Genus Suaeda (family Chenopodiaceae, subfamily Suaedoideae) has two structural types of Kranz anatomy consisting of a single compound Kranz unit enclosing vascular tissue. One, represented by Suaeda taxifolia, has mesophyll (M) and bundle sheath (BS) cells distributed around the leaf periphery. The second, represented by Suaeda eltonica, has M and BS surrounding vascular bundles in the central plane. In both, structural and biochemical development of C4 occurs basipetally, as observed by analysis of the maturation gradient on longitudinal leaf sections. This progression in development was also observed in mid-sections of young, intermediate, and mature leaves in both species, with three clear stages: (i) monomorphic chloroplasts in the two cell types in younger tissue with immunolocalization and in situ hybridization showing ribulose bisphosphate carboxylase oxygenase (Rubisco) preferentially localized in BS chloroplasts, and increasing in parallel with the establishment of Kranz anatomy; (ii) vacuolization and selective organelle positioning in BS cells, with occurrence of phosphoenolpyruvate carboxylase (PEPC) and immunolocalization showing that it is preferentially in M cells; (iii) establishment of chloroplast dimorphism and mitochondrial differentiation in mature tissue and full expression of C4 biochemistry including pyruvate, Pi dikinase (PPDK) and NAD-malic enzyme (NAD-ME). Accumulation of rbcL mRNA preceded its peptide expression, occurring prior to organelle positioning and differentiation. During development there was sequential expression and increase in levels of Rubisco and PEPC followed by NAD-ME and PPDK, and an increase in the 13C/12C isotope composition of leaves to values characteristic of C4 photosynthesis. The findings indicate that these two forms of NAD-ME type C4 photosynthesis evolved in parallel within the subfamily with similar ontogenetic programmes.
C4 photosynthesis evolved as a means of increasing the supply of CO2 to ribulose bisphosphate carboxylase oxygenase (Rubisco) and the C3 cycle under conditions that would otherwise be limiting for photosynthesis. Limitations can occur by reduced stomatal conductance (e.g. in response to a water deficit), and by warm temperatures where photorespiration and the Rubisco oxygenase/carboxylase activity ratio is increased.
C4 plants evolved independently multiple times, with convergence in a number of specialized features. These shared features include fixation of atmospheric CO2 into C4 acids within mesophyll (M) cells through phosphoenolpyruvate carboxylase (PEPC), and decarboxylation of C4 acids and donation of released CO2 to Rubisco and the C3 cycle within bundle sheath (BS) cells. Common to this process is the selective expression of the carboxylation phase of C4 enzymes in the M cells, with Rubisco and other C3 cycle enzymes accumulating in the chloroplasts, and glycine decarboxylase in the mitochondria, of the BS. This pathway is also correlated with distinctive cellular structural features that provide diffusive resistance to leakage of CO2, so that it can be effectively donated from the C4 cycle to Rubisco (Edwards and Walker, 1983; Hatch, 1987; Sage, 2004).
The independent evolution of C4 photosynthesis from C3 ancestors across 19 different plant lineages has led to considerable diversity in biochemistry, with three types of C4 cycle operating through the C4 decarboxylases, NADP-malic enzyme (NADP-ME), NAD-malic enzyme (NAD-ME), and phosphoenolpyruvate carboxykinase (PEP-CK), as well as diversity in structural features. There are at least 25 forms of Kranz anatomy, and two forms of single-cell C4 without Kranz (Edwards and Voznesenskaya, 2011). The forms of Kranz anatomy can be classified into two major groups: one in which two layers of Kranz chlorenchyma are formed around each vein (multiple Kranz units per leaf); the other in which layers of Kranz chlorenchyma surround all the veins, forming a single compound Kranz unit, according to the classification of Peter and Katinas (2003). Current information, though limited, indicates that there is also diversity in the ways different enzymes were recruited during evolution to function in the C4 cycles, and in the mechanism leading to control of selective expression of C4 in M versus BS cells, e.g. transcriptional, post-transcriptional, translation control (Hibberd and Covshoff, 2010; Berry et al., 2011; Gowik and Westhoff, 2011).
The family Chenopodiaceae includes many species, with great diversity in forms of C4 and C3 photosynthesis, including different anatomical, physiological, and biochemical features. Subfamily Suaedoideae is especially interesting as among ∼85 species there are four structural forms of C4 (two different forms having a compound Kranz unit, and two single-cell forms without Kranz anatomy, all NAD-ME biochemical subtype), as well as C3 species (Kadereit et al., 2003; Schütze et al., 2003; Voznesenskaya et al., 2007). In a classification organized according to sections within the genus Suaeda, three different types of anatomy were recognized among C3 species (Brezia, Vera, and Schanginia), and two C4 Kranz types named Salsina and Schoberia (Schütze et al., 2003).
The leaves of the two Kranz type Suaeda have a similar distribution of the vascular network. However, they differ in the arrangement of chlorenchyma and water storage tissue relative to the vascular bundles, as well as in the positioning of chloroplasts in BS cells. Salsina type has a compound Kranz unit with two layers of chlorenchyma characteristic of species with C4 photosynthesis, BS and M cells, occurring at the leaf periphery surrounding water storage tissue in which the vascular bundles are embedded. Schoberia type leaves have a compound Kranz unit characterized by two concentric layers of chlorenchyma cells, BS and M, encircling the vascular bundles, and by a layer of enlarged water storage hypodermal cells just beneath the epidermis. The chloroplasts in BS cells are located in the centripetal position in the Salsina type, and in the centrifugal position in the Schoberia type (Freitag and Stichler, 2000; Schütze et al., 2003; Voznesenskaya et al., 2007). Transmission electron microscopy (TEM) shows that chlorenchyma cells in species representing the two different C4 anatomical types have features characteristic of NAD-ME type C4 chenopods: the M cells contain chloroplasts with reduced grana, while BS cells have chloroplasts with well-developed grana and large, specialized mitochondria (Fisher et al., 1997; Voznesenskaya et al., 2007). Biochemical and physiological studies show that the availability and selective immunolocalization of key photosynthetic enzymes, CO2 compensation points, photosynthetic response curves to varying CO2 and light were similar for Salsina and Schoberia anatomical types, and these were typical for C4 plants (Voznesenskaya et al., 2007; Smith et al., 2009).
