with the global community. It is our hope that this new evidence will allow for a better understanding of the coexistence and distribution of the two Asian elephant clades.

2. Distribution of α and β Mitochondrial D-Loop Haplogroups in Asian Elephant Populations

The α and β clades diverged from a common ancestor from approximately 1.5 million years ago (MYA) [25,26,35,36], and the most diverse nucleotide sites were found in the conserved sequence blocks 1 (CSB1) region of the mt D-loop (discussed below). The β haplogroup is widely distributed in Sri Lanka and Sumatra, whereas the α haplogroup is common on the Asian mainland, with a few populations in Sri Lanka. Myanmar was previously thought to be the ancestral region of the α clade [25,26]. All β haplogroup Sumatran elephant populations belong to the β2 sub-clade, whereas the β1 sub-clade is distributed primarily in Sri Lanka and South and Central India, with a few populations in Myanmar [24,25,26]. A phylogenetic discontinuity of two distributions of mt D-loop haplogroups (α and β clades and β1 and β2 sub-clades) and their incomplete geographical partitioning has been previously discussed [25,26]. Three hypotheses have been put forth to explain the possible evolutionary circumstances: (i) introgression of mt D-loop sequences through hybridization with ancestrally related species; (ii) ancestral isolation in allopatry with secondary contact and admixture; or (iii) retention of divergent lineages owing to large population size and incomplete lineage sorting [17,19,25,26]. However, currently available haplotype information in public repositories, including data from Asian elephants derived from Thailand, shows a large diversity of β clades in addition to the β1 and β2 sub-clades; therefore, these three hypotheses may need revision. Nonetheless, here, we discuss the merits of each of these hypotheses.

The hypothesis of mt D-loop haplotype introgression through hybridization with a closely related species was proposed on the basis of evidence from ancestral elephant species, such as Elephas (Palaeoloxodon) namadicus (Falconerand and Cautley, 1846) [17,19]. E. namadicus frequently expanded its range in Asia during the Middle and Late Middle Pleistocene; however, the α and β clades split prior to the divergence of E. namadicus and E. maximus, approximately 0.15–0.70 MYA [25,37,38]. Therefore, the existence of a current mt D-loop haplotype gene pool in Asian elephants should be considered using other, more recent ancestral elephant species in the Late Pleistocene, such as E. hysudricus (Falconer and Cautley, 1845) [39], E. namadicus (Asian straight-tusked elephant, Falconer & Cautley, 1846) [40], and E. antiquus (straight-tusked elephant, Falconer & Cautley, 1847) [25,37,38,41,42]. On the basis of fossil evidence, the current Asian elephant is thought to have evolved from E. hysudricus. E. hysudricus is considered to have given rise to E. maximus, in which the α haplogroup originated [25]. By contrast, the emergence of the β clade remains controversial and may involve introgression. One hypothesis postulates that the introgression of the mt D-loop haplotype from E. antiquus gave rise to the β haplotype in today’s Asian elephants (Figure 1); however, fossils of E. antiquus have been recovered only in Europe. The time of divergence between E. antiquus and the current Asian elephant is estimated at 7 MYA [42], which is longer than the divergence time between the α and β clades at the beginning of the Pleistocene, approximately 1.5 MYA, and is long enough for introgression.

Alternatively, the second hypothesis postulates the allopatric distribution of the α clade on the mainland. Southern and Central India are home to the β1 sub-clade, concurring with its distribution in Sri Lanka (Figure 2). How the allopatric signals between Sri Lanka and India share the two haplogroups also needs to be considered. The interconnection between Sri Lanka and India would have allowed the gene flow of Asian elephants with the α and β1 sub-clades. The formation of a single landmass and land bridges during the Pleistocene, caused by glaciation events, may have prevented genetic differentiation between the two allopatric zones [25,26,44]. The coexistence of both the α and β clades in Sri Lanka and the predominance of the α clade on the mainland may be explained by the direction of gene flow from the larger mainland population to the smaller Sri Lankan population during more recent events of reconnection, dependent on climate alteration, reflecting the genetic admixture of haplogroups in the Sri Lankan population [25,26]. Considering again the first hypothesis, the mt D-loop haplotype from E. antiquus and the β haplogroup may have undergone introgression and genetic drift via the founder effect, resulting in the formation of the β haplogroup in the central and southern regions of India and Sri Lanka (Figure 2 and Figure 3).

