Mitochondrial DNA CoI haplotype variation in sibling species of rough periwinkles

13 Pages·2000·0.64 MB·English

Heredity85(2000)62–74 Received12April1999,accepted28February2000
Mitochondrial DNA CoI haplotype variation
in sibling species of rough periwinkles
C. S. WILDING, J. GRAHAME & P. J. MILL
SchoolofBiology,TheUniversityofLeeds,LeedsLS29JT
Three sibling species of rough periwinkles are currently recognized: Littorina arcana, L. compressa
andL. saxatilis.Certain formsof L.saxatilisarealso argued bysome to deserve speciesstatus, such
as the barnacle-dwelling ‘L. neglecta’ and the lagoonal ‘L. tenebrosa’. Relationships between these
taxa, and between and within representative populations, are investigated using sequence analysis
and restriction fragment length polymorphism of a mitochondrial DNA fragment spanning the
cytochromeoxidaseIandcytochromeoxidaseIIgeneboundary.Thesedatashowthatthereissome
sharing of haplotypes between species, with L. arcana haplotypes paraphyletic with respect to
L. saxatilis haplotypes, and L. compressa haplotypes paraphyletic to both L. arcana and L. saxatilis
haplotypes. Such sharing of mtDNA haplotypes could be a consequence of either persistent
hybridizationorepisodesofhybridization,orincompletelineagesortingofancestralpolymorphisms.
Onthebalanceofevidenceitissuggestedthatthelatter,ratherthanhybridizationevents,isthemore
likelycausalagentoftheobserveddistribution.Intraspecificvariationisextensiveanditissuggested
thatthepatternsofintraspecificpolymorphismareexplainablebyacombinationofhistoricalfactors
(the impact of the Pleistocene ice-age) and contemporary restrictions to gene flow. It is argued that
Littorina haplotypes evolved in at least two separate glacial refugia and became scattered by the
subsequentrangeexpansionaroundmostofthecoastline.Recentfactorssuchasrestrictedgeneflow
andbottleneckswouldthenbecapableofa(cid:128)ectingthehaplotypedistribution,resultinginthepattern
observed.
Keywords: hybridization, lineage sorting, littorina, mitochondrial DNA, phylogeography, post-
glacial recolonization.
(1996) suggest that a time period of approximately
Introduction
1.73 Myr (SE(cid:136)0–3.55) is likely for this group.
Periwinkles of the genus Littorina are familiar compo- Thethreeroughperiwinklespeciesshowconsiderable
nents of rocky shore communities. Around northern allozyme polymorphism, which is apparently greatest
European coastlines six species are recognized. in L. saxatilis (Ward, 1990). They lack a planktonic
L.littorea(L.),theedibleperiwinkle,isadistantrelative dispersal phase, and therefore show greater interpopu-
to the other five species, probably sharing an ancestor lation genetic heterogeneity than do related snails with
12–15Ma(Reidet al.,1996).Thesefiveareallmembers such dispersal (Janson, 1987). Further, their distribu-
of the subgenus Neritrema, including two flat periwin- tion on coasts di(cid:128)ers: L. arcana is absent from the
kles L. fabalis (W. Turton) and L. obtusata (L.), and eastern two thirds of the south coast of England and
three sibling species of rough periwinkles, namely L. compressa is particularly patchy in distribution; in
L. saxatilis (Olivi), L. arcana Hannaford Ellis and contrasttothesetwo,L.saxatilisisrelativelyubiquitous
L.compressaJe(cid:128)reys.Thesefivespecieshaveexistedfor on rocky coasts (Mill & Grahame, 1992). These distri-
amaximumof3.5–4MyrwhentheirancestralLittorina butions, together with the levels of inherent variation,
crossed from the Pacific into the Atlantic following the population size and history, and low dispersal and gene
opening of the Bering Strait. flow, should have an influence on the genetic structure
The age of the rough periwinkles is di(cid:129)cult to of the populations.
estimate, but reviewing allozyme evidence Reid et al. Allozyme electrophoresis studies of these species have
failed to detect diagnostic loci, although there are
frequency di(cid:128)erences. Phylogenies based on allele fre-
*Correspondence.E-mail:[email protected] quencies are usually in agreement with morphological
62 (cid:211)2000TheGenetical Society ofGreat Britain.
