The control of source to sink carbon flux during tuber development in potato

Authors


*For correspondence (fax +1865 275074;
e-mail steven.hill@plants.ox.ac.uk).

Summary

We have used top-down metabolic control analysis to investigate the control of carbon flux through potato (Solanum tuberosum) plants during tuberisation. The metabolism of the potato plant was divided into two blocks of reactions (the source and sink blocks) that communicate through the leaf apoplastic sucrose pool. Flux was measured as the transfer of 14C from CO2 to the tuber. Flux and apoplastic sucrose concentration were varied either by changing the light intensity or using transgenic manipulations that specifically affect the source or sink blocks, and elasticity coefficients were measured. We have provided evidence in support of our assumption that apoplastic sucrose is the only communicating metabolite between the source and sink blocks. The elasticity coefficients were used to calculate the flux control coefficients of the source and sink blocks, which were 0.8 and 0.2, respectively. This work suggests that the best strategy for the manipulation of tuber yield in potato will involve increases in photosynthetic capacity, rather than sink metabolism.

Introduction

Our aim with this work was to provide quantitative information concerning the control of carbon fluxes through an intact plant. Specifically, we applied the methods of top-down control analysis to the problem of carbon partitioning at the whole-plant level.

Carbon flow through an intact plant starts with the fixation of carbon dioxide in the photosynthetic portions of the plant (the source). Carbon, usually in the form of a non-reducing sugar, often sucrose, is then translocated to the growing regions of the plant (the sinks), where it is utilised for the biosynthesis of cellular constituents, the synthesis of storage compounds, or respired to provide energy (Farrar 1992). Discussion of the control of this process has largely concentrated on whether the rate of carbon import into growing sinks is source- or sink-limited (Farrar 1996). This approach has led to a range of conclusions, with some authors providing evidence of source-limitation, others sink-limitation and some suggesting that whether carbon metabolism is source- or sink-limited depends on the growth conditions of the plant (see Wardlaw 1990 for review).

A second approach to this problem has centred on determination of the parameter ‘sink strength’. This parameter is ill-defined (Farrar 1996) and has been the subject of much debate (Farrar 1993 and accompanying papers). In essence, it is taken to be the property of a sink that determines the rate of carbon import into that sink. In some cases, specific enzyme activities have been identified that correlate with sink activity (Jenner & Hawker 1993;Sung et al. 1989), while other authors have suggested that properties such as the rate of synthesis of specific plant growth regulators are measures of sink strength (Kuiper 1993). Several workers have argued that the concept of ‘sink-strength’ is not useful (Farrar 1996;Marcelis 1996), and the suggestion that single enzyme activities can be responsible for sink strength has been particularly criticised because of the shared nature of metabolic control (Stitt 1993).

In recent years, the manipulation of specific enzyme activities in transgenic plants has been used to probe the control and regulation of assimilate flow (Herbers & Sonnewald 1996). The major conclusions from this work are that carbon fluxes are tightly regulated, that this regulation can occur on a range of time scales, and that regulation varies under differing environmental conditions (Stitt & Sonnewald 1995). For example, reduction of the rate of sucrose export from leaves by expression of a yeast-derived invertase in the leaf apoplast of tobacco, potato and Arabidopsis results in a major alteration in the rate of root growth, and a down-regulation of photosynthetic capacity (Heineke et al. 1992;Sonnewald et al. 1991;von Schaewen et al. 1990). The latter is achieved via changes in gene expression that are brought about by accumulation of sugars within the leaves (Krapp et al. 1993). Investigation of potato tuber growth (Engels & Marschner 1986a) has led to the suggestion that the flux of carbon is primarily determined by the ability of the sink to metabolise carbon. In the long term, source activity is thought to be co-ordinated with sink demand as a result of the effects of sucrose on expression of photosynthetic genes.

While this conclusion may hold for the developmental time scale, there is evidence to suggest that determination of export rates by the sink is not the case in the short term. For example, when the rate of sucrose synthesis in illuminated potato leaves is decreased through the reduction of cytosolic fructose-1,6-bisphosphatase activity, carbon flux to the tubers over the diurnal cycle is unaffected because of increased leaf starch breakdown and sucrose export during the night (Zrenner et al. 1996). In this instance, carbon partitioning within the source leaves is altered to compensate for the manipulation. Although these transgenic plants have similar sinks to the wild type, the carbon flux from source to sink is very different, suggesting that the short-term control and regulation of carbon flux may be distinct from that on the long-term, developmental time scale.

