| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Anatomy and Morphology |
Department of Botany, University of Toronto, Toronto, Ontario, M5S 3B2, Canada
Received for publication September 16, 2006. Accepted for publication January 10, 2007.
ABSTRACT
Kranz anatomy and C4 vein pattern are required for C4 biochemical functioning in C4 plants; however, the evolutionary timing of anatomical and biochemical adaptations is unknown. From the genus Flaveria, 16 species (C3, C4, intermediates [C3–C4, C4-like]) were analyzed, novel anatomical and vein pattern characters were analyzed and key anatomical differences among photosynthetic groups were highlighted. A stepwise acquisition of anatomical and vein pattern traits prior to derived biochemistry was outlined on the basis of the phylogeny of Flaveria. Increased vein density represents a potential "precondition" contributing to lower ratios of photosynthetic tissues (mesophyll, bundle sheath) and precedes further anatomical and biochemical modifications observed in derived C3–C4 intermediates. In derived Flaveria species, bundle sheath volume is modified through cell expansion, whereas mesophyll volume is altered through mesophyll cell expansion, reductions in the number of ground tissue layers, and increased vein density. Results demonstrated that key anatomical features of C4 plants are also required for C3–C4 biochemical intermediacy, and anatomical and biochemical alterations acquired during evolution of intermediacy may predispose a species for evolution of C4 photosynthesis. C4-like species are similar to C4 species, demonstrating that Kranz anatomy is fully evolved before complete C4 biochemistry is achieved.
Key Words: bundle sheath • C3–C4 photosynthesis • C4 photosynthesis • C4-like photosynthesis • interveinal distance • mesophyll • parallel evolution • vein density
C4 photosynthesis represents an excellent example of convergent evolution in the angiosperms and has multiple evolutionary origins spread throughout 19 families (Sinha and Kellogg, 1996
; Kadereit et al., 2003
; Sage, 2004
; Muhaidat et al., 2007
, this issue). The C4 pathway is a complex suite of features, which evolved in response to the oxygenase function of ribulose 1,5-bisphosphate carboxylase oxygenase (Rubisco) and results in decreased photosynthetic efficiency and carbon loss from photorespiration. C4 plants have a common functionality in concentrating CO2 near the site of Rubisco to circumvent photorespiration, and conditions promoting photorespiration (e.g., aridity, heat, salinity, and low atmospheric CO2) are considered driving factors in the evolution of C4 traits (Sage, 2004
). In addition to C4 biochemistry, some form of anatomical modification (Kranz anatomy) to support CO2 concentration and C4 metabolite cycling is required in all known terrestrial C4 species (Dengler and Nelson, 1999
).
The term Kranz was first used to describe the conspicuous wreath-like tissue around vascular bundles in Cyperus longus (Haberlandt, 1882
), but Kranz anatomy now most often refers to a suite of anatomical characteristics associated with C4 photosynthesis (Dengler and Nelson, 1999
; Dengler and Taylor, 2000
). While remarkable variation in Kranz anatomy exists among C4 species (Carolin et al., 1975
; Hattersley, 1984
; Dengler et al., 1994
; Soros and Dengler, 1998
; Kadereit et al., 2003
; Voznesenskaya et al., 2001
, 2002
; Muhaidat et al., 2007
, this issue), certain features are common, including (1) specialization of anatomically and functionally distinct photosynthetic tissue types, mesophyll (M, site of primary carbon assimilation), and bundle sheath (BS, site of photosynthetic carbon reduction); (2) arrangement of M peripheral to BS and adjacent to intercellular space (ICS); (3) high vein density and low M to BS tissue volume ratio resulting in short diffusion pathways for C4 metabolites; (4) extensive contact between M and BS cells enhancing metabolite cycling, minimizing exposure of BS to ICS, and reducing diffusive loss of CO2. In many C4 species, M tissue is arranged so that each cell is in contact with BS tissue, but M cell structure remains relatively unaltered compared to counterparts in C3 species. By contrast, BS cells in C4 plants are enlarged and contain numerous chloroplasts, mitochondria, and peroxisomes compared to BS cells in C3 plants. Further modifications include a high density of plasmodesmata to facilitate rapid intercellular metabolite diffusion between the carbon fixation and carbon reduction cycles. Another key feature of Kranz anatomy is close vein spacing (Dengler and Nelson, 1999
). Many C4 species alter vein pattern to achieve a low M to BS tissue ratio by increasing vein density, thereby lowering the relative volume of interveinal M tissue and increasing the relative volume of vein-associated BS tissue per unit leaf surface (Hattersley, 1984
; Dengler et al., 1994
; Dengler and Nelson, 1999
; Dengler and Taylor, 2000
). Numerous studies of vein patterning in C4 species exist (summarized in Dengler and Nelson, 1999
; Ogle, 2003
), and recent work suggests the shift in vein density in C4 grasses is related to minor and cross veins, rather than major veins (Ueno et al., 2006
). This combination of features is present in nearly all known terrestrial C4 plants, with the exception of two unusual genera (Bienertia and Borszczowia, Chenopodiaceae) that operate single-cell C4 photosynthesis (Voznesenskaya et al., 2001
, 2002
). These C4 anomalies do maintain certain aspects of Kranz anatomy, however, including an intracellular separation of an outer carbon assimilation region and interior carbon reduction region, qualitatively low volume ratio of these subcellular regions, and limited exposure of the carbon reduction region to ICS.
