Generic phylogeny, historical biogeography and character evolution of the cosmopolitan aquatic plant family Hydrocharitaceae.

Abstract

Hydrocharitaceae is a fully aquatic monocot family, consists of 18 genera with approximately 120 species. The family includes both fresh and marine aquatics and exhibits great diversity in form and habit including annual and perennial life histories; submersed, partially submersed and floating leaf habits and linear to orbicular leaf shapes. The family has a cosmopolitan distribution and is well represented in the Tertiary fossil record in Europe. At present, the historical biogeography of the family is not well understood and the generic relationships remain controversial. In this study we investigated the phylogeny and biogeography of Hydrocharitaceae by integrating fossils and DNA sequences from eight genes. We also conducted ancestral state reconstruction for three morphological characters.
Phylogenetic analyses produced a phylogeny with most branches strongly supported by bootstrap values greater than 95 and Bayesian posterior probability values of 1.0. Stratiotes is the first diverging lineage with the remaining genera in two clades, one clade consists of Lagarosiphon, Ottelia, Blyxa, Apalanthe, Elodea and Egeria; and the other consists of Hydrocharis-Limnobium, Thalassia, Enhalus, Halophila, Najas, Hydrilla, Vallisneria, Nechamandra and Maidenia. Biogeographic analyses (DIVA, Mesquite) and divergence time estimates (BEAST) resolved the most recent common ancestor of Hydrocharitaceae as being in Asia during the Late Cretaceous and Palaeocene (54.7-72.6 Ma). Dispersals (including long-distance dispersal and migrations through Tethys seaway and land bridges) probably played major roles in the intercontinental distribution of this family. Ancestral state reconstruction suggested that in Hydrocharitaceae evolution of dioecy is bidirectional, viz., from dioecy to hermaphroditism, and from hermaphroditism to dioecy, and that the aerial-submerged leaf habit and short-linear leaf shape are the ancestral states.
Our study has shed light on the previously controversial generic phylogeny of Hydrocharitaceae. The study has resolved the historical biogeography of this family and supported dispersal as the most likely explanation for the intercontinental distribution. We have also provided valuable information for understanding the evolution of breeding system and leaf phenotype in aquatic monocots.

Full text

Page 1

RESEARCH ARTICLEOpen Access
Generic phylogeny, historical biogeography and
character evolution of the cosmopolitan aquatic
plant family Hydrocharitaceae
Ling-Yun Chen1,2,3, Jin-Ming Chen1,2, Robert Wahiti Gituru4and Qing-Feng Wang1,2*
Abstract
Background: Hydrocharitaceae is a fully aquatic monocot family, consists of 18 genera with approximately 120
species. The family includes both fresh and marine aquatics and exhibits great diversity in form and habit including
annual and perennial life histories; submersed, partially submersed and floating leaf habits and linear to orbicular
leaf shapes. The family has a cosmopolitan distribution and is well represented in the Tertiary fossil record in
Europe. At present, the historical biogeography of the family is not well understood and the generic relationships
remain controversial. In this study we investigated the phylogeny and biogeography of Hydrocharitaceae by
integrating fossils and DNA sequences from eight genes. We also conducted ancestral state reconstruction for
three morphological characters.
Results: Phylogenetic analyses produced a phylogeny with most branches strongly supported by bootstrap values
greater than 95 and Bayesian posterior probability values of 1.0. Stratiotes is the first diverging lineage with the
remaining genera in two clades, one clade consists of Lagarosiphon, Ottelia, Blyxa, Apalanthe, Elodea and Egeria;
and the other consists of Hydrocharis-Limnobium, Thalassia, Enhalus, Halophila, Najas, Hydrilla, Vallisneria,
Nechamandra and Maidenia. Biogeographic analyses (DIVA, Mesquite) and divergence time estimates (BEAST)
resolved the most recent common ancestor of Hydrocharitaceae as being in Asia during the Late Cretaceous and
Palaeocene (54.7-72.6 Ma). Dispersals (including long-distance dispersal and migrations through Tethys seaway and
land bridges) probably played major roles in the intercontinental distribution of this family. Ancestral state
reconstruction suggested that in Hydrocharitaceae evolution of dioecy is bidirectional, viz., from dioecy to
hermaphroditism, and from hermaphroditism to dioecy, and that the aerial-submerged leaf habit and short-linear
leaf shape are the ancestral states.
Conclusions: Our study has shed light on the previously controversial generic phylogeny of Hydrocharitaceae. The
study has resolved the historical biogeography of this family and supported dispersal as the most likely explanation
for the intercontinental distribution. We have also provided valuable information for understanding the evolution
of breeding system and leaf phenotype in aquatic monocots.
Keywords: Hydrocharitaceae, Phylogeny, Historical biogeography, Dispersal, Vicariance, Morphological character
evolution
* Correspondence: qfwang@wbgcas.cn
1CAS Key Laboratory of Aquatic Botany and Watershed Ecology, Chinese
Academy of Sciences, Wuhan 430074, Hubei, P. R. China
Full list of author information is available at the end of the article
Chen et al. BMC Evolutionary Biology 2012, 12:30
http://www.biomedcentral.com/1471-2148/12/30
© 2012 Chen et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

