The Plant Journal (2010) 61, 122–133
doi: 10.1111/j.1365-313X.2009.04040.x
C4 acid decarboxylases required for C4 photosynthesis are active in the mid-vein of the C3 species Arabidopsis thaliana, and are important in sugar and amino acid metabolism Naomi J. Brown1, Ben G. Palmer2, Susan Stanley1, Hana Hajaji1, Sophie H. Janacek1, Holly M. Astley1, Kate Parsley1, Kaisa Kajala1, W. Paul Quick2, Sandra Trenkamp3, Alisdair R. Fernie3, Veronica G. Maurino4 and Julian M. Hibberd1,* 1 Department of Plant Sciences, Downing Street, University of Cambridge, Cambridge, CB2 3EA, UK, 2 Department of Animal and Plant Sciences, University of Sheffield, Sheffield, S10 2TN, UK, 3 Max-Planck-Insitu¨t fu¨r Molekulare Pflanzenphysiologie, Am Mu¨hlenberg 1, D-14476 Potsdam-Golm, Germany, and 4 Botanisches Institut, Universita¨t zu Ko¨ln, D-50931, Cologne, Germany Received 2 September 2009; revised 17 September 2009; accepted 28 September 2009; published online 4 November 2009. * For correspondence (fax 44 1223 333953; e-mail
[email protected]).
SUMMARY Cells associated with veins of petioles of C3 tobacco possess high activities of the decarboxylase enzymes required in C4 photosynthesis. It is not clear whether this is the case in other C3 species, nor whether these enzymes provide precursors for specific biosynthetic pathways. Here, we investigate the activity of C4 acid decarboxylases in the mid-vein of Arabidopsis, identify regulatory regions sufficient for this activity, and determine the impact of removing individual isoforms of each protein on mid-vein metabolite profiles. This showed that radiolabelled malate and bicarbonate fed to the xylem stream were incorporated into soluble and insoluble material in the mid-vein of Arabidopsis leaves. Compared with the leaf lamina, mid-veins possessed high activities of NADP-dependent malic enzyme (NADP-ME), NAD-dependent malic enzyme (NAD-ME) and phosphoenolpyruvate carboxykinase (PEPCK). Transcripts derived from both NAD-ME, one PCK and two of the four NADP-ME genes were detectable in these veinal cells. The promoters of each decarboxylase gene were sufficient for expression in mid-veins. Analysis of insertional mutants revealed that cytosolic NADP-ME2 is responsible for 80% of NADP-ME activity in mid-veins. Removing individual decarboxylases affected the abundance of amino acids derived from pyruvate and phosphoenolpyruvate. Reducing cytosolic NADP-ME activity preferentially affected the sugar content, whereas abolishing NAD-ME affected both the amino acid and the glucosamine content of mid-veins. Keywords: NADP-dependent malic enzyme (NADP-ME), NAD-dependent malic enzyme (NAD-ME), phosphoenolpyruvate carboxykinase (PEPCK), amino acid metabolism, C4 photosynthesis, Arabidopsis.