The anatomical differences between C3 and C4 plants are complex, and the evolution of C4 anatomy requires developmental modifications at three levels including (1) overall tissue pattern (vein spacing and photosynthetic tissue proportions), (2) cell pattern (cellular arrangement within BS and M tissue), and (3) specialized BS and M cell ultrastructure (Dengler and Taylor, 2000
). Specific information on the role of vein pattern alteration between C3 and C4 species remains unknown, despite being considered a vital and basic step in the evolution of C4 photosynthesis (Kellogg, 1999
). Aspects of M and BS patterning and cellular specialization have been well described, yet it remains unclear whether the evolutionary modifications of these two photosynthetic tissues occurred separately, or in combination, at any developmental level. Furthermore, the role of anatomical "preconditioning" has been proposed, but remains unstudied (Sage, 2001
, 2004
). To link the modifications to anatomy and vein patterning involved with C4 photosynthesis requires a comparative study of anatomical traits between C3 progenitor and C4-derived species within an evolutionary framework (see Sinha and Kellogg, 1996
; Soros and Dengler, 2001
).
The genus Flaveria (Asteraceae) represents an excellent model for research into the evolution of Kranz anatomy within a monophyletic eudicot taxon (McKown et al., 2005
). Flaveria is a photosynthetically diverse genus with species that are C3, C4 (NADP–ME type), or intermediate (C3–C4 and C4-like). C3 photosynthesis is the ancestral condition in Flaveria, and intermediate and C4 photosynthesis are both derived conditions (Fig. 1). C4 photosynthesis is present in clade A only, whereas intermediate photosynthesis (C3–C4 and C4-like) is found in both clades A and B, with at least two origins of each type in the phylogenetic tree. A few intermediate species are evolutionary lineages between C3 and C4 photosynthesis, whereas other intermediate species likely represent recent developments of intermediate photosynthesis.
|
; Ku et al., 1991
; Drincovich et al., 1998
; Hatch, 1999
; Monson, 1999
; Westhoff and Gowik, 2004
). Similar to C4 photosynthesis, C3–C4 intermediate photosynthesis reduces carbon loss from photorespiration through a cycling of metabolites between M and BS resulting in the release of photorespired CO2 by glycine decarboxylase to be refixed in BS tissues (Monson and Rawsthorne, 2000
). C3–C4 intermediate photosynthesis can be further divided into two subcategories based on additional C4 cycling between M and BS: type I with little to no C4 cycle and type II with a partial C4 cycle (Edwards and Ku, 1987
). By contrast, C4-like intermediate photosynthesis is characterized by plants that assimilate the majority of carbon through the C4 cycle, but lack complete compartmentation of photosynthetic enzymes between M and BS (Moore et al., 1988
; Ku et al., 1991
). Biochemical and physiological features of intermediates have been well studied, but Kranz anatomy has only been described qualitatively and in general terms (Holaday et al., 1984
). One study presented quantitative data for M and BS cell ultrastructure in C3, C4, and intermediate Flaveria species (Brown and Hattersley, 1989
); however, Kranz anatomy has not been formally assessed in the majority of Flaveria species. The evolutionary framework and diversity of photosynthetic types in Flaveria make this genus ideal for research into understanding the evolution of the complex adaptation of C4 photosynthesis and the relationships between Kranz anatomy, vein patterning, and photosynthesis. Using the phylogenetic framework of Flaveria, we describe detailed qualitative and quantitative anatomical and vein pattern traits between ancestral and derived states in 16 Flaveria species. We explore a potential stepwise acquisition of key anatomical traits in relation to photosynthetic type, investigate the hypothesis of anatomical "preconditioning" where specific anatomical traits may precede the evolution of C4 photosynthesis and offer insight into specific developmental events necessary to achieve Kranz anatomy in intermediate and C4Flaveria species.
MATERIALS AND METHODS
Plant material
Sixteen Flaveria species were studied from the basal group and derived clades and include representatives of each photosynthetic type (species with asterisks in Fig. 1). Plants were grown either from seed (donated by H. Bauwe, P. Westhoff, and S. von Caemmerer to N. G. Dengler or R. F. Sage) or vegetative cuttings (donated by A. M. Powell to R. F. Sage), and cultivated under natural light in University of Toronto greenhouses during winter and an open-air rooftop during summer (F. Kocacinar, Kahramanmaras Sutcu Imam University, Turkey, personal communication). Additional plant material from Oaxaca, Mexico, was used for F. kochiana and F. ramosissima (donated by E. Sudderth, Harvard University, to A. D. McKown). All specimens were identified using morphology and gene markers (McKown et al., 2005
), and voucher specimens of mature shoots, leaves and flowers were collected and are housed at the Vascular Plant Herbarium, Royal Ontario Museum (TRT) (Appendix 1).