Despite the progress made in understanding regulation of carbon fluxes in the long term, we still have little understanding of the mechanisms that regulate carbon fluxes in the short term. In addition to the work on potato described above, the occurrence of regulation on the latter time scale has also been demonstrated for barley, using 11CO2 labelling. Alteration of source metabolism (by changing light intensity) or sink metabolism (by changing root temperature) leads to changes in carbon flux and partitioning within minutes (Farrar & Minchin 1991;Minchin et al. 1994). These short-term effects appear to be related to changes in the rate of sucrose supply and demand (Farrar et al. 1995), but the role of source and sink reactions in determining carbon flux in the short term remains unknown for any system. While it is possible to manipulate crop yield through an understanding of long-term regulation of partitioning, understanding short-term regulation will be important for prediction of the response of allocation to variation in environmental conditions, such as shading and water stress. As a first step to understanding regulation of resource allocation in the short term we are interested in identifying the distribution of control of carbon fluxes through the plant.

We decided to use top-down metabolic control analysis to study the control of carbon flux from source to sink in potato plants during tuberisation. Top-down control analysis was first described by Brand and co-workers (Brown et al. 1990;Hafner et al. 1990), and allows the measurement of the contribution of large groups of reactions to the control of flux. A block control coefficient can be measured for each group of reactions, which is equal to the sum of the individual control coefficients within the group (Brown et al. 1990). We have split the reactions involved (fixation of CO2, synthesis of sucrose, transport to the sinks and subsequent metabolism) into two blocks, to allow us to provide a quantitative answer to the question ‘Is carbon flux source- or sink-limited?’. These blocks were taken to communicate only via the leaf apoplastic sucrose pool and we have provided evidence in support of this assumption.

Potato was chosen because of the availability of suitable transgenic lines, the presence of a single major sink (the tubers) during tuberisation, and its commercial importance. We have used transgenic plants with reduced activities of chloroplast fructose-1,6-bisphosphatase (cpFBP) or the phloem-associated sucrose transporter (αSUT), and variation in light intensity to change the leaf apoplastic sucrose concentration. By measuring the effect that this has on the flux of carbon, as estimated from the transfer of 14C from CO2 to the tuber, we can then determine the effect of varying the apoplastic sucrose concentration on the flux through the source and sink blocks. This allows the experimental determination of elasticity coefficients, which can then be used to calculate flux control coefficients for each block.

Results and discussion

The experimental system

Top-down control analysis requires that the system under consideration be divided into a small number of blocks of reactions, which communicate through a small number of metabolic intermediates. We have divided the metabolism of the potato plant into two blocks, as illustrated schematically in Fig. 1. The first block, termed the source block, consists of the all the reactions of leaf metabolism which lead to the fixation of CO2 and the delivery of sucrose into the leaf apoplast. The second block, termed the sink block, consists of all the reactions in the plant responsible for the subsequent metabolism of the apoplastic sucrose. This will include loading of sucrose into the phloem, phloem translocation and unloading, and the metabolism of sucrose in the various sinks of the plant. During tuberisation the major sink for carbon in the potato plant is the developing tuber. We chose to use apoplastic sucrose as the communicating metabolite because there is strong evidence that phloem loading in potato proceeds via an apoplastic route (Gamelai 1989;Heineke et al. 1992;Riesmeier et al. 1993;Riesmeier et al. 1994), so that all the carbon that is exported from the leaf passes through this pool.

Figure 1.

Schematic representation of the system under consideration.

Leaf apoplastic sucrose is produced by the ‘source block’ which consists of all the steps from atmospheric CO2 to leaf apoplastic sucrose, and consumed by the ‘sink block’ which consists of all the steps after leaf apoplastic sucrose, including uptake and metabolism of sucrose in the tuber. The system flux is J. The flux through the source block is JSo and through the sink block is JSi.

In order to apply top-down metabolic control analysis we need to be able to measure the flux of carbon through the system and apply block-specific perturbations that allow the determination of elasticity coefficients with respect to the communicating metabolite (Brown et al. 1990). These elasticity coefficients can then be used for the calculation of flux control coefficients using the summation and connectivity theorems of metabolic control analysis (Brown et al. 1990). We have measured the flux of carbon through the plant using a 14CO2 labelling approach. In order to perturb the system we have used three methods. To determine the effect of perturbation of source metabolism, flux was measured, either at a range of light intensities in wild-type plants, or in transgenic potato plants that have reduced activity of chloroplast fructose-1,6-bisphosphatase (cpFBP) as a result of the expression of the cDNA encoding this enzyme in the antisense orientation (Kossmann et al. 1994). The effect of perturbation of sink metabolism was investigated in transgenic plants that have reduced activity of the transporter (αSUT) that is responsible for loading of sucrose into the phloem as a result of expression of the cDNA encoding this protein in the antisense orientation (Riesmeier et al. 1994). In the case of the transgenic manipulations, we have used lines with minimal pleiotropic effects (cpFBP-12 (Kossmann et al. 1994); cpFBP-16 (J. Kossmann, unpublished data);αSUT-13 and αSUT-17 (Riesmeier et al. 1994)). Although these transgenic manipulations involve long-term changes in the source or sink blocks, they are suitable for measuring the short-term distribution of control, provided that there are no secondary changes in the block distinct from the manipulation (see below).