The establishment of complex C4 systems requires highly coordinated expression of many genes during development, under the control of regulatory factors that must function concurrently with differentiation of Kranz anatomy (Hibberd and Covshoff, 2010; Berry et al., 2011). Studies on development of leaf structure relative to enzyme expression have been performed on a limited number of C4 species belonging to different structural and biochemical subtypes. Except for Salsola richteri, which has a structural form of C4 called Salsoloid type with a compound Kranz unit and NADP-ME biochemistry (Voznesenskaya et al., 2003b), most studies on C4 development have been on C4 species that have multiple simple Kranz units with chlorenchyma surrounding individual veins. Among the eudicots studied this includes Atriplex rosea (Dengler et al., 1995) and Amaranthus hypochondriacus (Ramsperger et al., 1996), which have atriplicoid type leaf anatomy and the NAD-ME biochemical subtype. In monocots, this includes developmental studies on Arundinella hirta (Wakayama et al., 2003) and Zea mays (Langdale et al., 1988a; Brutnell et al., 1999; Covshoff et al., 2008; Majeran et al., 2010), which belong to the NADP-ME type in family Poaceae, and on NADP-ME and NAD-ME representatives of the family Cyperaceae (Soros and Dengler, 2001). All Kranz type C4 plants studied are characterized by a gradual pattern of anatomical transition from morphologically similar chlorenchyma cells to structural and biochemical dimorphism between BS and M cells (Liu and Dengler, 1994; Dengler et al., 1996; Voznesenskaya et al., 2003a, b; Wakayama et al., 2003). Structural differentiation of M and BS cells is closely associated with the C4 gene and enzyme expression, although some differences have been reported in temporal relations between the level of chlorenchyma structural differentiation, the establishment of C4 enzyme expression, and their distribution within the two cell types (Wang et al., 1992; Dengler et al., 1995; Soros and Dengler, 2001). These processes are often found to be dependent on environmental (e.g. light) and/or developmental factors, and it has been suggested that positional information during vascular tissue differentiation may control expression of Kranz around veins (Langdale et al., 1988a; Dengler et al., 1995; Hibberd and Covshoff, 2010).
The purpose of this study was to characterize and compare the structural and biochemical development of photosynthesis in two C4 representatives of the genus Suaeda, Suaeda taxifolia (section Salsina) and Suaeda eltonica (section Schoberia). The anatomical and ultrastructural features, accumulation of key photosynthetic enzymes, and carbon isotope discrimination at different stages of leaf development are described. In addition, the cellular compartmentation of the BS and M cell C4 marker enzymes, Rubisco large subunit (LSU) and PEPC respectively, were investigated immunohistochemically, together with immunogold labelling TEM, to evaluate the relative amount of these peptides in different cell types and at different stages of leaf development. In situ mRNA hybridization procedure was used to reveal temporal and spatial patterns of Rubisco LSU gene expression in these two C4 species. Unlike previous studies on C4 development where Kranz anatomy occurs around individual veins, this is the first comprehensive study of C4 development in two forms of Kranz that have a single compound Kranz unit enclosing all the vascular tissue. It was shown that the pattern of C4 differentiation was not affected by vascular tissue positioning relative to BS, or by differences in the origin of BS cells. Spatial and temporal patterns of Rubisco LSU transcript and peptide accumulation indicate that control of transcript levels, by transcription or mRNA stability, has a role in determining very early C4 development (before differentiation and positioning of BS chloroplasts).
Material and methods
Seeds of two species of genus Suaeda, S. eltonica Iljin (provided by H. Freitag from W. Kazakhstan, 28.237, KAS) and S. taxifolia Standley (provided by Dr J. Thorsch, University of California Santa Barbara, CA, USA) were stored at 3–5 °C prior to use. Seeds were germinated on moist paper in Petri dishes at room temperature at a photosynthetic photon flux density (PPFD) of 20 μmol photosynthetic quanta m−2 s−1. The seedlings were then transplanted to 10 cm diameter pots with commercial potting soil and grown for 3 d under 30 PPFD and 25/15 °C day/night temperature regime. Established plants were then transferred to a growth chamber (model GC-16; Enconair Ecological Chambers Inc., Winnipeg, Canada) and were grown under a PPFD of 400 μmol quanta m−2 s−1, with a 16-h/8-h light/dark photoperiod and 25/15 °C day/night temperature regime, atmospheric CO2, and 50% relative humidity. For microscopy and biochemical analyses, leaf samples of different lengths (see below) were taken from vegetative branches on the plants that were ∼30–45 d old. As described in the Results, three stages of development of photosynthetic cells were established from studies on longitudinal sections of individual leaves (Figs 1, 5, 6, 8), and analyses of leaves of different ages (Table 1, Figs 2–4, 7). Voucher specimens are available at the Marion Ownbey Herbarium, Washington State University: S. eltonica (E. Voznesenskaya 38), April 2006, WS 369797 and S. taxifolia (E. Voznesenskaya 39) July 2006, WS 369802.
|Leaf age||Leaf length (cm)||S. taxifolia δ13C (‰)||S. eltonica δ13C (‰)|
|Leaf age||Leaf length (cm)||S. taxifolia δ13C (‰)||S. eltonica δ13C (‰)|
Sample preparation for light and electron microscopy
Study of C4 anatomy during leaf development was carried out on expanding leaves with lengths from 0.1 to 0.7 cm, and on mature leaves after full expansion. For structural studies, three samples of each leaf length were harvested from three independent plants of each species (a total of nine samples from each plant for each leaf length). Cross-sections were made in the middle part of leaves that were young (0.2–0.3 cm), intermediate (0.5–0.7 cm), or mature, fully expanded (2–3 cm). Vegetative shoot apices with several of the youngest leaf primordia, and intermediate age leaves (0.5–0.7 cm in length), were also sectioned longitudinally. Samples for structural studies were fixed at 4 °C in 2% (v/v) paraformaldehyde and 2% (v/v) glutaraldehyde in 0.1 M phosphate buffer (pH 7.2), postfixed in 2% (w/v) OsO4, and then, after a standard acetone dehydration procedure, embedded in Spurr's epoxy resin. Reichert Ultracut R ultramicrotome (Reichert-Jung GmbH, Heidelberg, Germany) was used for sectioning. For light microscopy, semi-thin sections were stained with 1% (w/v) Toluidine blue O in 1% (w/v) Na2B4O7, and studied under an Olympus BH-2 (Olympus Optical Co. Ltd) light microscope equipped with LM Digital Camera and Software (Jenoptik ProgRes Camera, C12plus, Jena, Germany). For TEM, ultra-thin cross-sections were stained with 4% (w/v) uranyl acetate followed by 2% (w/v) lead citrate. Hitachi H-600 and JEOL JEM-1200 EX TEMs were used for observation and photography.
For scanning electron microscopy (SEM), shoot apices with the youngest leaf primordia were fixed as described above for structural studies, postfixed in 2% (w/v) OsO4, and then dehydrated in an ethanol series to 100% ethanol, critical-point dried, attached to SEM mounts, sputter-coated with gold, and observed with a Hitachi S570 SEM (Hitachi Ltd, Tokyo, Japan).
To observe formation of the leaf vascular pattern, leaves of different ages, from youngest primordia to fully expanded leaves, were cleared in 70% ethanol (v/v) until chlorophyll was removed, bleached with 5% (w/v) NaOH overnight, and then rinsed three times in water. At least five vegetative shoot tips with different-aged leaves from two or three different plants were used. The leaves were mounted in water and examined under UV light (with DAPI filter) on a Fluorescence Microscope Leica DMFSA (Leica Microsystems Wetzlar GmbH, Germany) using autofluorescence of lignified tracheary elements of the xylem.
Immunolocalization using confocal and electron microscopy
Leaves of different ages were fixed at 4 °C in 2% (v/v) paraformaldehyde and 1.25% (v/v) glutaraldehyde in 0.05 M PIPES buffer, pH 7.2. The samples were dehydrated with a graded ethanol series and embedded in London Resin White (LR White; Electron Microscopy Sciences, Fort Washington, PA, USA) acrylic resin. Antibodies used (all raised in rabbit) were anti-Spinacia oleracea Rubisco LSU IgG (courtesy of B. McFadden) and commercially available anti-Z. mays L. PEPC IgG (Chemicon, Temecula, CA, USA). Preimmune serum was used in all cases for controls.
