Hynes 1996
Hynes 1996
Hynes 1996
A purified mitochondrial DNA (mtDNA) probe was used to examine restriction fragment
length polymorphisms produced by six restriction enzymes (Xba I, Eco RV, Ava II, Hinf I, Hae
III, Mbo I) in 915 brown trout from western Europe. A total of 20 composite haplotypes were
found with one to seven haplotypes in individual populations. Icelandic trout samples from
north, south, east, and west coast drainages showed only a single common haplotype in
contrast to the high level of polymorphism found in Irish and Scottish populations. The
phylogeny of mtDNA haplotypes and the pattern of haplotype distribution suggests that
post-glacial colonization of brown trout in NW Europe was more complex than the dual
colonization model which has been proposed on the basis of differential LDH-5* allele
distribution. For example, Lough Melvin (Ireland) appears to have been independently
colonised by three distinct types of trout. ? 1996 The Fisheries Society of the British Isles
Key words: brown trout; Salmo trutta; mitochondrial DNA; LDH-5*; post-glacial
colonization.
INTRODUCTION
Brown trout Salmo trutta L., has caught the attention of biologists by virtue of
the extensive variability and plasticity which it shows in many aspects of its
morphology, ecology and behaviour. From the time of Linnaeus about 50
species have been described for what is commonly referred to as brown trout,
including 10 species described by Günther for Britain and Ireland. The tendency
during this century has been for biologists to regard all of these forms as
belonging to a single polytypic species.
During the past two decades, electrophoretic analysis of enzyme variation has
provided extensive evidence of genetic differentiation within and among brown
trout populations (Ferguson, 1989; Skaala, 1992; Cross et al., 1992; Hansen
et al., 1993). Examination of loci common to various studies of brown trout
revealed some 60% of the observed variability to be distributed among popula-
tions in Atlantic drainages (Ferguson, 1989). Part of this considerable genetic
diversity of NW European brown trout populations could have evolved since
these waters were colonized at the end of the last Ice Age (10–15 000 years ago).
In addition, it could be the result of independent post-glacial colonization by
already distinct types of trout which diverged from each other during, or prior
to, the last glaciation. Based on the differential occurrence of two alleles at
LDH-5* it was suggested that NW Europe was independently colonized by two
races of brown trout (Ferguson & Fleming, 1983; Hamilton et al., 1989). The
first colonist was the ancestral race, so-called because of its possession of the
54
0022–1112/96/010054+14 $12.00/0 ? 1996 The Fisheries Society of the British Isles
m 55
ancestral LDH-5*100 allele. At a later time the modern race, typified by the
LDH-5*90 allele, colonized NW Europe and replaced the ancestral race in those
waters to which it was able to gain access. The ancestral race may have spread
postglacially from a refuge in Biscay drainages while the modern race may have
evolved in a refuge in the current Baltic or White sea area.
Studies on mtDNA, as compared with nuclear coded genes, have a distinct
advantage in studying population phylogeny. During population isolation (e.g.
following glacial advance), divergence occurs in both the nuclear and
mitochondrial genomes as a result of drift and selection. However, once the
populations come into secondary contact (e.g. on glacial retreat), nuclear
divergence is liable to be lost due to gene flow and recombination. By contrast
the asexual inheritance of mtDNA permits the organization of individuals into
matriarchal lineages.
To date relatively few studies of restriction fragment length polymorphism
(RFLP) of mtDNA variation have been carried out on brown trout populations
(Palva, 1986; Gyllensten & Wilson, 1987; McVeigh & Ferguson, 1988; Hynes
et al., 1989; Ovenden et al., 1993; Bembo et al., 1994). In general, these have
been preliminary investigations aimed primarily at developing and testing new
techniques. Nevertheless, they revealed considerable mtDNA variation within
and among brown trout populations, providing corroborative evidence of the
high genetic variability demonstrated by protein studies.
As a test of the dual colonization hypothesis, this study reports on mtDNA
diversity in brown trout populations from NW Europe, using a mtDNA probe
for the rapid detection of mtDNA RFLPs in total DNA samples restricted with
four, five and six base pair cutting enzymes.
through the hybridization of the probe with 50 ìg of competitor DNA (i.e. brown trout
total cellular DNA), in 1·5#SSPE (0·54 NaCl, 0·03 Na2H2PO4, 3 m EDTA Na2
salt; pH 7·7), for 45 min at 65) C. Prior to use the competitor DNA was extensively
sonicated to reduce the fragment size to <2 kb.