MITOCHONDRIAL DNAVARIATION IN LITTORINA 63
dataandanalysisofreproductivemodeinsuggestingthat Table 1 Locations of collection sites for Littorina samples
L. compressa is the least derived member of the group used in the present study
(Ward, 1990). However, from allozyme evidence Back-
Location Grid reference Species
eljau & Warmoes (1992) concluded that L. arcana was
basal.Whicheveriscorrect,itiscertainthatthespeciesare Ursholmen, Sweden 58(cid:176)50¢10¢¢N 0(cid:176)59¢4¢¢E S
closely related, and this, coupled with the existence of North Berwick OS: NT555857 A,C,N,S
contentiousforms,hasledtothetermL.saxatilisspecies St Abb’s OS: NT907692 A, C, S
complex as a descriptor of the group as a whole. Their Old Peak OS: NZ984021 A,B,N,S
closerelationshipsarecorroboratedbytheabilityofmale Filey OS: TA128816 N
Snettisham OS: TF649319 S
L. saxatilis and female L. arcana to hybridize under
Wells Next The Sea OS: TF909458
laboratory conditions (Warwick et al., 1990). Enzyme
(2 samples) OS: TF915456 S
studies have been applied to the analysis of some of the
Holkham OS: TF886451 T
contentious forms such as the barnacle-dwelling L. neg-
Cley OS: TG062448
lecta (Johannesson & Johannesson, 1990) and the lago-
(2 samples) OS: TG067447 S
onal L. tenebrosa (Janson & Ward, 1985), with the Alderton OS: TM363419 S
conclusion that there is no evidence for species status, St Margaret’s at Cli(cid:128)e OS: TR368444 S
since neither diagnostic loci nor frequency shifts were Folkestone OS: TR245369
evident. However, allozyme studies can be influenced by (2 samples) OS: TR244373 S
selectiononthelociinvolved,moreoverallozyme-inferred St Alban’s Head OS: SY959754 S
locimay require extensive time to reacha point at which Portland Bill OS: SY675683 S
East Fleet OS: SY799635 S
separate gene pools are detectable, leading to low
Fleet gravel OS: SY758665 S
resolution. DNA based studies are potentially more
Golden Cap OS: SY407918 S
powerful (Mitton, 1994), but few such studies have been
Pinhay Bay OS: SY318907 S
undertaken on Littorina. Reid et al. (1996) established a
Cargreen OS: SX436627 S
mtDNA (cytochrome b, 12SrRNA and 16SrRNA) phy-
The Lizard OS: SW699114 A, C, S
logeny of all extant Littorina but the resolution of the Cape Cornwall OS: SW349318 C
threeroughperiwinklespecieswaslow,andtherewasno Trevaunance OS: SW725519 A, C, S
indicationastohowinclusionofmultipleindividuals(and The Mumbles OS: SS632873 A, S
therefore potential intraspecific variation) a(cid:128)ected the The Gann flats OS: SM812069 S
phylogeny.Kyle&Boulding(1998)reportedontheuseof Musselwick (Gann) OS: SM819065 C
mtDNA cytochrome b variation to examine population St Ann’s Head OS: SM809028 S
West Dale Bay OS: SM797056 A,C,N,S
structure of L. subrotundata but, as yet, there have been
Porth Ysgo OS: SH207265 S
no attempts to use mtDNA to examine population level
Porth Llanllawen OS: SH266267 A, C, S
variability in North Atlantic Littorina species. Evidence
Trwyn Maen Melyn OS: SH138251 S
from maternally inherited mtDNA may allow analysis of
Porth Towyn OS: SH229376 C
introgressionorofdirectionalityofpossiblehybridization
PortStMary,IsleofMan OS: SC213680 S
between closely related species, especially where com- Nynian’s Cove,
parable nuclear data sets are available. Galloway OS: NW417364 S
Here we investigate the mitochondrial cytochrome Mizen Head IGR: V 074023 C, S
oxidase I gene with respect to phylogeny, phylogeog- Inismo´r, Aran Islands IGR: L 103221 C, S
raphy, evidence for monophyly in the three currently Ballynahown IGR: L 992202 A, C, S
recognized species, and the possibility of hybridization Golam Head IGR: L 826214 S, T
Seapoint Beach,
between them in the wild. We also ask what evidence
near Kittery, ME, USA S
these data provide concerning the taxonomic status
of some contentious forms of L. saxatilis such as A,Littorinaarcana;B,L.saxatilis‘b’;C,L.compressa;
L. neglecta (e.g. Caley et al., 1995) and L. tenebrosa N,L.neglecta;S,L.saxatilis;T,L.tenebrosa;OS,Ordnance
(e.g. Barnes, 1993). Surveygridreference;IGR,Irishgridreference.
between July 1996 and August 1998 (Table 1). Animals
Materials and methods
were collected from shores where locally abundant.
Samples were taken over an area of at least 2 m2 in an
Sample collection
e(cid:128)ort to avoid sampling of single families. A single
Samples of Littorina were collected from locations specimen of L. sitkana from Akkeshi, Hokkaido, Japan
around the United Kingdom and Eire over the period was utilized as an outgroup.