Longitudinal sections, 0.8–1 μm thick, were dried from a drop of water onto gelatin-coated slides and blocked for 1 h with TBST+BSA (10 mM TRIS–HCl, 150 mM NaCl, 0.1% v/v Tween 20, 1% w/v BSA, pH 7.2). These were then incubated for 3 h with either preimmune serum diluted in TBST+BSA (1:100), anti-Rubisco (1:500 dilution), or anti-PEPC (1:200 dilution). The slides were washed with TBST+BSA and then treated for 1 h with protein A–gold 10 nm (diluted 1:100 with TBST+BSA). After washing, the sections were exposed to a silver enhancement reagent for 20 min according to the manufacturer's directions (Amersham, Arlington Heights, IL, USA), stained with 0.5% (w/v) Safranin O, and imaged in a reflected/transmitted mode using a Zeiss Confocal LSM 510 Meta Laser Scanning Microscope (Carl Zeiss, Inc. Headquarters, Thornwood, NY, USA). The background labelling with preimmune serum was very low, although some infrequent labelling occurred in areas where the sections were wrinkled due to trapping of antibodies/label (results not shown).
For TEM immunolabelling, thin cross-sections (∼70 nm thick) on Formvar-coated nickel grids were incubated for 1 h in TBST+BSA to block non-specific protein binding on the sections. They were then incubated for 3 h with either the preimmune serum diluted in TBST+BSA, or anti-Rubisco (1:50), or anti-PEPC (1:20) antibodies. After washing with TBST+BSA, the sections were incubated for 1 h with Protein A–gold (10 nm) diluted 1:100 with TBST+BSA. The sections were washed sequentially with TBST+BSA, TBST, and distilled water, and then post-stained with a 1:4 dilution of 1% (w/v) potassium permanganate and 2% (w/v) uranyl acetate. Images were collected using a Hitachi H-600 TEM.
The density of labelling was determined by counting the gold particles on digital electron micrographs using an image analysis program (UTHSCSA, Image Tool for Windows, version 3.00) and calculating the number per unit area (μm2). Standard errors were determined and analysis of variance (ANOVA) was performed using Statistica 7.0 software (StatSoft, Inc.). Tukey's honest significant difference (HSD) test was used to analyse differences between amounts of gold particles at different stages of leaf development. All analyses were performed at the 95% significance level.
In situ localization of mRNAs encoding Rubisco LSU protein
Sense and antisense RNA probes for Rubisco LSU were generated from pBlsl (which contains a 600-bp HindIII fragment from the central coding region of the amaranth LSU gene) using a modification of procedures described in Wang et al. (1992). Sense and antisense transcripts were synthesized and labelled in vitro with Biotin-11-UTP (Roche) using T7 or T3 polymerase (Roche). Leaf samples were fixed in FAA (50% ethanol, 5% glacial acetic acid, 10% formalin) fixative at room temperature overnight. After ethanol and t-butyl alcohol dehydration, samples were embedded in Paraplast Plus. The paraffin-embedded samples were sectioned (thickness 5–10 μm) using a rotary microtome; sections were mounted to poly-L-lysine-coated slides, dried, and stored at 4 °C overnight. After deparaffinization by xylene, the sections were rehydrated through an ethanol series, incubated in 0.2 M HCl for 20 min at room temperature (RT), and rinsed in H2O. Slides were then incubated in 2×SSC (1×SSC is 0.15 M NaCl, 0.15 M sodium citrate) at 70 °C for 30 min and rinsed by H2O. Treatment of sections with proteinase K (1 μg ml−1 in TE: 100 mM TRIS pH 8.0, 50 mM EDTA) for 15 min at 37°C was followed by a brief rinse in PBS (10 mM phosphate buffer, 2.7 mM KCl, 137 mM NaCl, pH 7.4), blocking in glycine (2 mg ml−1 in PBS) for 2 min at RT, and then fixation for 20 min in 4% formaldehyde. Prehybridization was accomplished in prehybridization medium overnight at 50 °C after incubation in 2×SSC for 10 min at RT. The probes were first heated at 75 °C for 30 s and mixed with prehybridization medium with a final transcript concentration of 0.5 μg ml−1. Hybridizations with the Biotin-labelled transcripts were performed at 50 °C overnight in a moist chamber. The slides were then subjected to the following series of washes: prewarmed at 37 °C 2×SSC, 10 min; 2×SSC, 1 h, RT; 1×SSC, 1 h, RT; 0.5×SSC, 30 min, 42 °C; 0.5×SSC, 30 min, RT. Hybridized transcripts were detected by streptavidin–alkaline phosphate conjugate (NeutrAvidin; Pierce) using blocking buffer (100 mM TRIS–HCl, pH 7.5, 150 mM NaCl). The final detection step was carried out by nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate (Sigma). Reactions were stopped by placing slides in 1×PBS, and the sections were mounted in 50% glycerol in PBS. Slides were studied under Olympus BH-2 light microscope equipped with LM Digital Camera and Software.
The colour intensity of chloroplast labelling at base, middle part, and tip of leaves was quantified from images collected under the same settings and magnification using an image analysis program (ImageJ 1.37v; Wayne Rasband, National Institutes of Health, Bethesda, MD, USA) and expressed relative to level at the tip, which was set as 100%.
Western blot analysis
Total proteins were extracted from leaves (n=2 for each species) by homogenizing 0.5 g of tissue in 1 ml of extraction buffer [100 mM TRIS–HCl, pH 7.5, 5 mM MgSO4, 10 mM DTT, 5 mM EDTA, 0.5% (w/v) SDS, 2% (v/v) β-mercaptoethanol, 10% (v/v) glycerol, 1 mM phenylmethylsulfonyl fluoride, and 25 μg ml−1 each of aprotinin, leupeptin, and pepstatin]. After centrifugation at high speed for 3 min in a microcentrifuge, the supernatant was collected, and protein concentration was determined using a Bradford protein assay (Bio-Rad), with BSA as a standard. Protein samples (10 μg) were separated by 12% SDS–PAGE, blotted onto nitrocellulose, and probed with anti-A. hypochondriacus NAD-ME IgG against the 65 kDa α subunit (Long and Berry, 1996) (1:5 000), anti-Z. mays PEPC IgG (1:10 000), anti-Z. mays pyruvate, Pi dikinase (PPDK) IgG (courtesy of T. Sugiyama) (1:5 000), or anti-S. oleracea Rubisco LSU IgG (1:10 000) overnight at 4 °C. Goat anti-rabbit IgG–alkaline phosphatase-conjugated secondary antibody (Bio-Rad) was used at a dilution of 1:50000 for detection. Bound antibodies were visualized by developing the blots with 20 mM nitroblue tetrazolium and 75 mM 5-bromo-4-chloro-3-indolyl phosphate in detection buffer (100 mM TRIS–HCl, pH 9.5, 100 mM NaCl, and 5 mM MgCl2). The intensities of bands in Western blots were quantified with an image analysis program (ImageJ 1.37v) and expressed relative to level in the mature stage of development, which was set as 100%.