Measurement of flux

In order to determine the rate of carbon movement through the plant we supplied 14CO2 to the leaves and measured the rate of appearance of 14C in the tuber. We supplied 14CO2 to the plants for a 2-h pulse and measured the flux in a subsequent 4-h chase. The rate of carbon entry into the tuber was linear for the latter period and there was no detectable label in the tuber tissue at the beginning of the chase (Fig. 2). In determining the rate of incorporation of 14C into the tuber we have neglected 14CO2 release because preliminary work showed that this was below 5% of the total label incorporated for all lines (data not shown). We also ignored incorporation into other sinks (e.g. roots, growing point of shoot) as, at the stage of development investigated, incorporation was very low (data not shown). The distribution of label within the tuber (data not shown) was characteristic of normal tuber metabolism (e.g. Merlo et al. 1993), suggesting that the experimental treatment did not significantly alter the behaviour of the plant. In order to convert the incorporation of 14C into metabolic flux, we measured the specific activity of the stolon sucrose pool at the end of the experiment. All of the sucrose that enters the tuber will pass through this pool, so that its specific activity can be used to convert 14C incorporated per unit time into the rate of sucrose metabolism. We have assumed that the stolon sucrose pool consists of entirely phloem-derived sucrose, and that any non-phloem sucrose pools at the start of the experiment are small compared to the phloem sucrose pool. We would argue that this is the case because the phloem sucrose concentration is high compared to that found in cells (Lohaus et al. 1995), and the phloem system accounts for a significant proportion of the stolon tissue (Engels & Marschner 1986b). The rate of sucrose entry into tuber metabolism in the wild-type plants was in the order of 7 μmol hexose equivalents min–1 plant–1. Assuming a growing season for the tubers of 3 weeks, this flux corresponds to a yield of around 30 g of carbohydrate per plant, which is of the correct order of magnitude for the conditions under which our plants were grown. To allow comparison of fluxes from different plants, the flux can be expressed relative to the total fresh weight of leaves, total fresh weight of tuber or total fresh weight of the plant. Whichever basis was used, the relative values of the flux between individuals and treatments were constant (data not shown). Since metabolic control analysis deals with relative fluxes (Fell 1997), the precise basis on which the flux is expressed is not important. All of the subsequent flux data have been expressed relative to the total fresh weight of the leaves.

Figure 2.

Linear rate of accumulation of 14C in the tubers of plants whose leaves were incubated with 14CO2.

Leaves of wild-type plants were incubated with 14CO2 by generating 14CO2 from NaH[14C]CO3 inside a clear polythene bag placed over the plant and sealed around the base of the stem. After 2 h the bag was removed and the experiment timed from this point. At the end of the experiment the tubers were removed, extracted in boiling ethanol and the amount of 14C in the extract determined. The specific activity of sucrose in the stolon was used to calculate the amount of hexose equivalents imported into the tuber. Values are the mean ± SEM of three plants.

Measurement of leaf apoplastic sucrose concentration

In order to measure the leaf apoplastic sucrose concentration, sap was extracted from the leaves by gentle centrifugation after vacuum infiltration. The dilution of the apoplastic sap was corrected by measuring apoplastic air and water volumes (Table 1).The values for wild-type leaves are similar to those reported previously (Leidreiter et al. 1995). The air and water volumes were not significantly different from the wild-type for the cpFBP lines, but there was a significant (P < 0.05) reduction in the apoplastic air volume in the αSUT antisense lines (Table 1). The reason for this difference is not clear, but the effect was taken into account in the calculation of apoplastic sucrose concentrations. In order to check for contamination of the apoplastic sap by cytosolic components, we measured the activity of malate dehydrogenase in the apoplastic sap. Only 0.3% of the total leaf malate dehydrogenase activity was detected in the apoplastic sap, suggesting that contamination from the cytosol is negligible. Furthermore, the sucrose:hexose ratio in the apoplastic sap was 3:1, very different from the ratio found in either the cytosol or vacuole (Leidreiter et al. 1995). There was no significant contamination of the apoplastic sap with either invertase or sucrose synthase activity, and the sucrose concentration in the sap remained constant for up to 4 h at room temperature (data not shown). Based on the evidence outlined, we would suggest that our measurement of apoplastic sucrose concentration are good estimates of the concentration in vivo.

