Phylogenetic Positions of the Enigmatic Asiatic Fern Genera Diplaziopsis and Rhachidosorus from Analyses of Four Plastid Genes

2011·7.7 MB·English

by  Li Chunxiang

Phylogenetic Positions of the Enigmatic Asiatic Fern
Genera Diplaziopsis and Rhachidosorus from Analyses
Four Genes
of Plastid
Chunxiang
Li
State Key Laboratory of Palaeobiology and Stratigraphy of Nanjing Institute of Geology and
Academy
Palaeontology, Chinese of Sciences, Nanjing 210008, China
Lu Shugang
Institute of Ecology and Geobotany, Yunnan University, Kunming 650091, China
Sun Xiaoyan and Yang Qun
State Key Laboratory of Palaeobiology and Stratigraphy of Nanjing Institute of Geology and
Academy
Palaeontology, Chinese of Sciences, Nanjing 210008, China
The
fern family Woodsiaceae, as circumscribed in the most recent familial
classification (Smith et ah, 2006), comprises about 15 genera and more than
700 species distributed mainly from America Old World temperate,
tropical to
which by monomorphic monomorphic and
characterized or nearly leaves
is
vascular anatomy (Tryon and Tryon, 1982). The family exhibits an extensive
chromosome
dysploid series of base numbers, ranging from 31 to 42, mostly
=
X
40, 41, also 31 [Hemidictyum], 33, 38, 39 [Woodsia], and 42 [Cystopteris]
(Smith The monophyly Woodsiaceae Smith
et al, 2006). of of et (2006)
al. is
lacking in all broad analyses (Hasebe et al., 1995; Sano et al., 2000; Schneider
2004; Schuettpelz and Pryer, 2007). The more recent fern global
et
al.,
showed Woodsiaceae Smith
phylogenetic analyses of et (2006) consists of
al.
and Gymnocarpium
four well-supported clades: together, Cystopteris are
sister
eupolypods Hemidictyum and
to the rest of sister to the asplenioid ferns;
II; is
Woodsia and
sister to a large clade of onocleoid, blechnoid, athyrioid ferns
is
and most
(Schuettpelz Pryer, 2007). This the inclusive analysis of
is
leptosporangiate fern relationships conducted to date, in which three plastid
genes (rbcL, atpA, atpB] from 400 leptosporangiate fern species were utilized.
However, Homalosorus and
the taxomically problematic genera Diplaziopsis,
CHUNXIANG
ET PHYLOGENETICS OF AND
AL.: DIPLAZIOPSIS RACHIDOSORUS
Rhachidosorus were
not included
in this groundbreaking study, and the
phylogenetic
affinities of these taxa are unclear. Smith
et (2006) tentatively
al.
placed
these genera Woodsiaceae and
in the
suggested, "further sampling
will
likely shed additional on and
light this subject, the recognition of several
may
additional families be
w^arranted."
Diplaziopsis a problematic genus, which has undergone many
is
systematic
Bower
changes.
(1928) treated provisionally along with "Asplenoid
his
it
while
ferns", Christensen
(1938) revised his opinion to give a generic status
it
and
treats as a group of ferns with areolate veins under Diplazium
Sw..
Copeland
(1947) defined Diplaziopsis by pinnate leaves, thin lamina
its
texture, anastomosing and
veins, sausage-like comprising two
sori, species
(D.
and
javanica D. cavaleriana). Ching added two new
(1964a) species from
China to Diplaziopsis, which were emended
later into three species in the flora
China
of (i.e., D. javanica, D. cavaleriana, and D. brunoniana) (Chu
et al,
1999). In geographical
distribution, Diplaziopsis an Old World
essentially
is
genus mainly from and
Eastern Southeastern Asia
(Ching, 1964a).
The
monotypic genus Homalosorus was
by
established Pichi Sermolli
(1973)
with H. pycnocarpos
distributed only in the temperate eastern North America,
but other authors have included in the genus Athyrium Roth (Kramer and
it
Kato, 1990) or the genus Diplazium (Tryon and
Tryon, 1982; Kato and
A
Iwatsuki, between
1983). relationship H. pycnocarpos and
Diplaziopsis
(Tryon and Tryon, Kato and
1973; Iwatsuki, 1983) has been suggested from
pinna
their similar shape, rachis-grooving, and
indusia, spores, although they
have lamina
different apex, venation, and chromosome numbers
and
(Kato
Darnaedi, The
1988; Price, 1990). sister relationship between H. pycnocarpos
and
Diplaziopsis has been supported by rbcL (Sano and
2000) rbcL+rps4
et
al.,
(Wei
trees et al, 2010), while a previous trnL-F study support
lent to the
placement
of Diplaziopsis cavaleriana Diplazium (Wang
in
et 2003).