INTRODUCTION In the bundle sheath (BS) of C4 plants C4 acid decarboxylases are critical in generating high concentrations of CO2 around ribulose 1,5-bisphosphate carboxylase oxygenase (RuBisCO) (Hatch, 1987; Brown et al., 2005a). In different lineages of C4 plants, three distinct decarboxylases known as NADP-dependent malic enzyme (NADP-ME), NADdependent malic enzyme (NAD-ME) and phosphoenolpyruvate carboxykinase (PEPCK) have been co-opted into this process. The oxygenase activity of RuBisCO (Bowes et al., 1971) is reduced in C4 plants through the involvement of these decarboxylases in the CO2 concentrating mechanism. Because oxygenation by RuBisCO becomes increasingly important at high temperatures (Jordan and Ogren, 1984), 122
C4 plants dominate tropical and subtropical ecosystems (Sage, 2002), and the fact that C4 species are amongst the most productive on the planet (Brown, 1999) has led to the proposal that placing characteristics of C4 photosynthesis into C3 crops would increase yield (Matsuoka et al., 2001; Leegood, 2002; Hibberd et al., 2008). However, there are many aspects of C4 photosynthesis that are poorly understood. In particular, the functions of the C4 acid decarboxylases in C3 plants remain poorly defined. Improving our understanding of the role of these enzymes in C3 plants is likely to be important to avoid generating futile cycles and unexpected pleiotropic effects as parts of the C4 pathway are placed into C3 crops. ª 2009 The Authors Journal compilation ª 2009 Blackwell Publishing Ltd
C4-like photosynthesis in Arabidopsis veins 123 C3 plants contain both cytosolic and chloroplastic NADPMEs (Edwards and Andreo, 1992; Drincovich et al., 2001). The cytosolic NADP-ME may provide NADPH to the oxidative pentose phosphate pathway and lignin biosynthesis (Walter et al., 1990; Schaaf et al., 1995), and also may control cytosolic pH via malate concentration (Martinoia and Rentsch, 1994; Lai et al., 2002). The chloroplastic isoform of NADPME is proposed to play a role in lipid metabolism (Smith et al., 1992; Eastmond et al., 1997; Gerrard-Wheeler et al., 2005). Additionally, both forms of NADP-ME are implicated in plant defence (Schaaf et al., 1995; Casati et al., 1999; Lai et al., 2002). NAD-ME functions as a heterodimer (Grover and Wedding, 1982; Willeford and Wedding, 1987; Long et al., 1994) and homodimer (Tronconi et al., 2008), and it can regulate the flux of carbon into both the tricarboxlic acid (TCA) cycle and glycolysis (Jenner et al., 2001; Tronconi et al., 2008). The third C4 acid decarboxylase, PEPCK, has a clearly defined function in germinating seedlings of C3 plants, where it allows the conversion of lipids to sugars in gluconeogenesis (Leegood and ap Rees, 1978; Rylott et al., 2003; Penfield et al., 2004; Malone et al., 2007). PEPCK may also control the metabolism of nitrogenous compounds and regulate pH (Walker et al., 2001; Chen et al., 2004; DelgadoAlvarado et al., 2007). In summary, these studies indicate that the products of each decarboxylase are likely to be used for different purposes, although none of the enzymes seem to function in supplying CO2 to photosynthesis, as they do in C4 plants. However, we previously reported that cells associated with veins of tobacco petioles possess high activities of each C4 acid decarboxylase (Hibberd and Quick, 2002). It is possible that the pre-existing ability to generate high activities of C4 acid decarboxylases around veins of C3 plants facilitated the polyphyletic evolution of C4 photosynthesis. In addition, the presence of the decarboxylases around C3 veins implies that malate in the transpiration stream could be decarboxylated, and the CO2 released used in photosynthesis by these cells. Because high activities of pyruvate, orthophosphate dikinase (PPDK), which is able to phosphorylate pyruvate to generate phosphoenolpyruvate (PEP), were also detectable, we proposed that the decarboxylases provide CO2 for photosynthesis, and PEP to the shikimate pathway (Hibberd and Quick, 2002). A significant flux of carbon into the shikimate pathway could be required in veins during periods of high lignin production. In accordance with this proposal, lines of Arabidopsis in which photosynthesis was removed from cells around veins showed reduced levels of shikimate and alterations to the abundance of transcripts that encode proteins involved in pathways that generate PEP (Janacek et al., 2009). Although the above studies are consistent with this C4-like pathway being involved in providing PEP to the shikimate pathway in veins of C3 plants, there is little information on the role of C4 acid decarboxylases within this process. In