Table 1.  Measurement of sucrose concentration in the leaf apoplast of cpFBP and αSUT transgenic plants
Line
Apoplastic air volume
(μl g–1 fresh weight) [Vair][Vwater]
Apoplastic water
volume (μl g–1 fresh weight)
(Vair + Vwater)/Vwater
Concentration of sucrose
in leaf apoplast (mm)
  1. Apoplastic air volume was estimated from the uptake of 100 mm KCl during vacuum infiltration of fully hydrated leaves. Apoplastic water volume was estimated by measuring the uptake of [14C] polyethylene glycol 4000 into leaf discs. Values for apoplastic air volume are the mean ± SEM of six leaves. Values for apoplastic water volume are the mean ± SEM of 12 discs, two from each of six leaves. Leaf apoplastic sap was extracted by vacuum infiltration of leaves with 100 mm KCl followed by gentle centrifugation and the sucrose concentration in the extracted sap was determined. Dilution of the apoplast during infiltration was corrected for using the ratio of the apoplastic air and water volumes.

Wild-type319 ± 1478 ± 85.09 2.13 ± 0.29
αSUT-13123 ± 1473 ± 42.6817.27 ± 9.97
αSUT-17259 ± 1367 ± 64.8634.52 ± 6.43
cpFBP-12339 ± 2286 ± 34.94 1.15 ± 0.10
cpFBP-16328 ± 1293 ± 84.53 1.25 ± 0.12

The apoplastic sucrose concentration measured assumes that sucrose is distributed evenly through the apoplast, but there is evidence to suggest that this assumption is not correct (Van Bel 1993). The application of metabolic control analysis only requires the measurement of relative changes in sucrose concentration, so that even if there is an uneven distribution of sucrose within the leaf apoplast our measurements are still sufficient, provided that this distribution does not change dramatically in the transgenic lines that we have used. If the distribution of sucrose within the apoplast changes in the transgenic lines relative to the wild type, then the accuracy of the calculated flux control coefficients may be affected (see below).

The apoplastic sucrose concentrations in the wild-type, cpFBP and αSUT lines are shown in Table 1. In wild-type plants the leaf apoplastic sucrose concentration was between 1 and 2 mm (Table 1), which is in agreement with previously reported values for potato (Leidreiter et al. 1995). There is a modest reduction in the apoplastic sucrose concentration in the cpFBP lines, while in contrast there is a dramatic increase in the apoplastic sucrose concentration in the αSUT antisense lines.

Characterisation of the transgenic lines

In order to apply top-down control analysis it is essential that the manipulations used are specific to one block (Brown et al. 1990). We therefore checked for changes in enzymes involved in tuber metabolism in the cpFBP antisense lines and enzymes involved in leaf metabolism in the αSUT antisense lines. All enzyme assayed were fully optimised, and enzymes were extracted using methods that have previously been demonstrated to give a good recovery of activity (Sweetlove et al. 1996a). In the cpFBP antisense lines we measured a range of enzymes in tuber extracts and found no significant differences between the transgenics and the wild-type (Table 2;P > 0.05 for all enzymes). We were unable to detect plastid FBPase in tuber extracts, supporting the data of Entwistle & ap Rees (1990), and the activities that we measured for other enzymes were in keeping with published values for potato tubers (Merlo et al. 1993). For line αSUT-13, Riesmeier et al. (1994) reported no significant changes in the activities of ADPglucose pyrophosphorylase, chloroplast FBPase and sucrose phosphate synthase in the leaves and, since αSUT-17 has a smaller reduction in sucrose transporter activity (Riesmeier et al. 1994), any changes in these enzymes are unlikely for this line. Riesmeier et al. (1994) did report a change in the steady state content of the mRNA encoding sucrose synthase for αSUT-13, so we measured the activity of this enzyme in leaf extracts. We found activities of 49 ± 5, 35 ± 7 and 50 ± 14 nmol min–1 gFW–1 (mean ± SEM, n =6) for the wild-type, αSUT-17 and αSUT-13, respectively. There was no significant difference between the transgenics and the wild-type (P > 0.05). Based on the evidence described we suggest that the transgenic manipulations that we have used are specific to a particular block in the top-down analysis.

Table 2.  Maximum catalytic activities of enzymes of carbohydrate metabolism in tubers from transgenic plants with decreased activity of chloroplastic fructose-1,6-bisphosphatase (cpFBP)
Enzyme activity (nmol min–1 per g fresh weight)
  Wild-type cpFBP-16cpFBP-12
  1. Developing tubers were harvested from 8-week-old plants. Within 30 min of their removal from the plant, the tubers were thinly sliced, frozen in liquid nitrogen and stored at –80°C as a frozen powder until required. Samples of the frozen powders were extracted, clarified by centrifugation, desalted and then assayed for enzyme activities. Each value is the mean ± SEM of six tubers. N.D. = not detected.