Filters were prehybridized for a minimum of 3 h in 1·5#SSPE, 6% polyethylene glycol
(PEG), 0·5% dried milk and 1% SDS (pH 7·7), at 65) C (Dalgleish, 1987). The total 10 ng
of oligolabelled mtDNA probe was added to hybridization solution and used to
simultaneously probe one to four filters. Following overnight hybridization at 65) C,
filters were washed once in 2#SSC (0·3 NaCl 0·03 Na3 citrate; pH 7·0) followed by
two washes in 0·2#SSC, 0·1% SDS. All washes were carried out for 20–30 min at 65) C.
Filters were finally rinsed in 1#SSC at room temperature and autoradiographed for 2–5
days at "70) C in the presence of intensifying screens.
Data on LDH-5* allele frequencies were taken from Hamilton et al. (1989) or, where
not available, samples were screened by the method described in that paper.
MtDNA restriction fragment data were analysed using the REAP (Mc Elroy et al.,
1991) and PHYLIP (Version 3.3; provided by J. Felsenstein, Department of Genetics,
SK-50, University of Washington, Seattle, WA 98195, U.S.A.) computer packages using
individual programs and methods as identified in the results section.
RESULTS
Table I lists the LDH-5*100 allele frequencies in the populations examined for
mtDNA RFLP variation. Ninety mtDNA restriction fragments were observed
among the 915 individuals sampled, representing a total of 394 nucleotides. This
corresponds to sampling some 2% of the mitochondrial genome. Patterns for
each enzyme were assigned upper case letters in order of discovery. A total of 20
composite haplotypes were found in the populations screened (Table II). Other
than for the unique haplotypes present in the Spanish and Mediterranean French
samples, there is no obvious geographical pattern in the occurrence of
haplotypes.
The haplotype and nucleotide diversity values within each of the populations
were calculated using the REAP DA program (Table III). While the majority of
populations examined exhibit several haplotypes, some, e.g. Melvin ferox and all
Icelandic brown trout populations, exhibited only a single haplotype. The
nucleotide divergence between haplotypes was calculated from restriction frag-
ment data using the REAP D program, following the method of Nei & Li (1979).
These values are summarized in the form of a dendrogram (Fig. 1) produced by
UPGMA cluster analysis. As expected, the most distinct haplotypes are XVIII
and XIX from the Mediterranean sample. Haplotype XII is also distinct and
was found in trout from the northern Scottish lakes, in several Irish Sea rivers
and in the Polish sample. The remaining haplotypes form three groups with each
group comprising a mixture of populations of both high and low LDH-5*100
frequency. Haplotypes found in gillaroo occur in all three groups and sonaghen
haplotypes in two of them.
Presence/absence fragment data were bootstrapped (Felsenstein, 1985) using
the PHYLIP SEQBOOT program, and the resulting data subsequently used to
construct unrooted Wagner parsimony trees of haplotype phylogeny using the
PHYLIP MIX program. Consensus trees were obtained for each bootstrapped
sample using the PHYLIP CONSENSE program. The resulting majority-rule
consensus tree is shown in Fig. 2 with associated bootstrap values. Most
groupings show low bootstrap support, i.e. they occur less than 50% in the
m 57
LDH-5*100
Country Locality Population Notes
frequency
Notes:
1. Rivers not known to be stocked.
2. Hatchery established in 1968 from eggs imported from England and Denmark.
3. Ferox, sonaghen and gillaroo are sympatric reproductively isolated brown trout populations in
Lough Melvin. (See Ferguson and Taggart, 1991, for details.)
4. Trout populations isolated above impassable water falls.
5. Trout populations consisting of a substantial anadromous (sea trout) component.
6. The Kallsjön and Fälpfjälltjärnarna fish were maintained in a hatchery and planted out in Lake Lilla
Bävervattnet from which these samples were obtained.