al.,
Previously, plants here recognized as Rhachidosorus have been included
in
Athyrium
either (Tagawa, 1936) or Diplazium Kramer and
(Kato, 1977;
Kato,
Ching
but
1990), (1964b) later separated those plants into the genus
Rhachidosorus from South-east Asia, and determined genus
that the consists
The
of eight species. genus Rhachidosorus from both Athyrium and
differs
Diplazium
(or rather Allantodia R. Br.; most Diplazium species in China have
been placed in Allantodia] in having thick creeping rhizomes,
the scales near
the base of stipe and, above in the narrow semilunate and
all, sori indusia of
which
the asplenioid type, are never diplazioid nor and
athyrioid, in the spore
morphology
(Ching, 1964b). Based on previous rbcL (Sano and
2000)
et
al.,
(Wang
frnL-F analyses Rhachidosorus
et al., 2003), does not cluster with either
Athyrium Diplazium
or but occupies
a position isolated from the other taxa
in
eupolypods such was
the a conclusion suggested by Ching
II; (1964b; 1978).
DNA
we
more
In this study, use sequence data from four plastid genes
[rbcL,
make
atpB, atpA, rps4) to comparisons with the previous
studies,
to
the phylogenetic
investigate relationships of Diplaziopsis, Homalosorus, and
Rhachidosorus other eupolypods and
to ferns of
specifically address
to
II,
whether they belong within
athyrioid
ferns.
AMERICAN VOLUME NUMBER
FERN
JOURNAL:
101
3 (2011)
Materials and Methods
—
Taxon make
Sampling. In order to comparisons with other studies (Sano
et
we
and
al, 2000; Schuettpelz Pryer, 2007), assembled three data matrices
(Table of which included newly generated sequences and sequences
all
2),
A
new
obtained from GenBank. total of 98 sequences were generated
for this
study, the corresponding voucher specimens have been deposited
in the
Herbarium Yunnan
of the University (PYU). Taxa, vouchers, and accession
numbers are provided in Table The data matrix consisted of 59 rbcL
1. first
sequences, of which 24 were newly generated. The second matrix comprised
rhcL, atpB and atpA sequences of 59 taxa, which included 22 atpB and 20 atpA
sequences newly
generated in this study plus additional sequences from
GenBank. The
matrix comprised atpA and
third rbcL, atpB, rps4 sequences of
which
59 taxa, included the three sequences of 59 taxa from the second matrix
and newly
32 rps4 generated in this study plus additional sequences from
GenBank. Those taxa with incomplete sequences were included
in the
analyses of the combined data, and the unsequenced fragments were coded
missing
as data. In order to investigate the phylogenetic relationships of
Diplaziopsis, Homalosorus, and Rhachidosorus other genera, our sampling
to
included 14 of 15 recognized genera in Woodsiaceae as treated by Smith et
al.
The two
(2006). previously unincluded Asian genera in the study of
and
Schuettpelz Fryer Diplaziopsis and Rhachidosorus,
(2007), are represent-
ed by two or more species, with each species represented by one or more
we
specimens. examined
In addition, species of Aspleniaceae, Blechnaceae,
Onocleaceae, and Thelypteridaceae, which were included with
the
all
Woodsiaceae in the eupolypods clade of Smith Following the
II et al. (2006).
previously published molecular
systematic studies of leptosporangiate ferns
and which
(Schuettpelz Fryer, 2007), in well established eupolypods
that
is
it
we
most
eupolypods
II is closely related to selected Drynaria rigidula,
I,
Dryopteris uniformis and Polypodium vulgare outgroups.
as
—
DNA
DNA
extraction, gene amplification, and sequencing. Total genomic
was CTAB
from
extracted 2 g of fresh or 1 g of silica gel dried leaves using the
DNA
procedure and The
(Doyle Doyle, 1987). selected regions were amplified
with standard polymerase chain reaction (FCR). The protocols used amplify
to
four genes were identical and followed Li et (2004). For information on
al.
amplification and sequencing primers, see Table
3.
—
Sequence The
analysis. obtained sequences have been assigned GenBank
numbers
accession (Table Alignments of sequences were performed
all
1).
X
(Thompson
using Clustal et 1997) and subsequently edited manually in
al.,
BioEdit There were no
(Hall. 1999). insertions or deletions (indels) in the
protein-coding sequence alignments. Indels were introduced into the align-
ment which
of rps4-trnS spacer region, in ambiguously aligned regions were
excluded from all analyses. Fhylogenetic analyses were investigated by
maximum maximum
parsimony and
(MF), likelihood (ML), Bayesian inference
FAUF* FHYML
methods
(BI) in 4.0bl0 (Swofford, 2002), 3 (Guindon and
2. 4.
MP
Gascuel, 2003), and MrBayes 2 (Ronquist and Huelsenbeck, 2003). For
3. 1.
