Abstract

C4 photosynthesis is a plant adaptation to high levels of photorespiration. Physiological models predict that atmospheric CO2 concentration selected for C4 grasses only after it dropped below a critical threshold during the Oligocene (∼30 Ma), a hypothesis supported by phylogenetic and molecular dating analyses. However the same models predict that CO2 should have reached much lower levels before selecting for C4 eudicots, making C4 eudicots younger than C4 grasses. In this study, different phylogenetic datasets were combined in order to conduct the first comparative analysis of the age of C4 origins in eudicots. Our results suggested that all lineages of C4 eudicots arose during the last 30 million years, with the earliest before 22 Ma in Chenopodiaceae and Aizoaceae, and the latest probably after 2 Ma in Flaveria. C4 eudicots are thus not globally younger than C4 monocots. All lineages of C4 plants evolved in a similar low CO2 atmosphere that predominated during the last 30 million years. Independent C4 origins were probably driven by different combinations of specific factors, including local ecological characteristics such as habitat openness, aridity, and salinity, as well as the speciation and dispersal history of each clade. Neither the lower number of C4 species nor the frequency of C3–C4 intermediates in eudicots can be attributed to a more recent origin, but probably result from variation in diversification and evolutionary rates among the different groups that evolved the C4 pathway.

Introduction

C4 photosynthesis is a highly convergent trait that has evolved more than 60 times in at least 18 families of flowering plants (Sage et al., 2011a). It consists of anatomical and biochemical innovations that increase the internal CO2 concentration around the carboxylating enzyme Rubisco, thereby suppressing oxygenation of RuBP and photorespiration. C4 photosynthesis thus provides an advantage in all conditions where photorespiration levels are deleteriously high, especially in warm, dry, and/or saline habitats (Sage, 2004). However, the energetic requirements of the C4 pathway can represent a net cost in conditions where photorespiratory rates are low, most notably at low temperature or in atmospheres of elevated CO2. One index that effectively describes the relationship between the efficiency of C3 and C4 photosynthesis and the environment is the maximum quantum yield of photosynthesis (QY) measured at low light intensities (Ehleringer et al., 1997). Maximum QY (hereafter referred to as simply QY) is a measure of the maximum light use efficiency of photosynthesis, calculated as the number of CO2 molecules fixed per absorbed photon. As an index of photorespiratory inhibition, QY is useful in C3 and C4 comparisons both at low light where variation in light use efficiency is directly proportional to carbon gain, and at high light, where QY is inversely proportional to the ratio of photorespiration to photosynthesis (Sage and Kubien, 2003). Due to the increase in photorespiration, QY in C3 plants declines with decreasing CO2 levels and increasing temperature. By contrast, C4 QY is little affected by variation in CO2 and temperature, but is below that of C3 plants at high CO2 and/or low temperature due to the two ATP equivalents needed for each turn of the C4 metabolic cycle (Ehleringer et al., 1997).

Because of its simplicity and close relationship with the ratio of photorespiration to photosynthesis, QY has been widely used to compare C3 and C4 species across a range of environments, and to model conditions, both present and past, when the C4 pathway would be advantageous over the C3 pathway. Using a QY model, Ehleringer and co-workers (Ehleringer et al., 1991, 1997; Cerling et al., 1997) predicted that C4 plants would have greater evolutionary fitness under the low CO2 levels of recent geological time than C3 species, leading to the hypothesis that low CO2 was a selection factor for the rise of the C4 pathway. With improved estimates of paleo atmospheric CO2 and evidence of C4 plants occurring in the Mid-Miocene, the date proposed for the earliest C4 plants has been placed in the late Oligocene period, when global climate underwent important changes and atmospheric CO2 levels fell to near or below current levels (Sage, 2001, 2004; Pagani et al., 2005). This hypothesis has been empirically supported in C4 monocots (grasses and sedges), where phylogenetic modelling indicates a marked increase in the probability of C4 evolution after the Oligocene CO2 decline (Christin et al., 2008; Besnard et al., 2009). It has not been tested in the numerous eudicot lineages with C4 species, however. QY comparisons generally demonstrated lower values of light use efficiency in C4 eudicots than C4 monocots (Ehleringer and Pearcy, 1983). This observation led to the hypothesis that C4 evolution in eudicots would have only occurred at lower atmospheric CO2 levels than in monocots, constraining C4 evolution in eudicots to the Pleistocene (Ehleringer et al., 1997). This prediction can now be evaluated using phylogenetic information accumulated in the past 15 years for eudicot lineages where C4 photosynthesis appears.

Phylogenetic data have increased exponentially in recent years, leading to a better understanding of the relationships between C4 taxa and affiliated C3 taxa. The grasses (Poaceae) have received the lion's share of phylogenetic attention and the number and timing of C4 origins in this group are now relatively well understood, with a minimum of 17 C4 origins, starting around 30 million years ago (Ma) and continuing to more recent geological time (Giussani et al., 2001; Christin et al., 2008; Vicentini et al., 2008; Edwards et al., 2010; Roalson, 2011). This pattern, along with considerations of the ecological setting for C4 evolution, has led to the current view that Oligocene CO2 decline met an essential environmental precondition for the evolution of C4 photosynthesis, while individual C4 origins were probably driven by additional local factors, such as warmth, seasonality, and habitat openness (Sage, 2001, 2004; Roalson, 2008; Osborne and Freckleton, 2009; Edwards and Smith, 2010). Besides grasses, five C4 origins in Cyperaceae (sedges) were estimated to have occurred between 20 and 4 Ma (Besnard et al., 2009; Roalson, 2011). The C4 origins in sedges thus also occurred in a low CO2 world, perhaps as a response to increasing disturbance and fire frequencies in the wetter parts of the warm biomes (Linder and Rudall, 2005). Together, sedges and grasses encompass about 80% of all C4 species, but represent only a minority of C4 origins, the majority of which (about 60%) occurred in eudicots (Sage et al., 2011a). Some of these eudicot groups serve as model systems for the study of C4 genetics and evolution, including members of the Cleomaceae, Amaranthaceae, and, especially, Flaveria in the Asteraceae (Kadereit et al., 2003; Svensson et al., 2003; Brown et al., 2005; McKown et al., 2005; McKown and Dengler, 2007). It has long been postulated that C4 eudicots were much younger than C4 monocots, based mainly on the physiological models described above, but also on the presence of C3–C4 intermediates and the lower number of C4 species in eudicot groups (Ehleringer et al., 1997; Sage, 2004). Some well-resolved and relatively well-sampled phylogenies are now available for several groups of C4 eudicots (McKown et al., 2005; Kapralov et al., 2006; Sage et al., 2007; Feodorova et al., 2010; Kadereit et al., 2010; Ocampo and Columbus, 2010; Christin et al., 2011). However, accurate time calibration of these phylogenies has been hampered by the lack of fossil records for the studied groups.

The first attempt to date C4 eudicots used a phylogeny based on rbcL for Amaranthaceae/Chenopodiaceae, which suggested that the first C4 origins in this group could have occurred more than 14.5 Ma (Kadereit et al., 2003), challenging the hypothesis of a Pleistocene origin of C4 eudicots. This potential early origin of C4 Amaranthaceae/Chenopodiaceae has been confirmed by more densely sampled phylogenetic analyses (Kadereit et al., 2010; Kadereit and Freitag, 2011). Very recently, two more studies have used time estimates obtained for the major angiosperm lineages to calibrate phylogenies for Molluginaceae and Cleomaceae (Feodorova et al., 2010; Christin et al., 2011). Surprisingly, these C4 groups with very few species (two, and less than five, respectively) were found to be up to 10 million years (My) old. These new lines of evidence demonstrate the need for a re-evaluation of the timing of C4 evolution in eudicots and their relationship with past fluctuation of atmospheric CO2 levels.