Leaf anatomy
To examine internal leaf anatomy of Flaveria species, tissue was excised from the central portion of a mature leaf blade from three to six replicate plants for each species. Tissues were fixed in 70% FAA (formaldehyde-acetic acid-alcohol), rinsed in 70% ethanol, dehydrated in a graded ethanol series, and embedded in Spurr's resin following a graded infiltration series (modification from Berlyn and Miksche, 1976
). Cross sections (3 µm) were cut using a Porter–Blum MT–2 ultramicrotome, mounted on poly-l-lysine coated slides, and stained with 0.05% toluidine blue O in 50 mM citrate buffer (Ruzin, 1999
). Sections were observed under bright field microscopy using a Reichert Polyvar microscope (Reichert-Jung, Vienna, Austria) and photographed with a Nikon DXM 1200 digital camera and ACT-1 Nikon imaging software (Nikon Instruments, Melville, New York, USA).
Leaf cross-section images were compared qualitatively (e.g., for location of chloroplasts) and analyzed quantitatively with Image Pro Plus software (Media Cybernetics, Silver Spring, Maryland, USA) for M thickness, vascular tissue (VT) area, M area, ICS area, BS area, BS surface area exposed to ICS, and BS cell number per vein. The M thickness was obtained from an average of 10 measurements per replicate. Other quantitative variables were determined using all minor veins within a replicate sample and averaging values for each replicate. Minor veins were identified in cross section by VT area, cell number, and lack of resin canals (see "RC," Fig. 2). BS and M tissue areas (equivalent to tissue volumes, see Parkhurst, 1982
) were measured within a rectangular area centered on each minor vein, determined on one axis by M thickness and on the other by the interveinal distance (IVD) for each species (obtained from analysis of leaf clearings, described later). In two species (F. brownii, F. kochiana), measurements were taken recording M area either with or without hypodermal tissue (see "H," Fig. 2). ICS area was measured using a stereological grid of 300 random points superimposed on cross-section images and calculating the proportion of point counts in ICS compared to total point counts within the leaf tissue (Parkhurst, 1982
). The proportion of BS surface area in contact with ICS was measured by calculating the cumulative length of BS outer cell walls exposed to ICS compared to the total perimeter of BS outer cell walls, and BS cell numbers were determined counting the number of entire cells encircling each vein.
|
Leaf venation
Minor vein pattern was compared for all 16 Flaveria species with the same mature leaves from replicate plants used for the anatomical study. Tissue (25 mm2) was excised from the central portion of each leaf replicate, rinsed in distilled water, and bleached in 10% bleach. Leaf segments were then further cleared in 5% NaOH, rinsed briefly with distilled water, and transferred to saturated chloral hydrate. Cleared tissues were mounted in an 8 : 2 : 1 chloral hydrate : glycerol : water mounting solution for differential interference contrast microscopy with a Reichert Polyvar microscope. For each replicate, five non-overlapping images were taken of veins with a Nikon DXM 1200 digital camera and ACT-1 Nikon imaging software.
Quantitative data collected from these subsamples were averaged for each replicate sample. Leaf minor vein density was calculated with Image Pro Plus software by measuring the total length of all veins in the entire image area. Numbers of branching points and freely ending veinlets per sample area were also recorded. Using a stereological grid of 300 or 600 superimposed random points (selected according to vein density), IVD was sampled where a point contacted vascular tissue (modified from Parkhurst, 1982
). In determining IVD, all species were measured for the shortest distance between veins (analogous to short-distance C4 metabolite movement between M and vein-associated BS tissue), and multiple measurements were used to calculate an average IVD for each sample. This approach to assessing IVD is different than the traditional approach of measuring cross sections that can be used for study of parallel venation in grasses, but does not accurately reflect the reticulate nature of most eudicot plants.
Numerical data analysis
Quantitative data from all replicate samples were averaged to obtain representative values for each species (N = 16) and analyzed for possible correlations among either anatomical variables or vein pattern variables with Pearson product moment correlation tests using SigmaStat software (Systat Software, Richmond, California, USA). The potential positive and negative correlations between variables determined by the first method were then re-tested with independent contrasts to incorporate phylogenetic history. Species values (N = 16) were analyzed for correlations with Pearson product moment correlation tests corrected for species relationships and branch lengths of the phylogeny (McKown et al., 2005
) using Mesquite v. 1.11 software (Maddison and Maddison, 2006
) with the independent contrast PDAP: PDTREE module v. 1.07(1) (Midford et al., 2005
). Correlated variables were reconstructed on the phylogeny, and ancestral states were simulated assuming parsimony using Mesquite v. 1.11.
Quantitative anatomy and vein pattern data from all replicate values (N = 68) and averaged species values (N = 16) were standardized and tested in principal components analyses using NTSYS–PC v. 2.2 software (Exeter Software, Setauket, New York, USA). Scatter plots (plotted with two principal components) from these analyses were used to visualize the similarity of species (evolutionary history vs. photosynthetic type) when all variables from anatomical, vein pattern, or both anatomy and vein pattern were considered simultaneously.