Fructose-1,6-bisphosphatase (plastidic)N.D.N.D.N.D.
Sucrose synthase 1070 ± 165 1129 ± 50 1200 ± 155
Fructokinase 359 ± 33 369 ± 16 416 ± 20
Phosphoglucomutase 6713 ± 230 7315 ± 447 7004 ± 647
UDPglucose pyrophosphorylase14878 ± 57314824 ± 29313126 ± 336
ADPglucose pyrophosphorylase 372 ± 41 337 ± 29 295 ± 46

Sucrose is the only significant communicating metabolite between the source and sink blocks

In order to apply top-down control analysis to the system that we have described, it is essential that apoplastic sucrose is the only communicating metabolite between the blocks of reactions (Brown et al. 1990). To test whether this is the case we have used the method proposed by Brown et al. (1990): if two manipulations are applied to the same block then, if sucrose is the only communicating metabolite, the relationship between flux and apoplastic sucrose concentration should be the same for both manipulations. If there are other communicating metabolites, then different relationships will be observed with each manipulation. Therefore, we compared the relationship between flux and apoplastic sucrose concentration in response to varying cpFBP activity (transgenic lines) and light intensity (wild-type plants). Both of these manipulations are in the source block. The results are shown in Fig. 3. In both cases there was a significant (P < 0.05) correlation between flux and apoplastic sucrose concentration. Based upon an analysis of covariance, the lines fitted to the cpFBP-reduction data and the light-intensity data are not significantly different from one another (P > 0.05), providing evidence that sucrose is the only significant communicating metabolite under the conditions of the experiment. These data support the top-down analysis of the system as defined in Fig. 1.

Figure 3.

Leaf apoplastic sucrose is the only significant communicating metabolite between the source and sink blocks.

Plants were incubated with 14CO2 for 4 h. At the end of the experiment the apoplastic sucrose concentration was measured in four mature leaves. The tubers and stolons were removed and frozen in liquid nitrogen. Flux was estimated from the amount of 14C in the tuber tissue and converted from the amount of 14C to μmol hexose equivalents using the specific activity of sucrose in the stolon. Flux and leaf apoplastic sucrose concentration were measured in cpFBP transgenic plants and in wild-type plants at varying light intensities. The values are expressed as natural logs and the linear relationship between apoplastic sucrose and flux established by linear regression analysis. Each data point represents a single plant. □, wild-type plants at varying light intensity;•, cpFBP transgenic plants. The solid line is fitted to the light intensity data, the dashed line is fitted to the cpFBP transgenic data.

Top-down control analysis

In order to measure the elasticity coefficients of the source and sink blocks with respect to apoplastic sucrose concentration, we measured flux and apoplastic sucrose concentration in the αSUT and cpFBP antisense lines and in response to reductions in light intensity in the wild type (Fig. 4). For perturbations in either block there is a significant correlation (P < 0.05) between lnJ and ln[sucrose] (Fig. 4), and the elasticity coefficients can be estimated from the gradients of these lines as –0.42 ± 0.03 and 1.79 ± 0.07 for the source and sink blocks, respectively (mean ± standard error). The flux control coefficients can be calculated from the block elasticity coefficients using the summation and connectivity theorems of metabolic control analysis (Fell 1997), giving values of 0.81 ± 0.01 and 0.19 ± 0.01, respectively. The confidence limits on these flux control coefficients were calculated from the standard errors of the elasticities, using the method of Small & Fell (1990). These data suggest that control of carbon flux through the potato plants during tuberisation under the conditions of our experiments rests largely with the source reactions.

Figure 4.

Elasticities of source and sink blocks with respect to leaf apoplastic sucrose concentration.

Apoplastic sucrose concentration and flux were measured in wild-type plants at a range of light intensities and in cpFBP and aSUT transgenic plants. Manipulation of the sink block (αSUT transgenic plants) allows measurement of the elasticity of the source block. Manipulation of the source block (changing the light intensity or cpFBP transgenic plants) allows measurement of the elasticity of the sink block. For the light intensity experiment, values are expressed as a percentage of values at the highest light intensity and for the experiments with the transgenic plants values are expressed as a percentage of wild-type values. All data were then converted to natural logarithms. Data for cpFBP plants and light intensity experiment are as Fig. 3. Each data point represents a single plant. The linear relationship between apoplastic sucrose and flux was established by linear regression analysis. Elasticities are given by the slope of the fitted lines. □, sink block;•, source block.