order to gain further insight into the role of each C4 acid decarboxylase in veins of C3 plants, we investigated the activity of NADP-ME, NAD-ME and PEPCK in the Arabidopsis mid-veins, assessed which genes are responsible for the accumulation of each protein in mid-veins, and investigated whether regions 5¢ to each gene are sufficient to direct the expression of uidA in veinal cells. In addition, we used insertional mutants to investigate which genes are responsible for the activity of each decarboxylase in the mid-vein, and analysed the impact of removing the activity of each decarboxylase on metabolite profiles in this tissue. RESULTS Veinal cells of Arabidopsis show characteristics of C4 photosynthesis In cells associated with veins of tobacco petioles, carbon for photosynthesis can be supplied by organic acids from the transpiration stream, and high activities of the C4 acid decarboxylases allow the release of CO2 for photosynthesis (Hibberd and Quick, 2002). We first tested in Arabidopsis whether these cells can fix carbon from substrates fed to the transpiration stream, and then determined the activity of each decarboxylase enzyme in mid-veins. 14 C-labelled bicarbonate and malate were fed to the xylem stream via the petioles of rosette leaves. Because malate was specifically radiolabelled on the fourth carbon that is released during decarboxylation, incorporation of the isotope derived from malate represents fixation of CO2 derived from decarboxylation of malate. Radionuclide derived from these substrates was incorporated into insoluble material in the mid-vein and the main veins of leaves (Figure 1a–b). To investigate whether this fixation was light dependent, we fed leaves with radiolabelled malate in the dark, and in these conditions, incorporation was very low (Figure 1c). Quantification of counts present in leaves fed with malate in either light or dark showed that incorporation of 14C was stimulated by 85% in the light. Fixation of radiolabel was also observed in minor veins when 14C-labelled bicarbonate was fed to the xylem stream (Figure 1d). In contrast, the 14C label derived from malate was restricted to the mid-vein and main veins in the basal regions of leaves (Figure 1e). Quantitative analysis of incorporation around the mid-vein of 14C indicated that most malate was fixed at the base of the leaf (Figure S1), supporting the suggestion that the photosynthetic vascular cells scrub the xylem of malate (Hibberd and Quick, 2002). To investigate the fate of 14C derived from both malate and CO2 supplied to the xylem stream, we used thin-layer chromatography followed by scintillation counting to separate and quantify radiolabel in soluble compounds. Significant proportions of radiolabel derived from 14CO2 were found in soluble sugars, particularly hexoses, as well as amino acids such as glutamate and proline or hydroxyproline (Table 1). When 14C-malate was fed to leaves, a significant proportion
ª 2009 The Authors Journal compilation ª 2009 Blackwell Publishing Ltd, The Plant Journal, (2010), 61, 122–133
124 Naomi J. Brown et al.
(a)
(b)
(c)
(d)
(e)
Figure 1. 14C supplied to the transpiration stream is incorporated in a lightdependent manner into insoluble material in veinal cells of Arabidopsis. Labelled compounds were fed to the leaves of Col-0 plants in the light (a, b, d & e) or dark (c), fixed, and soluble material removed prior to imaging. Leaves were then imaged for 90 min. The substrates that were fed were bicarbonate (a & d) and malate specifically labelled on the fourth carbon (b, c & e). a–c show images of whole leaves, whereas d & e show close-up images to show that minor veins are able to fix 14CO2, but that they fix little radiolabelled malate supplied from the transpiration stream. Incorporation of 14C in the dark was very low, and so the edge of the leaf is marked with a dashed line.
of it remained as malate, but radiolabel was also detected in glutamate, glutamine, aspartate and asparagine (Table 1). These data indicate that radiolabelled malate is incorporated into insoluble material such as starch and cell walls in a lightdependent manner, and also that it is used for amino acid Table 1 Incorporation of radiolabel derived from 14C malate and HCO3 supplied to leaves of Arabidopsis thaliana. Malate was labelled on the fourth carbon, and both malate and HCO3 were supplied to leaves via the transpiration stream. Total incorporation into insoluble and soluble material in the leaf was determined, and the soluble metabolites were separated by thin-layer chromatography. The proportion of radiolabel in major spots was then quantified by scintillation counting. Data are shown as means and standard errors, and are depicted as the percentages of total radionuclide detected in each soluble metabolite Percentage of incorporation 14
14
40.0 2 60.0 2 2.1 1.5 11.1 3.2 28.8 8.8 3.7 1.0 5.8 0.4 6.6 2.3 5.1 0.2
