Molecular systematics and ...

Molecular systematics and phylogeny of the ‘Marbled Whites’ (Lepidoptera - Nymphalidae, wszystko, motyle i ...

[ Pobierz całość w formacie PDF ]
Systematic Entomology
(2009), DOI: 10.1111/j.1365-3113.2009.00493.x
Molecular systematics and phylogeny of the ‘Marbled
Whites’ (Lepidoptera: Nymphalidae, Satyrinae,
Melanargia Meigen)
VAZRICK NAZARI
1
, WOLFGANG TEN HAGEN
2
and G I A N
CHRISTOFORO BOZANO
3
1
Department of Integrative Biology, University of Guelph, Ontario, Canada,
2
Fruhlingstrasse 1, 63853 Momlingen, Germany
and
3
Viale Romagna 76, 20133 Milano, Italy
Abstract.
We investigated genetic divergence and phylogenetic relationships
amongst all known species of Palaearctic butterflies of the genus
Melanargia
using
sequence information from three genes [mitochondrial
cox1
barcode region (658 bp),
ribosomal 16S
rRNA
(
c
. 518 bp), and nuclear
wg
(404 bp)]. Results show a lack
of DNA divergence among several poorly characterized taxa, as well as deep diver-
gences within and between others. We corroborated the molecular information with
morphological and genitalic characters as well as with geographic data. We revise the
taxonomy of
Melanargia
, and propose a new systematic scheme for the group. We
revive some previous synonymies (
M. lucasi meadwaldoi
stat. rev.
,
M. ines fathme
stat. rev.
,
M. ines jahandiezi
stat. rev.
,
M. meridionalis tapaishanensis
stat. rev.
),
revise the status of some subspecies into species (
M. transcaspica
stat. nov.
,
M. lucida
stat. nov.
,
M. wiskotti
stat. nov.
) and of several species into subspecies of other taxa
(
M. evartianae sadjadii
stat. nov.
,
M. larissa hylata
stat. nov.
,
M. larissa grumi
stat.
nov.
,
M. larissa syriaca
stat. nov.
,
M. larissa titea
stat. nov.
,
M. lugens montana
stat.
nov.
,
M. epimede ganymedes
stat. nov.
), revise the status of subspecies and transfer
them to other species (
M. larissa lorestanensis
stat. nov.
,
M. larissa iranica
stat.
nov.
,
M. larissa karabagi
stat. rev.
,
M. larissa kocaki
stat. nov.
,
M. transcaspica
eberti
stat. nov.
), and propose new synonymies (
M. larissa titea
=
M. titea standfussi
syn. nov.
=
M. titea titania
syn. nov.
,
M. leda leda
=
M. leda yunnana
syn. nov.
,
M
.
lugens lugens
=
M. lugens ahyoui
syn. nov.
,
M
.
lugens hengshanensis
=
M.
lugens hoenei
syn. nov.
,
M. halimede halimede
=
M. halimede gratiani
syn. nov.
,
M. asiatica asiatica
=
M. asiatica dejeani
syn. nov.
,
=
M. asiatica elisa
syn. nov.
,
=
M. asiatica sigberti
syn. nov.
).
Introduction
1975). This phenotypic distinctiveness has warranted a mono-
typic tribe (Melanargiini Wheeler), but with further phylo-
genetic relationships with other Satyrines remaining unclear.
Previous molecular studies have suggested various members
of Erebiini, Maniolini or Satyrini (Martin
et al
., 2000; Yin
et al
., 2007), or the Asiatic
Orsotriaena
, Neotropical
Cyllop-
sis
or Afrotropical
Neocoenyra
(Pena
et al
., 2006) as clos-
est relatives of Melanargiini. Within
Melanargia
, three sub-
genera are often recognized:
Melanargia
(including the taxa
M. galathea
,
M. russiae
,
M. larissa
,
M. hylata
,
M. titea
,
M. syriaca
, etc.);
Argeformia
(including
M. arge
,
M. ines
,
M. occitanica
and
M. pherusa
); and
Halimede
(including
eight species confined to eastern Asia) (Oberth ur & Houlbert,
Butterflies of the genus
Melanargia
Meigen comprise 24
species distributed from Europe to the far east of Russia.