Page 2

Background
Hydrocharitaceae is a fully aquatic monocot family, con-
sists of 18 genera and approximately 120 species [1,2]
with a cosmopolitan distribution. The group is an
important component of aquatic ecosystems and occurs
in both freshwater and marine habitats. It also exhibits
great diversity in form and habit including annual and
perennial life histories; submersed, partially submersed
and floating leaf habits, and linear to orbicular leaf
shapes [1]. It includes a number of valuable aquarium
plants and some species (e.g., Hydrilla verticillata and
Elodea canadensis) provide fish breeding sites and fod-
der for the poultry raising industry [3]. Owing to habitat
destruction and unprecedented increase in commercial
shipping activities, several species (e.g., Ottelia alis-
moides and Blyxa japonica) are threatened, while some
other species (e.g., Hydrilla verticillata) have become
invasive weeds of great concern [4]. Similar to many
aquatic plants, the group displays considerable conver-
gent adaptations and morphological reduction, which
make the study of phylogenetics and taxonomy of the
group difficult [1,5-7]. There exists little consensus in
classification systems of the group based on morphologi-
cal characters [1,8].
Although molecular phylogenetic studies of Hydro-
charitaceae have created consensus on the relationships
between certain genera, disagreements on the relation-
ships of other genera still exist. The seagrass subclade
which includes Halophila, Enhalus and Thalassia was
resolved as sister to the subclade comprising Najas,
Hydrilla and Vallisneria by analyses using rbcL+ matK
[8] and rbcL+ cob + atp1 [9]. However, the seagrasses
was resolved as sister to the subclade comprising Necha-
mandra, Vallisneria and Maidenia by analyses using
morphological + rbcL+ matK+ trnK intron + ITS [1].
Similarly, Hydrilla, which was reported to be closely
related to Vallisneria by analyses using rbcL, matK [2,8],
rbcL, cob and atp1 [9], was reported to be closely
related to Najas by analyses using morphological char-
acter + rbcL+ matK+ trnK intron + ITS [1]. Further-
more, Stratiotes has been resolved as sister to the rest of
this family by analyses using rbcL [9] and mitochondrial
genes (nad5, cob, ccmB, mtt2, atp1) [9,10]. The position
is not in agreement with the results of two other studies
using chloroplast and nuclear sequences [1,2]. In addi-
tion, a single species was selected for each genus in pre-
vious studies. The phylogeny of this family could be
improved by a better sampling of taxon and DNA
sequence, and perhaps a careful outgroup selection [1].
The divergence time of Hydrocharitaceae is still a sub-
ject of debate, and two competing ages (one much more
recent than the other) have been proposed. Janssen and
Bremer (2004) [11] placed the crown node age of this
family in the Late Cretaceous (75 Ma) by analyses using
rbcL and fossil calibrations. Notably, in that study the
family was represented by only 16 terminals and there
was no internal calibration point. This could be
improved by better sampling and by incorporating inter-
nal fossil calibration points [12]. Kato et al (2003) [13]
dated the seagrasses within Hydrocharitaceae at 119 ±
11 Ma by analyses using the substitution rates of rbcL
and matK. However, this time overlaps with the gener-
ally accepted age of the order Alismatales thus putting
the validity of the results of that study into doubt [14].
He et al. (1991) [15] proposed that Ottelia had origi-
nated no later than the Cretaceous based on the distri-
bution of the genus and the predictions of the
continental drift theory. Fossils of Hydrocharitaceae
have been found in Europe including those of the extant
genera Vallisneria, Hydrilla, Ottelia, Thalassia, Stra-
tiotes, Hydrocharis and Najas from the Eocene, Oligo-
cene and Miocene [16-18]. The oldest fossil of
Hydrocharitacae (genus Stratiotes) is from the Late
Paleocene [19]. These fossils have prior to the present
study not been incorporated in divergence time
estimates.
The geographic origin of Hydrocharitaceae remains
unresolved. The diversity centre of the family has been
suggested to be in tropical Asia [20]. However, the
diversity centre of a taxon does not necessarily corre-
spond to its centre of origin. Numerous fossils of
Hydrocharitaceae have been found in Europe [17],
implying a possible European origin of the family. A
biogeographic analysis is required to elucidate the geo-
graphical origin of the family.
The transoceanic distribution of angiosperms has long
intrigued biologists. Two competing hypotheses have
been proposed to explain this phenomenon: the first
attributing it to dispersal [21-23] and the second to
vicariance (continental drift) [24]. Les et al. (2003) [21]
proposed that dispersal is the major factor accounting
for the disjunctive distribution of aquatic plants. This is
contrary to the traditional viewpoint which considered
vicariance [24] as the major cause of the disjunctive dis-
tribution in aquatic taxa such as Limnocharitaceae [25],
Ottelia (Hydrocharitaceae) [15] and Sagittaria (Alisma-
taceae) [26]. Hydrocharitaceae exhibits a wide transocea-
nic distribution at genera and species levels. The family
can serve as a suitable model to investigate transoceanic
distribution in aquatic monocots.
About 10% of all flowering plants have unisexual flow-
ers which have traditionally been regarded as a derived
state in angiosperms [27]. Most species of Hydrocharita-
ceae are unisexual while some are hermaphrodite. Her-
maphroditism is regarded as the ancestral condition
which gave rise to unisexual flowers [28]. However, her-
maphroditic flowers also occur in more recently evolved
genera such as Ottelia [1]. This suggests that the view
Chen et al. BMC Evolutionary Biology 2012, 12:30
http://www.biomedcentral.com/1471-2148/12/30
Page 2 of 12

