Content uploaded by Guillaume Besnard
Author content
All content in this area was uploaded by Guillaume Besnard on Jun 30, 2020
Content may be subject to copyright.
LETTER Photosynthetic innovation broadens the niche within a single
species
Marjorie R. Lundgren,
1
Guillaume Besnard,
2
Brad S.
Ripley,
3
Caroline E. R. Lehmann,
4
David S. Chatelet,
5
Ralf G. Kynast,
6
Mary Namaganda,
7
Maria S. Vorontsova,
6
Russell C. Hall,
1
John Elia,
8
Colin P. Osborne
1
* and
Pascal-Antoine Christin
1
*
Abstract
Adaptation to changing environments often requires novel traits, but how such traits directly
affect the ecological niche remains poorly understood. Multiple plant lineages have evolved C
4
photosynthesis, a combination of anatomical and biochemical novelties predicted to increase pro-
ductivity in warm and arid conditions. Here, we infer the dispersal history across geographical
and environmental space in the only known species with both C
4
and non-C
4
genotypes, the grass
Alloteropsis semialata. While non-C
4
individuals remained confined to a limited geographic area
and restricted ecological conditions, C
4
individuals dispersed across three continents and into an
expanded range of environments, encompassing the ancestral one. This first intraspecific investiga-
tion of C
4
evolutionary ecology shows that, in otherwise similar plants, C
4
photosynthesis does
not shift the ecological niche, but broadens it, allowing dispersal into diverse conditions and over
long distances. Over macroevolutionary timescales, this immediate effect can be blurred by subse-
quent specialisation towards more extreme niches.
Keywords
Adaptation, Alloteropsis,C
4
photosynthesis, ecological niche, evolution, phylogeography.
Ecology Letters (2015) 18: 1021–1029
INTRODUCTION
The ecological niche of organisms is shaped by the metabolic
and morphological adaptations acquired during their evolu-
tionary history (Kellermann et al. 2012; Ara
ujo et al. 2013;
Hertz et al. 2013). However, the relationships between adap-
tive traits and ecological niches are still poorly understood.
Some traits can evolve in situ, for example, as a response to
changes in the surrounding environment following migration
or external modification of the local habitat, which leads to a
shift in the ecological niche (Simon et al. 2009). Other traits
can modify the niche breadth to facilitate the colonisation of
novel habitats, as well as persistence in the ancestral ones,
with possible subsequent specialisation to the new habitats
(Ackerly 2004; Cacho & Strauss 2014). In plants, one impor-
tant determinant of the ecological niche is the efficiency of
photosynthesis in different environments. Photosynthetic effi-
ciency can be lowered by photorespiration, which occurs when
O
2
is fixed instead of CO
2
and requires energy to recycle the
resulting metabolites (Ogren 1984). This phenomenon can
retard net carbon-fixation in the ancestral C
3
photosynthetic
type by more than one third (Skillman 2008), and increases
under all conditions that limit the availability of CO
2
at the
active site of the carbon-fixing enzyme Rubisco. Intercellular
CO
2
decreases at low atmospheric CO
2
concentrations, but
also at high temperatures, where the solubility of CO
2
decreases faster than the solubility of O
2
, and Rubisco
becomes less able to discriminate between CO
2
and O
2
(Eh-
leringer & Bjorkman 1977). In addition, arid and saline condi-
tions promote stomatal closure and thereby reduce CO
2
input
from the atmosphere (Sage et al. 2012).
Several lineages of plants have evolved novel trait com-
plexes that decrease photorespiration. These include CO
2
-con-
centrating mechanisms such as C
4
photosynthesis, which
evolved independently as an addition to the C
3
pathway in
more than 60 lineages of flowering plants in response to past
decreases in atmospheric CO
2
(Sage et al. 2011; Christin &
Osborne 2014). C
4
physiology is assembled from a combina-
tion of anatomical and biochemical components that increases
CO
2
concentration at the active site of Rubisco (Hatch 1987).
The C
4
pathway nearly eliminates photorespiration (Skillman
2008), but requires extra energy such that the maximum
efficiency of photosynthetic light-use in C
4
photosynthesis
surpasses that of C
3
photosynthesis only when photorespira-
tion is high (Ehleringer & Bjorkman 1977). C
4
photosynthesis
is therefore predicted to provide an advantage in any environ-
ment that promotes photorespiration (Sage et al. 2012; Chris-
tin & Osborne 2014). Accounting for one quarter of terrestrial
1
Department of Animal and Plant Sciences, University of Sheffield, Western
Bank, Sheffield S10 2TN, UK
2
CNRS, Universit
e Toulouse III - Paul Sabatier, ENFA, UMR5174 EDB (Labora-
toire
Evolution & Diversit
e Biologique), 118 route de Narbonne, 31062 Tou-
louse France
3
Department of Botany, Rhodes University, Grahamstown 6139, South Africa
4
School of GeoSciences, University of Edinburgh, Crew Building,
The King’s Buildings, Alexander Crum Brown Road, Edinburgh
EH9 3FF, UK
5
Department of Ecology and Evolutionary Biology, Brown University, Provi-
dence RI, USA
6
Royal Botanic Gardens, Kew, Richmond Surrey TW9 3AB, UK
7
Department of Biological Sciences, Makerere University, PO Box 7062,
Kampala Uganda
8
National Herbarium of Tanzania, Arusha Tanzania
*Correspondence:E-mails: c.p.osborne@sheffield.ac.uk (or) p.christin@sh
effield.ac.uk
©2015 John Wiley & Sons Ltd/CNRS
Ecology Letters, (2015) 18: 1021–1029 doi: 10.1111/ele.12484
primary production (Still et al. 2003), plants using C
4
photo-
synthesis are globally ecologically important. In particular,
the productive C
4
grasses dominate savannas and grasslands
of warm regions, novel environments that expanded during
the Miocene and in which grazing ungulates and other
groups, including humans, diversified (Lehmann et al. 2011;
Sage & Stata 2014). The consequences of C
4
photosynthesis
for the ecological niche have primarily been investigated
through comparisons of species distributions, which show an
important effect of temperature on the distribution of C
4
grasses (Teeri & Stowe 1976; Ehleringer et al. 1997). However,
these investigations are biased by differences among phyloge-
netic groups (Taub 2000), and recent interspecific comparisons
accounting for phylogenetic structure have revolutionised our
understanding of C
4
evolutionary ecology (reviewed in Chris-
tin & Osborne 2014). In particular, phylogeny-based analyses
have shown that C
4
photosynthesis evolved in groups of
grasses inhabiting warm regions and facilitated shifts into
drier and more saline habitats (Osborne & Freckleton 2009;
Edwards & Smith 2010; Bromham & Bennett 2014). However,
the photosynthetic transitions investigated in these analyses
occurred many millions of years ago and there is often a gap
of several million years between C
3
and C
4
nodes in species
phylogenetic trees (Christin et al. 2011). These vast timescales
make it difficult to confidently reconstruct the conditions
under which C
4
photosynthesis evolved or the events that
occurred immediately after this physiological divergence.