24.0 4 76.0 4 26.3 7.3 1.7 0.4 3.4 0.6 2.3 0.9 5.6 1.8 6.4 0.4 6.0 0.4
CO2
Insoluble Soluble Malate Fructose Glucose Sucrose Alanine Asparagine Aspartate
C-malate
biosynthesis. The flux into glutamine may indicate that carbon supplied from malate is preferentially used for loading into the phloem. When radiolabelled malate or bicarbonate were fed to leaves of insertional mutants for each decarboxylase, we detected little alteration in the pattern or extent of incorporation of isotope (data not shown), indicating that the individual gene products have little impact on the ability to fix carbon from the xylem stream. We then assessed the maximum catalytic activities of NADP-ME, NAD-ME and PEPCK in isolated mid-veins (Figure 2a), and compared this with their activities in whole leaves. Enzyme-coupled assays revealed that NADP-ME activity on a chlorophyll basis is sixfold higher in veins, relative to whole leaf lamina, whereas NAD-ME and PEPCK are enriched by 10-fold (Table 2). Because veinal cells tend to possess fewer chloroplasts than mesophyll cells, we assessed the extent to which the increase in activity of each decarboxylase resulted from the inherently lower chlorophyll content of these cells. This showed that only part of the increase in the activity of the C4 acid decarboxylases is associated with lower quantities of chlorophyll (Table 2). In order to gain insight into the cellular location of each of the enzymes, we made longitudinal sections of mid-veins and viewed them with light microscopy (Figure 2b). This showed that mid-veins consisted of xylem vessels surrounded by xylem parenchyma, as well as phloem companion cells, sieve elements and phloem parenchyma, but that BS cells (Kinsman & Pyke, 1998) were not present. Transcripts derived from specific genes encoding the C4 acid decarboxylases are detectable in veinal cells, and their products are targeted to different subcellular compartments Transcripts derived from two (NADP-ME2 and NADP-ME4) of the four NADP-ME genes, both NAD-ME genes and the PCK1 gene were detectable in mid-veins (Figure 2c). The putative targeting peptide of NADP-ME4 was fused to the gene encoding GFP, and transient expression in tobacco protoplasts marked the chloroplasts. However, the fulllength NADP-ME4 protein has not been investigated, nor have the subcellular localization of the other decarboxylases in Arabidopsis. We therefore cloned full-length spliced coding regions from each gene, and generated translational fusions to the GFP gene. Each fusion was placed under the control of the CaMV35S promoter and Nos terminator, and Agrobacterium tumefaciens was used to transfer these constructs into Arabidopsis. Analysis of GFP accumulation within cells of primary transformants showed that the fulllength NADP-ME4 was targeted to chloroplasts, NADP-ME2 and PEPCK1 were both cytosolic, whereas NAD-ME1 and NAD-ME2 localized to mitochondria (Figures 2d and 3r). Therefore, in mid-veins of Arabidopsis, the activities of the C4 acid decarboxylase were likely due to both chloroplastic and cytosolic NADP-MEs, the mitochondrial NAD-ME and cytosolic PEPCK.
ª 2009 The Authors Journal compilation ª 2009 Blackwell Publishing Ltd, The Plant Journal, (2010), 61, 122–133
C4-like photosynthesis in Arabidopsis veins 125
(a)
(b)
XP XV P
(c) NADP-ME NAD-ME PCK MW 1 2 3 4 1 2 1 2 DECARBOXYLASE ACTIN8
GFP (d)
Chlorophyll Overlay (f) (e) AtNADP-ME2
(g)
(h)
(i) AtNADP-ME4
(j)
(k)
(l) AtNAD-ME1
(m)
(n)
(o) AtNAD-ME2
(p)
(q)
(r) AtPCK1
Figure 2. Transcripts derived from specific genes encoding the C4 acid decarboxylases accumulate in veinal cells and their gene products are targeted to separate subcellular compartments. Bundle of isolated mid-veins (a) and toluidine blue-stained longitudinalsection of an isolated mid-vein (b). Semi-quantitative RT-PCR was conducted on cDNA derived from veins. Specific primers were used for each gene family member, and PCR was performed as a duplex reaction with primers for ACTIN8. Transcripts from two of the NADP-ME, both of the NAD-ME, and one of the PCK genes are detectable in veins (c). Subcellular localization of decarboxylase::GFP fusion proteins (d–r). For each of the decarboxylase genes expressed in veins, a C-terminal fusion construct to the GFP gene under the CaMV35S promoter was generated, and stable Arabidopsis plants expressing these constructs were produced. Leaves from primary transformants were imaged using confocal microscopy. (d–f) NADP-ME2-GFP in guard cells; (g–i) NADP-ME4-GFP in mesophyll; (j–l) NAD-ME1-GFP in mesophyll; (m–o) NAD-ME2-GFP in epidermal pavement; (p–r) PEPCK1-GFP in guard cells. (d, g, j, m, p) GFP channel; (e, h, k, n and q) chlorophyll channel; (f, i, l, o, r) overlay image. Scale bars: (a) 1 cm; (b) 50 lm; (d–r) 5 lm.