(cid:211)The Genetical Societyof Great Britain,Heredity,85, 62–74.
64 C. S. WILDINGETAL.
screen for variation. Digestions were performed in
DNA extraction and PCR
15 lL volumes containing 5 lL PCR product, 1.5 mL
Total genomic DNA was extracted from head-foot 10· reaction bu(cid:128)er and 1–4 units enzyme at 37(cid:176)C.
tissue of individual animals using protocol 47 of Digestion products were separated on 2% agarose gels.
Ashburner (1989). DNA sample concentrations were Estimates of h (F analogue) were calculated using
ST
measured by fluorometry and adjusted to 10 ng lL–1. FFSSTTAATT (Goudet, 1995) with standard errors calculated
PCR was performed in 50 lL volumes containing using jackknifing procedures. AAMMOOVVAA and FST were
50 mMM KCl, 10 mMM Tris-HCl pH 9.0, 1.5 mMM MgCl2, performed in AARRLLEEQQUUIINN v. 1.1 (Schneider et al., 1997).
0.1% Triton X-100, 0.01% gelatin, 200 lMM each dNTP, For the AAMMOOVVAA analysis, six groups were identified on
25 pmol each primer, 25 ng DNA and 1 U Taq thebasisofregionaldi(cid:128)erencesinshellmorphologyand
(Supertaq,HTBiotechnologies).TheprimerpairsaxcoI on distributional breaks: south-east, south-west, South
and saxcoII, designed from sequence in Wilding et al. Wales, Irish Sea, Ireland and East coast. Contingency
(1999) (accession number AJ132137), was used. SaxcoI tests were implemented in GGEENNEEPPOOPP (Raymond & Rous-
covers bases 219–240 and saxcoII bases 1185–1166 of set, 1995).
Wilding et al. (1999), and amplifies 967 bp (inclusive of
primers) covering the 3¢ end of cytochrome oxidase I,
Results
30 bp of apparent noncoding sequence and the first
38 bp of CoII. PCR conditions of 1· 94(cid:176)C, 5 min; 35·
Sequence analysis
(94(cid:176)C, 1 min, 55(cid:176)C, 30 s, 72(cid:176)C, 1 min); 1· 72(cid:176)C, 5 min
were employed. Positive PCR reactions were cleaned PCR using the primer pair saxcoI–saxcoII generated a
by PEG precipitation (Ausubel et al., 1990) and resus- 967-bp fragment. The single primer saxcoII was then
pended in 20 lL. Subsequently 4 lL was used in an used to sequence one end of this fragment. Four
automated sequencing reaction with the amplification hundred base pairs of unambiguous sequence were
primer saxcoII, using Taq terminator sequencing mix, generated from each of 95 separate individual Littorina
and run on an ABI377. representing a range of recognized species including 89
rough periwinkles (L. arcana, L. compressa and
L. saxatilis and including some representatives of the
Sequence analysis
contentious taxa L. neglecta, L. tenebrosa and
Haplotype sequences were aligned in CCLLUUSSTTAALL WW L. saxatilis ‘b’), two L. fabalis, three L. obtusata and
(Thompson et al., 1994). Identical haplotypes were one specimen of Littorina sitkana (Table 2). Twenty-
identified through the construction of a distance matrix four distinct haplotypes were identified from these 95
inMMEEGGAAv.1.02 (Kumar et al.,1993) andremovedfrom individuals (Table 3). Sequences are deposited in Gen-
the dataset. Sequence variation was examined with bank with accession numbers AJ133311–133334. Of the
DnaSP v. 2.2 (Rozas & Rozas, 1997). Phylogenies were 400 bp of sequence, 56 sites exhibited variation, all
constructed using both distance and maximum likeli- involving individual substitutions with the exception of
hood (ML) methodologies with L. sitkana as an out- a single base indel di(cid:128)erence between the L. sitkana
group. Haplotype relationships were inferred by both sequence and all other sequences. When L. sitkana is
neighbour joining of Jukes–Cantor distances with 1000 removed there are 37 variable sites of which 25 are
bootstrapped data sets, implemented in MMEEGGAA v. 1.02 parsimony informative. If only the rough periwinkles
(Kumar et al., 1993) and by ML analysis performed by are considered these figures drop to 28 variable sites
quartetpuzzling(Strimmer&vonHaeseler,1996)using with 16 parsimony informative positions. All substitu-
PPUUZZZZLLEE with 1000 quartet puzzling steps (analysis tions are silent with the exception of the G fi A
performed at HTTP://bioweb.pasteur.fr/seqanal/inter- transition at position 938 which alters the amino acid
faces/Puzzle.html). A haplotype network was also from valine to isoleucine, and the T fi G transversion
manually constructed detailing the relationships of at position 1070 which changes the amino acid from
haplotypesandthenumberofmutationalstepsrequired serine to alanine. Because of the low number of
to interconvert haplotypes. parsimony informative sites, parsimony-based phylo-
geny reconstruction methods were considered inappro-
priateandhaplotypephylogenieswereconstructedusing
RFLP analysis
both a distance method and a maximum likelihood
From aligned sequences, polymorphic restriction variant. The resultant trees (Fig. 1A and B) are essen-
enzymeswerechosenforwhichtherecognitionsequence tially congruent, although the distance-based tree sug-
spanned a variable site. Three suitable enzymes (DdeI, gestssomedetailsofbranchingorderthatarenotseenin
DraI and HindIII) were identified and employed to the maximum likelihood tree. Bootstrap support values
(cid:211)TheGenetical Society ofGreat Britain, Heredity,85, 62–74.