The goal of the present study was to estimate the potential ages of the different eudicot C4 lineages described in the literature. First, a phylogenetic tree containing the major eudicot clades was inferred and dated and then these time estimates were used to calibrate a series of smaller and more detailed phylogenies inferred from fast-evolving markers. The resultant time-calibrated phylogenetic framework contains many of the postulated C4 origins in eudicots and is used to address variation in the probability of C4 evolution through geological time. This study also identified groups for which phylogenetic information is limited or completely absent, but are critical for understanding the timing, evolvability, and reversibility of transitions between C3 and C4 photosynthesis in angiosperms.

Materials and methods

General methodology and eudicot phylogeny

The GenBank database was screened for genes available for different C4 lineages (according to Sage et al., 2011a). Non-coding genes, such as nuclear internal transcribed spacers (ITS) and intergenic spacers from the plastid genome, are available for numerous C4 groups. Unfortunately, these fast-evolving markers cannot be unambiguously aligned among distantly related plant groups. Therefore, slower-evolving coding markers were selected that allowed the largest number of C4 taxa to be incorporated. The two markers settled upon were genes from the plastid genome encoding matK and rbcL. They have been used in various attempts to reconstruct the angiosperm phylogeny (Soltis et al., 2000; Hilu et al., 2003) and are relatively well sampled for several groups that contain C3/C4 transitions (Cuénoud et al., 2002; Kadereit et al., 2003; Sage et al., 2007; Christin et al., 2011). These two markers were retrieved for species spanning the main eudicot lineages, the purported sister group of eudicots (Ceratophyllum) and one monocot, which was used to root the tree (Acorus). Clades previously reported to contain C4 taxa were densely sampled and, where possible, the closest C3 relatives of each C4 group was included in order to improve the estimates of divergence times. This sampling was completed by de novo sequencing of matK and rbcL for several Caryophyllales species in order to improve the resolution of the phylogeny and, in particular, the relationships between C3 and C4 taxa (see Supplementary Document S1 at JXB online). These markers were amplified and sequenced with the primers developed in Christin et al. (2011), following the procedure described therein.

Sequences were aligned with ClustalW (Thompson et al., 1994) and the alignment was then manually edited. A phylogenetic tree was inferred simultaneously from rbcL and matK through a Bayesian procedure implemented in MrBayes 3.2 (Ronquist and Huelsenbeck, 2003). The substitution model used for the analyses was a general time-reversible model with a gamma shape parameter and a proportion of invariant sites (GTR+G+I). Two analyses, each consisting of four parallel chains, were run for 7 000 000 generations after a burn-in period of 3 000 000, and a tree was sampled each 1000 generations. The consensus phylogeny resulting from the 14 000 sampled trees was used for molecular dating with a Bayesian method that accounts for changes in rates of evolution among branches, following the recommendations of Rutschmann (2006). Model parameters were estimated with baseml (Yang, 2007) for the two genes separately. Branch lengths and the variance–covariance matrix were then optimized using estbranches (Thorne et al., 1998). A Bayesian MCMC procedure implemented in multidivtime (Kishino et al., 2001; Thorne and Kishino, 2002) approximated the posterior distributions of substitution rates and divergence times, given a set of time constraints. The MCMC procedure was run for 1 000 000 generations after a burn-in of 100 000 generations, with a sampling frequency of 100 generations. The outgroup (Acorus americanus) was removed during the analysis. Calibration points, based on reported fossils, were set as described in Christin et al. (2011), with lower bounds on the stems of Buxales at 102.2 Ma, of Malpighiales at 91.2 Ma, of Fabales at 59.9 Ma, of Malvales at 69.7 Ma, of Myrtales at 88.2 Ma, of Ericales at 91.2 Ma, a lower bound on the divergence of Polycarpon from the higher Caryophyllaceae at 34 Ma, a lower bound on the stem of eudicots at 120 Ma and an upper bound on the crown of eudicots at 130 Ma.

This eudicot phylogeny gave relatively good estimates for several C4 origins, but many groups were poorly represented. For this reason, additional phylogenies were reconstructed with fast-evolving markers for several groups containing C4 taxa. Since good fossils are not available for most groups at this taxonomic scale, time estimates for nodes of the eudicot phylogeny, together with their associated confidence interval, were used to calibrate these lower scale phylogenies. With this approach, the ages estimated for the different groups are not independent, increasing the probability of accurate estimates of the relative ages, even in the presence of misleading calibration points. For constrained nodes, the lower bound was set to the estimated divergence time minus the standard deviation and the upper bound to the estimated divergence time plus the standard deviation. All trees were inferred and calibrated with the method described above. Specific details for each group are detailed below.

To account for recent suggestions that angiosperms could be older than previously thought (Smith et al., 2010), dating of the eudicot tree was repeated by removing the upper constraint on the crown of eudicots and setting the maximal age of the root (eudicots+Ceratophyllum) to 200 Ma. Divergence times estimated in this analysis were then used to calibrate lower taxonomic scale phylogenies under the hypothesis of an earlier origin of angiosperms.

Detailed analyses of selected clades

The phylogeny for Flaveria was reconstructed using the nuclear ITS and the plastid intergenic spacer trnL-trnF. Sequences for Flaveria were extracted from the dataset of McKown et al. (2005) and other Asteraceae were added to allow for more calibration points. Only one accession per Flaveria species was considered. Scaevola aemula (Goodeniaceae) was used as the outgroup (removed during the dating analysis). The inferred phylogeny was calibrated with the estimated divergence times between Cichorioideae (represented by Lactuca in the eudicot phylogeny) and Asteroideae (i.e. Helianthus and Flaveria) and between Helianthus and Oyedaea. The photosynthetic types were determined by McKown et al. (2005) and McKown and Dengler (2007).

A phylogeny for the Chenopodiaceae was inferred from nuclear ITS and the non-coding plastid marker psbB-psbH, from data originating from previous studies (Kapralov et al., 2006; Akhani et al., 2007; Wen et al., 2010). The phylogeny was rooted with Chenopodioideae (Atriplex and Chenopodium, removed during the dating analysis), according to previous phylogenies (Kadereit et al., 2003; Kapralov et al., 2006). It was calibrated with the estimated divergence time between Salsoloideae (Salsola, Camphorosma, and Bassia) and Suaedoideae/Salicornioideae (Bienertia, Allenrolfea, and Halocnemum). Photosynthetic types were reported from the literature (Jacobs, 2001; Kadereit et al., 2003, and references therein, 2010; Kapralov et al., 2006; Akhani et al., 2007).

A phylogenetic tree for C4 and related C3 Nyctaginaceae was inferred from four markers, the nuclear ITS, the plastid coding gene ndhF, and the plastid non-coding genes rpl16 and rps16. These data come from a densely sampled phylogenetic dataset by Douglas and Manos (2007). The phylogenetic tree was rooted with Colignonia scandens, according to Douglas and Manos (2007). It was calibrated with the estimated divergence times between Bougainvillea glabra and the other species and between the Mirabilis genus and the group composed of Allionia/Okenia. Photosynthetic types were deduced from carbon isotope ratio measurements (RF Sage, unpublished data).

A phylogeny was inferred using ITS for Sesuvioideae (data generated by Hassan et al., 2005), other Aizoaceae, and Montiaceae (used as the outgroup, removed in the dating analysis). The resulting phylogeny was calibrated using the estimated divergence time between Galenia+Delosperma+Trichodiadema and Sesuvioideae and the estimated divergence time between Galenia and Delosperma+Trichodiadema. Photosynthetic types were reported from previous studies (Sankhla et al., 1975; Kocacinar and Sage, 2003; RF Sage, unpublished data).

GenBank was screened for markers representing Euphorbia species previously identified as C4. ITS was selected as it was available for a large number of C4 as well as C3Euphorbia. Euphorbia species spread across the phylogeny were selected together with Manihot esculenta and Jatropha curcas (used as the outgroup, removed during the dating analysis). The resulting phylogeny was calibrated with the estimated divergence time between Manihot esculenta and the Euphorbia genus. Photosynthetic types were reported from the literature (Pearcy and Troughton, 1975; Webster et al., 1975; Batanouny et al., 1991; Sage et al., 2011b).