Commonly known as ‘Marbled Whites’ owing to their che-
quered black and white wing pattern, they are characterized
also by a dilated vein 12 at the base of the forewing (Higgins,
Correspondence: Vazrick Nazari, Department of Integrative
Biology, University of Guelph, Guelph, Ontario N1G2W1, Canada.
E-mail: vnazari@uoguelph.ca
Unpublished for the purposes of zoological nomenclature (Art. 8.2,
ICZN)
© 2009 The Authors
Journal compilation
©
2009 The Royal Entomological Society
1
2
V. Nazari
et al
.
1922; Verity, 1953). Disagreement over the status and place-
ment of species within these subgenera (e.g. see Wagener,
1983) exists, compounded by electrophoretic evidence, which
placed
M. arge
in a separate clade, sister to all
Melanargia
(Mensi
et al
., 1990).
Despite a shared uniform general wing pattern, species of
Melanargia
vary, particularly in the intensity of black suffusion
on the upperside of the wings. Within each species, populations
comprise very light and/or very dark individuals, and the
geographical distribution of the lighter or darker forms shows
no correlation with regional humidity or altitude (Hesselbarth
et al
., 1995). However, this variation seems to be heritable:
in a rearing experiment, the offspring of
Melanargia galathea
resembled their parents (Roos, 1983).
Structural characters used traditionally in Lepidoptera pro-
vide little insight into species boundaries in
Melanargia
.
The larvae are indistinct and show wide individual varia-
tion (Jutzeler, 1994; Jutzeler & Leestmans, 1994; Nardelli
& Giandolfo, 1994; Jutzeler
et al
., 1995, 1996; Nardelli
et al
., 1998). The commonly used genitalic characters, the
shape and number of valval terminal teeth, do vary within
some species (Higgins, 1975; Bozano, 2002). Although micro-
scopic structure and sculpture of the eggshells can be
informative (Wagener, 1983), few studies exist. The few kary-
otyped species show no informative variation (
M. lachesis
and
M. galathea n
=
absence of sympatry, these taxa have been treated predomi-
nantly as valid species, each with subspecies that often are
used interchangeably. Past morphometric studies on this group
have been inconclusive. Wagener (1983) used differences in
egg morphology as evidence for the status of 12
Melanargia
species, including the five members of the
M. larissa
group. In
a multivariate analysis on morphometric measurements from
the male uncus and valva of 307 specimens of the five species
in the
M. larissa
group, M. Ercolino & V. Sbordoni (1997,
unpublished) provided some evidence for species-level sep-
aration of
M. grumi
and
M. syriaca
, but demonstrated that
M. larissa
and
M. hylata
show very few, if any, differences
(V. Sbordoni, personal communication).
The use of poorly defined and ‘fluid’ diagnostic characters
from wing elements, egg morphology or even genitalia in
Melanargia
has negatively impacted the taxonomy of the group
and has obstructed morphological phylogenetic studies of the
genus, and a genetic analysis of
Melanargia
has been supported
before (Hesselbarth
et al
., 1995). Here, for the first time, we
use both mitochondrial and nuclear DNA sequences to infer
the phylogeny of
Melanargia
and to evaluate the robustness
of the latest species-level taxonomy for the genus (Bozano,
2002) (Supporting Information SI1) against molecular data and
corroborating information from morphology (wing-pattern and
genitalia) and geography.
24,
de Lesse, 1960;
M. titea n
=
23, Larsen, 1975;
M. larissa iran-
ica n
=
19 Lorkovic, 1977
in litt.
, Hesselbarth
et al
., 1995).
Similarly, the host preference in
Melanargia
is uninforma-
tive, as all species feed on various grasses in one family
(Poaceae).