δ13C values, a measure of carbon composition, were determined on leaf samples of different ages taken from plants using a standard procedure relative to PDB (Pee Dee Belemnite) limestone as the carbon isotope standard (Bender et al., 1973). Plant samples from the growth chamber were dried at 80 °C for 24 h, and milled to a fine powder. This preparation (1–2 mg) was placed in a tin capsule and combusted using a Eurovector elemental analyser. The resulting N2 and CO2 gases were separated by gas chromatography, and admitted into the inlet of a Micromass Isoprime isotope ratio mass spectrometer (IRMS) for determination of 13C/12C ratio (R). δ13C values were determined where δ13C=[(Rsample/Rstandard) –1]×1000. δ13C is an expression of the 13C/12C of the plant sample relative to the standard (reported in parts per 1000), Rsample and Rstandard is the 13C/12C ratio of the plant sample and the PDB limestone, respectively.
Structure of shoot apex and leaf growth
Vegetative apical shoot meristems of both S. taxifolia and S. eltonica at the stage of active organogenesis are dome shaped (Fig. 1 A–D). These have a diameter at the base of ∼200 μm, and height of 140 μm in S. taxifolia (Fig. 1A, C), and 160 and 120 μm, respectively, in S. eltonica (Fig. 1B, D). In both species, the outermost layer (L1, also known as the first layer of the tunica) divides only anticlinally and gives rise to the epidermis of the leaf primordium (see Fig. 1D for S. eltonica). Cells of the inner layers (L2 and L3), which make up the body of the apical meristem, divide both anticlinally and periclinally, generating the inner tissues of leaf primordium. Leaf primordia in both species are initiated alternately and rather often (Fig. 1A, B), resulting in the formation of a compact apical bud with close positioning of leaves (Fig. 1E, F).
Figures 1E and F show the organization of the shoot tips in the two Suaeda species during active vegetative growth. In the bud, the emergence of leaf tips from the encircling superior leaves, and their subsequent greening, occurs earlier in S. taxifolia (length of leaf ∼0.2 cm) than in S. eltonica (length of leaf ∼0.3 cm), while in general the rate of leaf growth was higher in S. eltonica. The extended leaves turn aside earlier in S. taxifolia (Fig. 1E) while in S. eltonica (Fig. 1F) they are still tightly packed in the apical bud and maintain a vertical orientation until they reach ∼0.4 cm in length.
Light and transmission electron microscopy: developmental stages
Development of Kranz anatomy was studied at the light microscopy level using longitudinal sections of intermediate size leaves, 0.5–0.7 cm (Fig. 1G, H), and cross-sections of young (0.2–0.3 cm), intermediate (0.5–0.7 cm), and mature fully expanded (2–3 cm) leaves (Figs 2, 3). These observations indicate that in both species, cellular differentiation during the growth of young leaves proceeds in a basipetal direction, with the youngest undifferentiated (meristematic) cells at the leaf base and the most differentiated cells at the leaf tip. Bundle sheath and M cell progenitors can be recognized at the base of the young leaf by their position in comparison with adjacent tissues (Fig. 1G, H). Observations of longitudinal sections of young leaves at their base show differences in the origin of BS layers during the process of Kranz anatomy development. In S. taxifolia (Fig. 1G), periclinal divisions of subepidermal cells of ground meristem give rise to M and BS layers, while in S. eltonica, periclinal divisions of subepidermal cells generate future water storage hypoderm and M layers (Fig. 1H). Bundle sheath cells in S. eltonica originate from the central layer of ground meristem around the procambial strands (Figs 1H, 3A). Development of the vascular system in Suaeda species also takes place in the basipetal direction. Lignin autofluorescence was used to investigate the sequence of development of xylem tracheary elements in young leaves. Clearly the xylem in the midvein is established first (developing acropetally, not shown) followed by development of lateral veins with their connecting loops, beginning from the leaf tip (Fig. 1I, J).
Following the initial formation of M and BS cells from progenitor cells, three developmental stages were characterized in the progression towards the formation of the Kranz syndrome. Structural features of these stages are shown in cross-sections of young, intermediate, and mature leaves (designated stages 1, 2, and 3, respectively), as observed by light and electron microscopy of S. taxifolia (Fig. 2) and S. eltonica (Fig. 3). In young leaves (0.2–0.3 cm in length), in S. taxifolia two clearly defined layers of chlorenchyma cells are located at the leaf periphery next to the epidermal cells (Fig. 2A); in contrast, the two chlorenchyma layers are located under the hypodermal cells between the veins in S. eltonica (Fig. 3A). Precursors of M and BS cells are distinct in shape and size; M cells are more palisade shaped due to anticlinal divisions especially in S. eltonica, where for each future BS cell, on average there are two future M cells (Fig. 3A).
In stage 1, BS cells in both species have a developing central vacuole, and chloroplasts are distributed around the cell periphery. In M cells, the nucleus has a central position in S. taxifolia (Fig. 2A), while in S. eltonica, the nucleus is adjacent to the middle part of the radial cell wall due to earlier formation of the central vacuole (Fig. 3A). At this stage of leaf development, the cells of all tissues are tightly packed in both species; minor intercellular air spaces exist between M and epidermal cells in S. taxifolia, while more developed airspaces occur between hypodermal cells in S. eltonica (Figs 2A, 3A). The TEM observations confirm that small chloroplasts are evenly distributed throughout the cytosol at the cell periphery, and sometimes around the nucleus. These have a relatively well-developed thylakoid system with numerous small grana and intergranal thylakoids that are structurally similar in both M and BS cells (Figs 2B, D, 3B, D). Mitochondria in chlorenchyma cells are few in number, small, and they have crescent-like cristae (see Figs 2C, 3C for BS cell). The vascular system is not fully differentiated at this stage, although the midvein is already clearly morphologically evident with fully developed phloem and xylem elements. The developing peripheral vascular bundles usually have only one differentiated xylem vessel. They still have periclinal cell divisions at the phloem end, especially in the most lateral bundles, in accordance with gradient of maturation: the farther from the central bundle, the less degree of minor vein differentiation (Figs 2E, 3E). Mature xylem vessels lack cytoplasm and show secondary wall thickening; mature sieve tubes appear to lack nucleus, tonoplast, and ribosomes, and they contain only a peripheral layer of cytosol, with rare occurrence of mitochondria, plastids, and smooth endoplasmic reticulum, when viewed at higher magnification (not shown).
Stage 2 was observed in cross-sections from the middle part of intermediate leaves (∼0.5–0.7 cm in length). For both species, at this stage M cells show a well-developed central vacuole with organelles distributed in a thin cytoplasmic layer adjacent to the cell walls (Figs 2F, 3F). The characteristic positioning of chloroplasts in BS cells for each species has occurred and is almost complete: plastids are distributed centripetally in BS cells of S. taxifolia (Fig. 2F) and centrifugally in BS cells of S. eltonica (Fig. 3F). Intercellular air spaces throughout the M cell layer are formed in both species (Fig. 2F, 3F). Chloroplasts in chlorenchyma cells are bigger and more numerous than in the previous stage; their thylakoid systems are more developed, but there is similar grana development in BS and M chloroplasts (Figs 2G, I, 3G, I). Bundle sheath mitochondria, which are colocalized with chloroplasts, show a tendency to locate next to the inner periclinal cell walls in S. taxifolia, and the outer periclinal cell walls in S. eltonica (not shown). At this stage of leaf development, the peripheral vascular bundles have two to four mature xylem vessels and one or two mature sieve elements, depending on the distance from the central vein (Figs 2J, 3J). The pattern of maturation of peripheral vascular bundles at this stage still maintains a gradient, with less differentiated minor veins situated closer to the leaf margins.
