京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第22282号 / 理博第4596号 / 新制||理||1659(附属図書館) / 京都大学大学院理学研究科生物科学専攻 / (主査)教授 朝倉 彰, 准教授 下村 通誉, 教授 曽田 貞滋 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

It is often hypothesised that marine species with mobile planktonic phases are capable of widespread dispersal and may therefore be genetically homogenous throughout their distribution. Studies that have demonstrated positive correlation between duration of plankton phase and levels of gene flow reinforce the prediction that life history characteristics of marine species determine the potential extent of genetic and demographic connectivity throughout their distributions. This prediction has however been challenged by studies that have employed genetic markers highly sensitive to both historical and contemporary demographic changes. Disparities between dispersal potential and measured levels of gene flow have been demonstrated both among historically disconnected ocean basins and within semi-enclosed areas of strong hydraulic connectivity. These studies and others highlight a need for greater focus on factors that may influence population structure and distribution for marine species. In this thesis, I have examined genetic structure within and among populations of an estuarine species of mud crab Scylla serrata (Forskal, 1775) using a number of genetic markers and methods. The species is widely distributed throughout mangrove and estuarine habitats of the Indo - West Pacific (IWP); it is generally assumed that life-history characteristics of S. serrata promote high levels of population admixture and gene flow throughout its distribution. Alternatively, factors that have promoted population genetic structure for a variety of IWP marine species may also have affected S. serrata populations. By investigating genetic structure at several spatial scales of sampling, I was able to address a variety of hypotheses concerning the species distribution, dispersal, and genetic structure. Episodic changes to marine habitat and conditions experienced within the IWP during the Pleistocene may have affected genetic structure for a broad variety of marine taxa. The relative strength of this hypothesis may be assayed by comparative genetic studies of widespread IWP taxa with high dispersal capacity. In order to ascertain levels of historical and contemporary gene flow for S. serrata, I investigated the phylogeographic distribution of mitochondrial DNA haplotypes sampled throughout the species range. Adults were sampled from three west Indian ocean locations (N=21), six west Pacific sites (N=68), and two sites from northern eastern Australia (N=35). Temperature gradient gel electrophoresis and sequencing of 549 base pairs of the mitochondrial cytochrome oxidase I (COI) coding gene identified 18 distinct haplotypes. Apart from that seen in northern Australia, haplotype diversity was low (h < 0.36) at each of the locations. Total nucleotide diversity in the entire sample (excluding northern Australian locations) was also low (p = 0.09). Haplotypes clustered into two clades separated by approximately 2% sequence divergence. One clade was widespread throughout the IWP (clade 1) whereas the other was strictly confined to northern Australia (clade 2). Genealogical assessment of sequenced haplotypes relative to their distributions suggested that a historical radiation of clade 1 S. serrata throughout the IWP occurred rapidly and recently (<1Myr bp) from a west Pacific origin. The evidence of fixed unique haplotypes at the majority of locations suggested that contemporary maternal gene flow between trans-oceanic sites was limited. Contrary to reports for other widespread species of IWP taxa, there was no evidence of lengthy periods of regional separation between Indian from Pacific Ocean populations. However, results may indicate a separation of northern Australian crabs from other locations before and during the IWP radiation. I speculated that this isolation might have resulted in the formation of a new species of Scylla. Additional sampling of mud crabs from the Australian coastline allowed an examination of the diversity and distribution of clade 1 and 2 haplotypes among recently formed shelf-connected coastal locations, and across a historical bio-geographic barrier. Over 300 individuals were sampled from multiple locations within coastal regions (western, northern and eastern) of Australia and analysed for mutational differences at the COI gene. Analysis of molecular variance partitioned by sampling scale (Among regions, within regions, and within all locations) indicated mitochondrial haplotypes were structured regionally (P < 0.001), which contrasted with evidence of genetic panmixia within regions. Regional genetic structure broadly correlated with hydrological circulation, supporting the contention that release and transport of propagules away from the estuary may allow genetic connectivity among widespread shelf-connected S. serrata populations. That similar patterns of maternal gene flow were absent among trans-oceanic populations may indicate that the spatial scale of effective dispersal for this species is generally limited to areas of coastal shelf. The two clades of haplotypes were