Regions 5¢ to each gene are sufficient for mid-vein expression, and 3¢ regions increase the amount of expression of NADP-ME genes We wished to determine the extent to which accumulation of the decarboxylases in veinal cells is specifically mediated by
sequences upstream of each gene. It is known that upstream regions of the NADP-ME2&4 and NAD-ME1&2, combined with the first two exons and the first intron of each gene, are able to direct the accumulation of GUS in veins of Arabidopsis petioles (Gerrard-Wheeler et al., 2005; Tronconi et al., 2008). However, for a number of genes that have been recruited into the C4 pathway, promoter regions alone are sufficient to generate BS-specific expression (Marshall et al., 1997; Kausch et al., 2001; Gowik et al., 2004). To investigate whether this type of regulation is associated with genes encoding C4 acid decarboxylases in the C3 plant Arabidopsis, we generated a series of promoter fusions to the uidA gene encoding the reporter b-glucoronidase (Figure 3a–j). The 5¢ region was defined by the next upstream gene, and included the 5¢ untranslated region up to and including the translational start codon (Figure 3a–e). All promoters were found to direct the expression of uidA in the mid-vein, and to varying extents in minor veins, trichomes, hydathodes and the mesophyll (Figure 3k–o). For the NADP-ME2&4 and the NAD-ME1 fusions, the strongest accumulation of GUS was in petioles of mature leaves, whereas for NAD-ME2 and PCK1 fusions the highest GUS activity was detected in senescing cotyledons. The BS-specific expression of an NADP-ME gene recruited into C4 photosynthesis in Flaveria bidentis is significantly enhanced by regions 3¢ to the gene. (Marshall et al., 1997; Ali and Taylor, 2001). To investigate whether this is the case for genes encoding C4 acid decarboxylases in Arabidopsis, we generated constructs containing each gene-specific downstream 3¢ region in addition to the promoter (Figure 3f–j). GUS staining of stable transformants containing these constructs indicated that 3¢ regions for the Arabidopsis NADP-ME2&4 genes appeared to increase the amount of expression, but that for NAD-ME1&2 and PCK1, the 3¢ regions had little effect. Quantitative methylumbelliferyl glucuronide (MUG) assays were therefore carried out on seedlings containing the uidA gene under the control of the NADP-ME2&4 promoters alone, or each promoter and the relevant 3¢ region. The 3¢ regions of NADP-ME2&4 increased the accumulation of GUS by three- and fourfold, respectively (Figure 3p–q). The NADP-ME2 gene is responsible for most NADP-ME activity in veins, whereas both NAD-ME genes are required for high NAD-ME activity Enzyme-coupled assays were carried out for each C4 acid decarboxylase on mid-vein tissue isolated from previously characterized insertional mutants for each gene (Penfield et al., 2004; Gerrard-Wheeler et al., 2005; Tronconi et al., 2008). This showed that NADP-ME activity in mid-veins of the nadp-me2 and nadp-me4 insertional mutants were 20 and 80% of that in the wild type, respectively (Figure 4a). A very low residual NAD-ME activity was detected in mid-veins from nad-me1.1 (Figure 4b), indicating that
ª 2009 The Authors Journal compilation ª 2009 Blackwell Publishing Ltd, The Plant Journal, (2010), 61, 122–133