Despite new ‘species’ of
Melanargia
having been described
as recently as 2006 (
M. sadjadii
Carbonell & Naderi, 2006),
considerable uncertainty remains over species diversity. The
eight East Asian species, although less intensively stud-
ied, seem to be well differentiated in morphology (par-
ticularly of the male genitalia) and stable in their taxo-
nomic status since Wagener (1956). The relationship and sta-
tus of European
M. galathea
,
M. lachesis
and
M. lucasi
has been debated (Higgins, 1969, 1975; Tilley, 1983, 1986;
Wagener, 1983; Mazel, 1986), but a general consensus may
exist (Bozano, 2002). The taxon
pherusa
, usually considered a
subspecies of
M. occitanica
(cf. Bozano, 2002), sometimes has
been regarded as a good species on the grounds of subtle differ-
ences between the early stages (Stauder, 1926; Jutzeler
et al
.,
1996), although enzyme electrophoresis showed very close
affinity between the two (Mensi
et al
., 1990). The taxon
meda
,
described originally as a separate species (Wagener, 1976), is
now a subspecies of
M. teneates
based on the shape of male
genitalia, the distribution ranges and the existence of interme-
diate populations (Bozano, 2002).
Another problem concerns the species boundaries and tax-
onomic relationships in a western Palaearctic complex com-
prising several closely related ‘species’, namely
M. larissa
,
M. hylata
,
M. syriaca
,
M. grumi
and
M. titea.
Despite a
lack of definitive diagnostic characters and a nearly complete
24, Lorkovic, 1941;
M. russiae n
=
Materials and methods
Taxon sampling
A total of 353 specimens from all 23 species (
sensu
Bozano,
2002) and several subspecies of
Melanargia
,aswellas
M.
sadjadii
(Carbonell & Naderi, 2006) were selected from pri-
vate collections of the authors or were received as donations
(Supporting Information SI2). The voucher data are publicly
available through the published project ‘
Melanargia of the
Wo r l d
’ (MEL) on the Barcode of Life Database (BOLD;
www.barcodinglife.com). Also included in the analysis were
11
Melanargia
sequences from GenBank (Martin
et al
.,
2000: AF214589, AF214603, AF214621; Pena
et al
., 2006:
DQ338706 – 8, DQ338843 – 5; Yin
et al
., 2007: EF545701 – 2).
Outgroups for the phylogenetic analyses were chosen from pre-
vious phylogenetic studies (Martin
et al
., 2000; Pena
et al
.,
2006) and sequences obtained from GenBank (Supporting
Information SI2).
Genitalia preparations
Previously published genitalia figures were re-examined
(Wagener, 1959 – 1961; Bozano, 2002). When molecular data
suggested a need for taxonomic revision at species or sub-
species level, selective dissections of single specimens of
Melanargia
were carried out by GCB and WtH (Supporting
Information SI3). Dissected specimens were not used in the
molecular analysis. The genitalia were fixed in Euparal glyc-
erin and photographed in lateral view with the frontal valva
©
2009 The Authors
Journal compilation
©
2009 The Royal Entomological Society,
Systematic Entomology
,
doi: 10.1111/j.1365-3113.2009.00493.x
Molecular systematics and phylogeny of the ‘Marbled Whites’
3
removed in order to have a better view of the inner side of the
other valva. The frontal valva was embedded separately. The
aedeagus sometimes had to be removed.
addition sequences. Bootstrapping of 100 replicates was con-
ducted under the parsimony criterion with the default set-
ting starting with a random seed and the tree bisection-
reconnection (TBR) branch-swapping algorithm. Bremer sup-
port values were calculated using
TREEROT
3 (Sorenson &
Franzosa, 2007). The ML trees were generated using
PHYML
online (Guindon & Gascuel, 2003), with the parameters of the
best-fit model (TIM
+
I
+
G) selected previously under
MOD-
ELTEST
3.0 (Posada & Crandall, 1998), and 100 bootstrap repli-
cates. Haplotype diagrams were constructed in
TCS
1.21, with a
95% confidence limit for parsimony (Templeton
et al
., 1995).
Shorter fragments of
cox1
barcodes or those with ambiguous
bases were excluded from haplotype analyses.