Page 3

that hermaphroditism is the ancestral state needs to be
re-examined. Leaf habit in Hydrocharitaceae varies from
aerial, aerial-submerged to fully submerged, and leaf
shape varies from circular, linear, to ribbon like [1].
Ancestral state reconstruction is useful in understanding
the evolutionary history of reproductive system and leaf
morphology and their significance in adapting the plants
to the aquatic environment.
In this study, we sampled 17 genera of Hydrocharita-
ceae, using DNA sequences from 8 genes. The aims of
our study were: 1) to reappraise the generic relation-
ships of Hydrocharitaceae; 2) to estimate the divergence
times; 3) to investigate the area of origin and the major
factor(s) shaping the global distribution; and 4) to inves-
tigate the evolution of reproductive system and leaf
morphology in this family.
Results
Sequence characteristics
Details of the voucher and DNA sequence information
are provided in Additional file 1. We successfully gener-
ated 117 sequences. Other sequences used in this study
were downloaded from GenBank. Seventeen of the 18
genera in Hydrocharitaceae were sampled. Genus Apper-
tiella was not included because of unavailability of
molecular data. The supermatrix dataset, which resulted
from assembling the DNA sequences of the 8 genes,
was 8,086 nt in length (2,035 mt, 4,406 chl, 1,645 nu).
The dataset consisted of 38 terminals of Hydrocharita-
ceae and was submitted to TreeBASE (accession number
S12110). The dataset comprised about 29% missing
characters mainly due to unavailability of some
sequences.
Phylogenetic analyses
Butomus and (Butomus + Alisma (Alismataceae) + Cym-
odocea (Cymodoceaceae) + Hydrocleys (Limnocharita-
ceae) + Potamogeton (Potamogetonaceae)) were
independently selected as outgroup. Both Maximum
likelihood (ML) and Bayesian analysis using the super-
matrix dataset resulted in completely identical relation-
ships and strong support (bootstrap value (BS)> 95,
Bayesian posterior probability (PP) = 1.0) for most
branches (Figure 1). ML analysis involving partitioning
the supermatrix dataset into eight genes and no parti-
tion resulted in slightly different support values, but
identical topologies. Stratiotes was resolved as the first
diverging lineage of Hydrocharitaceae with strong sup-
port (BS = 100, PP = 1.0). Other genera formed two
clades (BS = 99, PP = 1.0; Figure 1). Clade A (BS = 95,
PP = 1.0) included 10 genera. Hydrocharis-Limnobium
was resolved as the first diverging lineage of this clade;
the seagrasses formed a well supported clade (BS = 100,
PP = 1.0) which was resolved as sister to the subclade
(BS = 96, PP = 1.0) formed by Najas, Hydrilla, Necha-
mandra, Vallisneria and Maidenia (Figure 1). Clade B
(BS = 100, PP = 1.0) consisted of Lagarosiphon, Ottelia,
