Royal Society Publishing

Low worldwide genetic diversity in the basking shark (Cetorhinus maximus)

A Rus Hoelzel, Mahmood S Shivji, Jennifer Magnussen, Malcolm P Francis


The basking shark (Cetorhinus maximus) is found in temperate waters throughout the world's oceans, and has been subjected to extensive exploitation in some regions. However, little is known about its current abundance and genetic status. Here, we investigate the diversity of the mitochondrial DNA control region among samples from the western North Atlantic, eastern North Atlantic, Mediterranean Sea, Indian Ocean and western Pacific. We find just six haplotypes defined by five variable sites, a comparatively low genetic diversity of π=0.0013 and no significant differentiation between ocean basins. We provide evidence for a bottleneck event within the Holocene, estimate an effective population size (Ne) that is low for a globally distributed species, and discuss the implications.


1. Introduction

The basking (Cetorhinus maximus) shark is a planktivorous species seasonally found in shoals nearshore, feeding near the surface. It is the second largest fish, sometimes exceeding 10 m in length and inhabits temperate regions in both hemispheres. Basking sharks mature slowly, requiring approximately 12–20 years and females have long gestation periods (approx. 1–3 years) after which they give birth to few offspring. These characteristics make this species especially vulnerable to over-exploitation (Compagno 2001). Basking sharks have been exploited for meat, fins, liver oil (Compagno 2001) and cartilage (Hoelzel 2001). They were listed by the IUCN as vulnerable worldwide (IUCN 2004), and in 2002 on Appendix II of the convention on international trade in endangered species. Despite the conservation concern for this species, there are few data on regional abundance, no estimates for abundance worldwide and no good data on population trends. The classification as ‘vulnerable’ is based largely on the rapid depletion of some populations subject to coastal harpoon fisheries, especially in the North Atlantic (Compagno 2001).

Here, we assess genetic diversity at the non-coding mtDNA control region and test the hypothesis that the species has been and remains sufficiently abundant to maintain levels of diversity comparable to those seen in other broadly distributed elasmobranch species. Expected levels of diversity depend in part on the substitution rate. Martin et al. (1992) investigating the mitochondrial Cytb and COI genes (coding loci) in elasmobranchs calculated a mutation rate that was 7–8 times slower than had been calculated for primates and ungulates.

However, recent studies have shown high levels of diversity at the mtDNA control region for some shark species. For example, a comparison of white sharks (Carcharodon carcharias) from Australian/New Zealand and South African waters showed a level of diversity that is comparable to other widely distributed, pelagic marine species (table 1), with substantial genetic differentiation between the two regions (FST=0.81, p<0.0001; Pardini et al. 2001). In contrast, we show that basking shark genetic diversity is exceptionally low worldwide.

View this table:
Table 1

Diversity at the mtDNA control region among pelagic marine vertebrate species. (Multiple geographical regions represented except as indicated; WNA, western North Atlantic; SA, South Africa; M, Mediterranean Sea. All studies based on the whole control region except those marked by *. Blank lines separate sharks, teleost fishes, loggerhead turtle and cetaceans. Species or populations proposed to have been through a population bottleneck marked by .)

2. Material and methods

Tissue samples were acquired from bycatch and strandings. The numbers of samples per region are given in table 2. DNA was extracted by standard phenol chloroform methods, and the full mtDNA control region amplified using primers set in the flanking tRNAThr and tRNAPhe genes, designed on the basis of aligned sequences available in GenBank. The PCR reaction mixture consisted of 0.20 μM each for the forward (5′-GACCTTGTAAGTCGAAGA) and reverse (5′-TCTTAGCATCTTCAGTGC) primers, 100 μM dNTPs, 1.5 mM MgCl2, 10 mM Tris–HCl pH 8.4, 50 mM KCl and 0.02 U (μl)−1 Taq polymerase. The PCR cycling profile was 5 min at 95 °C, 35 cycles of 45 s at 94 °C, 1.5 min at 48 °C and 1.5 min at 72 °C, followed by 8 min at 72 °C. PCR products were purified with QIAgen PCR purification columns and sequenced directly using the ABI dye-terminator method. All samples were sequenced in both directions using the amplification primers, and internal primers when necessary (forward: 5′-GCACATTACTCATCTCGACTACATCAC, 5′-GAAGCAATCGCTATCAATCGAA; reverse: 5′-CTGGTCAATTGGTGGGGATCAACCG, 5′-CGTTTATTGCGAATTTGTCCCCGGGG). Resulting sequences were aligned using ClustalX.

View this table:
Table 2

Haplotypes and their frequency from sampling areas: NZ, New Zealand; TW, Taiwan; NOR, Norway; SCO, Scotland; WNA, western North Atlantic; MED, Mediterranean Sea; CAR, Caribbean; SA, East coast of South Africa. (Position numbers are with reference to light-strand sequence along the 1085 bp sequences submitted to GenBank.)

Measures of haplotypic (h) and nucleotide (π) diversity and theta (θS) and differentiation (ϕST using the Kimura 2-parameter model and FST using haplotype frequencies) were calculated using Arlequin 2.000 (Schneider & Excoffier 1999). The Kimura 2-parameter model was chosen because the amount of variation was small, distributed across the locus and consisted of only transitions. Proportional haplotype frequencies were compared using a Fisher exact test (with a Markov chain of 10 000). Arlequin was also used to construct a minimum spanning tree (after Rohlf 1973), estimate Tajima's D (Tajima 1989) and construct a mismatch distribution (Rogers & Harpending 1992). The latter two analyses can help determine if a population has undergone a rapid expansion (possibly as a result of a population bottleneck). Tajima's D-tests for departure from mutation–drift equilibrium. The mismatch distribution will be multi-modal in stable populations (if the generation of new mutations is offset by random drift), and unimodal for expanding populations (if new mutations are accumulated faster than loss due to drift). The time of a possible population expansion (t) can be calculated using the formulation: τ=2ut (Rogers & Harpending 1992), where τ is the mode of the mismatch distribution and u is the mutation rate of the sequence (such that u=2μk, where μ is the mutation rate per nucleotide and k is the number of nucleotides). The time (t) is measured in generations.