RESULTS
Qualitative leaf anatomy
Flaveria species can be separated by qualitative anatomical features into three photosynthetic categories, regardless of phylogeny: C3, C3–C4 and C4-like/C4. Species classified as C3 all have BS cells that are indistinct (either by shape or size) from surrounding M cells and contain very few chloroplasts (Figs. 2, 3, Table 1). The C3 species F. cronquistii is sister to all other Flaveria species, but is unlike the other C3 species (F. pringlei, F. robusta) because it possesses leaves with unifacial M having both adaxial and abaxial palisade layers and internal nonchlorenchymatous M layers. Leaves of F. pringlei and F. robusta have typical C3 bifacial M (adaxial palisade only) and likely represent the closest C3 leaf "body plan" to intermediate and C4Flaveria species.
|
|
C4-like and C4 species (clade A, clade B) have typical atriplicoid-type Kranz anatomy in which BS cells are isodiametric and uniform in shape, and almost all M cells are radially arranged to retain contact with BS tissue (Figs. 2, 3, Table 1). BS tissues in the C4-like and C4 species have numerous chloroplasts located near VT and qualitatively appear to have thicker cell walls than C3 and C3–C4Flaveria species. The M tissue of clade A C4-like species F. vaginata is qualitatively similar to its sister-species C4F. kochiana (not including hypodermal tissue) and other clade A C4 species (F. australasica, F. bidentis, F. trinervia). M cells in these species contain relatively few chloroplasts, and cell sizes are much smaller than M counterparts in C3 and C3–C4Flaveria species. By contrast, C4-like F. brownii in clade B has M cells (not including hypodermal tissue) with numerous chloroplasts that are not as reduced in size as M cells in clade A species. Among Flaveria species, only C4F. kochiana (clade A) and C4-like F. brownii (clade B) have large, elongate nonchlorenchymatous hypodermal cells (see "H," Fig. 2). These hypodermal cells are generally vacuolated, contain very few chloroplasts and are not in contact with BS tissues. In F. brownii, hypodermal cells are adjacent to both the adaxial and abaxial epidermis, whereas in F. kochiana these cells are solely abaxial.
Quantitative leaf anatomy
Quantitative measurements of Flaveria leaf anatomy yield several important patterns and emphasize subtle anatomical characteristics of BS and M tissues (Fig. 4, Appendix S1). Patterns of these features vary among species within each clade and also within photosynthetic types, especially among intermediate species.
|
The observed increases in BS tissue area are not achieved by modifying BS cell numbers, because variability in BS cell numbers is low among Flaveria species (Appendix S1) and correlated with minor vein size (P < 0.01, data not shown). By contrast, there is greater variability in BS cell size among Flaveria species (Fig. 4b). Patterns of cell sizes observed in Flaveria species are similar to the patterns of BS tissue area (Fig. 4a, b). The positive association between greater BS area and larger individual BS cells is observed in C3–C4 intermediate species, including F. sonorensis, clade A F. ramosissima and clade B F. anomala, F. chloraefolia, F. floridana, and F. linearis. This association is also reflected in the elongate BS cell shape of C3–C4F. sonorensis and clade B C3–C4 species (Figs. 2, 3, Table 1).
In contrast to the single layer of BS tissue, the multilayered M tissue has diverse patterns within Flaveria (Fig. 4c, Appendix S1). Among the basal grade species, M tissue area per vein is highest in C3F. cronquistii and F. pringlei but much lower in C3F. robusta and C3–C4F. sonorensis. In clade A, M area is intermediate in C3–C4F. ramosissima between the basal grade Flaveria species and the derived C4 clade A species. The clade A C4-like and C4 species have similar M area values and are among the lowest values for the genus. In contrast to clade A, there is no pattern of reduction in M area among clade B species. Clade B C3–C4F. angustifolia has similar M area to basal grade species C3F. robusta and C3–C4F. sonorensis. Among the more derived clade B species, two C3–C4 species (F. anomala, F. chloraefolia) have less M area per vein than F. angustifolia; however, two other C3–C4 species (F. floridana, F. linearis) have the greatest area of M per vein among all clade B species. Of all clade B species, C4-like F. brownii is the only species that has a comparable M area per vein to clade A C4-like and C4 species.
Unlike BS tissue, the observed modifications to M area associated with each vein are associated with alterations both to M cell number and expansion (Figs. 4d, 5). Cell numbers in M tissue are modified between the adaxial and abaxial surfaces of the leaf (dorsiventral axis) by changing the number of ground tissue layers and from margin to margin (medial axis) through altered vein spacing (discussed later). Among the basal grade species, all C3Flaveria species have eight layers of ground tissue in the mid-portion of developing leaves, whereas C3–C4F. sonorensis has six layers (Fig. 5, Appendix S1). Within clade A, all species, regardless of photosynthetic type, have five layers. By contrast, clade B has more variability as C3–C4F. angustifolia maintains eight layers, C3–C4F. anomala has five layers, and all other derived clade B intermediates (F. brownii, F. chloraefolia F. floridana, F. linearis) have six layers of ground tissue. These reductions in ground tissue layers among derived Flaveria species are all expressed as the loss of a single layer of palisade M and up to two layers of spongy M. A reduction in ground tissue layers results in fewer cells contributing to M tissue area overall; however, cell expansion also contributes a substantial role in altering M area. This is evident among the eight-layered species, because the C3 species F. cronquistii and F. pringlei have thicker M tissues than C3F. robusta and C3–C4F. angustifolia (Fig. 4d, Appendix S1). This is also supported by the six-layered species, such as C3–C4F. sonorensis and clade B species (F. brownii, F. floridana, F. linearis) that have thicker M tissues than eight-layered C3F. robusta and C3–C4F. angustifolia, despite the loss of two cell layers. The thinnest M tissues are observed in all five-layered clade A species and clade B C3–C4F. anomala; however, the variability observed in M thicknesses among these five-layered species suggests a role of M cell expansion as a key factor in determination of overall M tissue area.
