Population Genetic Structure of Tropical eel Anguilla bicolor in Indonesia Water
habitat or both, during their life cycle Aoyama 2009, questionning the origin of species distinction.
Most of freshwater eels are considered to form randomly mating panmictic populations. The mechanism producing panmixia is that individual from different
area spawn in common area in the ocean Avise et al. 1986. Two or more species have been regarded as completely panmictic, each comprising a single
and randomly mating population A. anguilla and A. rostrata as temperate species; A. borbneensis at least, as tropical species. However, recent evidence
from studies of microsatellite variation of European eel has challenged this model Wirth and Bernatchez, 2001; Daemen et al. 2001; Maes Volckaert
2002. Wirth and Bernatchez 2001 found that populations of European eel show slight but significant, genetic differentiation and genetic distance between
population, this was interpreted as geographical heterogeneity and geographic distance. This implies restricted gene flow and lack of random mating between
eels from different population. But this hypothesis has then been contreadicted by considering that the observed disequilibrium should be due to temporal
differences between samples Dannewitz et al. 2005 ; Maes et al. 2006; Pujolar et al. 2006. This question is still discussed.
Ege 1939 has described A. bicolor as composed of two subspecies, A. bicolor bicolor in the Indian Ocean and A. bicolor pacifica in the western Pacific
Ocean and the sea around the northern part of Indonesia. Allopatric distribution has led to differences in the morphology of both subspecies, Ege 1939 noticed
slight differences in their means and ranges of total number of vertebrae. Indonesia is an important biogeographic crossroad for ichthyofauna, the
western part of Indonesia is connected with Indian Ocean and the eastern one was connected with the Pacific Ocean. Concerning the widespread distribution of
A. bicolor, this exceptional geographic distribution and structure made this species an important model for eel biology understanding. While investigated at
its whole distribution scale Minegishi et al. 2011, nothing is known in Indonesian waters, such as the exact distribution and lineages differentiation
among Indonesian rivers. Beside, this species was also an important commercial fish because of its distribution and abundance.
In recent years, scientists have begun using microsatellites as a way to study within species evolution. Microsatellites are widely used as markers for
studying genetic mapping, population structures and evolutionary genetics Whirt 55
and Bernatchez 2001. This aims at revealing the genetic connectivity and population structure of shortfinned A. bicolor in Indonesian waters using
microsatellites.
Material and Methods
Specimens A total of 180 specimens of A. b. bicolor were collected from ten sampling
stations along Indonesian waters that represent the sea connecting Pacific Ocean and Indian Ocean. As follow, were considered here: Ace 24 individuals,
Padang 24, Bali 24, Pelabuhan Ratu 24, Obi 7, Poigar 4, Donggala 18, Pangandaran 21, Bengalon 11 and Bengkulu 23 Figure 9 and Table 14. All
specimens were collected between 2008 and 2011. Total genomic DNA was extracted by using gSYNC Mini Kit Tissue
Geneaid. The protocol, according to the manufacturer’s recommendations, is as follow: approximately 25 mg of eel tissue were incubated with lysis buffer and
Proteinase-K at 60
o
C for 1 hour or until lyses was completed, and then incubated with a second lysis buffer at 70
o
C. The next steps were binding DNA with ethanol and separate DNA from undesirable material by using GD column. The last step
was washing the DNA twice by different wash buffers. The genomic DNA was eluted by Nuclease Free-Water. The quantity of DNA was checked by
electrophoresis.