126 Naomi J. Brown et al. Table 2 Mid-veins of Arabidopsis thaliana have high activities of the C4 acid decarboxylases that are recruited into C4 photosynthesis. Mid-vein material was removed from petioles of Arabidopsis and enzyme-coupled assays were used to determine the maximum catalytic activity of each enzyme. Because veinal cells tend to possess fewer chloroplasts than mesophyll cells, we also calculated the degree to which the amount of chlorophyll in mid-veins was lower than that in whole leaf material. This was then used to correct the maximum catalytic activities for the inherently lower chlorophyll content of mid-veins, compared with leaves. This showed that only part of the increase in the activity of the C4 acid decarboxylases is associated with lower levels of chlorophyll. Data are shown as means one standard error of the mean
Leaf tissue Vein tissue Relative enrichment (vein/leaf) Relative enrichment corrected for lower chlorophyll in veins (vein/leaf)
NADP-ME activity (lmol min)1 mg chl)1)
NAD-ME activity (lmol min)1 mg chl)1)
PEPCK activity (lmol min)1 mg chl)1)
1.09 0.07 5.99 0.94 6.0-fold 4.0-fold
0.28 0.06 2.73 0.18 10.0-fold 6.6-fold
0.05 0.01 0.48 0.04 10.0-fold 7.6-fold
enrichment, respectively. This implies that the chloroplastic NADP-ME4 protein is significantly enriched in veins compared with the cytosolic NADP-ME2. Enzyme-coupled assays for each decarboxylase in the individual insertional mutants did not show any compensatory increases in the activity of the other decarboxylases (data not shown).
NAD-ME2 can form active homodimers in the mid-vein, at least when NAD-ME1 (the alpha subunit) is not present. In contrast, nad-me2.1 had essentially no activity in the midvein (Figure 4b). Very low activities of PEPCK were detected in veins of the pck1-2 insertional mutant, confirming that PEPCK1 alone is likely to be responsible for the PEPCK activity in veins (Figure 4c). We also carried out enzyme-coupled assays with whole leaf material (Table S1). For NADP-ME, this allowed us to calculate the enrichment in veins of each isoenzyme relative to whole leaf tissue. In wild-type plants we measured a sixfold enrichment (vein/leaf), whereas in the nadp-me2&4 insertional mutants we recorded a nine- and fourfold
ATG
pNADP–ME4 –1700
(c)
ATG
pNAD–ME1 ATG
pNAD–ME2 ATG
pPCK1 –2500
(f)
uidA TGA
TGA
uidA
–2000
(e)
TGA
uidA
–2000
(d)
TGA
uidA
+277
NosT
AtNADP-ME2 –850
(b)
ATG
pNADP–ME2
TGA
uidA ATG
pNADP–ME2
+277
pNADP–ME4
+277
NosT +277
NosT +775
3'NADP–ME2
–1700
(h)
–2000
(i)
ATG
+900
3'NADP–ME4
TGA
uidA
pNAD–ME2
+700
3'NAD–ME1
TGA
uidA
+700
3'NAD–ME2
10
AtNADP-ME2
(p)
8 6 4 2 0
–2000
(j)
ATG
pNAD–ME1
TGA
uidA
pPCK1
ATG
TGA
uidA
(o)
(n)
+277
GUS activity (nmol MU min–1 mg protein–1)
(g)
ATG
AtPCK1
AtNAD-ME2
(m)
NosT
12 –850
AtNAD-ME1
AtNADP-ME4
(l)
(k)
NosT
TGA
uidA
It was suggested that high activities of C4 acid decarboxylases in mid-veins of tobacco are important in supplying pyruvate and PEP to the shikimate pathway (Hibberd and
+1000
5
GUS activity (nmol MU min–1 mg protein–1)
–2500
(a)
Removing the activity of individual decarboxylases affects amino acids derived from pyruvate and phosphoenolpyruvate
AtNADP-ME4
(q)
4 3 2 1 0
Tnos
NADP-ME2 3’
Tnos
NADP-ME4 3'
3'PCK1
Figure 3. Promoter regions are sufficient for the expression of C4 acid decarboxylase genes in mid-veins, and 3¢ regions to NADP-ME2&4 increase expression. The promoter for each of the C4 acid decarboxylase genes expressed in mid-veins was fused to the uidA gene encoding b-glucoronidase. Each promoter was defined by the position of the next upstream gene if this was