Identifying the selective factors that promoted the gradual
assembly of C
4
photosynthesis within populations requires
investigations within species complexes that vary in photosyn-
thetic phenotype. Groups with such variation are rare, and
the grass Alloteropsis semialata (R.Br.) Hitchc. is the only
known species that encompasses both C
4
and non-C
4
individ-
uals (Ellis 1974). This taxon is spread throughout a diversity
of habitats across multiple continents and therefore consti-
tutes an excellent system to investigate the evolutionary ecol-
ogy of C
4
photosynthesis. The history of photosynthetic
transitions within the Alloteropsis genus is not resolved with
confidence. The reconstruction of photosynthetic types as bin-
ary characters on the species phylogeny would lead to the
most parsimonious hypothesis of a single C
4
origin followed
by a reversal to an ancestral non-C
4
type in A. semialata
(Ibrahim et al. 2009). Such an approach, however, would fail
to acknowledge the complexity of the C
4
trait and, when indi-
vidual components are analysed independently, a more com-
plex scenario emerges (Christin et al. 2010). Indeed, the
various C
4
species within the Alloteropsis genus use different
tissue types for the segregation of photosynthetic reactions
and different C
4
biochemical subtypes (Christin et al. 2010),
and the genetic determinism for key C
4
enzymes differs among
A. cimicina (L.) Stapf, A. angusta Stapf and C
4
populations of
A. semialata (Christin et al. 2012). The most likely scenario
given current data therefore involves multiple C
4
optimisa-
tions from an ancestor with C
4
-like or C
3
-C
4
intermediate
characters (Christin et al. 2012).
Here, we capitalise on the photosynthetic diversity within A.
semialata to reconstruct the environments in which photosyn-
thetic types diverged, and examine the consequences of photo-
synthetic innovation for the ecological niche. We sample
individuals spread across the whole geographic range, and
characterise their phenotype as well as their habitat. We then
apply phylogenetic methods to markers from the chloroplast
genome, which are maternally inherited, to reconstruct the
history of expansion into new geographic areas and environ-
mental conditions via seed dispersal. Based on this time-cali-
brated phylogeographic hypothesis, we quantify the rates of
dispersal across geographical and environmental spaces, and
compare these among clades, also supported by nuclear
markers, that differ in their photosynthetic phenotype. This
first intraspecific investigation of C
4
evolutionary ecology
demonstrates that C
4
photosynthesis does not shift the ecolog-
ical niche but broadens it, leading to the rapid colonisation of
diverse habitats and dispersal over large geographic distances.
MATERIALS AND METHODS
Plant sampling, photosynthetic pathway and habitat
Collection locations for 309 A. semialata specimens were col-
lated from several sources, as described in the Supporting
Information Methods online (Table S1). Photosynthetic type
was determined using stable carbon isotopes, which unam-
biguously differentiate individuals that grew using C
4
photo-
synthesis from those that grew without fixing the majority of
carbon via phosphoenolpyruvate carboxylase (PEPC; Sup-
porting Information Methods). This latter category can
include C
3
individuals as well as several types of C
3
–C
4
inter-
mediates (von Caemmerer 1992; Sage et al. 2012). In addition
to photosynthetic type, ploidy level, seed size, culm height
and flowering phenology data were collected for several acces-
sions (Supporting Information Methods).
Characterisation of the environment
Information on the environmental conditions at the collection
location of the 309 A. semialata accessions was obtained by
overlaying geographic coordinates onto high-resolution raster
layers of environmental variables predicted to affect the sort-
ing of C
3
and C
4
plants (reviewed in Christin & Osborne
2014; Table S2; see Supporting Information Methods). As
multivariate analyses on distribution data provide an estimate
of the abiotic component of the ecological niche (Petitpierre
et al. 2012), a principal component analysis (PCA) was per-
formed to summarise the environmental variation among the
collection localities of A. semialata using eight environmental
variables (Table S2) with the FACTOMINER package (L^
e
et al. 2008) in R. In addition, localities were classified as being
in open or wooded habitats, based on descriptions provided
on herbarium sheets, when available.
Sequencing and phylogenetic analyses
Besides the two congeners A. cimicina (one accession) and
A. angusta (two accessions), a total of 66 accessions assigned
to A. semialata and representing 55 different populations were
sampled for phylogenetic analyses (Table S1). These were
selected to encompass the largest possible diversity of geo-
graphical origins and photosynthetic types. Five plastid
©2015 John Wiley & Sons Ltd/CNRS
1022 M. R. Lundgren et al. Letter
regions (trnK-matK,rpl16,ndhF,rpoC2 and trnL-trnF) were
isolated via PCR or retrieved from previous studies (Ibrahim
et al. 2009; Grass Phylogeny Working Group II 2012). In
addition, the nuclear-encoded ITS marker was isolated from a
subset of accessions (Supporting Information Methods).