(cid:211) Table 2 Numbers of haplotypes 1–24 encountered at various locations around the UK and Ireland through sequencing of 400 bp of mtDNA CoI from
T individual Littorina
h
e
G Haplotype
e
n
etic Species Sampling site N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
a
l L. saxatilis N. Berwick 1 1
S
o L. saxatilis St Abb’s 1 1
c
iety L. saxatilis Old Peak 4 2 2
L. saxatilis Snettisham 2 2
o
f L. saxatilis St Margaret’s 2 1 1
G
r L. saxatilis Folkestone 11 1 2 7 1
e
at L. saxatilis Portland Bill 5 1 4
B L. saxatilis East Fleet 1 1
r
ita L. saxatilis St Albans 2 1 1
in L. saxatilis Pinhay Bay 5 5
,
H L. saxatilis Cargreen 1 1
e
re L. saxatilis Lizard 5 2 3
d
ity L. saxatilis Porth Ysgo 4 1 1 2
, L. saxatilis Gann 2 2
8
5 L. saxatilis Isle of Man 2 2
,
6 L. saxatilis Mizen Head 1 1
2
–7 L. saxatilis Aran Islands 1 1
4
. L. saxatilis Maine, USA 1 1
L. neglecta N. Berwick 2 2
L. neglecta Peak Steel 2 2
M
L. neglecta Filey 1 1
IT
L. saxatilis ‘b’ Peak Steel 2 1 1 O
L. tenebrosa Golam 2 2 C
H
L. arcana St Abb’s 2 2 O
N
L. arcana Old Peak 2 2 D
L. arcana Lizard 4 3 1 R
IA
L. arcana Porth Towyn 1 1
L
L. arcana Ballynahown 1 1 D
N
L. compressa N. Berwick 3 2 1
A
L. compressa St Abb’s 4 1 2 1 V
L. compressa Lizard 3 2 1 A
R
L. compressa Cape Cornwall 1 1 IA
L. compressa Gann 1 1 T
IO
L. compressa Porth Llanllawen 3 1 1 1
N
L. compressa Porth Towyn 2 2 IN
L. compressa Mizzen Head 1 1
L
L. compressa Aran Islands 1 1 IT
L. fabalis Old Peak 2 2 TO
L. obtusata Old Peak 3 1 2 R
L. sitkana Akkeshi, Japan 1 1 IN
A
6
5
6
6
C
.
S
.
W
Table 3 Polymorphicnucleotidepositionsdefiningthe24mitochondrialCoIhaplotypesuncoveredinLittorina.Positionnumbersrelatetothemitochondrial IL
D
sequence of Wilding et al. (1999). Full stops indicate identity to this sequence IN
G
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 E
T
7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 N 1 1 1
A
Happlo- 6 7 7 9 1 1 2 3 4 5 5 6 6 6 6 8 8 9 0 1 1 1 2 3 5 6 7 7 8 9 0 1 3 3 5 5 6 6 7 7 8 8 9 0 1 1 1 1 2 2 2 2 2 D 3 3 3 L
.
type 9 5 8 0 4 7 9 5 7 3 6 2 3 8 9 3 9 8 4 0 3 9 5 8 2 4 0 9 2 8 9 8 3 6 1 7 0 9 0 8 1 7 9 8 1 7 8 9 0 1 6 7 9 * 1 3 8
T T G C T C T G T T G T C C T A T C T C C G T G C C C G T T G T T G A T T A T C T A A T T A C C T G A A G – T CA
1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 . . . . . . . A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 . . . . . . . . . . . . . T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 . . . . . . . . . . . . . . . . . . . . . . C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . . . . . . . . . . . . . . . . . . . . . . . . . .