As mentioned earlier, the accuracy of our estimates of block flux control coefficients is dependent on the assumption that sucrose is either distributed uniformly in the apoplast or that the distribution of sucrose does not change in the transgenic lines relative to that in the wild type. If this assumption does not hold, our estimates of the elasticity coefficients may differ from the actual values. It is difficult to assess the impact of any divergance in these estimates because, to our knowledge, nothing is known about the distribution of sucrose in the leaf apoplast. However, in order to examine the dependence of the distribution of control on the values obtained for elasticity coefficients we have calculated control coefficients assuming either a twofold over- or under-estimation of elasticity coefficients. This degree of error in the elasticity coefficients would only arise if there were substantial differences in the relationship between the total apoplastic sucrose concentration (measured) and the sucrose concentration at the site of loading. For example, if there is a non-linear error that results in a twofold underestimation of the sucrose concentration at the point of loading in the most extreme sucrose transporter antisense lines, then the true elasticity of the source block would be around –0.37, as compared to the measured values of –0.42.

The calculated values for the flux control coefficient of the source block, when the elasticity coefficients are varied by a factor of two, are shown in Table 3. It can be seen that despite the large changes in the magnitude of the elasticity coefficients, only when the magnitude of the source block elasticity is increased while that of the sink block is decreased does the flux control coefficient of the source block approach 0.5. On the basis of these calculations we argue that, even if the assumption that we have made regarding the distribution of apoplastic sucrose does not hold, the general conclusion that reactions of the source block dominate the control of flux is valid.

Table 3.  The effect of changing elasticity coefficients on the flux control coefficient of the source block with respect to source to sink carbon flux
Flux control coefficient of the source block if the source block elasticity coefficient is:
Elasticity coefficient of the sink block-0.21-0.42-0.84 
  1. The flux control coefficient of the source block was calculated using the summation and connectivity theorems (see Experimental procedures). The experimentally obtained elasticity coefficients were –0.42 and 1.79 for the source and sink blocks, respectively, and the value calculated from these values is shown in bold. The effect of increasing or decreasing the elasticity coefficients by a factor of two was investigated. The flux control coefficient of the sink block is 1 less the control coefficient of the source block.

0.8950.810.680.52 
1.790.900.810.68 
3.580.940.900.81 

Our finding that source metabolism dominates the control of source to sink carbon flux is consistent with a range of experimental observations. For example, substantial reductions of ADP-glucose pyrophosphorylase activity in potato plants, while reducing the rate of starch synthesis in the tuber, only reduce the tuber dry weight per plant to 60–70% of the wild type (Müller-Röber et al. 1992), suggesting that the rate of translocation of carbohydrate into the tubers is not dramatically affected. Similarly, despite a major increase in tuber size and the fresh weight yield per plant, expression of yeast invertase in the apoplast of potato tuber does not consistently change the dry weight of tubers per plant (Sonnewald et al. 1997). In contrast, manipulations that affect source reactions do appear to influence the rate of carbon movement through the plant. There is much evidence to suggest that increased CO2 concentration and light intensity lead to increased tuber yield in potato (Collins 1976;Wheeler et al. 1991;Yandell et al. 1988). Antisense reduction in cpFBP activity in potato leaves significantly reduces the rate of export from the leaves (Kossmann et al. 1994). Reduction in leaf apoplastic sucrose concentration in a range of species, by expression of yeast invertase in the apoplast, leads to a significant reduction in sink growth rates (Heineke et al. 1992;Sonnewald et al. 1991;von Schaewen et al. 1990).

There are two examples of manipulation of reactions in our sink block that affect sink growth. Inhibition of phloem loading, either by antisense reduction in the activity of the phloem-associated sucrose transporter in potato (Riesmeier et al. 1994), or by expression of E. coli pyrophosphatase in the companion cells of tobacco (Kühn et al. 1996;Lemoine et al. 1996) leads to a significant reduction in sink yield. These experiments suggest that much of the control that we have attributed to the sink block may reside in the process of phloem loading.

Conclusions and implications

We have demonstrated that the approach of top-down control analysis can be used to determine the distribution of control of whole-plant carbon fluxes. It is important to remember, however, that the experimental result that we present here only gives the distribution of control for the conditions of the experiments and for plants grown in specific conditions. In order to investigate this further we are currently determining the effect of a range of environmental variables, applied both during plant growth and during the determination of flux, on the distribution of control. We are also investigating whether the distribution of control varies during the diurnal cycle. Our conclusions are also restricted to the short-term control of flux, and do not take into account any long-term regulation of enzyme activity through changes in gene expression. However, the short-term regulation of flux will potentially interact with the long-term regulation, through changes in sucrose pool sizes. For example, while our experiments predict that changes in sink reactions will have little effect on flux in the short term, if a consequence of those changes is an increase or decrease in sucrose pool sizes in the leaf, then this may lead to changes in the capacity of photosynthetic components (Jang & Sheen 1994), which we predict will affect the carbon flux.