Modelling of photosynthetic transition probabilities

Two models of transition between C3 and C4 photosynthesis were optimized on the eudicot phylogeny inferred from rbcL and matK, with branch lengths estimated from molecular markers and incorporating the estimated divergence times. The null model allows different transition rates from C3 to C4 and from C4 to C3, but these rates are constant through time. The alternative model allows for one change in the transition rates, independently optimizing transition probabilities between C3 and C4 before and after a given time threshold (Christin et al., 2008). The two models are nested and can be compared through likelihood ratio tests. Species were coded as C4 or C3 (including C3–C4 intermediates) and all parameters were estimated from the data, using MLtree software (Christin et al., 2008).

Results

Eudicot phylogeny

The phylogeny inferred from matK and rbcL (see Supplementary Fig. S1A, B at JXB online) is congruent with the known relationships among angiosperms (APG III, 2009; Brockington et al., 2009), and our age estimates for major clades are consistent with recent estimates (see Supplementary Fig. 1A at JXB online; Bell et al., 2010; Moore et al., 2010). However, to account for recent suggestions that angiosperms could be older than previously thought (Smith et al., 2010), the dating analysis was repeated with relaxed constraints on the maximal age of the eudicots and the root. Removing these constraints affected the estimated ages of the major clades, but did not significantly change those for nodes closer to the tips, which included all potential C4 origins (see Supplementary Table S1 at JXB online).

Our sampling of rbcL and matK incorporated C4 taxa forming a total of 31 C4-like or C4 groups (Table 1). Of the 36 C4 lineages of eudicots hypothesized by Sage et al. (2011a), eight are missing because data in GenBank were insufficient (Blepharis, Anticharis, Heliotropium, Pectis, Chrysanthellum/Isostigma, Polycarpaea, and two C4 groups of Cleome). The two C4Atriplex, for which only rbcL was available, were not monophyletic, congruent with previous analysis of this marker (Kadereit et al., 2003). However, more recent analyses with other markers and a denser sampling have shown the monophyly of the C4Atriplex (Kadereit et al., 2010). All other relationships are perfectly congruent with previous phylogenies for C4 eudicots (Kadereit et al., 2003; Kapralov et al., 2006; Akhani et al., 2007; Sage et al., 2007; Christin et al., 2011).

Table 1.

Estimated ages (in millions of years, with standard deviation in brackets) of stem and crown nodes for the different photosynthetic transitions, assuming no reversal

Cladea Transition Stem age Crown age Remarks 
Portulacaceae 
Portulaca C3 to C4 28.8 (5.0) 9.7 (3.3) For details, see Ocampo et al. (2010) 
Molluginaceae 
Mollugo cervianaC3 to C3–C4 17.1 (3.5) 7.0 (2.0) For details, see Christin et al. (2011) 
Hypetertelis spergulacea 
Mollugo cerviana group C3–C4 to C4 7.0 (2.0) 0.5 (0.4) For details, see Christin et al. (2011) 
Mollugo fragilis group C3–C4 to C4 5.3 (1.6) 1.4 (0.6) For details, see Christin et al. (2011) 
Gisekiaceae 
Gisekia C3 to C4 4.8 (3.6)  
Amaranthaceae 
Gomphreneae C3 to C4 8.6 (2.5) 7.0 (2.2)  
Tidestromia C3 to C4 13.0 (3.3) 3.2 (1.7)  
Alternanthera C3 to C3–C4 10.3 (2.9) 7.5 (2.4)  
Alternanthera C3–C4 to C4 7.5 (2.4) 5.9 (2.1)  
Aerva C3 to C4 14.5 (3.7) 2.5 (1.5)  
Amaranthus C3 to C4 15.4 (4.1) 1.8 (1.2)  
Chenopodiaceae 
Atriplex C3 to C4 12.9 (2.4) 4.7 (2.3) Recent evidence suggests one C4 origin in Atriplex 14.1-10.9 Mya (Kadereit et al., 2010
Suaeda sect. Salsina C3 to C4 9.9 (3.1) 7.0 (2.4)  
Suaeda sect. Schoberia C3 to C4 5.6 (2.1) 4.5 (1.8)  
Bienertia C3 to single- 20.8 (3.9) 2.6 (2.0)  
 celled C4    
Suaeda aralocaspica C3 to single- 7.7 (3.2)  
 celled C4    
Halosarcia indica C3 to C4 6.5 (3.9)  
Camphorosmeae C3 to C4 16.7 (4.6) 13.1 (4.1) Recent evidence suggests two C4 origins, 21.3-8.3 and 13.8-9.8 Mya (Kadereit and Freitag, 2011
Caroxyloneae C3 to C4 24.6 (3.4) 22.1 (3.6)  
Salsola kali group C3 to C4 22.9 (3.5) 17.9 (4.0)  
Halothamnus C3 to C4 21.2 (3.5) 7.2 (3.3)  
Haloxylon/Anabis C3 to C4 13.7 (2.9) 12.7 (2.8)  
Nyctaginaceae 
Boerhavia C3 to C4 4.7 (0.9) 2.2 (0.5)  
Allionia C3 to C4 6.1 (1.0)  
Aizoaceae 
Trianthema C3 to C4 22.1 (4.9) 20.2 (4.7) Postulating all Trianthema are C4 
Zaleya C3 to C4 20.1 (4.8) 4.2 (3.4) Postulating all Zaleya are C4 
Sesuvium C3 to C4 4.8 (3.3)  
Cypselea C3 to C4 11.3 (4.1)  
Polygonaceae 
Calligonum C3 to C4 19.9 (5.2) 1.2 (1.1)  
Asteraceae     
Flaveria C3 to C3–C4 3.1 (0.9) 2.9 (0.8)  
Flaveria sonorensis C3 to C3–C4 2.8 (0.8)  
Flaveria brownii C3–C4 to C4-like 0.4 (0.2)  
Flaveria clade A1 C3–C4 to C4-like 1.8 (0.6) 1.3 (0.5)  
F. campestris C4-like to C4 0.5 (0.3)  
F. trinervia group C4-like to C4 1.0 (0.4) 0.2 (0.2)  
F. bidentis C4-like to C4 0.6 (0.3)  
F. kochiana C4-like to C4 0.3 (0.2)  
Cleomaceae 
Cleome gynandra C3 to C4 17.7 (4.1) Recent evidence suggests five C4 origins in Cleome in the last 10 My (Feodorova et al., 2010
Zygophyllaceae 
Zygophyllum C3 to C4 8.6 (5.3)  
Tribulus C3 to C4 63.5 (7.6) 13.8 (7.2)  
Euphorbiaceae 
Chamaesyce C3 to C4 10.4 (2.7) 7.4 (2.2)  
Cladea Transition Stem age Crown age Remarks 
Portulacaceae 
Portulaca C3 to C4 28.8 (5.0) 9.7 (3.3) For details, see Ocampo et al. (2010) 
Molluginaceae 
Mollugo cervianaC3 to C3–C4 17.1 (3.5) 7.0 (2.0) For details, see Christin et al. (2011) 
Hypetertelis spergulacea 
Mollugo cerviana group C3–C4 to C4 7.0 (2.0) 0.5 (0.4) For details, see Christin et al. (2011) 
Mollugo fragilis group C3–C4 to C4 5.3 (1.6) 1.4 (0.6) For details, see Christin et al. (2011) 
Gisekiaceae 
Gisekia C3 to C4 4.8 (3.6)  
Amaranthaceae 
Gomphreneae C3 to C4 8.6 (2.5) 7.0 (2.2)  
Tidestromia C3 to C4 13.0 (3.3) 3.2 (1.7)  
Alternanthera C3 to C3–C4 10.3 (2.9) 7.5 (2.4)  
Alternanthera C3–C4 to C4 7.5 (2.4) 5.9 (2.1)  
Aerva C3 to C4 14.5 (3.7) 2.5 (1.5)  
Amaranthus C3 to C4 15.4 (4.1) 1.8 (1.2)  
Chenopodiaceae 
Atriplex C3 to C4 12.9 (2.4) 4.7 (2.3) Recent evidence suggests one C4 origin in Atriplex 14.1-10.9 Mya (Kadereit et al., 2010
Suaeda sect. Salsina C3 to C4 9.9 (3.1) 7.0 (2.4)  
Suaeda sect. Schoberia C3 to C4 5.6 (2.1) 4.5 (1.8)  
Bienertia C3 to single- 20.8 (3.9) 2.6 (2.0)  
 celled C4    
Suaeda aralocaspica C3 to single- 7.7 (3.2)  
 celled C4    
Halosarcia indica C3 to C4 6.5 (3.9)  
Camphorosmeae C3 to C4 16.7 (4.6) 13.1 (4.1) Recent evidence suggests two C4 origins, 21.3-8.3 and 13.8-9.8 Mya (Kadereit and Freitag, 2011
Caroxyloneae C3 to C4 24.6 (3.4) 22.1 (3.6)  
Salsola kali group C3 to C4 22.9 (3.5) 17.9 (4.0)  
Halothamnus C3 to C4 21.2 (3.5) 7.2 (3.3)  
Haloxylon/Anabis C3 to C4 13.7 (2.9) 12.7 (2.8)  
Nyctaginaceae 
Boerhavia C3 to C4 4.7 (0.9) 2.2 (0.5)  
Allionia C3 to C4 6.1 (1.0)  
Aizoaceae 
Trianthema C3 to C4 22.1 (4.9) 20.2 (4.7) Postulating all Trianthema are C4 
Zaleya C3 to C4 20.1 (4.8) 4.2 (3.4) Postulating all Zaleya are C4 
Sesuvium C3 to C4 4.8 (3.3)  
Cypselea C3 to C4 11.3 (4.1)  
Polygonaceae 
Calligonum C3 to C4 19.9 (5.2) 1.2 (1.1)  
Asteraceae     
Flaveria C3 to C3–C4 3.1 (0.9) 2.9 (0.8)  
Flaveria sonorensis C3 to C3–C4 2.8 (0.8)  
Flaveria brownii C3–C4 to C4-like 0.4 (0.2)  
Flaveria clade A1 C3–C4 to C4-like 1.8 (0.6) 1.3 (0.5)  
F. campestris C4-like to C4 0.5 (0.3)  
F. trinervia group C4-like to C4 1.0 (0.4) 0.2 (0.2)  
F. bidentis C4-like to C4 0.6 (0.3)  
F. kochiana C4-like to C4 0.3 (0.2)  
Cleomaceae 
Cleome gynandra C3 to C4 17.7 (4.1) Recent evidence suggests five C4 origins in Cleome in the last 10 My (Feodorova et al., 2010
Zygophyllaceae 
Zygophyllum C3 to C4 8.6 (5.3)  
Tribulus C3 to C4 63.5 (7.6) 13.8 (7.2)  
Euphorbiaceae 
Chamaesyce C3 to C4 10.4 (2.7) 7.4 (2.2)  
a