CHUNXIANG
ET PHYLOGENETICS OF AND
AL. DIPLAZIOPSIS RACHIDOSORUS
145
:
analysis, unweighted analyses were performed by
heuristic searches with
tree-
bisection-reconnection (TBR) branch swapping, MulTrees
the
in
effect,
steepest descent off using 1000 random taxon-addition and one
replicates,
tree held at each step during stepwise addition. Bootstrap analyses
(Felsenstein, 1985) were conducted examine
to the relative level of support
for individual clades on the cladograms of each search (MPBS), using 500
bootstrap and same
replicates the search procedure
tree as described above.
ML
For the and BI analyses, the model sequence
best-fitting of evolution
for
each was
data identified with the Akaike Information Criterion in Modeltest
3.07 (Posada and Crandall, The SYM+I+G model was
1998).
selected for the
rbcL data and the GTR+I+G model was combined
set, selected for the data sets
Once ML
(Table the best sequence evolution model was determined,
2). the
was
analysis performed each
for data the parameters such
set, as base-
composition, Gamma-shape, and
were
ratio of invariable sites also estimated
ML
ML
during each analysis. Nodal robustness on the was estimated by
tree
the nonparametric bootstrap (500 replicates, MLBS). BI was conducted using
MrBayes
3.1.2 with appropriate evolutionary models determined
as described
We
above and
two
the default priors. ran concurrent analyses, each with four
chains Markov Monte
of the chain sampling one
Carlo, every 100
tree
X
generations of 2 1,000,000 generations, starting with a random The
tree. first
25%
of the samples (5000 trees) were discarded as "burn-in". At this point, the
standard deviation of split frequencies was <0.01, indicating convergence
that
to a stationary distribution had been achieved. The
posterior probability
(PP)
was
used
estimate nodal
to robustness.
Results
The
alignment length and number
the of included characters for the three
data sets are presented in Table The aligned rbcL matrix contained 1308
2.
ML
which
characters, of 417 were MP. and
variable. BI analyses of rbcL matrix
resulted in nearly identical topologies, with several minor differences the
at
genus
level (results not shown). Strong support was lacking along the
backbone of the rbcL tree. The 50% majority-rule consensus revealed
tree that
eupolypods
II fall into nine hneages (Fig. athyrioids (Woodsiaceae
1):
I),
Blechnaceae + Onocleaceae, Woodsiaceae
[Woodsia, Prowoodsia and
II
Cheilanthopsis), Thelypterdiaceae, Woodsiaceae
[Cystopteris, Acystopteris
III
and Gymnocarpium],
Woodsiaceae IV [Diplaziopsis and Homalosorus,
in
shadow V
in Fig. Woodsiaceae {Rhachidosorus, shadow
1), in in Fig.
1),
Aspleniaceae, and Woodsiaceae VI [Hemidictyum].
In the rbcL four
tree, all
Rhachidosorus specimens
united two
are in a single clade; species of
Diplaziopsis form another monophyletic clade with Homalosorus;
the three
genera Woodsia, Prowoodsia and
of Cheilanthopsis (Woodsiaceae
are
II)
united and
in a single clade; the three genera of Cystopteris, Acystopteris and
Gymnocarpium
(Woodsiaceae
are united in another one. All four clades
III) are
isolated from other genera in the family. Hemidictyum in Woodsiaceae
is sister
= MPBS <
to Aspleniaceae with low support (PP 0.90 and MLBS. 50%);
the
m
m
i|!
I
i
i!
I
11
ill
ii! ilii
i!
!
i
ii !i
! 1
ii
ill iiii
i i
.1-
ill
ii
1 ilill
!l 1!
i
i
iiiil
i
!
ili
liiiiiii
!! !i i!
iil
ili
ii
li III ill!
m
illliiifllllii
ii I!
51-11
!

CHUNXIANG
ET PHYLOGENETICS OF DIPLAZIOPSIS AND RACHIDOSORUS
AL.:
Table 2. Statistics for the Three Data Sets Analyzed in This Study.
SYM+I+G
GTR+I+G
GTR+I+G
athyrioids clade resolved in rbcL but support
is trees, for this relationship
is
= MPBS <
very low (PP 0.53 and MLBS, 50%).
The combined rbcL, atpB and atpA data matrix included 4092
characters,
v\^ith 1045 characters that were variable. The results from the three combined
showed
sequences better resolved and supported and
inter- intra-familial
relationships than that of rbcL not shown), Woodsia-
tree (results especially,
ceae IV [Diplaziopsis and Homalosorus, shown shadow and
in in Fig.
1)
V
Woodsiaceae shown shadow
[Rhachidosorus,
in in Fig. the focus of our
1),
= MLBS=65%, MPBS =
study, are united in a single clade (PP 1.00, 71%), so
eupolypods
the trees reveal that fall into eight lineages.