6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . . . . . . . . . . . . . . . . . . . . . . . .
7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . T .
8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . . . . . T . . . . . . . . . . . . . . .
9 . . . . . . . . . . . . . . C . . . . . . . . . . . . . . . . . . . . C . . . . . . . . . . . . . . . . . . . T .
10 . . . . . . C . . . . . . . . . . . . . . . . . T . T . . C . . . . . . . . G . . . . . . . . . C . . . . . . . .
(cid:211) 11 . . . . . . . . . G . . . . . . . . . . . . . . T . T . . C . . . . . . . . G . . . . . . . . . C . . . . . . . .
T 12 . . . . . . . . . . . . . . . . . . . . . . . . T . T . . C . . . . . C . . . . . . . . . . . . . . . . . . . . .
h
e 13 . . . . . . . . . . . . . . . . . . . . . . . . T . T . . C . . . . . . . . G . . . . . . . . . C . . . . . . . .
G 14 . . . . . . . . . . . . . . . . . . . . . . . . T . T . . C . . . . . . . . . . . . . . . . . . C A . . . . . . .
e
ne 15 . . A . . T . . . . . . . . . . . . . . . . . . T . . . . . . . . . . . . . . . . . . . . . T . . . . . . . . . .
tic 16 . . A . . T . . . . . . . T . . . . . . . . . . T . . . . . . . . . . . . . . . . . . . . . T . . . . . . . . . .
a
l 17 A . . . . . . . . . . . . T . . . . . . . . . . T . . . . . . . . . . . . G . . . . T C . . . T . . . G . . . T .
S
oc 18 C. . . . . . . . . . . . . T . . . . . . . . . . T . . . . . . . . . . . . G . . . . T C . . . T . . . G . . . T .
iety 19 . . . . . . . . . . . . . T . . . . . . . . . . T . . . . . . . . . . . . G . T . . T C . . . T . . . G . . . T .
o 20 . . . . . . . . . . . . . T . . . . . . . . . . T . . . . . . . . . . . . G . . . . T C . . . T . . . G . . . T .
f
G 21 . C . . . . . . . . . . . . . . . . . . . . . . T . . . . . . . . . . . . G . . . . . . . . . T . . . G . . . T .
re 22 . . . . . . . . C C . C . T . G . . . . T . . A T . . . C . . . . . . C C C . . . . . . . G . T . . . . . . . . .
a
t 23 . . . . . . . . C C . C . T . G . . . . T . . A T . . . C . . . . . . C C C . . . . . . . G . T . . . . . . . . G
B
r 24 . . T T C . . A . . A . T T . . C T A T . A . A T T . A . . . C . A G . . . G . C . . . C . . T . . G . A C C . .
ita
in
, IND,indelevent.
H
e
r
e
d
ity
,
8
5
,
6
2
–
7
4
.
MITOCHONDRIAL DNAVARIATION IN LITTORINA 67
Fig.1 (A)Neighbour-Joiningtree of
Jukes–Cantordistances for the 24CoI
haplotypesof Littorina.Numbers at
nodesrefer to percentagebootstrap
supportfrom 1000replicates.(B).
Quartet-puzzling maximumlikelihood
treeconstructed usingthe HKY model
inPPUUZZZZLLEE.Numbers at nodesindicate
howmany timesthe cluster to the right
ofthe node wasfound in 1000quartet
puzzlingsteps.
are generally low, although there are some well sup- supported in both phylogenies (haplotypes 10–14). The
ported nodes. remaining sequences (1–9) form a single group in the
Disregarding the outgroup (L. sitkana, haplotype 24) distance method, whilst in the ML tree they are
and the sequences from flat periwinkles (haplotypes 22 unresolved, with the exception of a group of three
and 23), there are four main groups discernible. The (7–9). Haplotypes 1–9 are labelled L. saxatilis group a.
haplotypeslimitedtoL.compressa(17–21)formasingle Thelackofresolutioninthephylogenyofhaplotypesof
group, and haplotypes 15 and 16 which are mainly group a is a consequence of these sequences di(cid:128)ering
limited to L. arcana (and some east coast L. compressa) from each other by only singleton base changes.
also group together. The remaining L. saxatilis sequen- Although the support values within the L. saxatilis
ces (in a few instances these also occur in L. arcana or haplotypes are low, the two distinct groups of
L. compressa — see later for details) form two main L.saxatilishaplotypesarealsorevealedinthehaplotype
groups. One group (termed L. saxatilis group b) is well network (Fig. 2). Here the four separate haplotype
Fig.2 Haplotypenetworkshowingthe
numberofbasechangesbetween24CoI
haplotypesof Littorina.Barsoninter-
connectinglinesrepresentthenumberof
mutationalsteps. Dashed linesencircle
majorgroups (see text).