The β clade of Asian elephants may have migrated from peninsular India and Sri Lanka to the eastern and southeastern regions of Asia during warm periods [25,26]. Unrelated haplotypes and admixtures of the α and β haplogroups were observed in Asian elephant populations in Myanmar and Thailand, possibly resulting from a northward range expansion of the β clade haplotypes from Sri Lanka, followed by subsequent admixture. Movements and recolonization of elephants on the Asian mainland and southern areas may have frequently occurred when warmer and wetter conditions returned as a response to climatic conditions [25,26]. Another parsimonious explanation involves the introgression of the β haplogroup in Central India, followed by migration to both the east and south during warmer conditions and the coexistence of both α and β haplogroups (Figure 2). The key distribution factor for both clades was climate alteration [45]; thus, the clinal distribution of the two clades appears to be a consequence of inhabiting different refugia, with range expansions during warmer interglacial periods leading to varying distributional overlaps. Two divergent clades of haplotypes may have expanded from glacial refugia in Central and Southern India and possibly from Peninsular Malaysia or Indochina. Different divergent clades of haplotypes have also been found in rhesus macaques (Macaca mulatta) [46] in Asia. Clades α and β1 in Thai Asian elephant populations have also been observed, similar to the Myanmar populations; however, the occurrence of established historical trade and human-assisted transfer of elephants in Asia cannot be ruled out [19,47]. Extensive sampling of different Asian regions may help resolve this puzzle. Additionally, it is unclear why only two sub-clades exist within the β clade. Notably, the β clade identified by Vidya et al. [25,26] was observed in the remaining Thai elephants, where haplotypes assigned to the β clade do not fall into the β1 or β2 sub-clades (Figure 1, Supplementary Figure S1); these haplotypes differ by 12 mutational steps from the β1 or β2 sub-clades, and, therefore, may be designated as ‘β3’ (see detail information of sequences of β3 in Table S1 and Ariyaraphong et al. [23]). Moreover, Borneo elephants contain the β1 sub-clade, even though this sub-species is located on Borneo Island in Southeast Asia.

Perhaps the rise of two clades or two (or three) sub-clades was a result of nothing more than chance within a single population of lineage retention. However, several bottlenecks or founder effects in the past make lineage retention an unlikely explanation for the coexistence of the α and β clades or the β1 or β2 subclades. A survey of cytochrome b variation in captive Asian elephants revealed eight haplotypes clustered into two groups with attributes of incomplete lineage sorting [24]. Incomplete lineage sorting or the retention of divergent lineages may result from a large population size [19]; however, this does not mean that large populations of Asian elephants currently exhibit high genetic diversity. The decline in Asian elephant numbers is far too recent to detect decreased heterozygosity. The average life span of an Asian elephant is 50 years, and its generation time is 15–40 years [23]. Thus, large sequence variations or new haplogroups may be underestimated, and more data are required, especially from Thailand. The Asian elephant is a social animal, historically used by humans for heavy laborious economic purposes [48]. These two aspects, as well as climatic conditions, may have influenced mt D-loop haplotype diversity. Here, we investigate how mutational dynamics of the mt D-loop haplotype compare with other genetic markers to infer Asian elephant genetic diversity and whether this would be a good representative genetic marker for Asian elephant study.