|
In addition to modifications of BS and M tissues, there is a general pattern of increased leaf ICS among derived Flaveria species because the C3 species F. cronquistii and F. pringlei have a lower proportion of ICS compared to all other Flaveria species (Appendix S1). The increase in ICS indicates higher exposure of M cell surfaces to ICS; however, there is not a corresponding increase in exposure of BS tissue to ICS (Fig. 4f, Appendix S1). BS surface exposure to ICS is low among C3 species, clade A C4-like and C4 species, and certain clade B species (C3–C4F. angustifolia, C3–C4F. floridana, C4-like F. brownii). By contrast, many C3–C4 species, including F. sonorensis, clade A F. ramossisima, and clade B F. anomala, F. chloraefolia, and F. linearis, have a high percentage of BS tissue surface area exposed to ICS (up to 50% in some replicates). The increased surface exposure of BS cells to ICS in these C3–C4Flaveria species does not correlate with greater leaf ICS because other derived Flaveria species also have greater proportions of ICS but lack exposure of BS to ICS. Of note, no Flaveria species, including C4 species, has complete isolation of the BS tissue from ICS.
Testing anatomical variables for correlations without considering phylogenetic history indicates a number of significant correlations among anatomical variables (non-PC correlations, Table 2). With correction for phylogenetic relationships among Flaveria species, only M area to ICS, BS area to BS cell size, and BS cell size to BS surface exposure remain significantly correlated (PC correlations, Table 2). In Flaveria, these correlations exist regardless of evolutionary history, and it is evident that many of the positive correlations between M tissue and other anatomical variables are related through phylogenetic history (Table 2).
|
|
|
|
The quantitative venation characteristics are all highly correlated statistically at P 
0.001, even with correction for phylogenetic history (Table 3). Analysis of all vein pattern characteristics in a principal components multivariate analysis demonstrates that the data are one-dimensional (Fig. 6b), with only one significant principal component (87%) and each variable contributing equally to this component (data not shown). There is some cohesion of photosynthetic types among Flaveria species; however, there is little cohesion within clades. On either end of the PC1-axis in Fig. 6b, C3F. cronquistii and F. pringlei (black squares), and C4 species (white circles) form large clusters, with C3F. robusta (asterisk), C3–C4 species (light grey), and C4-like species (dark grey) and C4F. kochiana ("k") positioned between these groupings.
|
|
|
Coevolution of anatomy, venation, and photosynthetic function
In combination with a robust phylogeny of Flaveria, the results from this comprehensive anatomical study suggest a stepwise acquisition of anatomical traits preceding biochemical photosynthetic evolution in Flaveria. In addition, it is apparent from parallel evolutions of intermediate photosynthesis in this genus that similar patterns of anatomical features are required. The results also suggest that intermediate photosynthesis builds upon itself in stages and that C4 photosynthesis is likely a result of these, or similar, stepwise events (Fig. 10).
|
; Dai et al., 1996
).
C3F. cronquistii and F. pringlei species occur in high light and arid environments in southern Mexico alongside intermediate and C4Flaveria species (Powell, 1978
; Sudderth et al., 2007
). Leaves of these C3 species are much more succulent, suggesting that the ancestral Flaveria species may have survived a similar environment through water storage in M tissues. C3F. robusta occurs in similar habitats (Powell, 1978
) but is qualitatively more comparable to the more derived Flaveria species and has little indication of succulence. In C3F. robusta, the observed modifications to vein pattern may represent the earliest anatomical "precondition" in Flaveria (Fig. 10). This suggests that the common ancestor of F. robusta and all intermediate/C4 species possessed higher vein density and consequently a lower M to BS tissue ratio than C3F. cronquistii and F. pringlei. Shifting vein density acts to decrease M volume and increase BS tissue volume within the leaf as a whole and is hypothesized to improve hydraulic integrity by reducing the evaporative surface of photosynthetic cells (= M tissue) relative to the conduit capacity to resupply transpired water loss (Sage, 2001
, 2004
). This adaptation may be related to environmental factors such as increased illumination, low humidity, or limited soil water (Roth-Nebelsick et al., 2001
). Higher vein density can be associated with xerophytic conditions, although this is usually accompanied by higher freely ending veinlet density (Roth-Nebelsick et al., 2001
), which was not observed in F. robusta. An alteration of vein patterning has been suggested as an early adaptation in the evolution of C4 photosynthesis (Kellogg, 1999
; Sage, 2004
); however, this is the first study to demonstrate modifications to vein pattern and its effect on decreasing photosynthetic M tissue volume occurring at an early node in the phylogeny. This also provides evidence for a potential "preconditioning" role of venation in the evolution of derived photosynthetic pathways from C3 photosynthesis.