The complete chloroplast genomes of 13 of these samples
were subsequently obtained through genome skimming (Sup-
porting Information Methods). These samples were selected
because they represent different lineages, as determined from
preliminary analyses of the chloroplast markers. Genomic
DNA was isolated from silica-gel-dried material and
sequenced using Illumina technology. Complete chloroplast
genomes were assembled and aligned using in-house Perl
scripts. The same approach was used to assemble the complete
nuclear ribosomal DNA units (rDNA encompassing the ITS;
Supporting Information Methods). All sequences are avail-
able in NCBI database under the accession numbers
KT271570-KT271740 and KT281145-KT281168.
The 13 complete chloroplast genomes were added to an
alignment of grass genomes covering the whole family, and
the trimmed alignment was used to compute a time-calibrated
phylogenetic tree through Bayesian inference (Supporting
Information Methods). A second phylogenetic analysis was
conducted on A. semialata and A. angusta accessions only. All
markers obtained via PCR were aligned with the complete
chloroplast genomes obtained for these two species and a
time-calibrated phylogenetic tree was inferred using Bayesian
approaches, using relative divergence times in the absence of
fossils for the group. The ITS sequences isolated by PCR were
similarly added to the complete rDNA units, and a phyloge-
netic tree was inferred on these nuclear markers (Supporting
Information Methods).
Rates of ecological and geographical dispersal
The rates of dispersal across environmental and geographical
spaces were estimated for A. semialata by regressing geo-
graphic and environmental pairwise distances to divergence
times. Only one individual per population was selected, which
resulted in 55 A. semialata samples for which both phyloge-
netic and environmental information was available. The geo-
graphic distance across the Earth’s surface was calculated for
each pair of locations using the latitude and longitude coordi-
nates and the earth.dist function in the FOSSIL package
(Vavrek 2011). The environmental distances among these 55
accessions were calculated as Euclidian distances in the space
formed by the first four axes of the PCA produced on all
accessions (see above). Finally, the divergence time between
each pair of accessions was extracted from the phylogeo-
graphic tree, using the APE package (Paradis et al. 2004).
Environmental distances are potentially correlated with geo-
graphical distances (spatial autocorrelation) and, as such, par-
tial Mantel permutation tests, as implemented in the APE
package, were used to test for statistical associations between
the three matrices, and to correct for such spurious correla-
tions. These tests were conducted separately on the ABC and
DE sister groups, which were retrieved on both plastid and
nuclear marker trees, and differ in photosynthetic type (see
results). Linear regressions were subsequently used to calcu-
late the slope for significant relationships. In cases where all
relationships were significant, the relationship between the
part of environmental distances not explained by geographical
distances (that is, the residuals of the regression) and diver-
gence times was tested.
For illustration purposes, the history of seed dispersal
across the PCA space was inferred by mapping changes in the
scores along the first two axes onto the phylogenetic tree,
using ancestral state reconstructions as implemented in APE.
The same approach was used to reconstruct dispersal across
environments differing in their mean annual temperature
(MAT) and mean annual precipitation (MAP), two variables
commonly used to characterise global climate space and
selected in the past to compare C
3
and C
4
distributions (Teeri
& Stowe 1976; Edwards & Smith 2010).
RESULTS
Phylogenetic relationships and dispersal through geographical space
In the plastid phylogeny, all accessions assigned to the species
A. semialata based on morphological characters formed a
strongly supported monophyletic group, sister to the C
4
A.
angusta (Figs S1 and S2), confirming previous investigations
with fewer samples (Ibrahim et al. 2009; Grass Phylogeny
Working Group II 2012). The first split within A. semialata
separates some Tanzanian accessions, with carbon isotope
ratios indicative of C
4
photosynthesis (Clade F), from all other
individuals (Fig. 1). The remaining accessions form two sister
clades (ABC and DE; Fig. 1). The DE clade contains all acces-
sions identified as C
4
outside of the F clade, while the ABC
clade contains all the accessions for which a non-C
4
isotopic
Figure 1 Phylogenetic relationships among A. semialata accessions. This
tree was obtained through Bayesian inference on chloroplast markers, and
branch lengths are proportional to estimated divergence time, in arbitrary
time units. Branches leading to monophyletic C
4
groups are in red.
Geographic regions are delimited next to the tips. The main clades are
delimited on the right, and coloured according to photosynthetic type
with red denoting C
4
, and black non-C
4
, clades. Asterisks indicate nodes
with Bayesian support values above 0.95. The phylogenetic tree is detailed
in Fig. S1.
©2015 John Wiley & Sons Ltd/CNRS
Letter Photosynthetic innovation broadens niche 1023
signature was measured (Fig. 1; Table S1). Some members of
clade ABC have carbon isotope ratios between the classical C
3
and C
4
ranges (Table S1), which might indicate the occurrence
of a weak C
4
cycle, although this requires further investigation.
Based on complete chloroplast genomes of A. semialata incor-
porated within a grass-wide data set, the divergence of clades
ABC and DE is estimated at 2.42 Ma (95% CI =1.42–3.77),
the first split within clade ABC at 1.53 Ma (95% CI =0.71–
2.7) and the first split within clade DE at 1.25 Ma (95%
CI =0.7–1.98; Figs S2 and S3). The split between C
4
and non-
C
4
lineages of A. semialata is consequently more recent than all
other origins of monophyletic C
4
groups (Christin et al. 2011).
This divergence occurred after the Miocene emergence of the
C
4
grassy savanna biome (Edwards et al. 2010), but falls within
the Pliocene interval when C
4
grasses became increasingly dom-
inant in African savannas (Hoetzel et al. 2013). The phyloge-
netic tree based on complete nuclear rDNAs for A. semialata
supports similar relationships, although the E clade is para-
phyletic (Fig. S4). The ITS marker contained few informative
sites, and the nuclear phylogenetic tree based on 37 A. semi-
alata accessions was poorly resolved (Fig. S5), which might be
partially caused by recurrent pollen-mediated gene flow after
the habitat expansion via seed dispersal. The C
4
and non-C
4
accessions however still sort into two distinct clades (Fig. S5),
which suggests that gene flow between clades ABC and DE
was limited over the last million years, and the photosynthetic
types remained tightly associated with the plastid lineages,
despite overlapping geographic distributions and flowering
periods (Fig. S6).