In mature leaves (stage 3), the structural features of Salsina and Schoberia types of Kranz anatomy are completely formed. In S. taxifolia (Salsina type), the chlorenchyma is composed of two continuous layers at the leaf periphery, and all vascular bundles are embedded within water storage tissue and arranged in one longitudinal plane; BS chloroplasts are positioned centripetally (Fig. 2K). In S. eltonica (Schoberia type), the inner vascular bundles are surrounded by chlorenchyma, and water storage tissue is positioned on the outside; BS chloroplasts are located centrifugally (Fig. 3K). Bundle sheath cells of mature leaves in both species contain chloroplasts having numerous grana with 5–20 thylakoids in stacks (Figs 2L, 3L). Intergranal thylakoids are few and short in S. taxifolia (Fig. 2L), and longer and more numerous in S. eltonica (Fig. 3L). For both species, in comparison with BS chloroplasts, M chloroplasts have poorly developed grana which are interconnected by rather numerous, long intergranal thylakoids (but they are more numerous in S. eltonica compared with S. taxifolia, Figs 2N, 3N). In mature leaves, the BS cells have abundant, large mitochondria with cristae showing a distinct tubular type structure (Figs 2M, 3M). BS organelles maintain characteristic positioning in the cell, as was noted in stage 2. In stage 3, all minor veins in both species are fully differentiated and contain three or four xylem vessels and three or four sieve elements (Figs 2O, 3O).
Carbon isotope composition during leaf development
The δ13C values for leaves at different stages of development (young, intermediate, nearly mature, and mature) are presented in Table 1. In both Suaeda species, the isotope values become more positive as development progresses from young to intermediate to mature leaves. The values in young versus mature leaves in S. taxifolia were –15.2‰ versus –12.3‰, and in S. eltonica, –18.4‰ versus –14.2‰, respectively, indicating less expression of C4 in the younger leaves.
Accumulation of C4 pathway enzymes and Rubisco during leaf development
Figure 4 shows western blot analysis of PEPC, PPDK, NAD-ME, and Rubisco in the total soluble protein extracts from very young up to mature leaves of S. taxifolia and S. eltonica. In both Suaeda species, Rubisco and to a lesser extent PEPC were present at the youngest stage of leaf development (0.1 cm leaf length), and their levels steadily increased as leaf age progressed. In contrast, very few PPDK or NAD-ME polypeptides were detected in young leaves (0.3 cm long, corresponding to stage 1). Increased levels of these C4 enzymes were detected in intermediate size leaves, 0.5 cm in length, with maximum levels occurring in the mature leaves.
In situ immunolocalization of Rubisco and PEPC during leaf development
Intermediate size leaves (0.5–0.7 cm) of both Suaeda species have a basipetal longitudinal gradient in chlorenchyma cell development, with features of stage 1 located at the leaf base where cell divisions and expansions occur, stage 2 in the mid-section, and stage 3 at the tip. The pattern of Rubisco LSU and PEPC expression, by antibody labelling, along this gradient was studied in longitudinal sections by confocal microscopy (Figs 5A–C, J–L, 6A–C, J–L). For quantitative evaluation of selective enzyme accumulation, immunogold labelling and detection using TEM was applied to cross-sections of young, intermediate, and mature leaves representing similar stages of development (Figs 5D–I, M–R, 6D–I, M–R), with subsequent counting of the number of gold particles detected in the different cell compartments (Fig. 7). In both S. taxifolia (Fig. 5A) and S. eltonica (Fig. 6A), in stage 1, labelling for Rubisco was detected by confocal imaging after the start of BS cell vacuole extension, but before specific positioning of the BS organelles. At this stage, Rubisco was present mainly in the BS chloroplasts (Figs 5A, 6A), a finding that was confirmed by TEM immunolocalization (Figs 5D, E, 6D, E). Only in S. eltonica did M chloroplasts contain a noticeable quantity of labelling for Rubisco at this stage; however, levels were still only half that of BS chloroplasts (Fig. 7A).
Subsequently by stage 2, Rubisco LSU labelling was strongly associated with BS chloroplasts in both species (Figs 5B, F, G, 6B, F, G,), with the density of labelling in BS cell chloroplasts increasing >2-fold (Fig. 7A). At this stage, the density of gold particles in M chloroplasts dropped to half the density of the earlier stage (Fig. 7A).
At stage 3, the pattern of Rubisco LSU labelling distribution did not change much qualitatively; however, the density of gold particles in the BS chloroplasts reached its highest level, while in M chloroplasts LSU labelling did not differ significantly from background (Figs 5C, H, I, 6C, H, I). Quantification of these data in Fig. 7A indicates a very similar pattern of Rubisco accumulation in both S. taxifolia and S. eltonica.
In both species, in the earliest stage PEPC was undetectable by both immunolocalization methods applied (Figs 5J, M, N, 6J, M, N). Nevertheless, at stage 2 (from the mid-section of intermediate leaves), specific labelling of PEPC was observed in M cells (Figs 5K, O, P, 6K, O, P). In stage 3, labelling for PEPC was at its maximum throughout the cytosol of M cells, with little, if any, labelling observed in the BS cytosol (Figs 5L, Q, R, 6L, Q, R). The graph shown in Fig. 7B indicates a very similar trend in PEPC accumulation and partitioning at the different developmental stages for both Suaeda species.
Taken together, the two imaging methods, coupled with quantification of the number of gold particles in the different tissues and stages, have clearly revealed the spatial and temporal patterns of Rubisco and PEPC accumulation during leaf development (Fig. 7A), with very similar results for both Suaeda species.
Temporal and spatial accumulation of rbcL mRNA at different stages of leaf development
Using in situ hybridization, longitudinal sections of young and intermediate leaves (0.2–0.6 cm long) of S. taxifolia and S. eltonica were studied to examine the temporal and spatial distribution of Rubisco LSU mRNA. Specific hybridization of Biotin-11-UTP-labelled antisense RNA probes transcribed using T7 polymerase from a vector containing the central coding region of an LSU gene, are clearly visualized as a dark purple colour, as seen in Fig. 8. As a negative control, the sense probes transcribed using T3 polymerase from the same plasmid showed no purple hybridization signal (Fig. 8A). In both S. taxifolia and S. eltonica, a very low hybridization signal (slightly over background) in pre-BS and M chloroplasts was detected at the earliest stages of development at the very base of the young leaf, before expansion of vacuoles begins (Fig. 8B, F, respectively). Beginning at stage 1, there was a clear hybridization signal within the BS chloroplasts of both species (Fig. 8B, C, F, G), with no (in the case of S. taxifolia, Fig. 8C), or very low (in S. eltonica, Fig. 8G), signal in M cells. At stage 2 of leaf development (Fig. 8B, D, F, H) in both species, expression levels of the Rubisco LSU gene appeared enhanced, and were confined to BS chloroplasts (Fig. 8B, D for S. taxifolia and Fig. 8F, H for S. eltonica). Colour intensity quantification per individual chloroplast area (not shown) revealed no difference for S. taxifolia and a slight (1.1-fold) increase in labelling for S. eltonica at stage 2 comparing with stage 1 (at P<0.05). The intensity of the colour in the section increased from stage 1 to stage 2 due to increase in chloroplast size and number, and their arrangement towards one side of BS cells. There were no significant changes in the pattern of labelling or signal intensity (at P<0.05) in stage 3 at the leaf tip (Fig. 8B, E, F, I), indicating that the mRNA accumulating patterns established at stage 2 were maintained throughout the remainder of leaf development.