Type I C3–C4 intermediacy
The most closely related Flaveria species to the C3 species are the C3–C4 intermediates; however, there is a substantial evolutionary distance among some of these species limiting their interpretation as a cohesive group (Fig. 1). Flaveria sonorensis is the nearest relation to C3F. robusta, but is classified as a type I C3–C4 intermediate due to its lowered CO2 compensation point (
), reduced O2 sensitivity (Ku et al., 1991
; Dai et al., 1996
), and absence of any C4 cycle (Moore et al., 1987
). Although morphologically similar, C3–C4F. sonorensis is anatomically different from C3F. robusta. Both species have comparable M areas, despite fewer ground tissue layers and a higher vein density in F. sonorensis. BS area in F. sonorensis is more than twice the BS area of F. robusta (through increased BS cell size, not numbers) resulting in an M to BS tissue ratio in F. sonorensis less than half the ratio of F. robusta. Further modifications to BS tissue in F. sonorensis compared to F. robusta include an increase in BS chloroplast numbers and exposure of BS surface area to ICS. These results suggest developmental coordination of modifying venation, M and BS tissue volumes, and BS cell structure to support C3–C4 photosynthesis within F. sonorensis.
C3–C4F. angustifolia (clade B) represents the only intermediate Flaveria species at the basal node of a clade that is physiologically similar to C3Flaveria species. The and O2 inhibition values of C3–C4F. angustifolia are close to C3–C4F. sonorensis (Ku et al., 1991
; Dai et al., 1996
), and little to no C4 cycling exists in F. angustifolia, classifying it as a type I intermediate (Moore et al., 1987
). Anatomically, F. angustifolia is qualitatively comparable to C3F. robusta and both species have the same number of ground tissue layers. Leaves of F. angustifolia are slightly thicker than F. robusta; however, M area is equivalent between both species, due to a higher vein density observed in F. angustifolia. By contrast, BS area is slightly greater in F. angustifolia than F. robusta, and results in a lower M to BS tissue ratio in F. angustifolia. The anatomical and vein pattern similarities between F. angustifolia and F. robusta (Fig. 6c) and the physiological data suggest that F. angustifolia may represent a C3–C4 intermediate closer to the common ancestor of clades A and B. The contribution of increased vein density to lower M to BS tissue ratio in F. angustifolia supports the hypothesis that alteration to vein patterning is among the earliest anatomical expressions of the changes from C3 to C3–C4 photosynthesis (Fig. 10). BS tissue in F. angustifolia has a qualitative increase in BS chloroplast numbers compared to C3 species, but has no further modifications observed in other C3–C4 intermediates, such as increased BS cell expansion or BS surface exposure to ICS.
The more derived type I C3–C4 intermediate Flaveria species in this study are clade B F. chloraefolia and F. linearis (Monson et al., 1986
; Moore et al., 1987
; Ku et al., 1991
; Dai et al., 1996
), and they represent a split in the evolution of clade B intermediate lineages (see clade B F. anomala discussed later). Both species are qualitatively more similar to F. sonorensis than F. angustifolia. In addition, vein densities and number of ground tissue layers (six) observed in F. chloraefolia and F. linearis are more comparable to F. sonorensis than to F. angustifolia. M area is equivalent between F. angustifolia and F. chloraefolia, whereas F. linearis has a much greater M area. By contrast, BS area and cell size are greater in F. chloraefolia and F. linearis than F. angustifolia, and subsequently, M to BS tissue ratios of both species are lower than in F. angustifolia and comparable to F. sonorensis. Furthermore, BS cells of F. chloraefolia and F. linearis are elongate in shape with a large number of chloroplasts, and both species have high BS surface exposure to ICS, as observed in F. sonorensis.
The correspondence of anatomical and physiological properties among type I C3–C4 species (little to no C4 cycle) indicates parallel evolution of intermediate photosynthesis within Flaveria. Among these species, modifications to vein pattern, and M and BS tissue volumes are not identical; however, the resulting M to BS tissue ratio and allocation of resources to augment the photosynthetic capacity of BS cells are common among all type I intermediates (Fig. 10). In the evolution of C3–C4 photosynthesis, BS tissues are hypothesized to increase in volume prior to gaining photosynthetic organelles (Sage, 2001
, 2004
). Within Flaveria, the reverse appears to have occurred, and the occurrence of photosynthetic capacity in C3 BS tissue may be linked to the overall increase in BS tissue volume within the leaf as vein density increases, but not to enlargement of individual BS cells. A greater volume of BS decreases available space for M on the lateral axis, and a decrease in photosynthetic M tissue volume could drive the nonphotosynthetic BS tissue to gain photosynthetic function. This supports models that outline modifications to BS tissue as an early step in the evolution of C3–C4 photosynthesis (Edwards and Ku, 1987
; Monson and Moore, 1989
; Monson and Rawsthorne, 2000
); however, the driving mechanisms in establishing BS as a photosynthetic tissue and modifying BS ultrastructure remain speculative.