Figure 9 Sampling location of “inter-ocean” Indian Ocean with black circles and Pacific Ocean red circle
56
Table 14 List of specimen used for microsatellite analyses in the present study Sampling Location
Code Sampling date
No
specimen
Indian Ocean
Estuary of Tadu, Aceh Ace
January 2010 24
Estuary of Bungus River, Padang Pad
August 2008 24
Estuary of Ketaun River, Bengkulu, Sumatera Ben
December 2009
23 Estuary of Cimadiri, Pelabuhan Ratu, Java
Pel August 2009
24 Estuary of Pangandaran, Java
Pang June 2009
21 Estuary of Mengereng River, Bali
Bal June 2009
24
Pacific Ocean
Estuary of Bengalon River, Borneo Beng
October 2010 11
Estuary of Donggala River, Celebes Dong
October 2010 18
Obi Island, Maluku Obi
November 2011 7
Estuary of Poigar, North Celebes Poig
January 2012 4
Total specimens 180
Microsatellite genotyping Eight microsatellite loci, which have been published for other anguillid
species, were genotyped as follow Table 15. Polymerase chain reactions PCR were performed on a Mastercycler Eppendorf, Hamburg, Germany in a
10 µl reaction volume containing 1,5 mM MgCl
2
, 0,2 mM of each dNTP, 1µ l 10x PCR buffer Promega, Fitchurg WI, USA, 0,5 µM of forward and reverse
fluorescent-labelled primers, 0,2 units of Taq polymerase, and 2,0 µl of template DNA. Amplification parameters were 35 cycles of denaturation at 95
o
C for 30 s, annealing for 30 s see Table 2 for annealing temperatures and extension at 72
o
C for 30 s after heating at 95
o
C for 30 s. Labelled PCR product were electrophoresed with a DNA ladder Promega on an 8,0 acrylamide gel, which
was later scanned on a FMBIO II Multi-View Hitachi, Tokyo, Japan scanner. Fragment size were determined using the software FMBIO Analysis 8,0 Hitachi.
Each molecue size produced by the software is then assimilated to an allele after checking the gel by eyes
Statistical Analyses Most statistical analyses were performed using the program GENETIX
4.05 Belkhir et al. 1998. The software GENETIX 4.05 was used for two types of classical parameters calculations: i calculation of the number of alleles, allele
57
58 frequencies, observed heterozygosity Ho, expected heterozygosity He for
each locus in each locality for all samples, ii estimation of Fis Wright 1965 and its statistical significance following Weir and Cockerham 1984 and iii the
linkage disequilibrium among the loci analyses for all samples in this study significance estimated with 5000 permutation test. Then calculation of the exact
test deviations from Hardy-Weinberg equilibrium HWE conducted under POPGENE 32.
The genetic differentiations between localities and Oceans were characterized using hierarchical F-statistic theta Fst of Wright and genetic
distance characterized under Nei 1972 as implemented in POPGENE 32. The presence of intraspecific genetic structure was tested using the model-based
clustering method of Pritchard et al. 2000 implemented in STRUCTURE VERSION 2.1 for each value of K, which is the number of genetically distinct
populations. The Markov chain Monte Carlo scheme was run with a burn-in period of 5000 steps and a chain length of 50000 replicates following the
admixture model.
Result
Eight pairs of primer were tested and seven of them gave easily interpretable and highly polymorphic patterns. The PCRs were carried out using
the annealing temperature proposed in the original paper for each pair of primers. Alleles frequencies at each locus are presented in Annex 4. The linkage
disequilibrium LD allow to detect non-random associations of alleles at two or more loci. LD was calculated only on samples of more than 15 specimens, three
populations Obi, Poigar and Bengalon which less than 15 specimens were not included in this analyse. There is no linkage disequilibrium between genotypes of
the seven microsatellite loci among seven populations, indicating that all loci are independent. Linkage disequilibria were observed at p0,05 for seven pairs of
loci. A total 172 alleles were observed across seven microsatellite loci with a
number by locus ranging from 3 in AjTR 12 to 22 in Aro63 for all populations. The population parameters at the 7 microsatellite loci is presented in Table 16.