Molecular techniques
Two dry legs from each adult specimen were detached and
stored in individual vials. The extraction of total genomic
DNA, amplification and sequencing were performed in the Bio-
diversity Institute of Ontario using previously described pro-
tocols (Hajibabaei
et al
., 2005). Initially, full-length mtDNA
barcode sequences (658 bp) were obtained for nearly all
specimens, and, based on results from sequence similar-
ity (neighbour-joining) analyses and the quality of DNA, a
subset was selected for additional gene sequencing. Failed
samples were targetted for smaller fragments of
cox1
(132
bp) using mini-barcode primers and protcols described pre-
viously (Meusnier
et al
., 2008). Ribosomal 16S
rRNA
and
nuclear wingless (
wg
) genes were also obtained using primers
and protocols described previously (Brower & DeSalle, 1994;
Aubert
et al
., 1999). Amplified DNA from all specimens
was sequenced in both directions for each gene, and final
sequencing products were run on an ABI 3730
®
DNA ana-
lyzer (Applied Biosystems, Foster City, CA). Complementary
strands were assembled into contigs and edited manually, and
primers were removed using S
EQUENCHER
4.5 ( Gene Codes
Corporation, Ann Arbor, MI). Sequences were aligned using
C
LUSTALX
2.0 (Thompson
et al
., 1997), evaluated by eye and
converted to Nexus using
SE-AL
2.0a11 (Rambault, 2002). New
sequences were deposited in GenBank, and accession num-
bers are given in Supporting Information SI2.
cox1
barcode
sequences are also available publicly through the published
project ‘
Melanargia of the World
’ (MEL) on the Barcode of
Life Database (BOLD; www.barcodinglife.com).
Results
No insertions or deletions were observed in mitochondrial
cox1
and nuclear
wg
genes. Several small indels were present
in 16S
rRNA
sequences, but the C
LUSTAL
X alignments for
this gene were unambiguous. These indels were treated as
gaps. Gene partitions were found to be homogenous (sum
of lengths for original partition
=
1210,
P
=
0
.
01); however,
to avoid any undetected incongruence, analyses were also
conducted independently on each partition. Nucleotide base
frequencies differed significantly in
cox1
and 16S but not
in
wg
, and parsimony-informative characters in
cox1
and
wg
were typically in the third codon positions. The combined
dataset of selected taxa had 1582 characters, of which 332
were parsimony-informative (Table 1).
A neighbour-joining tree of barcode haplotypes is given in
Fig. 1, with bootstrap and Bremer values given for supported
nodes. Species-level identification of several samples was
revisited by re-examination of vouchers following barcode
results. Some identification errors were corrected, including
one specimen of
M. larissa
identified originally as
M. russiae
(DNAwthmel 026), and one specimen of
M. teneates
identified
originally as
M. evartianae
(DNAwthmel 041). Phylogenetic
hypotheses for each of the three genes, as well as for the
combined dataset, are presented in Fig. 3. The ML analysis for
the combined data (Fig. 3d) recovers three previously proposed
subgenera,
Melanargia
,
Argeformia
and
Halimede.
Sequence data analysis
Neighbour-joining (NJ) trees for barcode data were con-
structed initially using the
QUICKTREE
algorithm (Howe
et al
.,
2002) and under the Kimura two-parameter (K2P) model
(Kimura, 1980). Additional NJ and maximum parsimony (MP)
analyses was conducted in
PAUP
*4.0
β
10 (Swofford, 2003).
Heuristic searches for MP analysis were carried out with
all characters equally weighted and under the tree bisection-
reconnection (TBR) swapping algorithm with 100 random
Clade 1: subgenus Argeformia
Our results do not support
M. arge
as sister to all
Mela-
nargia
, as suggested previously (Mensi
et al
., 1990). Instead,
Table 1.
Summary of sequence statistics.
Informative sites
Empirical base frequencies (%)
Length
Variable
Informative
First
Second
Third
Gene
(bp)
sites
sites
codon
codon
codon
A
C
G
T
cox1
658
228
187
28
2
157
0.2967
0.16335
0.14579
0.39416
barcodes
16S
521
97
68
–
–
–
0.42831
0.08062
0.12588
0.36519
wg
404
140
77
8
4
65
0.19896
0.32938
0.32601
0.14565
2009 The Authors
Journal compilation
©
2009 The Royal Entomological Society,
Systematic Entomology
,
doi: 10.1111/j.1365-3113.2009.00493.x
©
4
V. Nazari
et al
.