Clade names follow Sage et al. (2011a).

For each C4 group, C4 photosynthesis could have originated at any time between the divergence of the C4 group and its C3 sister group (stem group node), and the earliest evidence of diversification within the C4 group (crown group node; Table 1). Thus, the limited species sampling in our rbcL/matK tree produces large uncertainties about the timing of some C4 origins. In addition, many groups are only represented by a single species, which makes it impossible to define a minimum age bracket with a crown group. The upper estimate for the C4 origin in Portulaca is at 28.8 Ma (with a minimum age for this origin at 9.7 Ma; Table 1). Dates for the Portulacineae suborder (which contains Portulaca) are older than in a previous study that calibrated the phylogeny with the appearance of the Hawaiian islands (Ocampo and Columbus, 2010). C4 photosynthesis could be 63.5 My old in the Tribulus group (with a minimum age of 13.8 Ma), but the only markers available for these species were rbcL or a small fragment of matK, decreasing the accuracy of the age estimation and leading to very wide confidence intervals (Table 1). All other C4 origins are estimated to have occurred during the last 30 My (Table 1).

Detailed analyses of selected clades

According to our time estimates, the appearance of C3–C4 photosynthesis in Flaveria occurred either once between 3.6 and 3.1 Ma (with a reversal in F. robusta) or twice, between 2.8 Ma and the present in F. sonorensis, and between 3.1 and 2.9 in the common ancestor of clades A and B (see Supplementary Fig. S1C at JXB online). This C3-C4 type was co-opted twice to evolve a C4-like trait, between 0.4 Ma and the present in F. brownii and between 1.8 and 1.3 Ma in clade A (Fig. 1). The number of C4 origins from the C4-like type are difficult to infer, since different scenarios are equally probable; in all cases, they occurred in the last 2 My (Table 1).

Fig. 1.

Temporal representation of the gradual C3 to C4 transition in Flaveria. According to McKown and Dengler (2007), the ancestor of the Flaveria group evolved increased vein density that represent a C4 precondition (light grey). This was co-opted twice to evolve C3-C4 photosynthesis (grey), which in turn was co-opted twice to evolve a C4-like trait (dark grey). The C4-like ancestors, in turn, gave rise to multiple C4 descendants (black). Note that this scenario assumes strictly directional C3 to C4 transitions but reversals between the different states cannot be excluded. For each group, triangles or thin bars represent the interval between the stem and crown nodes (interval in which the trait could have evolved).

Fig. 1.

Temporal representation of the gradual C3 to C4 transition in Flaveria. According to McKown and Dengler (2007), the ancestor of the Flaveria group evolved increased vein density that represent a C4 precondition (light grey). This was co-opted twice to evolve C3-C4 photosynthesis (grey), which in turn was co-opted twice to evolve a C4-like trait (dark grey). The C4-like ancestors, in turn, gave rise to multiple C4 descendants (black). Note that this scenario assumes strictly directional C3 to C4 transitions but reversals between the different states cannot be excluded. For each group, triangles or thin bars represent the interval between the stem and crown nodes (interval in which the trait could have evolved).

The reconstruction of C3/C4 transitions is highly ambiguous in the inferred Chenopodiaceae phylogeny and the most parsimonious scenarios would imply C4 to C3 reversals (see Supplementary Fig. S1D at JXB online). Such reversions have been hypothesized (Pyankov et al., 2001), but other studies concluded that the family contains mainly C4 origins based on anatomical variation (Kadereit et al., 2003, 2010; Kadereit and Freitag, 2011). According to our estimate, the evolution of single-celled C4 photosynthesis (Edwards et al., 2004) occurred between 20.8 and 2.6 Ma in Bienertia and between 7.7 Ma and the present in Suaeda aralocaspica (Table 1). If two transitions from C3 to C4 are assumed in Suaeda, they would have occurred between 9.9 and 7.0 Ma and between 5.6 and 4.5 Ma, respectively. In Salsoloideae, a single origin followed by reversals would be estimated between 28.3 and 26.2 Ma. In the case of multiple C4 origins in Salsoloideae, these would be spread across the last 25 My (Table 1). The oldest lower estimate for a C4 origin in the eudicots is in the Caroxyloneae at 22.1 Ma. Dates for C4 Chenopodiaceae are slightly younger than previous estimates based on phylogenetic trees encompassing only Amaranthaceae/Chenopodiaceae (Kadereit et al., 2003, 2010; Kadereit and Freitag, 2011).

In Nyctaginaceae, C4 taxa form two groups (see Supplementary Fig. S1E at JXB online). If they correspond to two C4 origins, the Allionia lineage evolved the C4 pathway between 6.1 Ma and the present, while C4 appeared between 4.7 and 2.2 Ma in the Boerhavia/Okenia group (Table 1).

The evolutionary history of photosynthetic types in Sesuvioideae is difficult to reconstruct (see Supplementary Fig. S1F at JXB online). The observed distribution of C4 taxa is compatible with a single C4 origin, which would have occurred between 29.5 and 22.1 Ma, followed by two reversals to a C3 state. Alternatively, in the case of multiple C4 origins, the Trianthema genus would have acquired its C4 trait between 22.1 and 20.2 Ma and the other genera during the last 20 My (Table 1). In either scenario, C4 photosynthesis in Sesuvioideae is at least 20 My old.