7. Samples from Harkjar Hatchery, Sejbaek, Rabis Baek and Trevad Mollebaek.
T II. Frequency of mitochondrial DNA haplotypes in brown trout populations
Haplotype* I II III IV V VI VII VIII IX X XI XII XIII XIV XV XVI XVII XVIII XIX XX
Population N
Ferox 37 1·00
Sonaghen 55 0·25 0·29 0·22 0·04 0·16 0·04
Gillaroo 47 0·13 0·04 0·23 0·46 0·06 0·06 0·02
Movanagher 50 0·54 0·04 0·04 0·02 0·14 0·18 0·04
Glenariff Upper 50 1·00
Glenariff Lower 1 50 0·42 0·42 0·16
Glenariff Lower 2 51 0·08 0·06 0·06 0·74 0·06
Glynn 102 0·92 0·01 0·07
Carnlough U 26 0·69 0·31
Carnlough L 28 0·64 0·15 0·21
Glencloy 15 0·40 0·40 0·20
Glenarm 39 0·51 0·10 0·39
Sutherland 32 1·00
Loch Ness 15 0·07 0·93
Iceland 130 1·0
Spain 7 0·43 0·43 0·14
Denmark 30 0·53 0·47
Sweden A+B 20 1·00
Sweden D 5 0·60 0·40
Erne 50 1·00
Leba, Poland 18 0·55 0·06 0·06 0·22 0·11
Crumlin 56 0·59 0·11 0·30
Tes, France 2 0·50 0·50
Total 915 0·19 0·02 0·02 0·002 0·01 0·39 0·02 0·003 0·02 0·06 0·01 0·18 0·05 0·01 0·003 0·003 0·001 0·001 0·001 0·02
Haplotypes comprise Xba I, Ava II, Hinf I, Hae III, Mbo I and Eco RV morphs respectively.
I, BAAAAA; II, ABAAAA; III, BADAAA; IV, BADAAB; V, ACEAAA; VI, ACAAAA; VII, ABAABA; VIII, ABAACA; IX, ACCBAB; X, ACCBAA; XI,
BABAAA; XII, ADACCB; XIII, BAADAA; XIV, ACFAAA; XV, AEACDA; XVI, CGCCAA; XVII, AGHAAA; XVIII, DFGEEA; XIX, DGGEEA; XX,
AGICAA.
m 59
Nucleotide divergence
I
XIII
III
IV
XI
II
VII
VIII
V
VI
XIV
IX
X
XVI
XVII
XV
XX
XII
XVIII
XIX
F. 1. Dendrogram based on UPGMA cluster analysis of nucleotide divergence values between mtDNA
haplotypes.
T IV. Presence of mtDNA lineages in populations grouped into high, middle and low
LDH-5*100 allele frequencies
simulation method of Roff & Bentzen (1989), using the MONTE program from
the REAP package.
The genetic relationships of the populations examined were calculated
using the REAP DA program and are shown in Fig. 3. Again other than for
m 61
V
XIV
VII
VI
(63) II
X (40)
(17) XII
IX
VIII
(60)
(87) (55)
(33)
(93)
(45)
(25)
XX (38)
I
(25)
(50)
XVII (28) XV
(91)
(59) (47) IV
III
XVI
(100) XIII
XVIII
XIX XI
F. 2. Unrooted Wagner Parsimony majority-rule consensus tree of mitochondrial haplotypes. Values
given in brackets represent the % number of times that grouping occurred out of 1000 bootstrap
replicates.
DISCUSSION
The application of a mtDNA probe for the detection of mtDNA RFLP
variation permitted the use of total genomic DNA. This was readily extracted
from minimal amounts of frozen or alcohol preserved tissue, thus dispensing
with the need for fresh tissue. One disadvantage, however, of this approach
was the inability to detect small fragments. This meant that it was not possible
to determine restriction site changes for all enzymes and consequently only
fragment data were available for use in estimating phylogenetic relatedness.
The development of PCR amplification and subsequent RFLP analysis of
specific mitochondrial regions undoubtedly surpasses the speed and ease of use
of the probe method. Initial results (authors’ unpubl. data) suggest that 15 of the
20 haplotypes reported in this study can be detected through RFLP analysis of
the NADH dehydrogenase-1 (ND-1) and ND-5/6 mitochondrial regions, using
the primers of Cronin et al. (1993). One disadvantage of the PCR approach,
however, is that it does not examine variation across the entire mitochondrial
genome, consequently variants outside the selectively amplified regions remain
62 . . .
Ferox
Sweden A and B
Erne
Sonaghen
Movanagher
Sweden D
Poland
Gillaroo
Glynn
Iceland
Carnlough Upper
Carnlough Lower
Glenarm
Crumlin
Glenariff Lower 1
Glencloy
Loch Ness
Denmark
Spain
Glanariff Upper
Sutherland
Glenariff Lower 2
Mediterranean
F. 3. Dendrogram based on UPGMA cluster analysis of percent sequence divergence estimates between
brown trout populations.
of 1·1% being observed between the two groups (Patarnello et al., 1994). Thus,
overall the RFLP approach may be more appropriate for population studies
unless sequencing can be directed at the regions of the mitochondrial genome
where variation is present.