Expected He and observed Ho heterozygosities of each locus range from 0,594 to 0,921 and from 0,250 to 1,000 respectively. Significant deviation from
Table 15. List locus that using during this study
Locus Name
Repeat motifs
Primer Name
Primer sequence Range bp
Primer labeling
Anneling temperature
oC Reference
FAro054 CTCAACTCCAGCACACTGGA
RAro054 ACAAAATAGCTCCGTAACAC
FAro95 GGCTGTTATTCTGGACGTCG
RAro95 CCCTAAGTATCCTACATACAG
FAro063 CCAGATACCTTGACAACGGC
RAro063 TCAAGAGCTTCCTGACCCTC
FAjTR04 CACCCTTGCCCTATTTTGATA
RAjTR04 GCGTAGTCATGATCACCTGT
FAjTR12 AACGTTAGTCCCTAGGTTCC
RAjTR12 TAAGGGTGTTATATGTTCAG
FAjTR45 GCGCATGGAGAACTCTAAT
RAjTR45 CAATAGAGTGAGGACAGTAGA
FAjTR37 AGACCTTATGTCACCTTATGCT
RAjTR37 AAGATGTTAAATTCAATTGTGC
FAJMS10 TGTCTAACACTAAGAAAAGGAGAGG
RAJMS10 GGCTGCCAGTATCTTCTCAAAG
Tseng et al 2001 Wirth Bernatchez
2001 Wirth Bernatchez
2001
Ishikawa et al 2001 Ishikawa et al 2001
Ishikawa et al 2001 Wirth Bernatchez
2001
Ishikawa et al 2001
AjTR-04 GT
15
167-217 5-CY3
55---57 Aro054
136-200 5-fluorescein
55---58
AjTR-45 CT
3
T TC
2
CCT
6
141-165 5-CY5
55---60 AjTR-12
AG
14
154-214 5-fluorescein
55
AjTR-37 TG
14
188-202 5-CY3
55 Aro95
86-136 5-CY5
55---57 Aro063
160-230 5-CY5
58
AJMS-10 GA
17
137-175 5-fluorescein
58
59
Hardy-Weinberg equilibrium HWE were detected in 26 out of 70 cases 7 loci x 10 populations. AjTR 04 and Aro 63 were more polymorphic than the five other
loci, with higher average numbers of alleles, 13.7 and 13.6 respectively and average of heterozygote were also higher than other loci, there are 0.885 and
0.886 respectively. In addition the meaningful of locus of AjTR04 and Aro63 also was demonstrated with the significant of deviation HWE, there were two of ten
populations non-significance of deviation HWE with both loci above. The ten samples are presented here in two groups: the Indian Ocean IO
and the Pacific Ocean PO groups. The Indian Ocean group is represented by Aceh, Padang, Bengkulu, Pelabuhan Ratu, Pangandaran and Bali and the Pacific
Ocean group by Donggala, Poigar, Bengalon and Obi. These two clusters correspond respectively to A. b. bicolor Indian Ocean part and A. b. pacifica
inter-oceans zone. The average number of alleles, He and Ho of Indian Ocean groups were from 13 to 15, from 0,860 to 0,891 and from 0,702 to 0,777
respectively, while group of Pacific Ocean were from 5 to13, from 0,767 to 0,873 and from 0,594 to 0,671 respectively. That means that the populations from
Indian Ocean group are globally more polymorphic than those belonging to the Pacific Ocean group.
F
ST
values, giving the genetic divergences between populations that revealed geographical differentiation and structuration. F-statistic F
IS
, F
IT
and F
ST
analysis of all locus among ten populations we present in Table 17. Inbreeding rate between F
IS
and within F
IT
populations are 0.194 19.4 and 0.209 20.9 respectively, whereas genetic differentiation F
ST
between population is 0.018 1.8. Accordance to Wright 1978 has suggested the
guidelines for the interpretation of Fst that the range value of F
ST
between 0 to 0,05 indicated little genetic differentiation. The same result was also showed by
the structuration under
STRUCTURE
programme analyses, there is no structure among 10 population of A. bicolor in Indonesia water Figure 10.