II
The combined
four data matrix atpB, atpA and included 5227
[rbcL, rps4]
ML
characters, with 2127 characters were MP, and
that variable. BI analyses
from combined
the four sequences resulted in nearly identical topologies, with
most
differences the support Because
at statistical values. the resultant
topologies for relationships of eupolypods from each of the datasets were not
II
with one
in conflict another, the phylogenetic relationships presented here
are
based on analyses of the four combined data The 50% majority-rule
set.
ML
consensus tree resulting from MP, and BI analyses of the four combined
shown
sequences data These
set is in Fig. analyses yielded an almost robust
1.
phylogeny with the exception of a few nodes. Together, Hemidictyum and
Aspleniaceae are sister to the rest of eupolypods Hemidictyum
sister to
is
II,
DNA
Table Primers Used Amplifying and Sequencing
3. for of This Study.
ATGTCACCACAAACGGAGAC
rbcLFl
CCA TTC ATG CGY TGG AGA
rbcLF631 G
CTC TCC ARC GCA TGA ATG
rbcLR631 G
rbcLR1369 GGA CTC CAC TTA CWA GCT TC
atpBFwood ATG AGT GCC AC A GAC GG
CCA GGA AGA ATC ATT TG
atpBRwood
AtpARF200 GAA TCK GAT AAT GTT GGG
CAG CCA CCT GTT TCA TAG
AtpARll40
C
ATG TCC CGT TAT CGA GGA
Rps4F CC
TAC CGA GGG TTC GAA TC
trnSR

CHUNXIANG
ET AL.: PHYLOGENETICS OF DIPLAZIOPSIS AND RACHIDOSORUS
151
the asplenioid ferns; then the clade of Woodsiaceae IV and Woodsiaceae V
[Rhachidosorus, Diplaziopsis and Homalosorus, shadow
in in Fig. the
1);
clade Woodsiaceae and
of the large clade of athyrioids, Blechnaceae +
III;
Onocleaceae, Woodsiaceae and
Thelypterdiaceae.
II,
Discussion
Phylogenetic
relationships eupolypods comparisons
of with previous
II,
studies.—
Generally, our phylogenetic
results are compatible with previous
on among
studies the relationships fern genera in eupolypods The
overall
II.
eupolypods shown
relationships
II in Fig. 1 are not in conflict with the results
Sano Our
of
phylogenetic
et al. (2000). analyses of multiple chloroplast genes
confirmed those resuUs and showed
and
better resolved supported and
inter-
intra-familial relationships than that of the rbcL With more
extensive
tree.
sampling
Chinese
of Woodsiaceae, a relationship between
sister Diplaziopsis,
Rhachidosorus and
the North American Homalosorus was
moderately
supported by
the four chloroplast gene and
data, the three genera of
Acystopteris, and Gymnocarpium
Cystopteris, were monophy-
resolved
as a
with
letic lineage strong support. Both were
statistical early diverging lineages
We
eupolypods and away
in from mapped
II, far athyrioid ferns. the three
enigmatic genera
[Diplaziopsis, Rhachidosorus, and Homalosorus]
onto the
Schuettpelz and
Fryer
(2007) global fern phylogenetic framework. However,
there is a major point of difference between our study and
theirs; their
phylogeny found
the clade of Cystopteris and Gymnocarpium
the
sister to rest
we
eupolypods
of while found
the asplenioid clade (including Hemi-
II,
dictyum]
sister to the rest of eupolypods the most basal-most lineage of
II, i.e.,
eupolypods
II is the clade of asplenioid ferns. It is possible that the sampling of
markers
different or their combinations caused the topological difference.
DNA
Plastid is inherited as an intact unit, and differences between
trees
constructed from
separate regions can be due and
to functional constraints
We
evolution (Wendel and
rates Doyle, 1998). can correct both by
for factors
combining
directly these separate because combined
regions, analyses
confidently resolved the conflicts between the single gene analyses, enhanced
phylogenetic and were
resolution, supported by
better morphological
information (Gontcharov
et 2004).
al.,
Phylogenetic and
relationships Homalosorus.—
Diplaziopsis
of
Cladisiic
analysis of four plastid gene [rbcL, atpA, atpB and rps4) sequences provided
strong evidence that Diplaziopsis and Homalosorus form monophyletic
a
and
lineage are clearly separated from Diplazium. The
relationship agrees with
the results of rhcL analyses (Sano and
et 2000) a recent study based on
al.
rbcL+rps4
analyses (Wei et 2010). Diplaziopsis and Homalosorus were
al.,
members
formerly treated as of Diplazium (Christensen, 1938; Kato, 1977; Kato
and
Iwatsuki, 1983). with which they shared such
features as linear
sori,
similar stipe base and frond axes (Kato, 1977). Diplaziopsis and Homalosorus
differ from Diplazium by lamina simple
to once-pinnate, veins anastomosing