Fig. 1.
Neighbour-joining tree of
cox1
barcodes of
Melanargia
(including GenBank sequences) examined in this study. Numbers in parentheses
represent number of individuals in each cluster; numbers on branches are bootstrap values from maximum parsimony analysis followed by Bremer
support values. The four-letter codes correspond to species names. The ‘
larissa
’and‘
halimede
’ haplogroups each represent multiple species, with
the complete haplotype complements presented in Fig. 2. Haplotypes in bold are those used in the phylogenetic analysis (Fig. 3).
©
2009 The Authors
Journal compilation
©
2009 The Royal Entomological Society,
Systematic Entomology
,
doi: 10.1111/j.1365-3113.2009.00493.x
Molecular systematics and phylogeny of the ‘Marbled Whites’
5
M. arge
appears to be closely related to
M. occitanica
,which
together with
M. ines
forms a sister-clade to all other
Melanar-
gia
(Fig. 3d). Specimens of
arge
from Puglia and Campania
(Italy) show limited (
cox1
,
wg
) or no (16S) genetic divergence
(Supporting Information SI4).
We found a deep split in the
M. ines
clade separating the
African specimens (haplotype INES4) from those in mainland
Europe (Fig. 1). This gap was more pronounced in the
cox1
barcodes (4.4%) than in 16S (1.4%) or
wg
(1.3%) genes.
Examination of genitalia showed some structural differences
between these two lineages: in Spanish populations, the distal
tip of the valva is rounded and there are four teeth of similar
size arranged in a row, whereas in the population from High
Atlas in Morocco there are five teeth on the valva and the
ventral tooth is much larger than the others (Fig. 4a, b).
Average genetic variation within these two clades was low,
with the exception of 16S among African populations (1.8%),
where individuals of the previously synonymized taxon
M. ines
ssp.
jahandiezi
(
n
=
3) clustered very distant from other
ines
(Fig. 3b). Their genitalia also showed differences, with several
smaller teeth instead of a single fifth tooth at the lower edge
of the valva (Fig. 4c).
As with
M. ines
,in
M. occitanica
we observed two dis-
tinct lineages, one consisting of specimens of the nominal
subspecies from France and Spain (OCCI5 – 6) and a second
comprising ssp.
pherusa
from Italy (OCCI4) together with all
African populations (OCCI1 – 3). The genitalia of specimens in
these two groups are very similar: the distal tip of the valvae in
both populations has three or four teeth arranged on a common
base (Fig. 4d, e). Within the African clade, the taxon
pherusa
stands out with a notably longer branch length (Fig. 1).
Iran, and another comprising all other populations from Europe
to Central Asia (RUSS1 – 9). In the second
M. russiae
clade,
two specimens of ssp.
cleanthe
(RUSS9) were somewhat
distant from others; however, despite a wide geographic
coverage, variation within the remaining
russiae
was relatively
small. Similarly,
M. parce
divided into two distinct groups, one
of the nominal subspecies (PARC) and a second comprising
ssp.
lucida
(LUCI). In our phylogenetic reconstructions, the
taxon
lucida
was always sister to all (except for
wg
,for
which no
lucida
sequences could be obtained); the nominal
M. parce
were closely related to the NNE Iran
M. russiae
(sspp.
transcaspica
and
eberti
), and these together were sister
to all other
M. russiae.
The genitalia in this group show
clear differences that coincide with their genetic divergence:
specimens of
lucida
have about 20 small teeth on the distal
end of the valva, but in the nominal
M. parce
there are only
about eight to nine claw-like teeth on the end of the valva,
in a somewhat different arrangement. In nominal
M. russiae
,
these claw-like teeth number about 11, and the ventral claw
is the largest, whereas the taxon
eberti
has about 7 shorter
teeth, which are less bent than in
M. russiae
, and the taxon
transcaspica
has about 16 teeth, which in size and form are
similar to those of
eberti
. The aedeagus in
M. russiae
is shorter
than in the taxa
eberti
and
transcaspica
(Fig. 4k – p).