While nuclear markers are important to detect pollen-medi-
ated gene movements, the colonisation of new habitats by
plants is caused by seed movements and consequently, better
inferred from plastid markers. With the exception of the wide-
spread A. cimicina, the three remaining congeners are of cen-
tral African origin, where members of the early diverging
clade F were also found, leading to the inference of a central
African origin for A. semialata (Fig. 2). All members of clades
B and C are also from central Africa, suggesting limited dis-
persal. However, all members of clade A are from southern
Africa, which implies a single migration to southern latitudes
at the base of clade A (Fig. 2). This strongly contrasts with
clade DE, which, despite a more recent common ancestor,
covers the tropical and subtropical regions of Africa, Asia
and Oceania (Fig. 2). In this group, clade E is endemic to
mainland Africa, with early splits separating central African
accessions and more recent splits leading to southern, western
and eastern African accessions (Figs 1 and 2). The first split
in clade D separates Madagascan from Asian and Oceania
accessions, suggesting a single migration outside of mainland
Africa (Figs 1 and 2). Long distance dispersal across the
Indian Ocean is often observed and might have occurred via
previously emerged islands (Warren et al. 2010).
Statistical comparisons among pairwise geographic distances
and divergence times revealed patterns of isolation by distance
in both the C
4
clade DE (P<0.00001) and the non-C
4
clade
ABC (P<0.00001). However, the slope of the regression of
geographic distances against divergence times is nearly six
times steeper in clade DE than in clade ABC (9992 km per
time unit vs. 1617 km per time unit; Fig. 3), which indicates
that, while dispersal is limited in both clades, the limitation is
stronger in the non-C
4
clade ABC. All analyses were repeated
with topologies sampled from the posterior distribution, and
the results remained unaltered (Fig. S7).
Dispersal through the environmental space
The distribution of C
4
individuals in the first four PCA axes,
which together explain 87.69% of the environmental variation
in the data set, overlaps with that of non-C
4
individuals.
However, the habitat space of non-C
4
accessions is smaller
and represents a subset of the conditions inhabited by C
4
accessions (Figs 4 and S8). The subset of accessions included
in the phylogeny covers most of the diversity seen in the sam-
ple of 309 populations (Fig. S8), and therefore constitutes an
accurate representation of the ecological diversity of the spe-
cies. Focusing on the accessions included in the phylogeny,
non-C
4
individuals from central Africa (clades B and C) are
clustered near the centre of the PCA, together with the early
diverging C
4
clade F (Fig. 4). On the other hand, the southern
African non-C
4
clade A spread towards negative values on the
first axis, into cool and dry atmospheric environments (Figs 4
and S8; Table S3). The broad habitat of the C
4
clade DE
encompasses the extremes along both PCA dimensions, with-
out clear distinction between geographical regions, as C
4
accessions from different continents can be found in environ-
ments with similar abiotic characteristics (Fig. 4). Similar pat-
(a)
(b)
Figure 2 Distribution of sampled Alloteropsis semialata individuals and
inferred dispersal events. (a) The six main clades are represented by
different symbols, with the C
4
accessions in red and the non-C
4
accessions
in black. (b) The phylogeographic tree is approximately projected on the
geographical space, with dispersal indicated by lines (ends of lines as in
panel a). The branch from the root is in grey, and other branches are
coloured by photosynthetic type (C
4
in red and non-C
4
in black).
©2015 John Wiley & Sons Ltd/CNRS
1024 M. R. Lundgren et al. Letter
terns are observed for the commonly used MAP and MAT
variables (Fig. S9).
According to reconstructions based on the phylogeographic
tree, the ancestors of all A. semialata accessions and clade
ABCDE occurred near the centre of the PCA space, where
members of the clades B, C and F are still located (Figs 4 and
5). The ancestors of each of the C
4
clades D and E and non-
C
4
clades B and C are inferred in the same location in envi-
ronmental space (Fig. 5), suggesting that the divergence of
photosynthetic types was not immediately followed by signifi-
cant changes on the PCA axes, or in MAT or MAP (Fig. S9).
Most of the environmental diversification therefore occurred
after the divergence of the C
4
and non-C
4
clades. Members of
the non-C
4
clades B and C remained in the same area of the
PCA, in relatively warm areas (Figs 4 and 5). However, a
strong departure from this type of environment occurred in
the ancestor of clade A (Fig. 5), corresponding to a migration
to temperate grasslands (Fig. S9; Table S3). The progressive
changes within the non-C
4
clade ABC contrast strongly with
those observed within the C
4
clade DE. Indeed, extreme val-
ues along both axes are randomly spread in clade DE (Fig. 5),
indicating repeated migrations across a wide range of precipi-
tation, temperature, fire and light environments that can be
tolerated by these C
4
plants (Fig. 4), in addition to different
tree covers (Table S4).
Mantel tests confirm that rates of dispersal across the envi-
ronmental spaces differ statistically between the C
4
and non-
C
4
clades. Environmental distances are significantly correlated
to divergence times within the non-C
4
clade ABC (P<0.001),
indicating a gradual migration into different conditions
(Fig. 3). However, these environmental distances are also cor-
related with geographic distances (P<0.00001). The relation-
ship between environmental distances and divergence times
remains significant once this spatial autocorrelation is taken
into account (P<0.005), which shows that lineages within
clade ABC transitioned gradually into different environments
as they adapted to slightly different conditions through natu-
Figure 3 Comparison of geographical and environmental distances and
divergence times. These analyses are based on distances between pairs of
non-C
4
individuals from clade ABC (black) and between pairs of C
4
individuals from clade DE (red). Regression lines forced to the origin are
shown for significant relationships, identified by Mantel tests.