All the Euphorbia species reported in the literature as C4 and for which genetic data were available in GenBank form a single clade in the phylogeny (see Supplementary Fig. S1G at JXB online), sister to a group that contains C3 and C3-C4 taxa (Sage et al., 2011b). According to our estimates, C4 photosynthesis originated between 10.4 and 7.4 Ma in this group (Table 1).

Modelling of photosynthetic transition rates

The model allowing variation of transition rates through time was marginally better than the null model (log-likelihood=-86.65; χ2=7.49; df=3; P value=0.058). The significance of this alternative model should be re-evaluated when more data are available, but the present result, with limited taxon sampling, suggests that transition probabilities were not constant through time. The optimal threshold for the change of transition probabilities was at 28.8 Ma (Fig. 2), essentially identical to what was estimated for grasses (27.6 Ma; Christin et al., 2008). According to this model, the probability of C4 evolution was almost zero before 28.8 Ma, but strongly increased after this threshold. After 28.8 Ma, the probability of C4 to C3 reversal was positive in eudicots while it was null in grasses (Christin et al., 2008).

Fig. 2.

Comparison of transition models for eudicots and grasses. The likelihood of the model allowing different rates of C3/C4 transitions after a time threshold is presented for all possible thresholds between 50 and 0 million years ago (Ma), for (A) eudicots and (B) grasses. The graph for grasses was redrawn from Christin et al. 2008, and reprinted by kind permission of Elsevier Ltd © 2008. The vertical grey bar shows the overlap between the optima for grasses and eudicots.

Fig. 2.

Comparison of transition models for eudicots and grasses. The likelihood of the model allowing different rates of C3/C4 transitions after a time threshold is presented for all possible thresholds between 50 and 0 million years ago (Ma), for (A) eudicots and (B) grasses. The graph for grasses was redrawn from Christin et al. 2008, and reprinted by kind permission of Elsevier Ltd © 2008. The vertical grey bar shows the overlap between the optima for grasses and eudicots.

Discussion

Timing of C4 origins in the eudicots

C4 eudicots have been proposed to be of recent origin compared to C4 monocots (grasses and sedges), based on arguments such as a low number of C4 species and the existence of numerous C3–C4 intermediates (Ehleringer et al., 1997; Kellogg, 1999; Sage, 2004). A combination of physiological models and estimations of past climatic conditions even led to suggestions of a Pleistocene origin of C4 eudicots (Ehleringer et al., 1997). While our phylogenetic sampling was limited, which hampered an accurate estimate of the timing of C4 origins in several eudicot lineages, narrow confidence intervals were still obtained for others. Our analysis provided little evidence for Pleistocene origins of C4 photosynthesis in the eudicots. The only C4 origin for which the maximal bound was inferred in the Pleistocene is Flaveria (Table 1). Several origins, within Portulaca, Euphorbia, and most C4 groups of Aizoaceae, Amaranthaceae, and Chenopodiaceae, are estimated to have occurred in the Miocene or even the late Oligocene, supporting previous estimates (Kadereit et al., 2003, 2010; Ocampo and Columbus, 2010). Excluding the Chloridoideae subfamily of grasses (first estimated C4 origin 32–25 Ma; Christin et al., 2008), C4 origins in monocots are also spread throughout the last 25 My and largely overlap with C4 origins in eudicots (Fig. 3). C4 eudicots should thus no longer be considered young compared to monocots.

Fig. 3.

Comparison of the ages of C4 monocots and C4 eudicots. The estimated age is indicated for each C4 origin where the interval between the stem and crown nodes is smaller than 5 My, and for the putative oldest C4 origin in the subfamily Chloridoideae of grasses. For each group, thick bars represent the interval between stem and crown nodes, and thin bars the confidence interval. Monocots are in grey and eudicots in black. For Flaveria, only one of the transitions to the C4-like state is represented. Ages of monocots are based on Christin et al. (2008) and Besnard et al. (2009). Names on the right are numbered for eudicots, sedges, and grasses; e1, Gisekia; e2, Sesuvium sesuvioides; e3, C4 and C4-like Flaveria; e4, Boerhavia/Okenia; e5, Suaeda sect. Schoberia; e6, C4Alternanthera; e7, Gomphreneae; e8, Suaeda sect. Salsina; e9, Chamaesyce; e10, Haloxylon/Anabasis; e11, Camphorosmeae; e12, Kali; e13, Trianthema; e14 Caroxyloneae; s1, Eleocharis vivipara; s2, C4Rhynchospora; s3, C4 Cypereae; g1, Neurachne munroi; g2, Panicum prionitis group; g3, Eriachne; g4, Mesosetum clade; g5, main C4 Paniceae; g6, Andropogoneae; g7, Chloridoideae.

Fig. 3.

Comparison of the ages of C4 monocots and C4 eudicots. The estimated age is indicated for each C4 origin where the interval between the stem and crown nodes is smaller than 5 My, and for the putative oldest C4 origin in the subfamily Chloridoideae of grasses. For each group, thick bars represent the interval between stem and crown nodes, and thin bars the confidence interval. Monocots are in grey and eudicots in black. For Flaveria, only one of the transitions to the C4-like state is represented. Ages of monocots are based on Christin et al. (2008) and Besnard et al. (2009). Names on the right are numbered for eudicots, sedges, and grasses; e1, Gisekia; e2, Sesuvium sesuvioides; e3, C4 and C4-like Flaveria; e4, Boerhavia/Okenia; e5, Suaeda sect. Schoberia; e6, C4Alternanthera; e7, Gomphreneae; e8, Suaeda sect. Salsina; e9, Chamaesyce; e10, Haloxylon/Anabasis; e11, Camphorosmeae; e12, Kali; e13, Trianthema; e14 Caroxyloneae; s1, Eleocharis vivipara; s2, C4Rhynchospora; s3, C4 Cypereae; g1, Neurachne munroi; g2, Panicum prionitis group; g3, Eriachne; g4, Mesosetum clade; g5, main C4 Paniceae; g6, Andropogoneae; g7, Chloridoideae.

The contemporaneous nature of C4 monocots and eudicots emphasizes that neither the number of species nor the presence of C3–C4 intermediates accurately predicts the age of the different C4 groups. A recent study suggests that C3–C4 intermediacy has been evolutionarily stable for more than 20 My in some Molluginaceae (Christin et al., 2011) and, according to our present results, the C3–C4 type evolved at least 7.7 Ma in Alternanthera (Table 1). With regard to the species richness of various C4 clades, the Trianthema genus contains fewer than 20 species, but could have appeared more than 20 Ma. Conversely, Cyperus and affiliated taxa in sedges acquired the C4 pathway at less than 11 Ma but encompass more than 550 C4 species (Bruhl and Wilson, 2007; Besnard et al., 2009). Thus, there appears to be no relationship between the age of a C4 clade and the number of species it contains. The larger number of C4 species in monocots may instead be the consequence of grass and sedge families being prone to high diversification rates, since both lineages also contain highly diversified and widespread groups of C3 species (Bruhl and Wilson, 2007; Edwards and Smith, 2010).

Ecological drivers of C4 photosynthesis in eudicots

Past variations of atmospheric CO2 levels have long been viewed as instrumental in driving the evolution of C4 photosynthesis (Ehleringer et al., 1991, 1997; Sage, 2004). Our results provide further evidence in support of this hypothesis; however, the proposal that C4 eudicots evolved in the Pleistocene, only after CO2 fell to lower levels than promoted C4 evolution in the grasses, is not supported by our data. The atmospheric CO2 concentration is estimated to have drastically decreased around 30 Ma to below current levels (Pagani et al., 2005). Further CO2 declines during the last 10 My are difficult to reconstruct with certainty (Edwards et al., 2010), although recent evidence suggests that atmospheric CO2 fell below 350 ppm between 15 and 8 Ma and reached its lowest levels during glacial episodes in the Pleistocene (Tripati et al., 2009). Several C4 eudicot lineages evolved more than 20 Ma (Table 1), and our modelling of C3/C4 transitions suggested that the probability of C4 evolution increased at the same time in monocots and eudicots, around 28 Ma (Fig. 3). These time estimates are consistent with a global effect of the decline in atmospheric CO2 during the Oligocene, and could explain the widespread C4 origins in many different geographic regions beginning around 30 Ma (Sage et al., 2011a).