Blyxa, Apalanthe, Elodea and Egeria (Figure 1). The two
outgroups selected resulted in slightly different support
values, but identical topologies of Hydrocharitaceae.
Dating analysis and ancestral area reconstruction
All BEAST MCMC runs yielded sufficient effective sam-
ple sizes (> 200) for all relevant parameters and con-
verged on topologies identical to the tree in Figure 1.
The crown node age of Hydrocharitaceae was dated at
65.2 Ma (95% HPD: 54.6-79.6 Ma). The mean diver-
gence between clade A and clade B, estimated to be
63.1 Ma. For the three calibration nodes, mean posterior
estimates were consistent with prior node ages, suggest-
ing that the calibration points were sufficiently concor-
dant [29,30].
Seven biogeographic areas were recognized according
to Morse [31] (Figure 2a) namely A, Nearctic area; B,
Neotropical area; C, West Palearctic area; D, Afrotropi-
cal area; E, Oriental area; F, Australasian area; and G,
East Palearctic area. Two strategies were applied in the
biogeographic analyses. One used genera as terminal
taxa, the other used species as terminal taxa. Results of
the analyses using genera as terminal taxa suggested
that the most recent common ancestors (MRCAs) of
both Hydrocharitaceae and clade A occurred in Oriental
area. The MRCA of clade B occurred in Oriental, Afro-
tropical and Neotropical areas (Figure 2b).
Results of the biogeographic analyses using species as
terminal taxa suggested that Hydrocharitaceae origi-
nated in Orient (Figure 2c). A minimum of 76 dispersal
events was inferred from DIVA to explain the current
distribution of Hydrocharitaceae. The ancestor of Stra-
tiotes was suggested to have been in Orient and dis-
persed to Europe during the Late Cretaceous and
Paleocene (Figure 2c, d, arrow 1). This route is similar
to that which has been reported for Alangiopollis (Alan-
giaceae) [32]. The analyses using genera or species as
terminal taxa yielded comparable results on the ances-
tral area of Hydrocharitaceae. This indicates that incom-
plete sampling may have little effect on the accuracy of
investigation into the geographical origin of Hydrochari-
taceae. However, the analyses using species as terminal
successfully resolved the ancestral areas for more genera
than those using genera as terminal.
The MRCAs of clade A, the seagrass genera and the sea-
grass subclade were shown to be of Oriental origin. The
taxa then diversified in Oriental region (DIVA, Mesquite;
Figure 2c). DIVA suggested that Vallisneria originated in
Oriental and Australasian regions. Mesquite suggested an
Oriental origin for Vallisneria, followed by diversification
and dispersal to other continents (Figure 2c). Long
Chen et al. BMC Evolutionary Biology 2012, 12:30
http://www.biomedcentral.com/1471-2148/12/30
Page 3 of 12