Figure 4 Ecological niche as inferred by principal component analysis
(PCA). In the left panel, dashed lines indicate the approximate
distribution of C
4
(red) and non-C
4
(black) accessions in the PCA space
(see Fig. S8 for the distribution of all points). The distribution of
individuals included in the phylogeny is shown with circles, squares and
triangles coloured by photosynthetic type. The location of the common
ancestor of clades ABC and DE as inferred along the phylogeny is
indicated by a grey circle. The right panel indicates the inferred changes
in the PCA space, with an environmental shift for the non-C
4
clade A
(black arrow) and extension of the C
4
niche in multiple directions (red
arrows).
Figure 5 Movements across the environmental space inferred along the
phylogeographic tree. Dot size is proportional to the absolute values
along the first two dimensions of the PCA, as observed for tips and
inferred for ancestral nodes. Negative values are in black and positive
values in pink. The main clades are indicated on the right.
©2015 John Wiley & Sons Ltd/CNRS
Letter Photosynthetic innovation broadens niche 1025
ral selection. The results are very different in the C
4
clade
DE, for which environmental distances are not correlated to
divergence times (P=0.77; Fig. 3). This shows that the
migration of C
4
accessions to diverse environments happened
rapidly, from their early diversification (Figs 2 and 3). Their
ecology is neither explained by the timing of dispersal nor by
their geographical proximity, which strongly supports the
hypothesis of a broad ecological niche from the outset. These
conclusions are not affected by phylogenetic uncertainty, as
the results of Mantel tests are confirmed across trees from the
posterior distribution (Fig. S7).
DISCUSSION
Photosynthetic diversification within A. semialata
Chloroplast markers retain the signature of seed dispersal,
and the phylogeographic hypothesis produced here indicates
the successive seed-mediated dispersal across geographical
and environmental spaces (Figs 2 and 5). Nuclear gene flow
is likely to differ, being more frequent and occurring across
longer distances in wind-pollinated species. There is, how-
ever, a tight association between the photosynthetic pheno-
type and the plastid lineages, and the nuclear-encoded ITS
also supports monophyletic C
4
and non-C
4
clades (Figs S4
and S5). This suggests that gene flow was limited following
the divergence of clades ABC and DE, despite overlapping
geographical distributions and flowering periods (Fig. 2). The
split of the sister groups ABC and DE consequently repre-
sents the physiological divergence between non-C
4
and C
4
plants.
The common ancestor of A. semialata clades ABC and DE
identified here indisputably represents the last ancestor with
both C
4
and non-C
4
descendants in the group. Variation other
than photosynthetic types exists in A. semialata as within any
species, and phenotypic variation was observed in both the
ABC and DE clades (Figs S6, S10 and S11). However, no
character other than C
4
photosynthesis consistently differed
among the clades. All individuals are perennial, and similar
plant height, gross morphology, flowering phenology and seed
size are present in the different chloroplast lineages (Figs S6,
S10 and S11). Earlier work suggested that C
3
A. semialata are
diploid while C
4
individuals are polyploid (Liebenberg & Fos-
sey 2001). However, this study included only South African
accessions. The geographically diverse accessions presented
here and in Ellis (1981) demonstrate that C
4
populations from
Asia, Australia and regions of Africa are diploid, with poly-
ploidy only detected in southern African C
4
accessions
(Fig. S1; Table S5), and the results of the Mantel tests remain
unchanged if the five individuals from the clade that contains
polyploids are removed. The divergence of clades ABC and
DE is therefore mainly characterised by a switch between
photosynthetic types. Based on dating analyses, the earliest
divergences identified within each of the C
4
and non-C
4
clades
happened shortly after their split and were followed in each
case by continued dispersals through geographical and envi-
ronmental spaces (Figs 1, 2 and 5). This short evolutionary
history, together with the diversity of ecological conditions
covered (Fig. 4), therefore provides a unique opportunity to
investigate the ecological causes and consequences of physio-
logical innovation.
Divergence of photosynthetic types is not followed by major
ecological shifts
Based on the phylogenetic relationships inferred here, the
common ancestor of the ABCDE clade originated from
wooded savannas in central Africa, and the early members of
clades ABC and DE persisted in this area for a considerable
length of time. The initial divergence of clades ABC and DE
might have been caused by geographic isolation, in a tectoni-
cally active region where mountain ranges, lakes and rifts pro-
vide barriers to dispersal. Interestingly, the divergence of
photosynthetic types did not directly lead to obvious modifi-
cations of the ecological niche, as assessed by climatic and fire
variables (Figs S8 and S9). Representatives of the different
clades and photosynthetic types can still be found in habitats
within central eastern Africa that match those inferred for
their common ancestor (Figs 4, 5 and S9). Indeed, some C
4
and non-C
4
members of clades B, C, E and F are found in
densely wooded savannas of Tanzania, Congo and Cameroon,
and individuals of clade D occur in similar habitats through-
out Asia and Madagascar (Table S4). In these savannas with
a high cover of deciduous trees, photorespiration is predicted
to vary throughout the year as leaf fall drastically increases
sunlight, temperature and aridity at ground level. The range
of open and wooded savannas in central Africa varied as a
function of the glacial cycles, but wooded savannas were con-
stantly present in this region from the Mioecene (Hoetzel
et al. 2013; Pound et al. 2014). Mutations providing a more
C
4
-like physiology might have been selected for in these habi-
tats where the persistence of more C
3
-like or intermediate phe-
notypes is still possible. Based on these investigations, we
speculate that C
4
physiology initially emerged in environments
that advantage different photosynthetic types across the sea-
sons or across small-scale ecological variations (e.g. densely
vs. lightly wooded habitats), where isolated populations could
explore different parts of the phenotypic landscape as a func-
tion of random mutations.