The similarity in the timing of C4 origins in eudicots and monocots is inconsistent with the physiologically-based predictions on the estimated lower QY of eudicots. A number of possibilities could explain this discrepancy. First, most C4 species occur in high light environments where QY is not directly limiting. At moderate to high light levels, photoprotective processes such as zeaxanthin quenching of excess light energy become engaged, reducing QY below the maximum values used in interspecies comparisons (Sage and Kubien, 2003). Hence, QY differences between C4 eudicots and monocots would not be directly related to carbon gain and fitness. Second, past CO2 levels cannot be reconstructed with certainty when lower than 500 ppm (Edwards et al., 2010), and post-Oligocene levels could have been below the threshold proposed for both the evolution of C4 monocot and eudicots, based on QY models (Ehleringer et al., 1997). Third, QY differences between C4 monocots and eudicots are small, and are not consistent between all groups (Skillman, 2008). While QY of C4 Amaranthaceae/Chenopodiaceae and Flaveria are clearly lower than those of C4 grasses analysed by the same authors, some C4Euphorbia had higher QY than many C4 grasses (Ehleringer and Bjorkman, 1977; Ehleringer and Pearcy 1983; Monson et al., 1986). In this light, QY differences among taxonomic groups could have been of limited importance for C4 evolution. Rather, ecological factors such as heat, aridity, and salinity might have been key determinants of C4 evolvability once the low CO2 precondition was met.

Length of the transition from C3 to C4

It has been proposed that the multiple transformations in anatomy and biochemistry required for C4 photosynthesis have been acquired gradually (Griffiths, 1989; Sage, 2004). The time required for the whole transition from a C3 ancestor to completely C4 descendants is difficult to estimate, since the precise moment at which the different C4 characteristics appeared cannot be easily mapped onto a phylogenetic tree (Christin et al., 2010). One approach is to consider the time that separates the first appearance of C4 characters from the first fully C4 node of the same lineage. In Flaveria, the first C3–C4 ancestor emerged between 3.1 and 2.9 Ma, but anatomical preconditions appeared earlier, between 3.6 and 3.1 Ma. The first C4 plants of Flaveria (if the possibility of reversals is excluded) evolved between 1 and 0.2 Ma (Table 1). This indicates that the transition from a C3 ancestor to C4 plants took at least 2 My in this clade (Fig. 1). In Molluginaceae, enlarged bundle-sheath cells evolved more than 20 Ma, probably in a C3 context, and were co-opted to evolve a C4 trait that was optimized in the last 1 My (Christin et al., 2011). This suggests that the complete transition from the appearance of C4 preconditions in a C3 context to completely developed C4 plants took at least 15 My (Fig. 4). The presence of anatomical C4 preconditions in taxa related to C4 lineages has also been suggested for Heliotropium and Cleome (Marshall et al., 2007; Vogan et al., 2007). Unfortunately, phylogenies are not available for Heliotropium and the taxonomic distribution of these preconditions is uncertain for Cleome (Feodorova et al., 2010).

Fig. 4.

Temporal representation of the gradual C3 to C4 transition in Mollugo. This diagram is drawn based on the conclusions of Christin et al. (2011). The ancestor of the group evolved enlarged bundle sheath cells that represent a C4 precondition (light grey) more than 20 Ma. This was co-opted to evolve C3–C4 photosynthesis (grey), which served twice to evolve a C4 trait (black). For each group, triangles or thin bars represent the interval between the stem and crown nodes (interval in which the trait could have evolved).

Fig. 4.

Temporal representation of the gradual C3 to C4 transition in Mollugo. This diagram is drawn based on the conclusions of Christin et al. (2011). The ancestor of the group evolved enlarged bundle sheath cells that represent a C4 precondition (light grey) more than 20 Ma. This was co-opted to evolve C3–C4 photosynthesis (grey), which served twice to evolve a C4 trait (black). For each group, triangles or thin bars represent the interval between the stem and crown nodes (interval in which the trait could have evolved).

Conclusions

By adopting a multi-faceted phylogenetic approach, it has been shown that the numerous C4 origins in eudicots have been spread across the last 30 My, and were contemporaneous with those in monocots. So far, no conclusive evidence of pre-Oligocene C4 plants have been found by molecular dating or analysis of fossilized organic remains (Edwards et al., 2010, Urban et al., 2010). The period extending from the Oligocene to the present has thus seen an exceptional burst of C4 origins in numerous distantly related groups of flowering plants (Sage et al., 2011a), suggesting the action of global environmental triggers. The decline of atmospheric CO2 probably increased the probability of C4 evolution by establishing photorespiration that inhibited C3 photosynthesis on the one hand, while on the other creating an internal pool of photorespired CO2 that could serve as resource for driving evolutionary innovation (Hylton et al., 1988; Sage, 2004). Low CO2 alone was probably insufficient, as shown by the relative lack of C4 species in habitats such as moist, shaded forests. Other environmental changes, notably increased seasonality, fire, and aridification possibly enhanced the probability of C4 evolution by opening up landscapes, reducing competition, and restricting stomatal aperture. Together with anatomical and genetic factors that enhanced C4 evolvability in some clades, these environmental factors might have been instrumental in driving the repeated origins of C4 photosynthesis in a variety of angiosperms, and profoundly changing the plant communities of many biomes across the planet.

Supplementary data

Supplementary data can be found at JXB online

Supplementary Document S1. List of sequences used for phylogenetic reconstructions.

Supplementary Fig. S1. Calibrated phylogenetic trees.

Supplementary Table S1. Comparison of estimates of stem and crown ages (in millions of years, with standard deviation in brackets) with those obtained with less calibration constraints.

This study was funded by the Swiss National Science Foundation grant PBLAP3-129423 and the Marie Curie IOF 252568 fellowship to PAC, and the National Science Foundation grant DEB-1026611 to EJE.