60
Table 16. Genetic variability of population parameters at 7 microsatellite loci on tropical eel in Indonesia water
Microsatellite locus
no allele Parameters
observed Ace
n=24 Pad
n=24 Ben
n=22 PelR
n=24 Pang
n=21 Bal
n=23 Obi
n=7 Poig
n=4 Dong
n=18 Beng
n=11
Average per locus
AjTR12
No. of alleles 8
15 15
13 13
15 7
3 8
9
11
23
Allele range 150-170
146- 186
140- 180
146- 172
142- 178
142- 174
146- 174
156- 170
146- 164
146- 168
140-186
He, 0,835
0,885 0,862
0,880 0,855
0,862 0,796
0,594 0,753
0,855 0,818
Ho, 0,636 0,750 0,864 0,667 0,810 0,783 0,857 0,250 0,389 0,455
0,646
H-W test n.s
n.s n.s
n.s n.s
AjTR45
No. of alleles 8 10
10 8 10 8 6 6 9 7 8,2
11
Allele range 135-149
135- 153
135- 155
137- 151
135- 155
135- 149
137- 151
137- 151
135- 151
137- 151
135-155
He, 0,833 0,846 0,838 0,848 0,816 0,824 0,694 0,750 0,825 0,744
0,802
Ho, 0,792
0,958 0,818
1,000 0,850
0,864 0,714
0,750 0,765
0,636 0,815
H-W test n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s
Aro63
No. of alleles 16
17 15
22 17
15 7
6 12
9
13,6
34
Allele range 168-22
172- 214
186- 230
160- 230
166- 230
170- 222
182- 208
188- 204
182- 224
182- 208
160-230
He, 0,898
0,921 0,906
0,933 0,910
0,914 0,806
0,813 0,909
0,845 0,885
Ho, 0,750 0,583 0,619 0,636 0,500 0,304 0,500 0,500 0,067 0,700
0,516
H-W test 0,0
0,0 n.s
0,0
AjTR37
No. of alleles 14 14 14 13 15 16 7 6 18 11
12,8
23
Allele range 188-218
188- 222
186- 220
188- 214
188- 218
188- 218
192- 222
188- 210
188- 234
192- 230
186-234
He, 0,879 0,888 0,878 0,885 0,900 0,912 0,816 0,813 0,923 0,888
0,878
Ho, 0,826
0,708 1,000
0,917 0,905
0,958 0,857
1,000 0,833
0,909 0,891
H-W test n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s
Aro54
No. of alleles 16
16 12
12 13
9 5
5 12
12
11,2
26
Allele range 148-194
136- 184
148- 184
148- 178
146- 194
150- 166
156- 172
156- 166
150- 200
152- 180
136-200
He, 0,902
0,893 0,830
0,687 0,859
0,815 0,781
0,778 0,897
0,884 0,833
Ho, 1,000 0,625 0,714 0,792 0,750 0,650 0,750 0,667 0,875 0,909
0,773
H-W test n.s
n.s n.s
n.s n.s
n.s n.s
n.s n.s
Aro95
No. of alleles 16 17 15 14 13 12 4 6 11 8
11,6
24
Allele range 94-130
86-122 100-
136 86-120
96-124 98-120
104- 114
92-118 96-120
96-130 86-136
He, 0,885 0,888 0,896 0,869 0,893 0,865 0,694 0,813 0,879 0,826
0,851
Ho, 0,958
0,875 0,682
0,696 0,619
0,833 0,286
1,000 0,563
0,455 0,697
H-W test n.s n.s
0,0 n.s
n.s n.s n.s n.s
AjTR04
No. of alleles 13
16 15
17 18
14 8
6 18
12
13,7
31
Allele range 187-225
177- 227
183- 221
181- 225
177- 229
183- 225
187- 233
187- 229
183- 241
179- 247
177-247
He, 0,891
0,918 0,899
0,921 0,904
0,899 0,806
0,813 0,926
0,888 0,886
Ho, 0,478 0,417 0,455 0,708 0,762 0,625 0,429 0,500 0,667 0,636
0,568
H-W test 0,0
0,0 0,0
n.s n.s
0,0
Average
He 0,875 0,891 0,873 0,860 0,877 0,870 0,770 0,767 0,873 0,847
Ho 0,777 0,702 0,736 0,774 0,742 0,717 0,628 0,667 0,594 0,671
No.of alleles
13 15 14 14 14 13 6 5 13 10
Note; Alelle range, n = number specimen on population, number under the loci name are the total number allele found, He = heterozygosities expected, Ho = heterozygosities
observed.
