Population structure and differentiation of the living elephants

The population structure of the African savanna elephant Loxodonta africana africana across eastern and southern Africa was first studied by a team led by molecular geneticist Nicholas Georgiadis. The team used a combination of blood, tiny amounts of skin, and tissue scrapings from tusks representing 270 elephants from five countries (Kenya, Tanzania, Zimbabwe, Botswana, and South Africa) for their analysis. A segment of mtDNA of 2,450 nucleotide base pairs spanning the ND 5-6 region was examined. Using several restriction enzymes, which cleave the DNA molecule at specific sites, they looked at variation in base sequences across this region of the mtDNA. They mapped 47 restriction sites; of these, 23 sites showed polymorphism or variation. There were 10 haplotypes at this segment of mtDNA among the elephants examined in this study (table 1.3).

Several unexpected patterns emerged from the mtDNA analyses. The two most common haplotypes (numbers 3 and 4) occurred throughout the range of countries sampled at varying frequencies. This clearly argued for protracted gene flow across the eastern and southern African elephant populations in the past. The other eight haplotypes were more localized, with some of them occurring only in eastern African populations and others in Zimbabwe and Botswana. Interestingly, the only two haplotypes found in the relatively isolated Kruger population of South Africa were the two most common ones. Thus, there was a marked differentiation of populations at the continental

Table 1.3

Frequency of mitochondrial haplotypes at ten locations in eastern and southern Africa.

Haplotype and Frequency (%) Sample -

Table 1.3

Frequency of mitochondrial haplotypes at ten locations in eastern and southern Africa.

Haplotype and Frequency (%) Sample -

Country

Location

Size

1

2

3

4

5

6

7

8

9

10

Kenya

Amboseli

29

37.8

3.5

55.2

3.5

Kenya

Tsavo

14

57.1

14.3

14.3

14.3

Tanzania

Tarangire

35

17.1

77.1

5.7

Zimbabwe

Sengwa

22

50.0

36.4

13.6

Zimbabwe

Hwange

24

12.5

29.2

29.2

25.0

4.2

Zimbabwe

Zambezi

28

14.3

3.6

3.7

75.0

3.6

Botswana

Chobe

40

50.0

20.0

30.0

Botswana

Savute

27

7.4

67.7

22.2

3.7

Botswana

Mashatu

17

11.8

88.2

South Africa

Kruger

34

26.5

73.5

Source: From Georgiadis et al. (1994). Reproduced with the permission of Nature Publishing Group, U.K.

Source: From Georgiadis et al. (1994). Reproduced with the permission of Nature Publishing Group, U.K.

scale, but no such differentiation was apparent at the regional scale of eastern or southern Africa.

The elephants sampled also separated into two exceptionally divergent mitochondrial "clades." Haplotypes representing both these divergent clades coexisted at the same location, not in just one location, but at several distant locations. This is indeed a puzzling result as the two major clades are thought to have diverged 4 My ago (in hindsight, probably an overestimate); this idea is based on standard calculations of evolutionary rates using mtDNA sequence divergence.

In other species, there are rarely such widely separated mitochondrial haplotypes coexisting in the manner seen in the African savanna elephant. One explanation for such a pattern is that populations were allopatric or reproduc-tively isolated at some distinct ancestral time and later came into contact. There is no evidence as to whether this occurred in L. a. africana. Another explanation, a more parsimonious one, is that these representatives of such differing mitochondrial clades have persisted by chance alone over the indicated evolutionary timescale of 4 My. Given the absence of any barrier to gene flow, these divergent haplotypes now coexist at distant locations. Calculations based on neutral coalescent theory indicated that the observed pattern of coexistence of divergent haplotypes could have occurred by chance alone. Given the long generation time of 20 years in elephants, a genetically effective population size of over 50,000 females, maintained by gene flow, needed for this chance persistence of a subdivided, but not isolated, population over 4 My was entirely plausible.

Another study of African elephant genetics, on a more localized scale, is that of Silvester Nyakaana and Peter Arctander. The savanna elephant population of Uganda had declined precipitously, from about 60,000 in the 1960s to only 1,200-2,000 by the 1990s (chapter 8). Nyakaana and Arctander sampled 72 elephants at three locations to look at the genetic patterns using both mtDNA (a 400-bp [base pair] segment of the control region) and nuclear DNA (microsatellite analysis of four variable loci). They distinguished 11 mtDNA haplotypes, of which 3 were represented by only one individual each. Of the remaining 8 haplotypes represented by multiple individuals, 6 were specific to certain localities, suggesting considerable genetic differentiation. Sequence divergence among the mitochondrial haplotypes varied from 0.25% (representing one substitution) to as much as 6.1% (representing 23 substitutions). The nuclear DNA also showed fairly high variation in the four microsatellite loci examined as well as differentiation among the populations, although to a lesser extent than the mtDNA, suggesting that gene flow was maintained largely by the males.