(cid:211)The Genetical Societyof Great Britain,Heredity,85, 62–74.
68 C. S. WILDINGETAL.
groups (the ‘compressa’ group, the ‘arcana’ group and AAAandhaplotype11isADA.Astylisedtreebasedon
L. saxatilis groups a and b) distinguishable in the Fig. 1 A and B and summarizing the composite haplo-
phylogeniescanbediscerned(surroundedbythedashed type information is depicted in Fig. 3. Table 5 and
lines). This haplotype network is not a minimum Fig. 4 together describe the distribution of composite
spanning tree in that individual haplotypes are not haplotypes within populations of the various species.
separated by single substitutions, and there are many Becausesamplesizesaremuchhigherthanthoseutilized
ambiguousconnections.Thisisprobablyaconsequence for the sequencing study, this RFLP analysis gives a
of not having uncovered all existent haplotypes, and much better picture of the geographical distribution of
perhaps also due to homoplasies in the data (Besansky, haplotypes. In L. arcana there is a definite east–west
1997;BrownGladdenet al.,1997),althoughsuchevents split, with all east coast members having the BAA
seem less likely when haplotypes of closely related haplotype and the majority of west coast animals being
species are being compared. AAA. Thus there is a geographical pattern to those
Most haplotypes are unique to a particular species, haplotypes that are shared with L. saxatilis. East coast
with the exceptions of haplotypes 1 and 10, which are animals (St Abb’s, North Berwick and Old Peak) all
seen in L. saxatilis, L. arcana and L. compressa, and havethe‘L.arcana’haplotypeBAA,whilstthoseonthe
haplotype 15, which has been encountered in both west coast often have haplotypes typical of L. saxatilis,
L. arcana and L. compressa. Representative L. neglecta, such as AAA. For L. compressa the picture is less clear
L. tenebrosa and L. saxatilis ‘b’ did not have distinct andvariationishigher thaninL.arcana.Inthisspecies
haplotypes, but shared haplotypes with L. saxatilis thereseemstobenoobviouspatterntothosehaplotypes
(Table 2). that are shared with either L. arcana or L. saxatilis.
L.compressafromtheLizard,TrevaunanceandInismo´r
are fixed for the ‘compressa’ haplotype, but elsewhere
Restriction digestion
animals with ‘saxatilis’ haplotypes (BCA, BBA, AAA)
From sequence information three informative restric- are common. Because of more intensive sampling, there
tion enzymes were identified, DdeI, DraI and HindIII. is more information on the geographical variation of
These released 1–3 fragments per enzyme (Table 4). haplotypes in L. saxatilis. Two things are apparent.
Individual haplotypes were designated by single letter Firstly,asforL.arcanathereisadi(cid:128)erencebetweeneast
codes and a three-letter composite haplotype was then coast and west coast animals, with the BBA haplotype
generated for each animal. There were insu(cid:129)cient much more frequent on the west coast, and a higher
informative enzymes to di(cid:128)erentiate every sequenced frequencyofBCAandAAAintheeast.Secondly,there
haplotype. However, the three enzymes used allowed
resolution into the main groups obvious from the
haplotype network and phylogenies. Six separate com-
posite haplotypes were uncovered. The L. compressa
group (17–21) are all BAB and the L. arcana haplotype
group (15–16) all BAA. For L. saxatilis group a,
haplotypes 2 and 5 are BCA and the remainder are
BBA, whilst for group b, haplotypes 10 and 12–14 are
Table 4 Sizes of fragments (in base pairs) generated by
restriction enzyme digestion with three separate restriction
enzymes (DdeI, DraI and HindIII)
Restriction enzyme Pattern Fragment sizes
DdeI A 333 634
B 257 333 377
C 195 333 439
D 966
DraI A 60 907
B 967
HindIII A 967
Fig. 3 Stylized dendrogramdepictingrelationships of RFLP
B 190 777
composite haplotypes ofLittorina.
(cid:211)TheGenetical Society ofGreat Britain, Heredity,85, 62–74.