61
Table 17. F
IS
, F
IT
and F
ST
for 7 loci among 10 population of A. bicolor in Indonesia water
Locus F
IS
F
IT
F
ST
AjTR12 0,210 ± 0,051 0,229 ± 0,055 0,0230 ± 0,010
AjTR45 -0,002 ± 0,036 0,020 ± 0,041 0,0219 ± 0,009
Aro63 0,440 ± 0,075 0,446 ± 0,074 0,0102 ± 0,005
AjTR37 0,037 ± 0,038 0,037 ± 0,039 0,0002 ± 0,005
Aro54 0,105 ± 0,061 0,156 ± 0,047 0,0576 ± 0,031
Aro95 0,188 ± 0,062 0,195 ± 0,062 0,0080 ± 0,008
AjTR04 0,382 ± 0,049 0,384 ± 0,050 0,0022 ± 0,004
Average 0,194 ± 0,053 0,209 ± 0,052 0,018 ± 0,010
The genetic distances between population were calculated according to Nei 1972, and are given in Table 18. The stemming dendrogram UPGM construction
with PHYLIP Ver.3.5 is showed on Figure 10. Genetic distance between population ranged from 0,238 between Pelabuhan Ratu and Bali to 0,869
between Obi and Poigar. Genetic distance between population within Indian Ocean group range from 0,238 to 0,458, within Pacific Ocean group 0,343 to
0,869 and betwen Pacific and Indian group 0,392 to 0,839. According to the dendogram on Figure 10, Anguilla bicolor is divided into three group: first group
consist of 5 populations Aceh, Padang, Pelabuhan Ratu, Pangandaran and Bali, second group consist of three populations Donggala, Bengalon and Bengkulu
and third group consist of two population Obi and Poigar. The geographic position of the samples are in agreement with the dendrogramme except the
Bengkulu sample Indian Ocean clustered with the Danggala one inter-oceans sample. Poigar and then Obi samples are positioned at the basis of the tree
Figure 11.
Figure 10. Clustering analyses K=2 for the A. bicolor data set 180 individu, 10 population and 7 loci performed using STRUCTURE programme Pritchard et al.
2000
62
Table 18. Genetic distance among population during present study calculated under Nei 1972
Ace Pad
Bal PelR
Obi Poig
Dong Pang
Beng Ben
Ace Pad 0.267
Bal 0.268 0.260 PelR 0.398 0.395 0.238
Obi 0.695 0.665 0.839 0.590
Poig 0.778 0.703 0.781 0.706 0.869
Dong 0.429 0.392 0.630 0.472 0.444 0.633
Pang 0.288 0.284 0.361 0.473 0.837 0.638 0.462
Beng 0.494 0.500 0.655 0.557 0.607 0.705 0.343
0.588 Ben 0.346 0.254
0.396 0.458 0.851 0.603 0.389 0.241 0.576
pop1 pop2
pop8
Aceh Padang
pop3 pop4
Pangandaran Bali
PelabuhanRatu
pop7 pop10
pop9
Donggala
pop6
Bengkulu
pop5
Bengalon Poigar
Obi
5
Figure 11. Dendogram based on genetic distances Nei 1972 between A. bicolor
samples, using UPGMA method.
Discussion
Microsatellite as marker Microsatellite loci have been used in recent studies of population structure
of anguillidae because of their higher degree of sensitivity to detect slight genetic differentiation. According to previous studies, A. bicolor is distributed from the
Indian Ocean to the Pacific Ocean Jespersen 1942, Ege 1939 and Watanabe et al. 2005 and is composed of two subspecies : Anguilla bicolor bicolor and
Anguilla bicolor pacifica. However, according to Watanabe et al. 2005 this taxonomy is not well supported by clear differences in genetics and morphology.
It was so expected that, using microsatellites as markers in the population
63
structure studies of A. bicolor, a more reliable picture of the populations structure will be obtained Wirth and Bernatchez 2001, Mank and Avise 2003.
Microsatellite have been successful to solve the long controversial issue of panmixia and population genetic structure of European eels Anguilla anguilla
and American eels Anguilla rostrata Dannewitz et al. 2005; Tomas et al. 2011; Mank and Avise 2003 and Wirth and Bernatchez 2001 because both of species
have sligh differences genetic and morphological characters. In the tropical area, using of microsatellite also have managed to solve scientific mystery of
population structure of wide geograpical ditribution of tropical eel A. marmorata Ishikawa et al. 2004; Minegishi et al. 2008 and Gagnaire et al. 2009.
The preliminary study of population structure of A. bicolor using microsatellite marker has been done by Minegishi et al. 2012. This study showed
that A. bicolor has diverged between the Indian Ocean and Pacific Ocean, which consisten with the classical subspecies designation, but apparently genetically
homogeneous in Indian Ocean. Population structure
The assignation analysis of population structure STRUCTURE software was unable to detect geographic organization among the ten samples analysed.