Genetic studies of Asian elephant populations also have generally kept pace with those of their African counterparts. A team led by geneticist Gu nther Hartl amplified mtDNA from hair follicles obtained from 53 elephants originating in Sri Lanka, southern and northeastern India, northern and southern Myanmar, northern and eastern Thailand, and Vietnam. A 480-bp sequence of the cyt b gene was screened for variation. Based on a 335-bp scored sequence of this gene, they found eight mitochondrial haplotypes. As in the continental African study, they found two common haplotypes, representing 18 and 12 individuals each of the 53 screened, distributed throughout the continent. The other six haplotypes had more restricted distributions. The eight haplotypes also clustered into two major clades, with one difference from the African savanna elephant: The two common haplotypes each went into a different clade.

A genetic study by Prithiviraj Fernando and associates was novel in that it used the dung of 118 elephants, mostly free ranging, in Sri Lanka, Bhutan-India, and Laos-Vietnam to extract DNA. They amplified a 630-nucleotide-bp sequence of mtDNA encompassing the highly variable control region as well as adjoining segments, including a part of the cyt b gene. They distinguished 17 mitochondrial haplotypes across these samples; these segregated into two assemblages or clades, one labeled a with 7 haplotypes and another labeled P with 10 haplotypes (fig. 1.9).

Two of these haplotypes, representing the most common haplotype in each of the two assemblages, were also the only haplotypes that were shared by elephant populations both in Sri Lanka and in mainland Asia. As many as 12 haplotypes were found within Sri Lanka, with high differentiation across the northern, central, and southern parts of the island. In comparison, fewer haplotypes were found in the mainland, although the data have to be interpreted cautiously because far fewer elephants were sampled there. The average divergence in the nucleotide sequences between the two haplotype assemblages

Figure 1.9

Parsimony network of Asian elephant haplotypes (letters) based on mitochondrial DNA analyses. Circles without a letter denote haplotypes assumed, but not observed, in this study. Shaded circles represent Asian mainland haplotypes; striped circles indicate those shared between the mainland and Sri Lanka; and open circles indicate those limited to Sri Lanka. Dashes between haplotypes represent mutational steps between haplotypes. Asterisks mark assumed instances of homoplasy. (From Fernando et al. 2000. Reproduced with the permission of Nature Publishing Group, U.K.)

Figure 1.9

Parsimony network of Asian elephant haplotypes (letters) based on mitochondrial DNA analyses. Circles without a letter denote haplotypes assumed, but not observed, in this study. Shaded circles represent Asian mainland haplotypes; striped circles indicate those shared between the mainland and Sri Lanka; and open circles indicate those limited to Sri Lanka. Dashes between haplotypes represent mutational steps between haplotypes. Asterisks mark assumed instances of homoplasy. (From Fernando et al. 2000. Reproduced with the permission of Nature Publishing Group, U.K.)

was 3.1%, about half the divergence between the Asian and the African elephant.

Using an estimated long-term sequence divergence rate of about 1% per million years, Fernando and associates suggested that the two clades had diverged about 3 My ago. Although there are several possible explanations for the observed patterns of mtDNA structure, they favored allopatric subspecia-tion of the two lineages, in Sri Lanka and the mainland, with subsequent migrations during the Pleistocene glaciations, when sea levels were lower, which would have resulted in regional mixing and coexistence of haplotypes.

A combination of cyt b and control region sequence was also used by Robert Fleischer and associates in their genetic survey of Asian elephant popu-

lations based on 57 captive animals of known origin. Their samples also separated into two major clades, with a broad north-to-south gradient across the continent. All the Sumatran and Malaysian elephants grouped into only one clade (termed the Indonesian clade), while elephants from other regions further north and west grouped into either clade. They speculated that the Indonesian clade haplotypes could have descended from Elephas hysudrindicus, the extinct Pleistocene species with fossils that are known from Java, while the other clade may have originated from the northern E. hysudricus, known from fossils in the Siwaliks in the Indian subcontinent. They also stressed that the historical trade in Asian elephants and subsequent escape of captive animals into the wild could have shaped present-day population genetic structure, in addition to natural dispersal facilitated by Pleistocene sea level fluctuations.

The distinct character of the Sumatran and Malaysian elephants is also supported by the more recent work of Prithviraj Fernando and Don Melnick. Their preliminary work on the elephants of Sabah, on the island of Borneo, has also shown a distinct mt haplotype not seen anywhere else in Asia (this finding has an important bearing on whether the elephants of Borneo are native or have descended from captive stocks presented to the Sultan of Sulu in A.D. 1750 and later released; the evidence now points to the former explanation but more data are needed). The elephants of the Indian subcontinent have been poorly represented in all these genetic studies. T.N.C. Vidya and I have been collaborating with Fernando and Melnick to rectify this deficiency. One interesting result that has emerged is that the southern Indian populations are represented by few mt haplotypes. In fact, the largest global population of Asian elephants (estimated at about 8,000 individuals or over 15% of total Elephas maximus in the wild), which ranges over the Nilgiri-Eastern Ghats region there, is represented by a single mt haplotype. The southern Indian mt haplotypes are also found in Sri Lankan elephants.

0 0

Post a comment