References

Akhani
H
Edwards
G
Roalson
EH
Diversification of the old world Salsoleae s.l. (Chenopodiaceae): molecular phylogenetic analysis of nuclear and chloroplast datasets and a revised classification
International Journal of Plant Sciences
 , 
2007
, vol. 
168
 (pg. 
931
-
956
)
APG
III
An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III
Botanical Journal of the Linnean Society
 , 
2009
, vol. 
161
 (pg. 
105
-
121
)
Batanouny
KH
Stichler
W
Ziegler
H
Photosynthetic pathways and ecological distribution of Euphorbia species
Oecologia
 , 
1991
, vol. 
87
 (pg. 
565
-
569
)
Bell
CD
Soltis
DE
Soltis
PS
The age and diversification of the angiosperms re-revisited
American Journal of Botany
 , 
2010
, vol. 
97
 (pg. 
1296
-
1303
)
Besnard
G
Muasya
AM
Russier
F
Roalson
EH
Salamin
N
Christin
PA
Phylogenomics of C4 photosynthesis in sedges (Cypercaeae): multiple appearances and genetic convergence
Molecular Biology and Evolution
 , 
2009
, vol. 
26
 (pg. 
1909
-
1919
)
Brockington
SF
Alexandre
R
Ramdial
J
Moore
MJ
Crawley
S
Dhingra
A
Hilu
K
Soltis
DE
Soltis
PS
Phylogeny of the Caryophyllales sensu lato: revisiting hypotheses on pollinations biology and perianth differentiation in the core Caryophyllales
International Journal of Plant Sciences
 , 
2009
, vol. 
170
 (pg. 
627
-
643
)
Brown
NJ
Parsley
K
Hibberd
JM
The future of C4 research: maize, Flaveria or Cleome?
Trends in Plant Science
 , 
2005
, vol. 
10
 (pg. 
215
-
221
)
Bruhl
JJ
Wilson
KL
Towards a comprehensive survey of C3 and C4 photosynthetic pathway in Cyperaceae
Aliso
 , 
2007
, vol. 
23
 (pg. 
99
-
148
)
Cerling
TE
Harris
JM
MacFadden
BJ
Leakey
MG
Quade
J
Eisenmann
V
Ehleringer
JR
Global vegetation change through the Miocene/Pliocene boundary
Nature
 , 
1997
, vol. 
389
 (pg. 
153
-
158
)
Christin
PA
Besnard
G
Samaritani
E
Duvall
MR
Hodkinson
TR
Savolainen
V
Salamin
N
Oligocene CO2 decline promoted C4 photosynthesis in grasses
Current Biology
 , 
2008
, vol. 
18
 (pg. 
37
-
43
)
Christin
PA
Freckleton
RP
Osborne
CP
Can phylogenetics identify C4 origins and reversals?
Trends in Ecology and Evolution
 , 
2010
, vol. 
25
 (pg. 
403
-
409
)
Christin
PA
Sage
TL
Edwards
EJ
Ogburn
RM
Khoshravesh
R
Sage
RF
Complex evolutionary transitions and the significance of C3–C4 intermediate forms of photosynthesis in Molluginaceae
Evolution
 , 
2011
, vol. 
65
 (pg. 
643
-
660
)
Cuénoud
P
Savolainen
V
Chatrou
LW
Powell
M
Grayer
RJ
Chase
MW
Molecular phylogenetics of Caryophyllales based on nuclear 18S rDNA and plastid rbcL, atpB, and matK DNA sequences
American Journal of Botany
 , 
2002
, vol. 
89
 (pg. 
132
-
144
)
Douglas
NA
Manos
PS
Molecular phylogeny of Nyctaginaceae: taxonomy, biogeography, and characters associated with radiation of xerophytic genera in North America
American Journal of Botany
 , 
2007
, vol. 
94
 (pg. 
856
-
872
)
Edwards
GE
Franceschi
VR
Voznesenskaya
EV
Single-cell C4 photosynthesis versus the dual-cell (Kranz) paradigm
Annual Review of Plant Biology
 , 
2004
, vol. 
55
 (pg. 
173
-
196
)
Edwards
EJ
Osborne
CP
Strömberg
CAE
Smith
SA
The origins of C4 grasslands: integrating evolutionary and ecosystem science
Science
 , 
2010
, vol. 
328
 (pg. 
587
-
591
)
Edwards
EJ
Smith
SA
Phylogenetic analyses reveal the shady history of C4 grasses
Proceedings of the National Academy of Sciences, USA
 , 
2010
, vol. 
107
 (pg. 
2532
-
2537
)
Ehleringer
J
Björkman
O
Quantum yields for CO2 uptake in C3 and C4 plants
Plant Physiology
 , 
1977
, vol. 
59
 (pg. 
86
-
90
)
Ehleringer
J
Pearcy
RW
Variation in quantum yields for CO2 uptake among C3 and C4 plants
Plant Physiology
 , 
1983
, vol. 
73
 (pg. 
555
-
559
)
Ehleringer
JR
Sage
RF
Flanagan
LB
Pearcy
RW
Climate change and the evolution of C4 photosynthesis
Trends in Ecology and Evolution
 , 
1991
, vol. 
6
 (pg. 
95
-
99
)
Ehleringer
JR
Cerling
TE
Helliker
BR
C4 photosynthesis, atmospheric CO2 and climate
Oecologia
 , 
1997
, vol. 
112
 (pg. 
285
-
299
)
Feodorova
TA
Voznesenskaya
EV
Edwards
GE
Roalson
EH
Biogeographic patterns of diversification and the origins of C4 in Cleome (Cleomaceae)
Systematic Botany
 , 
2010
, vol. 
35
 (pg. 
811
-
826
)
Giussani
LM
Cota-Sánchez
JH
Zuloaga
FO
Kellogg
EA
A molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis
American Journal of Botany
 , 
2001
, vol. 
88
 (pg. 
1993
-
2012
)
Griffiths
H
Luttge
U
Carbon dioxide concentrating mechanisms in relation to the evolution of CAM in vascular epiphytes
Ecological studies 76: Vascular plants as epiphytes
 , 
1989
Berling
Springer-Verlag
(pg. 
42
-
86
)
Hassan
NS
Thiede
J
Liede-Schumann
S
Phylogenetic analysis of Sesuvioideae (Aizoaceae) inferred from nrDNA internal transcribed spacer (ITS) sequences and morphological data
Plant Systematics and Evolution
 , 
2005
, vol. 
255
 (pg. 
121
-
143
)
Hilu
KW
Borsch
T
Muller
K
, et al.  . 
Angiosperm phylogeny based on matK sequence information
American Journal of Botany
 , 
2003
, vol. 
90
 (pg. 
1758
-
1776
)
Hylton
CM
Rawsthorne
S
Smith
AM
Jones
A
Woolhouse
HW
Glycine decarboxylase is confined to the bundle-sheath cells of leaves of C3–C4 intermediate species
Planta
 , 
1988
, vol. 
175
 (pg. 
452
-
459
)
Jacobs
SWL
Review of leaf anatomy and ultrastructure in the Chenopodiaceae (Caryophyllales)
Journal of the Torrey Botanical Society
 , 
2001
, vol. 
128
 (pg. 
236
-
253
)
Kadereit
G
Borsch
T
Weising
K
Freitag
H
Phylogeny of Amaranthaceae and Chenopodiaceae and the evolution of C4 photosynthesis
International Journal of Plant Sciences
 , 
2003
, vol. 
164
 (pg. 
959
-
986
)
Kadereit
G
Freitag
H
Molecular phylogeny of Camphorosmeae (Camphorosmoideae, Chenopodiaceae): implications for biogeography, evolution of C4 photosynthesis and taxonomy
Taxon
 , 
2011
, vol. 
60
 (pg. 
51
-
78
)
Kadereit
G
Mavrodiev
EV
Zacharias
EH
Sukhorukov
AP
Molecular phylogeny of Atripliceae (Chenopodioideae, Chenopodiaceae): implications for systematics, biogeography, flower and fruit evolution, and the origin of C4 photosynthesis
American Journal of Botany
 , 
2010
, vol. 
97
 (pg. 
1664
-
1687
)
Kapralov
MV
Akhani
H
Voznesenskaya
EV
Edwards
G
Franceschi
V
Roalson
EH
Phylogenetic relationships in the Salicornioideae/Suaedoideae/Salsoloideae s.l. (Chenopodiaceae) clade and a clarification of the phylogenetic position of Bienertia and Alexandra using multiple DNA sequence datasets
Systematic Botany
 , 
2006
, vol. 
31
 (pg. 
571
-
585
)
Kellogg
EA
Sage
RF
Monson
RK
Phylogenetic aspects of the evolution of C4 photosynthesis
C4 plant biology
 , 
1999
San Diego, CA
Academic Press
(pg. 
411
-
444
)
Kishino
H
Thorne
JL
Bruno
WJ
Performance of a divergence time estimation method under a probabilistic model of rate evolution
Molecular Biology and Evolution
 , 
2001
, vol. 
18
 (pg. 
352
-
361
)
Kocacinar
F
Sage
RF