3. Influence of Social Behavior and Camp Management on Mitochondrial D-Loop Clustering Deviation

Alteration of climatic conditions in different refugia would primarily affect the distribution of the two haplogroup clades of Asian elephants across the Asian continent. Social behavior and human-assisted transfer (trade), together with camp management, may also influence directional gene flow in different regions. Asian elephants are extremely sociable and form matriarchal families of adult females and their offspring [18], and haplotype diversity among these individuals was low. A clan consists of several family groups that have shared a dry season in the past decades, and changes in the mt D-loop haplotypes of clan members are identical by descent [49,50,51]. Thus, a dominant hierarchy among clans may lead to the marginalization of subordinate clans, and some mitochondrial haplotypes may undergo extinction through a genetic selective sweep process [52]. By contrast, independent family groups occupying small home ranges may support diverse gene pools of the mt D-loop haplotypes. The natural movement habits of Asian elephants also imply a current pattern of genetic variation [53]; however, anthropogenic disturbances, such as deforestation and wildlife habitat loss, have indirectly increased human–elephant conflict, thus promoting elephant translocation within the last several decades [54].

The continuation of elephant translocation reduces genetic fitness among wild elephant populations and affects their social organization [55,56]. Human translocation or trade networks between elephant camps are other factors causing potentially high levels of gene flow among Asian elephant populations [4,57,58]. In Thailand, many captive bulls are subfertile; thus, only one or a few bulls are used for breeding in each camp, with occasional transfers of elephants between camps for breeding [59,60]. Domestication of elephants may also contribute to the dispersal of the two haplogroups. Elephants have been domesticated for thousands of years, and the historical importation and exportation of elephants to and from the Asian continent have been recorded [61,62]. Domestic elephant escapes have also occurred [63,64], and elephant translocations have likely contributed genes to wild populations, with major impacts on the mt D-loop haplotype distribution. A major conservation problem relates to the capture and trade of wild elephants from Myanmar to Thailand and within Myanmar, Sri Lanka, and India [19,65,66]. These smuggled elephants are sold to tourist camps throughout Thailand, particularly in Chiang Mai, Phuket, and Surin [62]. Origin and habitat differences affect genetic diversity in Asian elephants, and animals from other countries may have genetic differences [32,62,67,68,69].

4. Intraspecific Variation of Mitochondrial D-Loop in the Asian Elephant

The mitochondrial control region is a major non-coding segment of the vertebrate mitochondrial genome that contains a D-loop. Nucleotide substitutions and indels are more frequently observed in mt D-loop sequences compared with other non-coding or coding regions of the mitochondrial genome [70] Analysis of intraspecies sequence divergence revealed that genetic diversity values in Asian elephants are lower than those estimated for the mt D-loop sequences of many large mammals. In Asian elephants, the overall haplotype and nucleotide diversities were 0.915 ± 0.005 and 0.056 ± 0.027, respectively, in the mt D-loop sequences (

Table 1. Mitochondrial D-loop sequence diversity of Asian elephants (Elephas maximus, Linnaeus, 1758).

Country	N 1	Number of Haplotypes (H) 2	Overall Haplotype 2	Nucleotide Diversity (π) 2
Thailand	229	56	0.919 ± 0.008	0.066 ± 0.032
Laos	6	6	1.000 ± 0.096	0.060 ± 0.035
Myanmar	35	17	0.931 ± 0.022	0.072 ± 0.036
India	35	15	0.876 ± 0.042	0.046 ± 0.023
Indonesia	5	4	0.900 ± 0.161	0.012 ± 0.008
Malaysia	8	2	0.536 ± 0.123	0.045 ± 0.025
Sri Lanka	14	9	0.923 ± 0.050	0.033 ± 0.018
Bhutan	4	3	0.833 ± 0.222	0.022 ± 0.016
Vietnam	6	5	0.933 ± 0.122	0.041 ± 0.025
Asia 3	76	11	0.842 ± 0.020	0.036 ± 0.018
Unclassified	71	17	0.896 ± 0.017	0.034 ± 0.017
All populations	489	79	0.915 ± 0.005	0.056 ± 0.027

¹ Partial nucleotide mitochondrial D-loop sequences were retrieved from the public data repository. Detailed information regarding each nucleotide sequence is provided in Table S1. ² Genetic diversity of partial mitochondrial D-loop sequences was analyzed following Ariyaraphong et al. [23]. ³ These samples were collected in Asia without specific sampling location information [82].