But C
4
photosynthesis enlarges the ecological niche and increases
dispersal success
The ecological similarity between the early members of the
non-C
4
and C
4
groups contrasts with the current distribution
of the two photosynthetic types. Indeed, extant accessions of
the C
4
clade DE inhabit environments ranging from the tropics
to southern latitudes and cover a broad range of temperatures,
precipitations, light intensities and fire regimes, as well as open
and wooded habitats (Figs 2 and 4, S8 and S9; Table S4). Elu-
cidation of the phylogeographic history shows that these varied
habitats were colonised rapidly after the divergence of photo-
synthetic types, while the otherwise similar non-C
4
members of
clade ABC remained confined to a narrower set of environ-
mental conditions over the same period (Figs 1, 3 and 5).
Moreover, members of clades D and E recurrently migrated
across the environmental space (Figs 3 and 5), indicating that
present distribution patterns are not due to specific groups of
©2015 John Wiley & Sons Ltd/CNRS
1026 M. R. Lundgren et al. Letter
C
4
accessions specialising to different habitats, but to a con-
stant movement across habitats, as attested by the lack of cor-
relation between environmental distances and divergence times
(Fig. 3). These results indicate that when other factors affect-
ing the ecology of individual plant species remain similar, C
4
photosynthesis acts as a niche opener, and does not simply
shift the ecological niche (Fig. 4). The main consequence of C
4
photosynthesis is to decrease photorespiration, and thus
increase the amount of CO
2
fixed per absorbed photon in con-
dition promoting photorespiration (Ehleringer & Bjorkman
1977). This enhances water- and nitrogen-use efficiencies
(Ehleringer & Bjorkman 1977; Pearcy & Ehleringer 1984),
which could facilitate the colonisation of drier and less fertile
habitats. However, it does not necessarily decrease success in
fertile and wetter environments, where it can provide a compet-
itive advantage by enabling faster growth (Monteith 1978;
Long 1999). In addition, the combination of different C
4
biochemical subtypes observed in A. semialata might con-
tribute to enlarging the ecological niche (Wang et al. 2014).
The diversity of ecological conditions tolerated by the C
4
accessions of A. semialata probably explains the more efficient
dispersal of these plants, as has been found across multiple
species of plants and animals (Slatyer et al. 2013). Indeed, the
capacity to survive in a broad range of environments following
long distance dispersal events likely facilitated the colonisation
of distant regions, leading to the spread of these plants across
three different continents (Fig. 2).
Other adaptations lead to ecological diversification
While the C
4
clade DE was quickly dispersing across geo-
graphical and environmental spaces (Fig. 3), members of the
non-C
4
clade ABC continued evolving, emphasising the
importance of considering the variation within each photo-
synthetic type when inferring evolutionary processes. Indeed,
non-C
4
lineages gradually came to colonise distinct environ-
ments independently of geography (Fig. 3). The gradual
migration towards distinct habitats implies a continuous pro-
cess of adaptation through natural selection. While clades B
and C remained in central Africa, in habitats that broadly
resemble those where the common ancestor of A. semialata
grew, members of clade A strongly deviated from these con-
ditions and colonised colder regions in southern Africa
(Figs 2, 4, 5 and S9). This southern dispersal also involved
the migration from wooded savanna habitats to open tem-
perate grasslands with leached, acidic soils, where non-C
4
A.
semialata are very successful, as attested by their local abun-
dance (Ellis 1981). The South African non-C
4
A. semialata
have acquired a cold adaptation mechanism for leaves to
resist freezing, enabling a leaf canopy to persist throughout
the winter (Osborne et al. 2008), and are able to maintain
photosynthetic capacity under drought conditions (Ripley
et al. 2007; Ibrahim et al. 2008). In addition, the non-C
4
A.
semialata completes its growing period during the cooler
periods in South African grasslands (Wand et al. 2002).
These adaptations may have contributed towards their suc-
cessful colonisation of southern latitudes. C
4
photosynthesis,
adopted by members of clade DE, and cold tolerance, pre-
sent in clade A, might represent alternative novelties that
allow the ecological expansion of tropical lineages. This pat-
tern is already evidenced for the grass family as a whole,
where distinct groups have evolved either C
4
photosynthesis
or cold tolerance, both of which strongly increased diversifi-
cation rates (Spriggs et al. 2014). Our intraspecific investiga-
tions show that, while C
4
photosynthesis broadens the niche
and allows rapid dispersal across environmental space, cold
adaptation might be an alternative but slower process that
leads to a narrower realized niche in otherwise similar
plants.
CONCLUSIONS
Capitalising on the variation that exists within a single spe-
cies complex, this study is the first to characterise the ecolog-
ical changes that directly follow the emergence of different
photosynthetic types. The joint analysis of geographical and
environmental dispersal histories within a phylogenetic con-
text shows that C
4
photosynthesis does not initially result in
a shift of the ancestral niche, but broadens this niche to
cover a wider range of conditions that encompass the ances-
tral ones (Fig. 4), enhancing the success of occasional long
distance dispersal events, and therefore increasing the geo-
graphic range. The variety of environments available to C
4
plants is also reflected in the ecological diversity observed
among C
4
species, with different C
4
taxa found in very dis-
tinct environments that promote photorespiration in different
ways (Sage et al. 2012). Interspecific phylogeny-based analy-
ses suggest that species using C
4
photosynthesis diversify
across a wider range of environments than closely related C
3
species (Christin & Osborne 2014). However, individual taxa
likely specialise in different environments after the initial evo-
lution of C
4
physiology, through differential integration of
the C
4
machinery with their growth and life-history traits
(Christin & Osborne 2014). Over time, this process leads to
some C
4
taxa becoming specialised to environments that dif-
fer strongly from those in which they evolved, inflating the
ecological differences between C
3
and C
4
photosynthesis and
blurring the initial effects resulting from differences in photo-
synthetic types.
