Modification of the catalytic properties of potato PEPC
Five attempts towards overexpression of ppc genes in plants have been described so far. Three groups expressed the maize PEPC cDNA or gene under control of different promoters (Hudspeth et al., 1992; Kogami et al., 1994; Ku et al., 1999) and increases in enzyme activity up to 100-fold were detected with the complete intact maize gene in rice (Ku et al., 1999). Recently, the overexpression of PEPC and malate dehydrogenase from Sorghum in potato has been described, but in this case expression levels were low and no modification in photosynthetic parameters was observed (Beaujean et al., 2001). To avoid any possible influence of improper post-translational modification, we have previously used a bacterial ppc cDNA under control of a modified 35S promoter in potato (Gehlen et al., 1996). This resulted in a clear increase in PEPC activity (Häusler et al., 2001) and slowed down growth in axenic culture with some minor impact on photosynthetic and metabolic parameters (Gehlen et al., 1996). Opposite effects were observed in plants with antisense stppc constructs and diminished PEPC activities (Häusler et al., 1999). In order to further increase the activity of PEPC in vivo under physiological conditions we introduced selected modifications into the coding sequence of the potato ppc cDNA. The exchange of an internal part of the sequence for the homologous stretch from the C4 plant Flaveria trinervia was initially thought to allow discrimination of the transgene and the endogenous protein by specific antibodies in Western analyses, but surprisingly resulted in a significant shift of the enzymatic properties in vitro. Blasing et al. (2000) applied domain shuffling to analyse regions responsible for C4-specific characteristics of Flaveria PEPC proteins. A region between aa 296 and 437 and a single serine residue at the carboxy-terminus were shown to control these characteristics. The domain exchanged in our study from the Flaveria trinervia to the potato PEPC is part of the identified region. Beside a decrease in the Km we observed an increase of the I50 for malate, a change towards values observed for C4 isoforms. Interestingly, the reciprocal exchange from the C3 enzyme to the C4 enzyme decreased the I50 for malate (Thomas Rademacher, unpublished data) supporting the importance of this region for the enzymatic properties of PEPC proteins. Blasing et al. (2000) used kinetic parameters and enzyme activation by Glc6P as indicative parameters instead. It will be interesting to test whether these parameters are also affected in the engineered potato protein.
Mutations at the N-terminal phosphorylation site of the engineered potato sequence also increased the affinity for PEP and decreased product inhibition of the enzyme. These alterations have previously been shown to reduce the malate sensitivity of C4 PEPC enzymes from Sorghum and maize (see also introduction). We made similar observations with a photosynthetic PEPC from Flaveria trinervia (unpublished data). Furthermore, the importance of this region for the activity of C3 PEPC enzymes was shown by in vitro and in vivo phosphorylation experiments (Duff and Chollet, 1995; Schuller and Werner, 1993; Zhang et al., 1995).
We obtained best effects with the potato enzyme carrying both the N-terminal and internal modifications. Allosteric inhibition of the engineered construct was in the range of values normally observed for photosynthetic isoforms and in parallel the affinity for PEP was further increased beyond the levels normally observed for anaplerotic isoforms. We expected this combination to be optimally suited for overexpression of PEPC with high and stable activities in vivo. This was tested by correlation of expression rates and the percentage of 14CO2 fixed into malate. Overexpression of PEPC clearly increased pulse labelling of malate. When the enzyme with both N-terminal and internal modifications was used, effects were clearly enhanced compared with the proteins carrying only the internal modification. The maximum rate of carbon fixation into malate in the transgenic lines was similar to the levels detected for C3-C4-intermediate species like Flaveria ramosissima where C3 and C4 cycles compete for carbon fixation (Rumpho et al., 1984). Thus, the degree of malate inhibition in vitro and the amount of malate labelled in vivo correlate well suggesting that product inhibition might be one limiting step for the enzymatic activity inside the cell.