A common feature of C4 development is the gradual differentiation of BS and M cell anatomy to produce the final functioning Kranz syndrome, starting from structurally uniform cells and ending with specialized C4 photosynthetic cells (Nelson and Langdale, 1989; Liu and Dengler, 1994; Dengler et al., 1996; Wakayama et al., 2003). To examine this progression, leaf development in C4 plants has been defined in various ways, e.g. according to leaf age (Liu and Dengler, 1994; Voznesenskaya et al., 2003a, b, 2005), distance from the leaf base (Wakayama et al., 2003; Majeran et al., 2010), development of leaf anatomy (e.g. vascular system differentiation, Dengler et al., 1986), chlorophyll content (Kirchanski, 1975), and changes in photosynthetic enzymes (transcripts and proteins) (Berry et al., 2011). This can make it difficult to assimilate information between studies in order to describe the sequence of events in C4 development. To determine the timing and order of development of the various C4 traits in a certain Kranz/biochemical type of C4, a number of analyses need to be made simultaneously (e.g. as in maize, Li et al., 2010; Majeran et al., 2010). In the present study we examined longitudinal and cross-sections and compared anatomical and ultrastructural features of chlorenchyma cells as well as vascular elements at different developmental stages of two Kranz type species of Suaeda. These data were complemented with quantitative expression analysis of several C4 photosynthetic enzymes, including the in situ localization of the two carboxylases, PEPC (protein) and Rubisco (protein and mRNA).
Sequence of events in development of two forms of Kranz in genus Suaeda
Basipetal maturation gradient. The two eudicot C4 species S. taxifolia and S. eltonica in subfamily Suaedoideae represent two forms of Kranz anatomy, Salsina and Schoberia, respectively. In both types, the development of lateral veins and cellular differentiation during growth of the semi-terete leaves proceeds in a basipetal direction along the base-to-tip maturation gradient, which is useful for studying the progression of differentiation to form C4. A similar progressive developmental gradient has been described for monocot leaves and was exploited for the study of photosynthetic tissue differentiation (Miranda et al., 1981; Langdale et al., 1988a; Nelson and Langdale, 1989; Soros and Dengler, 2001; Wakayama et al., 2003; Li et al., 2010; Majeran et al., 2010).
Origins of M and BS. Analysis of cell files near the leaf base on longitudinal and cross-sections revealed that both BS and M cells of the two Suaeda species originate early in leaf development from the ground meristem. There is, however, different positioning of the chlorenchyma tissue in the leaf, so that S. taxifolia has Kranz around the periphery, and S. eltonica has Kranz around the veins in a central plane. Study of tissue differentiation showed that the origin of the BS layer is restricted to different cell lineages in the two species: to subepidermal cells in S. taxifolia, and to central cells in S. eltonica. Bundle sheath and M cells in S. taxifolia represent sister cells separated by a single periclinal cell division (followed by anticlinal divisions of M cells). In S. eltonica, M cells are sister to hypodermal water storage cells as a result of periclinal division of subepidermal cells; hence in this species, BS cells originate earlier than M cells. According to several studies on leaf tissue origin by clonal analysis, in eudicots the epidermis is derived from the L1 layer of the shoot apical meristem; the L2 forms a single layer immediately beneath the epidermis, giving rise to future subepidermal M layers, and L3 is a progenitor of more central tissues (the vascular tissues and inner spongy M cells) (Telfer and Poethig, 1994; Leyser and Day, 2003). Thus, the most probable scenario is that from the very beginning, BS cells of the two Suaeda species are derived from different meristematic precursors. In general, BS cells of C4 species can ontogenetically originate from the procambium or from the leaf ground meristem, while M can be derived only from ground meristem (Esau, 1965; Dengler et al., 1985, 1986; Liu and Dengler, 1994; Dengler et al., 1995; Soros and Dengler, 2001). The BS cells in C4 plants are an example of convergent evolution of tissues having different ontogenetic origins, structures, and C4 biochemistry, yet still carrying out the common functions of decarboxylating C4 acids and donating the released CO2 to the C3 cycle.
Although BS cells of the two Suaeda Kranz types have different origins, and originate earlier in ontogeny in S. eltonica than in S. taxifolia, both species pass through similar anatomical (vacuolization, specific positioning of organelles, and organelle differentiation) and biochemical stages (timing of cell-specific Rubisco expression) during their development, which in both plants coincides with M cell differentiation. This is consistent with two previous studies which show that the differentiation of BS and M cells in C4 plants is tightly coordinated irrespective of the anatomical and biochemical types of C4 species or the ontogenetic origin of the BS (Dengler et al., 1986, 1996). In particular, in a study of two C4Panicum species, the origin of ground meristem-derived BS cells occurs earlier in development in Panicum effusum in comparison with the BS of Panicum bulbosum, which originates from the procambium; this disparity had little effect on the timing of BS cell differentiation (Dengler et al., 1986).
Stage 1 of C4 development. Anatomical and ultrastructural criteria were initially used to distinguish between the main stages of leaf development. Stage 1 was defined as lacking structural and ultrastructural differences between BS and M cells. Leaf vasculature at this stage consisted of a differentiated central vein, while minor veins had only one or two xylem vessels and no differentiated phloem elements. C4 was not established biochemically, as there was substantial Rubisco, low PEPC, and lack of expression of the other C4 enzymes PPDK and NAD-ME (western blots on extracts from young leaves). From the earliest detection of carboxylases by immunolocalization using TEM on leaf cross-sections there was preferential expression of Rubisco in BS cells, and PEPC was not detected in stage 1 (Fig. 7).
Stage 2 of C4 development. At this stage, the two chlorenchyma cell layers were clearly defined. Palisade M cells showed a well-developed central vacuole; BS cells were mainly characterized by organelles becoming positioned to one side of the cell (centripetal in S. taxifolia and centrifugal in S. eltonica). The ultrastructure of chloroplasts with respect to grana development does not differ between BS and M cells. In stage 2, BS mitochondria appeared more numerous and bigger than in stage 1; however, the structure of the cristae was similar to M cell mitochondria. This shows that the positioning of organelles in BS cells in both species occurs prior to the structural differentiation of chloroplasts and mitochondria which is observed in stage 3. This stage is also characterized by the first appearance of differentiated sieve elements in the lateral veins. Analyses of enzymes suggest that some capacity for biochemistry of C4 function is being established. The expression of PEPC, along with Rubisco, begins to reach high levels in stage 2 of chlorenchyma differentiation and low levels of the C4 enzymes PPDK and NAD-ME appear. Immunolocalization shows that Rubisco is selectively localized in BS cells and PEPC in M cells.
Stage 3 of C4 development. This occurs when the leaf tissue is nearly fully mature, and the structural features of the two Kranz types are completely formed with M and BS cells distributed around the leaf periphery in S. taxifolia, and with M and BS cells surrounding vascular bundles in the central lateral plane in S. eltonica. In both species, TEM shows that chloroplasts have become structurally differentiated in this final stage; M cells have chloroplasts with reduced grana, while BS cells contain chloroplasts with well-developed grana. Also, BS cells have large mitochondria that are specialized to function in C4 acid decarboxylation, a feature that is characteristic of NAD-ME type C4 chenopods. Western blots show that NAD-ME and PPDK and the carboxylases reach maximum levels at this stage.