MITOCHONDRIAL DNAVARIATION IN LITTORINA 69
Table 5 Frequency of composite haplotypes in populations of Littorina from locations around the UK and Ireland. Order
of enzymes in composite haplotype is HindIII, DdeI, DraI
RFLP haplotype
Sample site AAA BBA ADA BCA BAB BAA N
L. saxatilis Portland Bill 0 3 6 0 0 0 9
L. saxatilis East Fleet 0 8 2 0 0 0 10
L. saxatilis Lizard 0 0 6 4 0 0 10
L. saxatilis Golden Cap 0 0 20 0 0 0 20
L. saxatilis St Margaret’s 1 18 0 1 0 0 20
L. saxatilis Pinhay Bay 0 0 5 0 0 0 5
L. saxatilis Cargreen 5 0 0 5 0 0 10
L. saxatilis Fleet gravel 0 10 0 0 0 0 10
L. saxatilis St Albans 0 9 1 0 0 0 10
L. saxatilis Trevaunance 1 7 1 4 0 0 13
L. saxatilis Folkestone 1 5 0 13 0 0 19
L. saxatilis N. Berwick 8 2 0 0 0 0 10
L. saxatilis St Abb’s 8 1 0 1 0 0 10
L. saxatilis Galloway 11 1 0 6 0 0 18
L. saxatilis Mumbles 3 12 0 5 0 0 20
L. saxatilis Gann 15 3 0 2 0 0 20
L. saxatilis St Ann’s Head 37 1 0 2 0 0 40
L. saxatilis Aran Islands 10 0 0 0 0 0 10
L. saxatilis Old Peak 20 0 0 0 0 0 20
L. saxatilis Porth Ysgo 5 15 0 0 0 0 20
L. saxatilis Golam Head 2 8 0 0 0 0 10
L. saxatilis Snettisham 3 0 0 7 0 0 10
L. saxatilis Sweden 10 0 0 7 0 0 17
L. saxatilis Porth Llanllawen 13 0 0 4 0 0 17
L. saxatilis Trwyn Maen Melyn 4 6 0 0 0 0 10
L. saxatilis Wells next the sea 9 0 0 11 0 0 20
L. saxatilis Cley 42 0 0 0 0 0 42
L. saxatilis Alderton 0 0 0 10 0 0 10
L. saxatilis Ballynahown 1 19 0 0 0 0 20
L. saxatilis West Dale Bay 0 10 0 0 0 0 10
L. saxatilis ‘b’ Peak Steel 27 3 0 0 0 0 30
L. neglecta Peak Steel 70 2 0 0 0 0 72
L. neglecta N. Berwick 7 0 0 0 0 0 7
L. neglecta West Dale Bay 0 12 0 7 0 0 19
L. tenebrosa Holkham 0 0 0 10 0 0 10
L. tenebrosa Golam 18 2 0 0 0 0 20
L. arcana Lizard 6 0 0 0 0 4 10
L. arcana Trevaunance 10 0 0 0 0 0 10
L. arcana N. Berwick 0 0 0 0 0 10 10
L. arcana St Abb’s 0 0 0 0 0 10 10
L. arcana Mumbles 5 0 0 0 0 0 5
L. arcana Old Peak 0 0 0 0 0 15 15
L. arcana Porth Llanllawen 5 0 0 0 0 0 5
L. arcana Ballynahown 6 2 0 0 0 0 8
L. arcana West Dale Bay 10 0 0 0 0 0 10
L. compressa Lizard 0 0 0 0 10 0 10
L. compressa Trevaunance 0 0 0 0 11 0 11
L. compressa N. Berwick 3 0 0 0 1 6 10
L. compressa St Abb’s 1 0 0 0 5 4 10
L. compressa Gann 2 0 0 0 2 2 6
L. compressa Aran Islands 0 0 0 0 10 0 10
L. compressa Porth Llanllawen 0 1 0 7 2 0 10
L. compressa Ballynahown 0 0 0 0 9 1 10
L. compressa West Dale Bay 0 2 0 1 7 0 10
N,samplesize.
(cid:211)The Genetical Societyof Great Britain,Heredity,85, 62–74.
70 C. S. WILDINGETAL.
is a south-west enclave of L. saxatilis with a high 1998) and it is well documented that mtDNA can cross
frequency of the ADA composite haplotype. This haplo- species boundaries through hybridization or introgres-
type is seen nowhere else in the country. F-statistics sion (Avise, 1994). If hybridization were the causal
(h and F ) indicate substantial population subdivision mechanism of the observed haplotype distribution, this
ST
(Table 6) and this is corroborated by contingency would require viability of hybridization between all
tests which show significant di(cid:128)erences among popula- three species. Although Warwick et al. (1990) showed
tions of all three species (P(cid:136)0.000 (cid:139) 0.000). AAMMOOVVAA laboratory hybridization in one direction (female
analysis (Table 6) using the preordained groups identi- L.arcanawithmaleL.saxatilis),thereisnoinformation
fied on the basis of shell morphology and known on the mating preferences during back-crosses, and
disjunctions in littorinid distribution, indicates varying there were no successful hybridizations with other
degrees of evidence for mtDNA variability being par- species crosses. We suggest that this asymmetrical
titioned among these groups. Although for L. arcana haplotype distribution, with ‘L. saxatilis’ haplotypes
over 70% of the total variation is among these groups, observed in L. arcana and L. compressa but no
the figure is somewhat misleading, since virtually all ‘L. compressa’ haplotypes or ‘L. arcana’ haplotypes
variation is partitioned solely on an east coast–west found in L. saxatilis, rules out general hybridization
coast basis, and the choice of groups reflects this to an (i.e.occasionalgeneflowbetweenthespecies).Although
extent. For L. compressa and L. saxatilis, substantial introgressionofthemtDNAofonespeciesintoanother
variation is within populations, reflecting high levels of is not precluded, it seems unlikely on the basis of
polymorphism. For L. saxatilis, little of the remainder existing evidence.