However, the net rate of CO2 fixation into malate could additionally be dependent on reaction sequences metabolising malate. We have recently reported that the introduction of either bacterial PEPC or the modified potato PEPC described here induced an endogenous cytosolic NADP malic enzyme and increased the activity of mitochondrial NAD malic enzyme in potato leaves (Häusler et al., 2001). The decarboxylation reaction catalysed by these enzymes might liberate labelled carbon that had already been fixed and by this reduce the gross incorporation. This assumption is also supported by the increase in dark CO2 release observed in the PEPC overexpressing lines.
Our data indicate that the introduced modifications are necessary to induce enhanced in vivo PEPC activity in the light. This is in agreement with additional experiments where neither overexpression of the Corynebacterium glutamicum nor of the intact Flaveria trinervia PEPC in potato induced comparable rates of CO2 fixation into malate (Thomas Rademacher, unpublished data). However, similar observations concerning enhanced labelling of C4 acids have been made by Suzuki et al. (2000) following overexpression of phosphoenolpyruvate carboxykinase (PCK) in rice. The authors assume that these effects are due to an enhanced activity of the endogenous PEPC in response to increased PEP levels produced by the overexpressed PCK.
Leaf metabolism is perturbed by overexpression of PEPC
Striking differences were observed for plant growth and metabolite levels comparing plants overexpressing PEPC with and without modified phosphorylation site. Plants overexpressing the constructs with a modified phosphorylation site were stunted in growth and tuber yield was severely impaired. Growth retardation appeared to be correlated with changes in metabolite content suggesting a redirection in metabolic fluxes. The most striking effect was the increase in malate and total amino acids. Both changes fit very well into the proposed role of PEPC in C3 plants. Moreover, the increase in malate and amino acids was inversely proportional to the decline in starch and soluble sugars in the light. The loss of carbon in starch and soluble sugars of about 34 matom C·m−2 in line stppcS9D-C4 21 was compensated by a gain of 23 matom C·m−2 in malate, 12 matom C·m−2 in Glu/Gln and 1 matom C·m−2 in Asp/Asn. This underlines the redirection of carbon flow from starch and soluble sugars into organic acids. The decrease in starch content and the increases in soluble, osmotically active compounds could also account for the higher leaf water content.
Potato plants are often limited in nitrogen supply during growth. Thus, an increase in the amino acid content in the transgenic lines might be interpreted as an indirect consequence of slower growth rates attenuating the plants from nitrogen limitation. However, nitrate contents in the transgenic lines were considerably increased both in the light and dark ruling out nitrogen limitation at the site of nitrate supply. Contents of Glu and Gln as well as Asp increased in the PEPC overexpressors during illumination. Interestingly, in the dark Gln contents increased considerably from a very low value (2 µmol m-2) in the wild type to 200 µmol m-2 in the line stppcS9D-C4 21, whereas Glu contents remained unchanged (data not shown). Kaiser and Brendle-Behnisch (1995) have shown that nitrate reductase activity is enhanced in leaf discs kept in the dark up to levels normally observed in the light when the pH is changed into the acidic range. In the PEPC overexpressors, such an acidification might be brought about by the increased accumulation of malate and thus nitrate reductase activity might be enhanced compared with the wild-type providing additional reduced nitrogen for amino acid synthesis. It is likely that 2-oxoglutarate (2-OG) deriving from the anaplerotic action of PEPC is utilised as the corresponding carbon source for amino acid biosynthesis in the dark. This idea is supported by the observed decline in 2-OG contents with an increase in PEPC activity. The increase in Gln combined with unaltered Glu contents suggests a limitation at the site of plastidial Fd-GOGAT, which requires reducing equivalents supplied by photosynthetic electron transport in the light.