This work has important implications for the biotechnological manipulation of tuber yield in potato. While it may be possible to alter the partitioning of carbon within tubers by manipulating sink enzymes, our work suggests that in order to increase the total dry matter yield of tubers it will be necessary to increase the rate of carbon capture by the source. Because the control of photosynthesis within leaves is distributed evenly (Stitt & Sonnewald 1995), it will be difficult to increase the rate of photosynthesis by increasing one or a small number of steps. We suggest that the strategy most likely to succeed will be to increase the capacity of sucrose transport out of the leaves, which we predict will have two effects. First, it will lead to a modest increase in the rate of carbon flux directly, and second, but more significantly, it may result in an overall increase in source capacity, the latter effect being caused via a reduction in leaf sucrose pools, and a consequent increase in the expression of genes encoding photosynthetic components (Jang & Sheen 1994). The latter change would be predicted to increase carbon flux into the sinks substantially, at least under the conditions of our experiments. The extent to which these conclusions are general, either to potato grown under differing environmental conditions or to other crop species, awaits further experimentation.

In summary, this work emphasises that manipulation of crop yield by genetic engineering requires a clear understanding of the distribution of metabolic control and the metabolic regulation of gene expression. Furthermore, the work suggests that strategies aimed at increasing the yield of potato might productively focus on the large-scale manipulation of source activity.

Experimental procedures

Materials

Except where otherwise stated, all enzymes and substrates were from Boehringer-Mannheim UK (Lewes, Sussex, UK). ADPglucose, fructose-1,6-bisphosphate, glucose-1,6-bisphosphate, Na4P2O7, oxaloacetate and UDPglucose were from Sigma Chemical Co. (Poole, Dorset, UK). Radiochemicals were from Amersham International (Aylesbury, Bucks, UK). TLC plates (20 × 20 cm, polygram Cel 300) were from Aldrich Chemical Co. (Gillingham, Dorset, UK). Compost was from E.A. Goundry & Son Ltd. (Duns Tew, Oxon, UK).

Plants

Potato plants (Solanum tuberosum var. Desiree) were grown by planting sprouted tubers in 150 mm diameter pots containing a mixture of sand and compost (1:2, v/v). The plants were maintained in a greenhouse at 16–25°C with a 16 h photoperiod of natural daylight supplemented to give a minimum irradience of 100 μE m–2 sec–1. The following transgenic lines were used in this work: lines αSUT-13 and αSUT-17 contain cDNAs encoding the phloem-associated sucrose transporter expressed from the CaMV 35S promoter in the antisense orientation and have been characterised previously by Riesmeier et al. (1994); lines cpFBP-12 and cpFBP-16 contain cDNAs encoding plastid fructose-1,6-bisphosphatase expressed from the CaMV 35S promoter in the antisense orientation and have also been characterised previously (cpFBP-12, Kossmann et al. 1994; cpFBP-16, J. Kossmann, unpublished data).

Determination of sucrose concentration in the leaf apoplast

Leaf apoplastic sap was extracted by vacuum infiltration and centrifugation according to Luwe & Heber (1995). Leaves were washed in distilled water and vacuum infiltrated in cold (4°C) 100 mm KCl twice for 1 min each. The leaves were then blotted dry, placed between two funnels to hold them flat and centrifuged for 10 min at 1000 g, 4°C. The apoplastic sap (approximately 100 μl) was collected in 15 ml centrifuge tubes, flash-frozen in liquid nitrogen and stored at –80°C until required. Measurement of sucrose was as in Sweetlove et al. (1996b).

Apoplastic air volume (Vair) was measured by vacuum infiltration of fully hydrated leaves. Mature leaves were removed from the plant and incubated on distilled water for 16 h before infiltrating with 100 mm KCl as above. The apoplastic air volume was calculated from the increase in weight after infiltration.