Photosynthetic pathway alters xylem structure and hydraulic function in herbaceous plants
Plant, Cell and Environment
 , 
2003
, vol. 
26
 (pg. 
2015
-
2026
)
Linder
HP
Rudall
PJ
Evolutionary history of Poales
Annual Review of Ecology, Evolution and Systematics
 , 
2005
, vol. 
36
 (pg. 
107
-
124
)
Marshall
DM
Muhaidat
R
Brown
NJ
Liu
Z
Stanley
S
Griffiths
H
Sage
RF
Hibberd
JM
Cleome, a genus closely related to Arabidopsis, contains species spanning a developmental progression from C3 to C4 photosynthesis
The Plant Journal
 , 
2007
, vol. 
51
 (pg. 
886
-
896
)
McKown
AD
Dengler
NG
Key innovations in the evolution of Kranz anatomy and C4 vein pattern in Flaveria (Asteraceae)
American Journal of Botany
 , 
2007
, vol. 
94
 (pg. 
382
-
399
)
McKown
AD
Moncalvo
JM
Dengler
NG
Phylogeny of Flaveria (Asteraceae) and inference of C4 photosynthesis evolution
American Journal of Botany
 , 
2005
, vol. 
92
 (pg. 
1911
-
1928
)
Monson
RK
Moore
BD
Ku
MSB
Edwards
GE
Co-function of C3- and C4-photosynthesis pathways in C3, C4 and C3–C4 intermediate Flaveria species
Planta
 , 
1986
, vol. 
168
 (pg. 
493
-
502
)
Moore
MJ
Soltis
PS
Bell
CD
Burleigh
JG
Soltis
DE
Phylogenetic analysis of 83 plastid genes further resolves the early diversification of eudicots
Proceedings of the National Academy of Sciences, USA
 , 
2010
, vol. 
107
 (pg. 
4623
-
4618
)
Ocampo
G
Columbus
JT
Molecular phylogenetics of suborder Cactineae (Caryophyllales), including insights into photosynthetic diversification and historical biogeography
American Journal of Botany
 , 
2010
, vol. 
97
 (pg. 
1827
-
1847
)
Osborne
CP
Freckleton
RP
Ecological selection pressures for C4 photosynthesis in the grasses
Proceedings of the Royal Society B: Biological Sciences
 , 
2009
, vol. 
276
 (pg. 
1753
-
1760
)
Pagani
M
Zachos
JC
Freeman
KH
Tipple
B
Bohaty
S
Marked decline in atmospheric carbon dioxide concentrations during the Paleogene
Science
 , 
2005
, vol. 
309
 (pg. 
600
-
603
)
Pearcy
RW
Troughton
J
C4 photosynthesis in tree form Euphorbia species from Hawaiian rainforest sites
Plant Physiology
 , 
1975
, vol. 
55
 (pg. 
1054
-
1056
)
Pyankov
V
Ziegler
H
Kuzmin
A
Edwards
G
Origin and evolution of C4 photosynthesis in the tribe Salsoleae (Chenopodiaceae) based on anatomical and biochemical types in leaves and cotyledons
Plant Systematics and Evolution
 , 
2001
, vol. 
230
 (pg. 
43
-
74
)
Roalson
EH
C4 photosynthesis: differentiating causation and coincidence
Current Biology
 , 
2008
, vol. 
18
 (pg. 
R167
-
R168
)
Roalson
EH
Raghavendra
AS
Sage
RF
Origins and transitions in photosynthetic pathway types in the monocots: a review and reanalysis
C4 photosynthesis and related CO2 concentrating mechanisms
 , 
2011
Dordrecht, The Netherlands
Kluwer
(pg. 
319
-
338
)
Ronquist
F
Huelsenbeck
JP
MrBayes 3: Bayesian phylogenetic inference under mixed models
Bioinformatics
 , 
2003
, vol. 
19
 (pg. 
1572
-
1574
)
Rutschmann
F
Molecular dating of phylogenetic trees: a brief review of current methods that estimate divergence times
Diversity and Distributions
 , 
2006
, vol. 
12
 (pg. 
35
-
48
)
Sage
RF
Environmental and evolutionary preconditions for the origin and diversification of the C4 photosynthetic syndrome
Plant Biology
 , 
2001
, vol. 
3
 (pg. 
202
-
213
)
Sage
RF
The evolution of C4 photosynthesis
New Phytologist
 , 
2004
, vol. 
161
 (pg. 
341
-
370
)
Sage
RF
Christin
PA
Edwards
EJ
The C4 plant lineages of planet Earth
Journal of Experimental Botany
 , 
2011
, vol. 
62
 (pg. 
3155
-
3169
)
Sage
RF
Kubien
DS
Quo vadis C4? An ecophysiological perspective on global change and the future of C4 plants
Photosynthesis Research
 , 
2003
, vol. 
77
 (pg. 
209
-
225
)
Sage
RF
Sage
TL
Pearcy
RW
Borsch
T
The taxonomic distribution of C4 plant lineages of Amaranthaceae
sensu stricto. American Journal of Botany
 , 
2007
, vol. 
94
 (pg. 
1992
-
2003
)
Sage
TL
Sage
RF
Vogan
PJ
Rahman
B
Johnson
DC
Oakley
JC
Heckel
MA
The occurrence of C2 photosynthesis in Euphorbia subgenus Chamaesyce (Euphorbiaceae)
Journal of Experimental Botany
 , 
2011
, vol. 
62
 (pg. 
3183
-
3195
)
Sankhla
N
Ziegler
H
Vyas
OP
Stichler
W
Trimborn
P
Eco-physiological studies on Indian arid zone plants
Oecologia
 , 
1975
, vol. 
21
 (pg. 
123
-
129
)
Skillman
JB
Quantum yield variation across the three pathways of photosynthesis: not yet out of the dark
Journal of Experimental Botany
 , 
2008
, vol. 
59
 (pg. 
1647
-
1661
)
Smith
SA
Beaulieu
JM
Donoghue
MJ
An uncorrelated relaxed-clock analysis suggests an earlier origin of flowering plants
Proceedings of the National Academy of Sciences, USA
 , 
2010
, vol. 
107
 (pg. 
5897
-
5902
)
Soltis
DE
Soltis
PS
Chase
MW
, et al.  . 
Angiosperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences
Botanical Journal of the Linnean Society
 , 
2000
, vol. 
133
 (pg. 
381
-
461
)
Svensson
P
Bläsing
OE
Westhoff
P
Evolution of C4 phosphoenolpyruvate carboxylase
Archives of Biochemistry and Biophysics
 , 
2003
, vol. 
414
 (pg. 
180
-
188
)
Thompson
JD
Higgins
DJ
Gibson
TJ
ClustalW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position speicific gap penalties and matrix choice
Nucleic Acids Research
 , 
1994
, vol. 
22
 (pg. 
4673
-
4680
)
Thorne
JL
Kishino
H
Divergence time and evolutionary rate estimation with multilocus data
Systematic Biology
 , 
2002
, vol. 
51
 (pg. 
689
-
702
)
Thorne
JL
Kishino
H
Painter
IS
Estimating the rate of evolution of the rate of molecular evolution
Molecular Biology and Evolution
 , 
1998
, vol. 
15
 (pg. 
1647
-
1657
)
Tripati
AK
Roberts
CD
Engle
RA
Coupling of CO2 and ice sheet stability over major climate transitions of the last 20 million years
Science
 , 
2009
, vol. 
326
 (pg. 
1394
-
1397
)
Urban
MA
Nelson
DM
Jimenez-Moreno
G
Chateauneud
JJ
Pearson
A
Hu
FS
Isotopic of C4 grasses in southwestern Europe during the Early Oligocene-Middle Miocene
Geology
 , 
2010
, vol. 
38
 (pg. 
1091
-
1094
)
Vicentini
A
Barber
JC
Aliscioni
SS
Giussani
LM
Kellogg
EA
The age of the grasses and clusters of origins of C4 photosynthesis
Global Change Biology
 , 
2008
, vol. 
14
 (pg. 
2963
-
2977
)
Vogan
PJ
Frohlich
MW
Sage
RF
The functional significance of C3–C4 intermediate traits in Heliotiopium L. (Boraginaceae): gas exchange perspectives
Plant, Cell and Environment
 , 
2007
, vol. 
30
 (pg. 
1337
-
1345
)
Webster
GL
Borwn
WV
Smith
BN
Systematics and photosynthetic carbon fixation pathways in Euphorbia
Taxon
 , 
1975
, vol. 
24
 (pg. 
27
-
33
)
Wen
ZB
Zhang
ML
Zhu
GL
Sanderson
SC
Phylogeny of Salsoleae s.l. (Chenopodiaceae) based on DNA sequence data from ITS, psbB-psbH, and rbcL, with emphasis on taxa of northwestern China
Plant Systematics and Evolution
 , 
2010
, vol. 
288
 (pg. 
25
-
42
)
Yang
ZH
PAML 4: Phylogenetic analysis by maximum likelihood
Molecular Biology and Evolution
 , 
2007
, vol. 
24
 (pg. 
1586
-
1591
)

Comments

0 Comments