of the variation is among groups, indicating that these It is also the case that the animals with haplotypes
do not explain the variation encountered. In L. com- typical of another species have a restricted distribution.
pressanearly40%ofthevariationisamonggroups,but For instance, L. arcana with the ‘saxatilis’ haplotype
as for L. arcana, this is probably attributable to east– occur on the west coast of Britain, but have not been
west division, whereby most west coast L. compressa found on the east coast (Fig. 4, Table 5), and
show the ‘compressa haplotype’ whereas a higher L. compressa with particular ‘arcana’ or ‘saxatilis’
proportion of east coast animals display the ‘saxatilis’ haplotypes are patchy. In most cases these haplotypes
or ‘arcana’ type. are not the commonest ones, and in some cases are not
present sympatrically in the other species, e.g. all West
Dale Bay L. arcana have a ‘saxatilis’ haplotype AAA,
Discussion
but all West Dale Bay L. saxatilis are BBA. This again
suggests that localized hybridization is not the cause of
The phylogeny of Littorina mitochondrial CoI
theobservedhaplotypesharing,unlessthehybridization
haplotypes
happenedinthepastwhen,perhaps,haplotypefrequen-
The deduced branching order of the Littorina species cies were very di(cid:128)erent.
considered here is in agreement with current, consensus The alternative explanation, retention of ancestral
opinion (Reid, 1996). The low bootstrap support in the polymorphisms, has been documented for mtDNA
phylogeny at nodes leading to the three rough periwin- genesinothergroups(e.g.Helm-Bychowski&Cracraft,
kle species could be a consequence of a period of rapid 1993;Fehrer,1996;Schneider-Broussardet al.,1998),in
speciation, with only short time periods between speci- which shared haplotypes antedate speciation and are
ation events during which informative substitutions haphazardly maintained in certain populations through
could accrue. Periods of rapid speciation have similarly incomplete lineage sorting. Rapid and recent speciation
a(cid:128)ected phylogeny construction in finches (Fehrer, (see above) is consistent with this since there will have
1996) and corvids (Helm-Bychowski & Cracraft, 1993). been little time for sorting of ancestral haplotypes into
Paraphyletic haplotype distributions are also apparent the descendant taxa. This hypothesis also fits with the
with L. obtusata haplotypes paraphyletic with respect ‘directional’ nature of haplotype sharing. The least
to L. fabalis, L. arcana haplotypes paraphyletic with derived littorinid, L. compressa, displays haplotypes
respect to L. saxatilis, and L. compressa haplotypes representative of all groups currently recognized as
paraphyletic with respect to both L. arcana and ‘L. compressa type’, ‘L. arcana type’ and ‘L. saxatilis
L.saxatilishaplotypes.Suchparaphylymayhavearisen type’. A more limited range of haplotypes were passed
as a consequence of shared ancestral polymorphisms on to the descendant L. arcana and L. saxatilis, being
when (usually moderate to large) populations have not enriched with respect to what has become the ‘arcana
progressed to reciprocal monophyly during speciation. type’andthe‘saxatilistype’whilststillleavinghistorical
However introgressive hybridization may also be signatures — ‘arcana type’ and ‘saxatilis type’ haplo-
involved (Besansky, 1997; Schneider-Broussard et al., types in L. compressa and ‘saxatilis type’ haplotypes in
(cid:211)TheGenetical Society ofGreat Britain, Heredity,85, 62–74.
MITOCHONDRIAL DNAVARIATION IN LITTORINA 71
Fig.4 Geographicalvariation in CoI
RFLPpatterns for Littorinasaxatilis,
L.arcanaandL.compressa.Areasofpie
chartsrepresent compositehaplotype
frequency.
(cid:211)The Genetical Societyof Great Britain,Heredity,85, 62–74.