Besides Glu and Asp, the contents of Thr, Ala, Gly and Val were clearly increased in the PEPC overexpressors compared with the wild-type. Thr is derived from Asp, Gly can be synthesized from Glu, Val from pyruvate, and Ala either from aspartate or by transamination from pyruvate (Goodwin and Mercer, 1983). Increased contents of these amino acids suggest higher fluxes starting from the respective precursors. Similar changes in the levels of individual amino acids were observed in plants, which were shifted from low to high nitrate nutrition. The most pronounced increase was obtained for Asp/Asn, Glu/Gln, Ser, and Gly (Champigny, 1995). Moreover, if plants depleted in nitrate or mutated in genes necessary for nitrogen metabolism are exposed to a nitrate pulse, carbon partitioning is redirected towards organic acids (Duff and Chollet, 1995; Scheible et al., 1997; Stitt et al., 2002). Interestingly, this is accompanied by an increase in PEPC kinase activity protecting PEPC from product inhibition (Champigny, 1995). The mutations introduced into the phosphorylation site mimic exactly this physiological state supporting the idea that a diminished malate sensitivity is the decisive step for the increase of in vivo PEPC activities. The data are consistent with a relief of nitrogen limitation in the transgenic lines caused by the slower growth and deviations in the relative ratios of amino acids due to indirect impacts of PEPC overexpression on the amino acid metabolism.
There were clear changes in the contents of intermediates that undergo high metabolic fluxes, such as phosphorylated metabolites and 2-oxoacids both in the light and the dark. Most strikingly, contents of RubP, 3-PGA, PEP, Glc6P, Fru6P and AMP declined in the light, whereas 2-PGA, ADP, 2-OG and OAA increased and trioseP, Fru1,6P2, Glc1P, ATP and pyruvate remained unaffected. As in the light the majority of 3-PGA is localised in the stroma (Leidreiter et al., 1995; Wirtz et al., 1980), a decrease in 3-PGA would be consistent with a limitation of starch biosynthesis at the site of ADP-glucose phosphorylase (AGPase) because this enzyme strongly responds to the stromal 3-PGA : Pi ratio in that a decrease in this ratio limits the flux of carbon into starch (Preiss, 1982). In fact, starch levels were significantly diminished in the transgenic lines. Moreover, AGPase activity also responds negatively to nitrate levels that were enhanced in transgenic leaves and high nitrate levels correlate with reduced levels of 3-PGA and PEP (Scheible et al., 1997; Stitt et al., 2002) consistent with the situation observed in the PEPC overexpressors.
RubP levels have been shown to respond to the leaf intercellular CO2 concentration (Ci) and the photon flux density (PFD), in that the contents decline with an increase in Ci above ambient and increase with an increase in PFD (Badger et al., 1984). Thus, CO2 assimilation rates can switch from RubP saturating rates to RubP limiting rates and vice versa. However, despite the decline in RubP in stppcS9D-C4 21 by more than 50% of the wild-type, CO2 assimilation was not severely impaired suggesting that RubP content in the transgenic lines was still sufficient for the maintenance of CO2 assimilation rates comparable with the wild-type. Interestingly, 3-PGA, which has been shown to be inversely correlated to RubP over a wide range of Ci-values and PFDs (Badger et al., 1984), declined as well suggesting that the Calvin cycle intermediates are drained from the stroma, most likely by excessive Pi liberation from PEP in the cytosol. Furthermore, changes in the metabolite ratios of the reactions catalysed by phosphoglyceromutase (3-PGA : 2-PGA) and enolase (PEP : 2-PGA) in the transgenic plants showed clear deviations from the wild-type. As chloroplasts usually lack a full glycolytic pathway (because they are deficient in phosphoglyceromutase and enolase), these metabolite ratios should be indicative of processes taking place in the cytosol. Assuming that approximately 20% of the 3-PGA measured in the light is localised in the cytosol (Wirtz et al., 1980), the 3-PGA : 2-PGA ratio in wild-type plants is 4.4 and thus near to the reported Keq of 5.0. In the transgenic lines the distribution of 3-PGA between the stroma and the cytosol is unknown. The overall ratios of 3-PGA : 2-PGA were diminished from 22.0 in the wild-type to 2.2 and 6.6 in the lines stppcS9D-C4 72 and 21, respectively. A similar scenario holds true for the PEP:2-PGA ratio. Again, the ratio of 6.1 in the wild-type is near to the reported constant equilibrium (Keq = 6.7) whereas ratios of 0.88 (line 72) and 1.3 (line 21) suggest that the reaction catalysed by enolase is far away from equilibrium in the transgenic lines. Enolase activity appears to become flux limiting as PEP : 2-PGA ratios declined. However, it is not quite clear why 3-PGA : 2-PGA ratios decline in concert. It is conceivable that 3-PGA in the cytosol falls far below the Km value of phosphoglyceromutase for 3-PGA.