The early increase in organelle numbers in BS cells, as observed in F. angustifolia, could also act to further augment allocation of resources to BS tissue. A greater BS tissue volume overall in C3–C4 species coincides with an upregulation and activity of enzymes for refixation of photorespired CO2, such as glycine decarboxylase (Monson and Rawsthorne, 2000
); however, activities of key C4 enzymes, such as NADP–malic enzyme or phosphoenolpyruvate carboxylase, remain low (Ku et al., 1991
; Drincovich et al., 1998
; Monson and Rawsthorne, 2000
). The increased contribution of BS tissue to net photosynthesis is coupled with a developmental shift that decreases the number of ground tissue layers, resulting in fewer M cells. As F. angustifolia maintains the ancestral condition of eight ground tissue layers, it is likely that the developmental shift to fewer ground tissue layers follows higher vein density and established photosynthetic function in BS cells. The result of these changes to both photosynthetic tissues is a lower M to BS tissue ratio, compared to C3 species, and may be significant for type I intermediate photosynthesis in facilitating the functioning of the photorespiratory CO2 pump (Monson and Rawsthorne, 2000
). In addition, type I C3–C4Flaveria intermediates demonstrate that anatomical and vein pattern alterations that modify photosynthetic tissues to facilitate photorespiratory cycling occur prior to the establishment of a C4 cycle.
Type II C3–C4 intermediacy
Type II intermediacy has evolved in parallel at least three times in Flaveria and is present in both clades A and B. C3–C4F. ramosissima at the basal node of clade A and C3–C4F. anomala and F. floridana in clade B are type II C3–C4 intermediates with significantly lower and less O2 inhibition than C3 species (Ku et al., 1983
; Rumpho et al., 1984
; Ku et al., 1991
; Dai et al., 1996
) and have the ability to fix 50% of leaf CO2 through a C4 cycle (Moore et al., 1987
; Monson et al., 1988
). Compared to clade A F. ramosissima, clade B F. anomala and F. floridana can have slightly higher , greater O2 inhibition and less C4 cycle (Moore et al., 1987
; Monson et al., 1988
; Ku et al., 1991
; Dai et al., 1996
). Flaveria ramosissima is qualitatively comparable to F. anomala, whereas F. floridana more closely resembles clade B C3–C4F. chloraefolia and sister-species F. linearis; however, quantitative anatomical and vein pattern features indicate parallel evolution of anatomy and physiology and suggest common requirements for type II C3–C4 photosynthesis.
Both C3–C4F. anomala and F. ramosissima have similar M and BS areas, M to BS tissue ratio, vein density, vein spacing, and number of ground tissue layers. In both species, vein spacing is similar to C3F. robusta and C3–C4F. angustifolia; however, less M area per vein is observed suggesting that the decrease in number of ground tissue layers (five vs. eight) is primarily responsible for the decrease in M tissue in F. anomala and F. ramosissima. BS area and BS cell size in F. anomala and F. ramosissima are higher than F. angustifolia, but are equivalent or lower than BS area and BS cell size observed in other type I intermediate species. In addition, the ratio of M to BS tissue in F. ramosissima and F. anomala is comparable to type I intermediates (excepting F. angustifolia). BS surface exposure to ICS is high in F. anomala, similar to other derived clade B intermediates and is the only significant anatomical difference observed between F. anomala and F. ramosissima. This indicates less contact between M and BS cells, enhancing CO2 and HCO3– leakage from BS cells, reducing photosynthetic metabolite cycling efficiency, and may result in the higher and O2 inhibition observed in F. anomala compared to F. ramosissima (Ku et al., 1991
). Despite this difference, there is a general correspondence of anatomy and physiology between F. ramosissima in clade A and F. anomala in clade B. Arrangement of M and BS tissues promoting higher contact between cells likely increases the efficiency of photorespiratory cycling in type II intermediates, but also provides the necessary anatomical arrangement to support increased C4 cycling between M and BS cells.
Flaveria floridana represents a type II C3–C4 intermediate that has evolved in parallel with clade A F. ramosissima and clade B F. anomala, but with somewhat different anatomical conditions. Similar to its sister-species, type I C3–C4 intermediate F. linearis, F. floridana has a greater M area than in F. anomala and F. ramosissima and higher number of ground tissue layers (six vs. five). BS area is also greater, vein density higher, and vein spacing lower in F. floridana than in F. anomala and F. ramosissima, but the resulting M to BS tissue ratio in F. floridana is equivalent. BS cells in F. floridana are elongate like type I intermediates F. chloraefolia, F. linearis, and F. sonorensis; however, F. floridana has much less BS surface area exposed to ICS, emphasizing the necessity of high contact between M and BS tissues for efficient photorespiratory cycling between these tissues and integration of C4 cycling in type II intermediate photosynthesis.
Type I and type II C3–C4 intermediate Flaveria species appear somewhat qualitatively dissimilar; however, the anatomical variables in this study are comparable quantitatively (excepting F. angustifolia, see Fig. 6c). The overall similarities in anatomy and vein patterning between type I and type II C3–C4 intermediate species suggest that the anatomical conditions exist in type I intermediate Flaveria species to support biochemical evolution toward type II intermediate photosynthesis (integration of the C4 cycle) (Fig. 10). Anatomically, the switch from type I to type II intermediate photosynthesis likely involves only slight tissue arrangement modifications to enhance contact between M and BS tissues and lower exposure of BS cell surface to ICS. The efficiency of this tissue arrangement in photorespiratory metabolite cycling may then act to promote or enhance the presence and activity of C4 enzymes, as observed in type II C3–C4 intermediates F. anomala, F. floridana, and F. ramosissima (Ku et al., 1983
, 1991
; Bauwe, 1984
; Drincovich et al., 1998
).