Apoplastic water volume (Vwater) was measured by uptake of [14C] polyethylene glycol 4000. Discs of 9 mm diameter were removed from leaves using a cork borer, washed in distilled water and blotted dry. The leaf discs were then floated on 20 ml of 100 mm KCl and [14C] polyethylene glycol 4000 (9.25 kBq ml–1). After 10 min, the discs were blotted dry and extracted successively in four 1 ml batches of boiling 80% ethanol (v/v) for 10 min. Apoplastic water volume was calculated from the amount of 14C in the extract. The actual concentration of sucrose in the apoplast was given by:

image

Enzyme assays

Processing and extraction of tuber material and measurement of ADPglucose pyrophosphorylase [EC2.7.7.27], fructokinase [EC2.7.1.4], phosphoglucomutase [EC5.4.2.2], UDPglucose pyrophosphorylase [EC2.7.7.9] and sucrose synthase [EC2.4.1.13] were as described in Sweetlove et al. (1996a), except that all assay reaction volumes were 200 μl instead of 1000 μl. Fructose-1,6-bisphosphatase [EC3.1.3.11] was assayed according to Entwistle & ap Rees (1990). The reaction mixture contained 100 mm Tris–HCl (pH 7.1 or 8.1), 0.4 mm NAD+, 20 mm MgCl2, 0.7 units glucose-6-phosphate dehydrogenase (NAD+ dependent), 0.7 units phosphoglucose isomerase and 2 mm fructose-1,6-bisphosphate in a final volume of 200 μl. Malate dehydrogenase [EC1.1.1.37] activity in apoplastic sap was assayed according to Tetlow & Farrar (1993).

Measurement of flux

Flux was estimated by incubating the photosynthetic tissues with 14CO2 and measuring the rate of accumulation of label in the tubers. The experiments were carried out on 8-week-old plants at the beginning of the photoperiod. Throughout the experiment, plants were maintained at 20°C, 60% relative humidity and at an irradiance of 200 μE m–2 sec–1.Incubation with 14CO2. A clear polythene bag of volume 20 l, with an Eppendorf tube containing 1.48 MBq of NaH[14C]CO3 (200 kBq μmol–1) taped to the inside, was placed over the plant and sealed around the base of the stem. 14CO2 was released by injecting 200 μl of 75% (v/v) malic acid into the Eppendorf tube. After 2 h the bag was removed. After a further 4 h, four mature leaves were removed and the apoplastic sucrose determined as described previously. The tubers were removed, rapidly sliced and frozen in liquid nitrogen. Segments of stolon (2 cm in length) adjacent to the tuber were also removed and frozen in liquid nitrogen.

Determination of 14C in tuber tissue. The tuber tissue was extracted in 200 ml of 80% (v/v) boiling ethanol for 10 min and subsequently homogenised in 100 ml of water. The extracts were combined and 14C in 1 ml aliquots determined by liquid scintillation counting.

Determination of specific activity of sucrose in the stolon tissue. The stolon segments were successively extracted in three 10 ml portions of boiling 80% (v/v) ethanol, the extracts combined, reduced to dryness in vacuuo and resuspended in 5 ml of Na acetate buffer, pH 5.5. The amount of sucrose in the extract was measured as in Sweetlove et al. (1996b). To determine the amount of [14C] sucrose, the extract was separated into neutral, acidic and basic components by ion exchange chromatography (Quick et al. 1989). Sucrose in the neutral fraction was isolated by TLC (cellulose on polyester, developed twice with 2-methylpropan-2-ol: 2-butanone: formic acid: water (8:6:3:3)). Sucrose was identified by comparison with standards, scraped from the plate, suspended in 1 ml of water and the amount of 14C determined. More than 98% of the loaded label was recovered from both the ion exchange chromatography and TLC.

Control analysis and statistics

Elasticity coefficients for the source and sink blocks with respect to apoplastic sucrose were calculated from the relationship between sucrose concentration and flux when sink and source metabolism, respectively, were manipulated. The elasticity coefficients are defined as:

image

when sink metabolism is perturbed, and

image

when source metabolism is perturbed (Brown et al. 1990), where J is the flux and [sucrose] is the leaf apoplastic sucrose concentration. The block elasticity coefficients are related to the flux control coefficients, CJsource and CJsink, by the connectivity theorem of metabolic control analysis (Fell 1997):

image

The flux control coefficients are also related by the summation theorem (Fell 1997):

image

Using the experimentally determined values for the elasticity coefficients, eqns 2 and (3) can be solved for the flux control coefficients of the source and sink.

Statistical analysis of the data was carried out using Minitab 10.5 Xtra Power for Apple Macintosh (Minitab Inc.) and Microsoft Excel 5 for Apple Macintosh (Microsoft corporation), and use of the word ‘significant’ indicates a returned probability value of 0.05 or lower.

Acknowledgements

We thank Dr Susan Robertson for critically reading the manuscript. This work was supported by grants from the Biotechnology and Biological Sciences Research Council of the UK (Ref. No. P03706) and the Royal Society of London to SAH. Travel between Oxford and Golm was funded by a Ciba Award for Collaboration in Europe to SAH.

Ancillary