The genetic divergences between
M. sadjadii
(EVAR1) and
M. evartianae
(EVAR2), and between nominal
M. teneates
(TENE1, 3) and
M. teneates
ssp.
meda
(TENE2) were rela-
tively small. The genitalia of taxa
teneates
and
meda
have
minor differences. The aedeagus of
teneates
is somewhat larger
and has a bigger diameter; the valvae are triangular with about
10 small, straight teeth on the distal tip. In
meda
,thereare
up to 15 even smaller teeth on the valva. Specimens of
M.
teneates
from Gilan (TENE3) are divergent from others, but
the genitalia are similar and no other reliable morphological
character corroborated this pattern (Fig. 4q – s).
Among the 152 specimens in the
M. larissa
group (LARI)
(including
M. hylata
,
M. larissa
,
M. syriaca
,
M. grumi
and
M. titea
) and throughout the genes examined, with a few
exceptions we found no significant genetic differentiation
among populations of these taxa, and the exceptions were
not fully congruent with the current taxonomic arrangements
(Figs 1 – 3). Among the 128 full-length
cox1
barcode sequences
included in the haplotype analysis, the most common haplotype
shared between
M. hylata
,
M. larissa
,
M. grumi
,
M. syriaca
and
M. titea
(LARI01) occurred in 31 specimens (Fig. 2).
The taxon
wiskotti
(WISK,
n
=
7), normally recognized as
a subspecies of
M. titea
, was notably divergent from others
(Figs 1 – 3). Individuals of
wiskotti
are larger than nominal
M.
titea
but their genitalia are about the same size; there are
8 to 10 teeth with a common base on the end of valvae,
which are only slightly bent. In the nominal
M. titea
,these
teeth are arranged more regularly, in rows, whereas they are
arranged irregularly on a common base in the taxon
wiskotti
(Fig. 4t – x). Other distinct clusters included populations of
M.
larissa
from Greece (ssp.
larissa
, LARI24 – 29) and central
Turkey (ssp.
taurica
+
massageta
,LARI31–34),aswellas
M. hylata
from southern Iran (ssp.
iranica
, LARI17 – 22).
Clade 2: subgenus Melanargia
Our results show unambiguously that the taxon
lucasi
is
a well-differentiated species, distinct from
M. galathea
and
M. lachesis
(Figs 1, 3). The genitalia of
M. galathea
,
M. lach-
esis
and
M. lucasi
are more similar to one another than to
other species in the subgenus, although
M. lucasi
has a smaller
uncus and a broader triangular valva, which ends in five or
six teeth (as opposed to seven to eight in
M. galathea
and
M. lachesis
) on a common base at the pointed distal end
(Fig. 4f, g). The genetic variation within the Moroccan pop-
ulations is small; however, our single specimen from Tunisia
(LUCA3) is notably divergent. The genitalia of the Tunisian
M. lucasi
are very similar to those of the Moroccan specimens
both in form and number of valval teeth, although the wing
patterns show certain differences (cf. Tennent, 1996).
We observed consistent genetic difference between
M. galathea
and
M. lachesis
(
cox1
:0
.
70
±
0
.
10%; 16S:
0
.
34
±
0
.
22%;
wg
:2
.
59
±
0
.
64%). The average intra-specific
variation in these two species was also higher in
wg
than
in mitochondrial genes, a phenomenon that warrants further
scrutiny.
In all reconstructions,
M. russiae
was divided into two
distinct groups: one that included specimens from north (ssp.
eberti
, TRAN1) and north-eastern (ssp.
transcaspica
, TRAN2)
2009 The Authors
Journal compilation
©
2009 The Royal Entomological Society,
Systematic Entomology
,
doi: 10.1111/j.1365-3113.2009.00493.x
©
[ Pobierz całość w formacie PDF ]