Following morphological differentiation and final positioning of organelles within BS cells during stage 2, there is ultrastructural differentiation of chloroplasts and mitochondria in stage 3. This order of development has also been observed in some NADP-ME type C4 species. In Arundinella hirta, completely dimorphic chloroplasts (agranal in M, granal in BS, characteristic of NADP-ME type C4) appear only at the latter stage of development after specific positioning of organelles in BS cells (from analysis of figures in Dengler et al., 1996). Other NADP-ME species that have been studied also show specific differentiation of the thylakoid system relatively late in development (Laetsch and Price, 1969; Brangeon, 1973; Leech et al., 1973; Kirchanski, 1975; Dengler et al., 1986; Voznesenskaya et al., 2003b; Majeran et al., 2010). Also, in the two single-cell NAD-ME type C4 plants in subfamily Suaedoideae, Bienertia cycloptera (Voznesenskaya et al., 2005) and Suaeda aralocaspica (formerly Borszczowia) (Voznesenskaya et al., 2003a), completion of chloroplast differentiation occurs after positioning of organelles to two different compartments that are analogues of BS and M in Kranz types.
The structural differentiation that occurs in mitochondria in the Suaeda spp. in stage 3 coincides with increasing levels of mitochondrial NAD-ME. Similar findings have been reported for the single-cell C4 species S. aralocaspica (Voznesenskaya et al., 2003a), and for B. cycloptera (Voznesenskaya et al., 2005). In C4 plants NAD-ME is localized in mitochondria of BS cells (Hatch and Kagawa, 1974; Kagawa and Hatch, 1975; Edwards and Walker, 1983; Long et al., 1994) as is glycine decarboxylase, a photorespiratory enzyme (Rawsthorne, 1992) that is also associated with the completion of BS mitochondria differentiation (Voznesenskaya et al., 2003a). The findings that development of C4 biochemistry in BS mitochondria occurs as leaves reach maturity indicates that complete structural differentiation of these organelles, in coordination with other C4 developmental processes, is essential for full C4 function.
Analysis of carbon isotope composition of Suaeda species leaves shows that C4 type values are only obtained in mature leaves, which coincides with structural and biochemical evidence of full development of C4 in mature leaves. The δ13C values for C4 plants are typically between –10o/oo and –15o/oo; whereas values for C3 species are typically between –24o/oo and –30o/oo (Cerling, 1999). Young leaves of the Suaeda spp. have a carbon isotope composition that is 3–4o/oo more negative than mature leaves (a shift of 20–25% towards C3 type values), which is consistent with incomplete development of the C4 system. The degree of shift towards C3 type carbon isotope values in the biomass of young leaves may be limited by their import of some carbon from C4 photosynthesis in mature leaves. The late development of C4 biochemistry in the Suaeda spp. indicates that leaves will only function in full C4 mode as they mature.
Immunohisto- and cytochemistry of compartmentation of carboxylases in Suaeda versus other forms of C4
Immunohistochemistry using light microscopy on longitudinal sections, as well as analysis of gold particle density by TEM, revealed that the intercellular partitioning of Rubisco to BS cells, and PEPC to M cells, began to occur as soon as their expression became evident. There was no stage where either Rubisco or PEPC were equally labelled in BS and M cells. After their first appearance, both of the carboxylases underwent dramatic increases in abundance in coordination with leaf maturation. Preferential expression of Rubisco in BS chloroplasts at the earliest stages of leaf development has also been observed in the NAD-ME type eudicot Atriplex rosea, which has atriplicoid type anatomy (Dengler et al., 1995), as well as in NADP-ME type monocots Zea mays (Martineau and Taylor, 1985; Langdale et al., 1988,a) and A. hirta (Wakayama et al., 2003). In these cases, as in the present study, regulatory processes responsible for cell type specificity for Rubisco are active well before the final anatomical differentiation of the two photosynthetic cell types.
However, in other species there is evidence for expression of Rubisco in both chloroplasts before selective localization in Rubisco with leaf maturation. A different and more complex pattern of Rubisco accumulation in leaf development was shown for Amaranthus hypochondriacus, another NAD-ME type plant with atriplicoid type anatomy (Wang et al., 1992; Ramsperger et al., 1996; Patel and Berry, 2008). In the youngest leaves (2 mm in length), Rubisco peptide accumulation showed a C4 pattern confined to precursors of BS cells (Ramsperger et al., 1996); but, later (5-mm-long leaves), Rubisco protein was present in both M and BS cells in a C3-like pattern, and became completely BS specific in the C4 pattern only at a later stage of development (Wang et al., 1992). Also, in the single-cell C4 species in subfamily Suaedoideae, S. aralocaspica and B. cycloptera (NAD-ME type), Rubisco occurs initially within chloroplasts present in both cytoplasmic domains, indicative of an initial C3-like stage (Voznesenskaya et al., 2003,a, 2005). Another representative of the Chenopodiaceae family, Salsola richteri, an NADP-ME species with Salsoloid type of anatomy, showed a similar localization pattern during early leaf development, with low levels of Rubisco expression in plastids of all tissues of very young leaves, where cell divisions leading to the formation of pre-BS and M cells are evident (Voznesenskaya et al., 2003b). In cases where a C3-default stage of Rubisco localization occurs, timing of the transition to BS cell specificity for Rubisco is species dependent. This transition may occur early, in parallel with the initiation of anatomical differentiation, as in S. richteri (Voznesenskaya et al., 2003b), or later after differences between the two Kranz layers are clearly evident (in A. hypochondriacus) (Wang et al., 1992). The observation that a late transition similar to amaranth also occurs for the two intracellular compartments in single-cell C4 species (Voznesenskaya et al., 2003,a, 2005) during the structural differentiation of chloroplasts, indicates that similar developmental processes may work to establish C4-like localization in these plants as well. It is clear from this and previous findings that the timing of Rubisco expression during development of C4 photosynthesis does not follow a fixed pattern across the many species that utilize this pathway, suggesting the evolution of alternative regulatory mechanisms to control these convergent processes.
In the Kranz type Suaeda leaves expression of PEPC is initiated later in development than Rubisco, and is cell type specific at its first occurrence. This pattern is in agreement with most previous studies, regardless of C4 anatomical or biochemical type (Wang et al., 1992; Dengler et al., 1995; Ramsperger et al., 1996; Soros and Dengler, 2001; Voznesenskaya et al., 2003,b; Wakayama et al., 2003). However, three different anatomical and biochemical types of C4 species in the family Cyperaceae accumulate low levels of PEPC in both M and BS cells within the extension zone of young leaves; and, surprisingly, PEPC remains non-specific in Rhynchospora rubra (NADP-ME, rhynchosporoid anatomical type) even in the mature state (Soros and Dengler, 2001). In the single-celled C4 chenopods, PEPC is not strictly compartmentalized to one cytoplasmic domain, which is expected, since this is a cytosolic enzyme (Voznesenskaya et al., 2003a, 2005). Moreover, in S. aralocaspica there is a detectable amount of PEPC (immunolocalization by light microscopy) from the earliest stages of development (1-mm leaf length), simultaneously with Rubisco (Voznesenskaya et al., 2003a).