That was indication a lack of genetic differentiation within A. bicolor species in Indonesian waters. Low genetic differentiation Fst 0,05 for A. bicolor in both
inter-ocean Indian Ocean dan Pacific Ocean was observed in this study that cause there are no private locus or diagnostic locus of each subspecies.
Minegishi et al. 2012 stated based on bayesian demographic history reconstruction and neutrality tests that are indicated population growth in each
ocean after the Indo-Pacific divergence. However, the dendrogramme construction, based on Nei 1972 distance
clearly divided the samples into a Indian Ocean cluster 5 locations and a inter- oceans cluster of 3 locations, one of them Bengkulu being aberrant. In the
Indian ocean, this species appears to be divided into two clusters: the first, composed of Aceh, Padang and Pangandaran, positioned rather at the north of
the Indian Ocean cost. The second group, composed of Bali and Pelabuhan Ratu, represents the southern part of the Indian Ocean sampled zone. Despite
the observed geographic and genetic organization of these samples, they represent probably the same panmictic population of the west of Indonesian
64
waters. It has been questioned whether there were georaphically divergent populations in western Indonesia because the spawning area of this species
recently only know around of Mentawai sea, west Sumatera. These results seems to confirm this hypothesis.
Interestingly, the population of Bengkulu, based on genetic divergence and genetic distance, seems closely related to Donggala and Bengalon
populations. The phylogenetic study on chapter IV showed that the population Bengkulu has no close relationship with A. b pacifica see Figure 12. However
Minegishi et al. 2012 suggested that no absolute physical barriers to gene flow exists in this zone, limiting the population divergence in eel. Perhaps, the
aberrant classification of the Bengkulu sample corresponds to this possible gene flow. The total inefficiency of STRUCTURE assignations demonstrates that the
genetic differences between the samples and between the sub-species is very limited.
A. b. pacifica DONG.11 A. b. pacifica POIG.15
A. b. pacifica BEG.3 A. b. pacifica BEG.4
A. b. pacifica DONG.1.14.16 POIG.5 OBI.3.9 A.b.pacifica
A. b. pacifica BEG.1 A. b. pacifica DONG.2
A. b. pacifica DONG.4 A. b. pacifica OBI.7
A. b. pacifica BEG.8 A. b. bicolor PEL.425
A. b. bicolor BAL.16 A. b. bicolor ACE.32 PANG.11
A. b. bicolor PAD.73 PEL.238 A. b. bicolor BEN.63
A. b. bicolor ACE.5 A. b. bicolor PEL.85.96
A. b. bicolor PAD.9 A. b. bicolor CIL.13
A. b. bicolor PANG.3 A. b. bicolor PAD.87 PEL.35
A. b. bicolor POS.16 A. b. bicolor PEL.8
A. b. bicolor ACE.9 A. b. bicolor PEL.298
A. b. bicolor BAL.21 A. b. bicolor BAL.3
A. b. bicolor PEL.105 A. b. bicolor PEL.14
Cla de 1
A. b. bicolor PAD.48 A. b. bicolor ACE.17
A. b. bicolor CIL.5 A. b. bicolor PEL.360
A. b. bicolor MEN.32 A. b. bicolor PAD.17
A. b. bicolor ACE.11 A. b. bicolor PAD.96
A. b. bicolor PAD.14 A. b. bicolor AN.927 PEL.404
A. b. bicolor PEL.63 A. b. bicolor PEL.92
A. b. bicolor PAD.3 PEL.71 A. b. bicolor ACE.29
A. b. bicolor BAL.49 A. b. bicolor CIL.2
A. b. bicolor PEL.4 A. b. bicolor BEN.39
A. b. bicolor PANG.18 A. b. bicolor PAD.60
A.b.bicolor A. b. bicolor PANG.2
A. b. bicolor PEL.339 A. b. bicolor AN923.930 CIL.10 PAD.16.31.7.56 PANG.8 PEL.500.175.390 BAL 35
A. b. bicolor BAL 55 A. b. bicolor PAD.3.75
A. b. bicolor PEL.8.308
Cla de 2 99
86 95
0.000 0.002
0.004 0.006
0.008 0.010
Figure 12. Phylogenetic relationship of A. bicolor based on cytochrome b gene
Fahmi 2013 chapter IV
65