ACKNOWLEDGEMENTS
This work was funded by a University of Sheffield Prize
Scholarship to MRL and a Royal Society Research Fellow-
ship URF120119 to PAC. The authors thank the herbarium
of the Royal Botanic Gardens, Kew, for providing DNA,
Heather Walker for help with the carbon isotope analyses,
Emanuela Samaritani for seed mass measurements, and Roger
Ellis, Paul Hattersley and Christine Long for useful discus-
sions on the biology of Alloteropsis semialata, access to
unpublished data, and guidance in locating populations. Oli-
vier Bouchez and C
eline Jeziorski from the Genopole in Tou-
louse helped with the Illumina sequencing, and J
er^
ome Chave
made useful comments on previous versions of the manu-
script. Guillaume Besnard is member of the Laboratoire Evo-
lution and Diversit
e Biologique (EDB) part of the LABEX
entitled TULIP managed by Agence Nationale de la
Recherche (ANR-10-LABX-0041).
©2015 John Wiley & Sons Ltd/CNRS
Letter Photosynthetic innovation broadens niche 1027
AUTHORSHIP
MRL, CPO and PAC designed the study. MRL, GB, PAC
generated the data. MRL, CPO and PAC analysed the data
and wrote the paper, with the help of all the authors. MRL,
GB, BSR, CERL, DSC, RCH, MN, MSV, JE and PAC con-
tributed plant material, CERL contributed data on fire and
rainfall seasonality and RGK helped with cytological investi-
gations.
REFERENCES
Ackerly, D.D. (2004). Adaptation, niche conservatism, and convergence:
comparative studies of leaf evolution in the California chaparral. Am.
Nat., 163, 654–671.
Ara
ujo, M.B., Ferri-Y
a~
nez, F., Bozinovic, F., Marquet, P.A., Valladares,
F. & Chown, S.L. (2013). Heat freezes niche evolution. Ecol. Lett., 16,
1206–1219.
Bromham, L. & Bennett, T.H. (2014). Salt tolerance evolves more
frequently in C
4
grass lineages. J. Evol. Biol., 27, 653–659.
Cacho, N.I. & Strauss, S.Y. (2014). Occupation of bare habitats, an
evolutionary precursor to soil specialization in plants. Proc. Natl Acad.
Sci. USA, 111, 15132–15137.
von Caemmerer, S. (1992). Stable carbon isotope discrimination in C
3
–C
4
intermediates. Plant, Cell Environ., 15, 1063–1072.
Christin, P.A. & Osborne, C.P. (2014). The evolutionary ecology of C
4
plants. New Phytol., 204, 765–781.
Christin, P.A., Freckleton, R.P. & Osborne, C.P. (2010). Can
phylogenetics identify C
4
origins and reversals? Trends Ecol.Evol., 25,
403–409.
Christin, P.A., Osborne, C.P., Sage, R.F., Arakaki, M. & Edwards, E.J.
(2011). C
4
eudicots are not younger than C
4
monocots. J. Exp. Bot.,
62, 3171–3181.
Christin, P.A., Edwards, E.J., Besnard, G., Boxall, S.F., Gregory, R.,
Kellogg, E.A. et al. (2012). Adaptive evolution of C
4
photosynthesis
through recurrent lateral gene transfer. Curr. Biol., 22, 445–449.
Edwards, E.J. & Smith, S.A. (2010). Phylogenetic analyses reveal the
shady history of C
4
grasses. Proc. Natl Acad. Sci. USA, 107, 2532–
2537.
Edwards, E.J., Osborne, C.P., Str€
omberg, C.A.E. & Smith, S.A. & C4
Grasses Consortium. (2010). The origins of C
4
grasslands: integrating
evolutionary and ecosystem science. Science, 328, 587–591.
Ehleringer, J. & Bjorkman, O. (1977). Quantum yields for CO
2
uptake in
C
3
and C
4
plants. Plant Physiol., 59, 86–90.
Ehleringer, J.R., Cerling, T.E. & Helliker, B.R. (1997). C
4
photosynthesis,
atmospheric CO
2
, and climate. Oecologia, 112, 285–299.
Ellis, R.P. (1974). The significance of the occurrence of both Kranz and
non-Kranz leaf anatomy in the grass species Alloteropsis semialata.S.
Afr. J. Sci., 70, 169–173.
Ellis, R.P. (1981). Relevance of comparative leaf anatomy in taxonomic
and functional research on the South African Poaceae. Thesis,
University of Pretoria, South Africa, Dsc.
Grass Phylogeny Working Group II (2012). New grass phylogeny resolves
deep evolutionary relationships and discovers C
4
origins. New Phytol.,
193, 304–312.
Hatch, M.D. (1987). C
4
photosynthesis: a unique blend of modified
biochemistry, anatomy and ultrastructure. Biochim. Biophys. Acta, 895,
81–106.
Hertz, P.E., Arima, Y., Harrison, A., Huey, R.B., Losos, J.B. & Glor,
R.E. (2013). Asynchronous evolution of physiology and morphology in
Anolis lizards. Evolution, 67, 2101–2113.
Hoetzel, S., Dupont, L., Schefuss, E., Rommerskirchen, F. & Wefer, G.
(2013). The role of fire in Miocene to Pliocene C4 grassland and
ecosystem evolution. Nat. Geosci., 6, 1027–1030.
Ibrahim, D.G., Gilbert, M.E., Ripley, B.S. & Osborne, C.P. (2008).
Seasonal differences in photosynthesis between the C
3
and C
4
subspecies of Alloteropsis semialata are offset by frost and drought.
Plant, Cell Environ., 31, 1038–1050.
Ibrahim, D.G., Burke, T., Ripley, B.S. & Osborne, C.P. (2009). A
molecular phylogeny of the genus Alloteropsis (Panicoideae, Poaceae)
suggests an evolutionary reversion from C
4
to C
3
photosynthesis. Ann.
Bot., 103, 127–136.
Kellermann, V., Loeschcke, V., Hoffmann, A.A., Kristensen, T.N.,
Fløjgaard, C., David, J.R. et al. (2012). Phylogenetic constraints in key
functional traits behind species’ climate niches: patterns of desiccation
and cold resistance across 95 Drosophila species. Evolution, 66, 3377–
3389.