Page 4

distance dispersal (LDD) for Vallisneria from Oriental
area to Africa and Europe was inferred from V. spinulosa,
V. spiralis and V. denseserrulata (Figure 2c, e, arrow 1 &
2). LDD from Australasia to North America for this genus
was inferred from V. neotropicalis and V. americana (Fig-
ure 2c, e, arrow 7). This route is similar to that envisaged
for the taxon Scaevola (Goodeniaceae) [33]. LDD from
Australasia to Asia similar to that which has been
recorded for the plant family Cucurbitaceae [34] was
inferred from taxa including V. asiatica and V. natans
(Figure 2c, e, arrow 4). The genus Najas was inferred to
have originated in Oriental area during the Oligocene.
LDD from Asia to North America (inferred from N. gracil-
lima), to Europe (inferred from N. minor), to Africa
(inferred from N. minor) and to Australia (inferred from
subclade N. browniana + N. tenuifolia) was suggested (Fig-
ure 2c, e, arrow 6, 1, 2 & 5).
The ancestor of clade B was inferred to have dispersed
from Orient to Afrotropical region (Mesquite) or South-
ern hemisphere (DIVA) during the Eocene (Figure 2c, d,
arrow 2), followed by diversification in Southern hemi-
sphere during the Tertiary. MRCA of Ottelia was sug-
gested to have lived in Oriental and Afrotropical regions
(Figure 2c). LDD from West Africa into South America
for this genus was inferred from O. brasiliensis (Figure
2c, e, arrow 9). This route is similar to the one sug-
gested for the dispersal of Gossypium (Malvaceae) [35]
and is further supported by the fact that Ottelia in
South America is confined to the southeastern area [36].
LDD from S.E. Asia to Australasia was inferred from O.
ovalifolia (Figure 2c, e, arrow 5). Dispersal from Africa
into Asia was also inferred from O. cordata and O.
mesenterium (Figure 2c, e, arrow 3). This route is simi-
lar to that envisaged for the two genera Coccinia and
Figure 1 Phylogeny and divergence time estimates of Hydrocharitaceae based on combined 18S + rbcL+ matK+ trnK 5’ intron + rpoB
+ rpoC1 + cob + atp1 data set. Numbers above branches refer to the maximum likelihood bootstrap values (BS, left) and the posterior
probabilities (PP, right). Numbers in blue refer to the branches with BS> 95 and PP = 1.0. Gray bars indicate 95% highest posterior distributions,
and nodes labelled with stars refer to the positions of fossil calibration points.
Chen et al. BMC Evolutionary Biology 2012, 12:30
http://www.biomedcentral.com/1471-2148/12/30
Page 4 of 12

Page 5

Figure 2 Historical biogeography of Hydrocharitaceae. (a) Area delimitation in biogeographic analyses. (b) Analyses using genera as terminal
taxa, and (c) analyses using species as terminal taxa. Distribution of each genus or species is indicated along with taxon names. Results of
Mesquite are indicated by coloured circles; results of dispersal-vicariance analysis (DIVA) are indicated by capital letters (only the results of major
clades are shown); equally optimal ancestral distributions are indicated by pie-charts (Mesquite) or slashes (DIVA). The times inferred from
divergence time estimates were marked on the major clades, numbers represent millions of years before present. (d) & (e) Possible origin,
differentiation centres and dispersal routes of Hydrocharitaceae.
Chen et al. BMC Evolutionary Biology 2012, 12:30
http://www.biomedcentral.com/1471-2148/12/30
Page 5 of 12