A main objective of this study was to test the LDH-5* based dual post-glacial
colonization hypothesis. The phylogenetic tree of mtDNA haplotypes does not
show any dichotomy leading to haplotypes specific to populations of low and
high LDH-5*100 frequency. Indeed the only highly supported clade, of more
than two haplotypes, contains haplotypes present in populations fixed for
alternate LDH-5* alleles. The low bootstrap support for most other clades in the
tree (Fig. 2) makes it difficult to draw reliable inferences. Three haplotypes (I, VI
and XII) are common and widespread which suggests that they were present in
the common ancestor of the Atlantic colonists. Each of the supposed ancestral
(high LDH-5*100 frequency) NW European populations appears to have
become fixed for one of these mtDNA types due to random lineage sorting. The
geographically adjacent, but now isolated, Melvin ferox and Erne populations
are fixed for the same haplotype and it is possible that these were both derived
from a common subpopulation already fixed for this haplotype. Similarly the
Glenariff river of NE Ireland has the same haplotype as that of the isolated
Scottish Sutherland populations. Again these two areas may have been colo-
nized from a common west Scottish ancestor. These common ancestors may
have been transient populations in lakes which existed in the immediate
post-glacial period only, as a result of the reduced sea level. Alternatively,
this sharing of haplotypes may, for example in the case of ferox and Erne
populations, reflect river capture (Waters et al., 1994).
Overall there is a significant difference in the occurrence and frequency of
mtDNA haplotypes in the low (putative modern race) and high (putative
ancestral race) LDH-5*100 population groups. The low mitochondrial diversity
of the presumed ancestral populations is in keeping with their lower allozyme
diversity and with postulated colonization bottlenecks (Hamilton et al., 1989).
The isolation of ancestral populations from present-day migratory trout has
prevented the acquisition of further mtDNA haplotypes by gene flow. The
elapsed post-glacial period is too short for new mutants to be detectable with the
sample sizes and proportion of the mtDNA genome studied here. The low *100
frequency (putative modern race) populations have a much greater mtDNA
diversity, in keeping with their larger population sizes and opportunities for
spread of haplotypes by interpopulation migration.
The occurrence of a single invariant mitochondrial DNA haplotype in all of
the Icelandic populations examined is of particular interest, especially as these
come from large, widely separated north, south, east and west coast drainages.
This is in contrast to the situation found in Ireland, where even small river
populations possess several haplotypes. Given the sample size examined only
rare haplotypes would have been missed. Using the equation given by Grewe
et al. (1993) it can be calculated that the sample size gives greater than 99·9%
confidence that at least one individual should have been observed of those
haplotypes present at a frequency §0·05. The absence of variation would
suggest that all of these Icelandic populations were derived from a single
ancestral population in which a severe bottleneck had occurred, and that
64 . . .
References
Bembo, D. G., Weightman, A. J., Beverton, R. J. H. & Cresswell, R. C. (1994).
Mitochondrial DNA variation in River Usk brown trout, Salmo trutta. Journal of
Fish Biology 44, 717–723.
Bernatchez, L. & Dodson, J. J. (1990). Allopatric origin of sympatric populations of lake
whitefish (Coregonus clupeaformis) as revealed by mitochondrial DNA restriction
analysis. Evolution 44, 1263–1271.
Bernatchez, L., Guyomard, R. & Bonhomme, P. (1992). DNA sequence variation of
the mitochondrial control region among geographically and morphologically
remote European brown trout Salmo trutta populations. Molecular Ecology 1,
161–173.
Cawdery, S. A. H. & Ferguson, A. (1988). Origins and differentiation of three sympatric
species of trout (Salmo trutta L.) in Lough Melvin. Polskie Archiwum
Hydrobiologii 35, 267–277.
Cronin, M. A., Spearman, W. J., Wilmot, R. L., Patton, J. C. & Bickham, J. W. (1993).
Mitochondrial DNA variation in chinook salmon (Oncorhynchus tshawytscha) and
chum salmon (O. keta) detected by restriction enzyme analysis of polymerase chain
reaction (PCR) products. Canadian Journal of Fisheries and Aquatic Sciences 50,
708–714.
66 . . .