L^
e, S., Josse, J. & Husson, F. (2008). FactoMineR: an R package for
multivariate analysis. J. Stat. Softw., 25, 1–18.
Lehmann, C.E.R., Archibald, S.A., Hoffmann, W.A. & Bond, W.J.
(2011). Deciphering the distribution of the savanna biome. New
Phytol., 191, 197–209.
Liebenberg, E.J.L. & Fossey, A. (2001). Comparative cytogenetic
investigation of the two subspecies of the grass Alloteropsis semialata
(Poaceae). Bot. J. Lin. Soc., 137, 243–248.
Long, S.P. (1999). Environmental responses. In: C4 Plant Biology (eds
Sage, R.F., Monson, R.K.). Academic Press, San Diego, pp. 215–249.
Monteith, J.L. (1978). A reassessment of maximum growth rates for C
3
and C
4
crops. Exp. Agr., 14, 1–5.
Ogren, W.L. (1984). Photorespiration: pathways, regulation, and
modification. Annu. Rev. Plant. Physiol., 35, 415–442.
Osborne, C.P. & Freckleton, R.P. (2009). Ecological selection pressures
for C
4
photosynthesis in the grasses. Proc. R. Soc. B, 276, 1753–1760.
Osborne, C.P., Wythe, E.J., Ibrahim, D.G., Gilbert, M.E. & Ripley, B.S.
(2008). Low temperature effects on leaf physiology and survivorship in
the C
3
and C
4
subspecies of Alloteropsis semialata.J. Exp. Bot., 59,
1743–1754.
Paradis, E., Claude, J. & Strimmer, K. (2004). APE: analyses of
phylogenetics and evolution in R language. Bioinformatics, 20, 289–290.
Pearcy, R.W. & Ehleringer, J. (1984). Comparative ecophysiology of C
3
and C
4
plants. Plant, Cell Environ.,7,1–13.
Petitpierre, B., Kueffer, C., Broennimann, O., Randin, C., Daehler, C. &
Guisan, A. (2012). Climatic niche shifts are rare among terrestrial plant
invaders. Science, 335, 1344–1348.
Pound, M.J., Tindall, J., Pickering, S.J., Haywood, A.M., Dowsett, H.J.
& Salzmann, U. (2014). Late Pliocene lakes and soils: a global data set
for the analysis of climate feedbacks in a warmer world. Clim. Past, 10,
167–180.
Ripley, B.S., Gilbert, M.E., Ibrahim, D.G. & Osborne, C.P. (2007).
Drought constraints on C
4
photosynthesis: stomatal and metabolic
limitations in C
3
and C
4
subspecies of Alloteropsis semialata.J. Exp.
Bot., 58, 1351–1363.
Sage, R.F. & Stata, M. (2014). Photosynthetic diversity meets
biodiversity: the C
4
plant example. J. Plant Physiol., 172, 104–119.
Sage, R.F., Christin, P.A. & Edwards, E.J. (2011). The C
4
plant lineages
of planet Earth. J. Exp. Bot., 62, 3155–3169.
Sage, R.F., Sage, T.L. & Kocacinar, F. (2012). Photorespiration and the
evolution of C
4
photosynthesis. Annu. Rev. Plant Biol., 63, 19–47.
Simon, M.F., Grether, R., de Queiroz, L.P., Skema, C., Pennington, R.T.
& Hughes, C.E. (2009). Recent assembly of the Cerrado, a neotropical
plant diversity hotspot, by in situ evolution of adaptations to fire. Proc.
Natl Acad. Sci. USA, 106, 20359–20364.
Skillman, J.B. (2008). Quantum yield variation across the three pathways
of photosynthesis: not yet out of the dark. J. Exp. Bot., 59, 1647–1661.
Slatyer, R.A., Hirst, M. & Sexton, J.P. (2013). Niche breadth predicts
geographical range size: a general ecological pattern. Ecol. Lett., 16,
1104–1114.
Spriggs, E.L., Christin, P.A. & Edwards, E.J. (2014). C
4
photosynthesis
promoted species diversification during the Miocene grassland
expansion. PLoS ONE, 9, e97722.
Still, C.J., Berry, J.A., Collatz, G.J. & DeFries, R.S. (2003). Global
distribution of C
3
and C
4
vegetation: carbon cycle implications. Global
Biogeoch. Cy., 17, 6.1–6.14.
©2015 John Wiley & Sons Ltd/CNRS
1028 M. R. Lundgren et al. Letter
Taub, D.R. (2000). Climate and the US distribution of C
4
grass
subfamilies and decarboxylation variants of C
4
photosynthesis. Am. J.
Bot., 87, 1211–1215.
Teeri, J.A. & Stowe, L.G. (1976). Climatic patterns and the distribution
of C
4
grasses in North America. Oecologia, 23, 1–12.
Vavrek, M.J. (2011). Fossil: palaeoecological and palaeogeographical
analysis tools. Palaeontol. Electron., 14, 1T.
Wand, S.J.E., Midgley, G.F. & Stock, W.D. (2002). Response to elevated
CO
2
from a natural spring in a C
4
-dominated grassland depends on
seasonal phenology. Afr. J. Range Forage Sci., 19, 81–91.
Wang, Y., Br€
autigam, A., Weber, A.P.M. & Zhu, X.G. (2014). Three
distinct biochemical subtypes of C
4
photosynthesis? A modelling
analysis. J. Exp. Bot., 65, 3567–3578.
Warren, B.H., Strasberg, D., Bruggemann, J.H., Prys-Jones, R.P. &
Th
ebaud, C. (2010). Why does the biota of the Madagascar region
have such a strong Asiatic flavour? Cladistics, 26, 526–538.
SUPPORTING INFORMATION
Additional Supporting Information may be downloaded via
the online version of this article at Wiley Online Library
(www.ecologyletters.com).
Editor, John Pannell
Manuscript received 30 June 2015
Manuscript accepted 3 July 2015
©2015 John Wiley & Sons Ltd/CNRS
Letter Photosynthetic innovation broadens niche 1029