For transgenic rice plants overexpressing the complete maize PEPC gene phosphate limitation of photosynthesis has been proposed (Fukayama et al., 2001; Matsuoka et al., 2001) and was recently supported experimentally (Agarie et al., 2002). From our metabolite data, phosphate limitation of photosynthesis (Sharkey, 1985) appears to be unlikely in the potato system.
In the dark, the contents of 3-PGA, PEP, hexoseP, ATP, ADP, and all three 2-oxoacids decreased in the transformants compared with the wild-type, whereas trioseP and Fru1,6P increased significantly and UDPG and AMP remained unchanged. A possible scenario for the delivery of precursors for PEP carboxylation could be the phosphorolytic degradation of starch that will lead to triose phosphates as end products of the oxidative pentose phosphate pathway. This is consistent with the depletion in Glc6P and concomitant with an increase in relative Glc1P levels and the increase in trioseP, which are the major exchange components linking the pentose phosphate pathway to glycolysis.
Interestingly, sucrose biosynthesis seemed not to be affected severely. The key metabolite UDP-glucose (UDPG) exhibited neither altered absolute levels nor clear changes in the relative abundance compared to the total esterified phosphates. This is in agreement with sucrose levels in wild-type and transgenic lines, which were not severely affected by PEPC overexpression (Figure 5). As sucrose is the main carbohydrate transported in the plant, export of this compound from leaves might be reduced. This would again match the reduced deposition of carbohydrates in tubers.
Taken together, the data indicate significant changes in carbohydrate metabolism and a depletion of phosphorylated intermediates due to the overexpression of PEPC in the transgenic lines.
Growth retardation in the transgenic lines overexpressing functional PEPC is a yet unexplained phenomenon
Internodial growth was severely impaired in the transgenic lines overexpressing a functional PEPC in a gene dose dependent manner. Stunted growth was accompanied by a diminished total leaf area and lowered rates of CO2 assimilation per plant, particularly if it is considered that photosynthetically active leaves were at a larger distance to the light source since plants were smaller. The lower carbon fixation on the whole plant level might be in part responsible for the decreased tuber yield. For transgenic potato plants overexpressing the PEPC from C. glutamicum a moderate growth retardation was only observed in axenic culture (Gehlen et al., 1996) and there was no reported effect on growth parameters in rice plants overexpressing the complete maize PEPC gene under the control of its own promoter (Ku et al., 1999). Moreover, even if the increase in nocturnal CO2 release observed in the potato plants carrying the mutated potato PEPC is considered, this would not result in the growth retardation observed, particularly if it is taken into account that nocturnal carbon loss was diminished in the line stppcS9D-C4 72 by almost complete stomatal closure. Likewise N-limitation as an explanation for growth retardation appears to be unlikely as nitrate contents are even increased in the transgenic lines. However, at this stage there is no information available on the impact an increase in malate could have on cytosolic pH or regulatory properties of enzymes involved in nitrogen metabolism. It also appears likely that the growth retardation observed here is an indirect consequence of an imbalance of growth factors by the constitutive overexpression of functional PEPC in all tissues and cell types. In order to tackle this problem it is intended to overexpress the same PEPC construct under the control of a leaf specific promoter.
Our approaches towards overexpression of a physiologically active PEPC are finally aimed to install a C4-like pathway in C3 plants. In this publication, we were able to show that our engineered enzyme induces massive accumulation of malate inside the cell. Still, this does not lead to a reduction of the CO2 compensation point (Γ) characteristic for C4 carbon fixation (data not shown). We have previously reported that combination of PEPC with a decarboxylating activity inside the chloroplast is capable of modulating photosynthetic parameters in potato (Lipka et al., 1999). Plants expressing additional activities for PEP regeneration are currently under investigation in our laboratory. We are convinced that integration of the novel protein into such a C4-like pathway will avoid the observed site effects of increased PEPC activity in the cytoplasm and lead to the establishment of an effective HCO3–/CO2 pump in C3 plants.