). The genetic distances within haplogroups were 0.014 for the α haplogroup and 0.005 for the β haplogroup, while the genetic distance between haplogroups was 0.035 due to the presence of low levels of genetic variation that can occur through genetic drift, indicating small long-term effective population sizes, as predicted in Asian elephants. The control region can be divided into three domains: the extended termination associated sequences (ETAS) domain adjacent to the tRNA^Tyr gene, where synthesis of the heavy (H) strand pauses at the level of the ETAS; the central conserved domain; and the CSB domain adjacent to the tRNA^Phe gene, which contains the origin of the H strand replication, the two promoters, and the CSBs associated with initiation of H strand synthesis [71,72]. Our estimates of genetic diversity values in Asian elephants were high for CSB1, with approximately one-third of the variable sites between two haplogroups, similar to those of the snapping turtle (Chelydra serpentina, Linnaeus, 1758), Thai bow-fingered gecko (Cyrtodactylus peguensis, Boulenger, 1893), and European mink (Mustela lutreola) [73,74,75]. This region also contains a specific nucleotide that differentiates between the α and β clades. Thus, we predict a new haplogroup or sub-haplogroup with potential for mutation in the CSB1 region of the mt D-loop, as observed in the Thai Asian elephant population. Nuclear DNA markers and whole genome sequences of Asian elephant are required to further elucidate Asian elephant diversity, but scant library data are available for less than 250 Asian elephant individuals and only one origin (captivity) [76,77,78,79,80,81]. Therefore, it is still reasonable to use mt D-loop variations to determine Asian elephant ancestry congruent with climate change, especially given that a large mt D-loop dataset is available.

5. Asian Elephants and Sustainable Development Goals for Global Communities

One of the Sustainable Development Goals (SDGs) of the United Nations (UN) is the preservation of land ecosystems [83]. Conservation of the Asian elephant is important to attain this goal for several reasons. First, Asian elephants are a charismatic species that attract much public attention. Through this, public attention could be drawn to other less well-known wildlife conservation issues [84]. Second, being large animals, elephants require large protected areas for their populations to thrive. Large Asian elephant conservation reserves become an asylum for many other smaller plant and animal species, making their protection an important byproduct of elephant protection [85]. Keeping and breeding elephants in captivity is a very important conservation measure together with the creation of protected areas for wild Asian elephants. In Thailand, almost half of the existing Asian elephants are captive [23] and used to boost tourism, while some were illegally captured. Captive as well as wild elephant populations are experiencing inbreeding depression due to low animal numbers and genetic diversity [86]. To attain the SDGs, the elephant tourism industry in Asia must be reformed to eliminate elephant poaching and promote breeding and public education programs. Studying the genetic diversity and phylogeny of Asian elephants is crucial for the development of proper conservation strategies, both concerning the management of wild populations and development of sustainable breeding strategies that would reduce inbreeding pressure. Moreover, improved knowledge of elephant genogeography would identify the geographic origin of individual elephants and elephant derivates (such as ivory), thereby assisting in the mitigation of elephant poaching [87].

6. Final Considerations

The number of wild Asian elephants was estimated to be more than 200,000 as recently as 1900, but current estimates place the wild population at 48,323 to 51,680 [4,7]. Extirpated from approximately 85% of its historical range, the Asian elephant is now classified as ‘threatened’ [88] and exists in several fragmented and isolated populations in South and Southeast Asia. The large decline in the Asian population may be reflected by the evolutionary history of the two haplotype clades of mt D-loop sequences. Three hypotheses were evaluated by integrating mt D-loop data with fossil evidence. The introgression of mitochondrial haplotypes from other species through hybridization and/or allopatric divergence led to mutations and sequence divergence among geographically separated populations. However, how to predict the evolutionary direction and impact of the two clades is currently unclear. Changes occur gradually over geological time owing to the power of natural selection, but the fossil record may not be complete; thus, the entire spectrum of mt D-loop sequences should be examined in extant Asian elephants to better understand evolutionary relationships within this threatened species.