geographically separated either side of the Torres Strait, a narrow sea channel connecting the northern and eastern regions of coastal Australia. This pattern of historical genetic separation was concordant with a number of other marine species across northern Australia, and might indicate a shared history of vicariance induced by eustasy. Alternatively, differences in diversity and distribution of the clades may be evidence of two independent expansions of clade 1 and 2 crab populations into Australian regions following post-glacial estuary formation. Overall, despite evidence of genetic panmixia within extensive sections of the Australian distribution, there was also evidence of significant barriers to maternal gene flow with both shallow and deep regional phylogeographic assortment of mtDNA haplotypes. The presence of these barriers indicated both historical and contemporary factors have imposed limits to effective dispersal by this species among coastal habitats. A subset of the Australian sample (8 locations, N = 188) was also examined for variation at five microsatellite loci developed specifically here for S. serrata. I examined variation among samples at each of the loci to: a) independently verify regional structure among crab populations previously detected using the mtDNA analysis; b) test for evidence of co-distributed non-interbreeding stocks of S. serrata within Australian waters by examining samples for segregation of alleles within microsatellite loci concordant with the two mtDNA clades. The frequency and distribution of alleles for each of the highly polymorphic microsatellites were homogenous at all levels of sample partitioning and contrasted sharply with the instances of both weak and strong regional phylogeographic assortment of mtDNA haplotypes. These contrasting results between different genomic markers were examined in relation to the species life history, and to differences in mutational rate and inheritance of the genetic markers. Several hypotheses may explain the disparity, however it is most likely that rampant homoplasy and high rates of mutation at the microsatellite in conjunction with large Ne at locations may be concerted to delay equilibrium between genetic drift and migration among populations at these highly polymorphic nuclear markers. There was also no evidence that alleles at microsatellite loci were co-segregated with mtDNA clades and therefore no evidence of segregated breeding between the clades of crabs. Whether or not this result was also driven by homoplasy at the microsatellites remains unknown. Recently established mud crab populations (~ 3-4 years old) observed in a number of southwest Australian estuaries are almost 1000 kilometres south of their previously recorded distribution on the Western Australian coast. Colonisation of the southwest region may have occurred either by a natural range expansion from northwest Australian mud crab populations or by means of translocation from any number of mud crab sources within the Indo - West Pacific. I used mtDNA analysis to verify the species and determine the potential source population(s) of the colonists, by comparing sampled genetic material from the southwest (N = 32) against that previously described for the genus. I also compared levels of diversity at mtDNA and two microsatellite loci between the colonist and suspected source population(s) to qualitatively determine if the southwest populations experienced reductions in genetic diversity as a result of the colonisation process. All colonist samples had S. serrata mtDNA COI sequences identical to one previously described as both prevalent and endemic to northwest Australia. High levels of genetic diversity among source and colonist populations at two microsatellite loci contrasted to the mitochondrial locus which displayed an absence of variation among colonists compared to moderately diverse source populations. I argued that the southwest was recently colonised by large numbers of S. serrata propagules derived from the northwest of Australia, possibly due to an enhanced recruitment event coinciding with the reported strengthening of the Leeuwin Current during 1999. Contrasting levels of diversity among nuclear and mitochondrial loci may be attributed to a difference in response by the two genomes to the colonisation process. I predict that such differences may be generally prevalent among plankton-dispersed species. Finally, I discuss aspects of the species distribution and biogeography obtained as a composite of the various results and ideas expressed in this thesis. I propose that S. serrata populations in the IWP may have experienced several cycles of extinction and population retraction from temperate areas followed by subsequent periods of colonisation and rapid coastal expansion in response to the effects of glacial episodes on coastal habitats in the IWP. I propose that persistence of this species as remnant populations of clade 1 and 2 crabs at equatorial locations during low sea level stands provided source populations for later expansions by the species into a variety of coastal areas throughout the IWP. Further analysis is required to determine if mtDNA clade 1 and 2 crabs are non-interbreeding species of mud crab.