C4-like and C4 photosynthesis
C4-like photosynthesis has evolved in parallel at least twice in Flaveria in both clades A and B. The C4-like Flaveria species are the most advanced of the biochemical intermediate species and have the ability to fix between 65–94% of leaf CO2 into C4 acids (Monson et al., 1986
; Moore et al., 1987
, 1989
); however, C4-like Flaveria species may lack strict compartmentation of C4 enzymes, have lower levels of C4 enzyme activity, or have some O2 inhibition of photosynthesis (Reed and Chollet, 1985
; Monson et al., 1987
; Moore et al., 1987
, 1989
; Cheng et al., 1988
; Ku et al., 1991
; Dai et al., 1996
). C4-like Flaveria species range in their integration of the C4 cycle, and clade A C4-like species are very physiologically similar to true C4 species and could be interpreted as "nonoptimized" C4 species (Sudderth et al., 2007
).
The C4-like species (F. brownii, F. vaginata) and C4 species (F. australasica, F. bidentis, F. kochiana, F. trinervia) in this study are generally qualitatively similar, having uniformly shaped BS cells with chloroplasts located near VT and M cells arranged in a constellation pattern around the BS. Quantitatively, C4-like Flaveria species are comparable to C4 species in anatomical and vein pattern characteristics (Fig. 6c) and have a lower proportion of M area, higher vein density, and lower vein spacing compared to related C3–C4 species from clades A or B. With exclusion of hypodermal cells in clade A F. kochiana and clade B F. brownii, all C4-like and C4Flaveria species have equivalent "functional" M tissue areas, despite variability in the number of ground tissue layers (five in clade A vs. six in clade B). Hypodermal cells in F. kochiana and F. brownii have very few chloroplasts, likely contribute less to photosynthesis than the smaller M cells near BS tissues, and may have a different physiological function, such as water storage (Sudderth et al., 2007
). BS tissue areas are comparable among C3, C4-like and C4Flaveria species; however, M to BS tissue ratios are substantially lower in C4-like and C4 species compared to C3 species, reflecting decreases in "functional" M area. There is little BS surface exposure to ICS in C4-like and C4Flaveria species, and BS cells qualitatively appear to have a thicker cell wall than C3 and C3–C4 intermediate Flaveria species.
Within Flaveria, it is likely that the anatomical and biochemical properties of type II intermediate photosynthesis precede higher integration of C4 cycling and the evolution of C4-like photosynthesis (Fig. 10). The anatomical differences between type II C3–C4 intermediates and C4-like Flaveria species in clades A and B represent the final anatomical modifications necessary to support C4 photosynthesis and include decreasing M tissue volume through higher vein density, limiting BS cell expansion to a uniform, isodiametric shape, increasing BS cell wall thickness, and maintaining BS chloroplasts near the VT. These modifications result in a lower functional M to BS tissue ratio in C4-like species than related C3–C4 type II intermediates, and further facilitates C4 metabolite cycling. In conjunction with anatomical alterations, biochemical adaptations toward fully expressed C4 photosynthesis involve increasing C4 enzyme activity levels, and complete compartmentation of photosynthetic enzymes to fully separate C4 cycle functions between M and BS (Edwards and Ku, 1987
; Ku et al., 1991
; Monson and Rawsthorne, 2000
). It is evident that complete Kranz anatomy is not required for C3–C4 photosynthesis to function or to gain C4 metabolite cycling; however, the presence of C4 anatomical and vein pattern features in all C4-like species indicates that complete Kranz anatomy and vein patterning precede fully expressed and optimized C4 biochemistry.
Concluding remarks
The derived C3–C4, C4-like and C4Flaveria species differ from the C3 condition at the three hierarchical levels proposed by Dengler and Taylor (2000)
: (1) overall tissue pattern in which C3–C4, C4-like and C4 species have lower M to BS tissue ratios than C3 species by modifying the number of ground tissue layers, M and BS cell expansion, and vein density; (2) cell pattern in which C3–C4 species have BS tissue with greater surface exposure to ICS, and M and BS are arranged to promote cell contact in C4-like and C4 species; (3) specialized cell structure by which BS cells of C3–C4 species are expanded in size with increased numbers of chloroplasts, and BS cells of C4-like and C4 species are isodiametric with chloroplasts located near vascular tissue. Differences that distinguish intermediate and C4Flaveria species from the C3 species outline a potential stepwise acquisition of anatomical and vein pattern traits that precede derived biochemical traits (Fig. 10). The alteration observed in vein patterning may be one of the earliest anatomical modifications, or "preconditions," observed in Flaveria, followed by changes to BS organelle content and cell size. Alterations to M tissue development are also significant, and the loss of ground tissue layers has likely occurred at least three and possibly four times during the evolution of Flaveria. These results highlight the important anatomical modifications to M and BS in facilitating C3–C4 intermediate and C4 photosynthesis and support the hypothesis that modifications to vein pattern and M/BS tissues precede advancement in photosynthetic evolution. Furthermore, the acquisition of derived physiological features at different nodes along the Flaveria phylogeny suggests common anatomical requirements for the evolution of C3–C4 and C4 photosynthesis.
APPENDIX 1.
Taxa used in this study, collector's number, and herbarium. All voucher specimens are located in the Royal Ontario Museum, Toronto, Ontario, Canada
Taxon; Voucher.