For basking sharks, neither the generation time nor the mtDNA control region mutation rate is precisely known. However, data from Pauly (1978) suggests a generation time of about 16 years. Donaldson & Wilson (1999) estimate an average control region mutation rate of 3.6% per million years for a range of fish species, while the Martin et al. (1992) data would imply roughly 2.7% and Duncan et al. (2006) have recently suggested 0.8% based on population isolation in one shark species. We use an average of these three (2.4%).

3. Results and discussion

In a preliminary study, Hoelzel (2001) compared basking shark populations in the North Atlantic (N=11) and South Pacific (N=6) based on sequence data from 550 bp of the mtDNA Cytb locus. There were two Cytb haplotypes, and their frequency did not differ between regions. Here, we investigate the complete mtDNA control region from a much larger set of globally distributed animals. Although the results show more haplotypes (6) compared to Cytb, the overall level of diversity remains remarkably low. The haplotypic and nucleotide diversity values are given for the full sample (N=62) in table 1 and compared with values for various other globally widespread, pelagic marine species (including elasmobranchs, teleosts, a sea turtle and mammals). Genetic diversity values for basking sharks in each ocean basin considered separately were very similar to the worldwide values (Pacific: h=0.7344±0.0418, π=0.0013±0.0009, N=34; Atlantic: h=0.7169±0.0495, π=0.0014±0.0009, N=27).

Despite the global distribution of the samples, there were only five variable sites among the six haplotypes and 1085 bp in the sequence. The minimum spanning tree (figure 1) showed no evidence for structure and was dominated by two haplotypes (found in similar frequencies in each ocean basin; table 2). By contrast, the white shark (C. carcharias) had 77 variable sites defining 29 haplotypes for the 1149 bp control region sequence among a sample of 88 sharks (Pardini et al. 2001). Similarly, a study of young blacktip sharks (Carcharhinus limbatus) over a much smaller geographical range (part of the WNA) had 15 variable sites defining 23 haplotypes (N=323; 1067–1070 bp), with high genetic differentiation (ϕST=0.35, p<0.001) between sampled areas (Keeney et al. 2005).

Figure 1

Minimum spanning network for the six haplotypes. Size of circle reflects relative frequency.

There was no difference between Pacific and Atlantic basking shark samples grouped as putative populations (non-significant and negative ϕST and FST and non-significant Fisher's exact test: p=0.85). Local sample sizes were small in the North Atlantic, Mediterranean, and Indian Ocean and possible further population substructure could not be assessed. However, the differences among all haplotypes were very small (table 2). The lack of structure may suggest that a bottleneck event preceded expansion into the current distributional range. Alternatively, it could suggest female mediated gene flow over a wide geographical range, but this would not account for the very low diversity levels.

Tajima's D was not significantly different from random expectations, but the mismatch distribution was unimodal (figure 2). Estimated τ=0.56, and given our assumed parameters for generation time and mutation rate, this would suggest a population expansion beginning approximately 86 000 years ago, well before the influence of modern fisheries. The lack of transversion mutations is consistent with a bottleneck, and with this timeframe. Uncertainty about parameter estimates means that correlation with any specific event is not possible. However, events such as the warming periods preceding Holocene interglacials can theoretically cause considerable disruption to oceanic systems, and therefore to the resources that basking sharks depend on, and plausibly lead to a population bottleneck. Selective sweeps can also reduce diversity, but it is less likely for this to result in the same common haplotypes occurring in distant populations.

Figure 2

Mismatch distribution with the observed values as bars and a solid line representing the expected distribution according to the sudden expansion model.

The census number of basking sharks in the world's oceans is unknown. However, our data suggest an Ne of 8200, calculated from estimated Theta (θS=0.852±0.467; θ=2Neu). Again, given the underlying uncertainties in the parameter estimates, it is necessarily a rough approximation. Even so, given the global distribution of the species, it is surprisingly low. In a meta-analysis, Frankham (1995) identified a median ratio of Ne to census population size (Nc) of 10%, though the range was very broad. However, historical bottlenecks can depress Ne/Nc, especially in rapidly expanding populations (e.g. this has been proposed to explain a human Ne of 10 000; e.g. Harpending et al. 1993).

Although diversity at this locus is low for some other elasmobranch species assessed in local populations, or following known bottlenecks, none have so far shown such low diversity worldwide (see table 1). We suggest that, the most parsimonious explanation is that the basking shark suffered a population bottleneck affecting global diversity (cf. similar data for the killer whale (Orcinus orca); Hoelzel et al. 2002), and propose that the low diversity revealed should be a key consideration for the development of conservation and management strategies for this species, on the assumption that further loss of diversity could affect evolutionary potential.


This study was supported in part by DETR, the Pew Institute for Ocean Science, the Florida Sea Grant Program and the Guy Harvey Research Institute. We thank M. Affronte, R. Baird, L. Boren, J. Cassin, G. Cliff, T. Knott, D. Mattila, J. Morrisey, L. Natanson, Mark O'Connell, M. Preide, S. Fowler, T. Thom, R. Torres, S. Wintner and New Zealand Ministry of Fisheries observers for help with the acquisition of basking shark samples.


    • Received April 26, 2006.
    • Accepted June 5, 2006.


View Abstract