There are also differences between species as to when PEPC expression is first initiated relative to development of Kranz anatomy. In the two Kranz type Suaeda species, this occurs only after completion of specific BS organelle positioning in stage 2; however, it precedes chloroplast ultrastructural differentiation. For a C4 representative of the family Cyperaceae, Pycreus polystachyos (NADP-ME, chlorocyperoid anatomical type) the accumulation of PEPC was suggested also to occur only after cell structural differentiation (Soros and Dengler, 2001). However, this conclusion was made based only on light microscopy observations; no data were available on organelle positioning during development, or on their ultrastructure. In contrast, in Z. mays and A. rosea, it was shown that PEPC accumulation, which is M specific, occurs before morphological differentiation of BS and M cells (Langdale et al., 1988,a; Dengler et al., 1995). Finally in amaranth, abundant levels of PEPC were detected in M cells at the base of the youngest leaves (2 mm long), which have structurally similar BS and M cells (Ramsperger et al., 1996).
Cellular localization of rbcL mRNA and RbcL protein in Suaeda species
At all stages of leaf development in both Suaeda species, Rubisco rbcL mRNA was preferentially localized to BS cells, followed by corresponding expression of the LSU peptide. Accumulation of transcripts for Rubisco LSU in BS cells increased basipetally in parallel with the formation of Kranz anatomy in young leaves. The partitioning of mRNA and LSU peptide to BS cells did not correlate with chloroplast structural differentiation, since rbcL mRNA accumulation was observed in both stages 1 and 2 prior to chloroplast differentiation. The spatial and temporal patterns and co-localizing of Rubisco rbcL transcript and protein to BS cells during chlorenchyma development indicates that rbcL mRNA accumulation (either transcriptional control or stability of transcripts) has a role in targeting to the BS and in the very early development of C4 (before differentiation and positioning of BS chloroplasts).
A pattern similar to this for Rubisco gene expression was described for light-grown Z. mays, in which Rubisco mRNAs are expressed mainly in BS cells from the earliest stages of leaf development before structural differentiation of M and BS cells (Martineau and Taylor, 1985; Langdale et al., 1988,a). A more complex pattern of development was shown for A. hypochondriacus leaves (Wang et al., 1992; Ramsperger et al., 1996). Rubisco rbcL transcripts were present in both M and BS cells until later stages of development (leaves up to 5 and 10 mm in length), while Rubisco peptides accumulated in the youngest regions of 2-mm leaves in a C4 manner (only BS), followed by a C3-like distribution (in BS and M) in 5-mm leaves. At a more advanced developmental stage (leaves 10 mm in length), Rubisco was expressed exclusively in BS cells (Wang et al., 1992; Ramsperger et al., 1996). This two-step developmental pattern suggests post-transcriptional control of Rubisco expression at the earliest stages of amaranth leaf development (Ramsperger et al., 1996), and additional regulation at the level of mRNA accumulation (transcription or stability) in the following stages (Wang et al., 1992; Patel and Berry, 2008).
In both of the Suaeda species, there was an increase in Rubisco LSU mRNA accumulation which temporally preceded the peptide expression, since maximal signal intensity of mRNA was observed beginning from stage 2, while the peptide reached its highest labelling in mature tissue. A similar pattern was observed for Z. mays, where Rubisco LSU transcripts accumulated ahead of the peptides, and the level of Rubisco enzyme increased towards the tip of the leaf, while levels of the corresponding mRNAs peaked near the base of the blade and decreased towards the tip (Langdale et al., 1988,a; Nelson and Dengler, 1992). In contrast, in developing amaranth leaf primordium, Rubisco transcripts and proteins are detected simultaneously from the earliest stages (Ramsperger et al., 1996).
Relationships between development of vascular tissue and C4 traits
In this current study we have observed correlations between vascular tissue development, stages of chlorenchyma differentiation, and expression of C4 enzymes in the two Kranz type Suaeda species. The transitions to the C4 features that appear in stage 2, which include specific organelle positioning in BS cells, significant increase in the level of Rubisco in BS chloroplasts, and initiation of selective expression of C4 enzymes, coincide with phloem sieve element differentiation. Additional research will be required to determine the relationship between vascular tissue differentiation and development of C4 anatomy and biochemistry in leaves of C4Suaeda.
It is interesting to note that during leaf development in amaranth, changes in patterns of C4 photosynthetic gene expression are coordinated with the maturation of the smaller veins and sink-to-source transition; but do not correlate with the morphological development of Kranz anatomy, which occurs much earlier (Wang et al., 1992; Wang et al., 1993; Ramsperger et al., 1996).
Analysis of BS and M cells during development in monocots suggests that the veins have a crucial role in cell-specific expression of photosynthetic enzymes, and that the leaf vascular tissue provides a positional signal that induces C4 photosynthetic differentiation (Langdale et al., 1988b; Nelson and Langdale, 1989; Langdale and Nelson, 1991; Soros and Dengler, 2001). According to this hypothesis, each vein would control the differentiation of at least two cells in radius beyond the vascular tissue (Langdale et al., 1988,b; Nelson and Dengler, 1992).
However, in the C4 monocot A. hirta, the leaves have a special type of BS cell (so-called distinctive cells) that is not associated with vascular bundles. Thus, in these plants the C4 expression pattern does not depend on direct contact of BS cells with veins (Wakayama et al., 2003).
The two forms of Kranz in Suaeda have a single compound Kranz unit surrounding the vascular tissue. In the Schoberia Kranz type (as in S. eltonica), there is contact between most, but not all, of the BS cells and the vascular bundles, whereas in the Salsina type (as in S. taxifolia), there is no contact. Yet, in both cases there is a similar pattern in the development of C4 features. These observations support the idea that direct contact with veins is not essential for cellular differentiation and C4 pattern of gene expression. In this case, there would need to be a different distribution of vein-derived signals, compared with that with Kranz tissue surrounding individual veins. How positional signals could be transmitted and received in the case of a single compound Kranz unit is not clear.
Taken together, our findings indicate that these two types of Kranz anatomy, which are suggested to have evolved independently in subfamily Suaedoideae (Kapralov et al., 2006) with NAD-ME type C4 photosynthesis, have similar ontogenetic programmes for development of C4 anatomy and biochemistry. Additional studies are needed on representatives from different lineages where C4 has evolved to determine the pattern in C4 development and its regulation in relation to structural diversity in Kranz anatomy and evolution.
- δ13C values
a measure of the carbon isotope composition
pyruvate, Pi dikinase
photosynthetic photon flux density
ribulose bisphosphate carboxylase oxygenase
scanning electron microscope/microscopy
transmission electron microscope/microscopy
This material is based upon work supported by the National Science Foundation under Grants IBN-0236959, IBN-0641232, and MCB0544234, NSF Isotope Facility Grant DBI-0116203, partly by the Russian Foundation of Basic Research under Grant 08-04-00936 and a collaborative grant between the Russian Foundation of Basic Research, 10-04-95512 and the Civilian Research and Development Foundation, RUB1-2982-ST-10. We are grateful to the Franceschi Microscopy and Imaging Center of Washington State University for use of their facilities and staff assistance, and to C. Cody for plant growth management.