ABSTRACT The zebra shark Stegostoma fasciatum is an oviparous, demersal carpet shark that forms temporary aggregations and is distributed within shallow, coastal, subtropical and tropical waters of the Indo-West Pacific (IWP) region. The IWP region encompasses various densities of human habitation leading to differences in the levels of exploitation of chondrichthyan fishes, which are targeted for local consumption and export. This is reflected in the two current regional IUCN Red List classifications for zebra sharks of ‘Least Concern’ and ‘Vulnerable’. Despite the conservation concerns for zebra sharks in the vulnerable parts of their range, as well as importance within the ecotourism industry and collection for aquarium trade, little is known about this species in the wild. The aim of this thesis was to conduct the first intensive examination of population ecology in the zebra shark by: (i) investigating the population structure and environmental parameters associated with a relatively unexploited aggregation of zebra sharks in south-east Queensland, Australia; and (ii) investigating the genetic continuity between zebra shark populations in areas with differing levels of exploitation. To investigate the size and structure of the aggregation at the site The Group in south-east Queensland, photo-identification and mark-recapture methods were employed over a three year period (2003 - 2006). In total 327 individual zebra sharks were identified from 570 photographs. Numbered dart tags on 15 zebra sharks were used to confirm pigmentation patterns were unique and persistent in wild zebra sharks for up to 810 days. Pollock’s robust design resulted in an annual population estimate of 458 individuals (95% CI = 298 – 618). The mean number of zebra sharks observed on a single day was 8 (± 8 SE) and the maximum number of zebra sharks seen on a single day was 34. In total, 27% of the sharks were sighted in more than one summer aggregation period and males had greater re-capture probabilities than females. The aggregation consisted exclusively of large (>1800 mm total length) adults with an overall female sex bias of 3.8:1, though sex-ratios varied temporally. Passive acoustic telemetry techniques were employed to investigate the seasonality and residency of the sharks and the environmental parameters associated with their presence at the study site. Five VR2 acoustic receivers were deployed for a period of 21 months (December 2004 – August 2006) and recorded ten acoustically-tagged zebra sharks (seven females and three males). The sharks showed clear seasonality in their visitation patterns with greatest presence during the austral summer months of November through to February and no presence at all during the winter months from June through to October. However their presence did not coincide with peak water temperatures. Rather there appeared to be a critical minimum water temperature of approximately 22 ºC, above which zebra sharks would be present. The tagged zebra sharks displayed greater visitation during the calmest sea conditions. The sharks further displayed diel periodicity with greater during daylight hours, and may be using The Group as a resting site during the inactive diurnal period of the day. Complementing the ecological study on the aggregation, the population genetic structure of 180 zebra sharks from 13 locations throughout the IWP was investigated to test the concordance of IUCN zones with demographic units that have conservation value. A zebra shark specific microsatellite library was developed consisting of 14 polymorphic loci of which 13 were employed for the population genetic analysis along with mitochondrial ND4 sequence data. Mitochondrial and microsatellite data-sets from samples collected throughout northern Australia and south-east Asia concord with the regional IUCN classifications. However, evidence of genetic subdivision was found within these regions, including subdivision between locations connected by habitat suitable for migration. Further, parametric FST analyses and Bayesian clustering analyses indicated the primary genetic break within the IWP is not represented by the IUCN classifications but rather is congruent with the Indonesian throughflow current. This study has demonstrated that zebra sharks form a large, predictable, shallow coastal aggregation comprised entirely of mature adults as well as display high genetic subdivision in the Indo-West Pacific region with little to no recruitment into over-exploited populations from healthy populations. These findings indicate that, if zebra sharks form similar aggregations as that which occurs at The Group in regions of high fishing pressure then, without sufficient protection and management localised depletion and extirpation will eventuate for zebra shark populations in heavily-fished areas.