Evidence of a key role for photosynthetic oxygen release in oil storage in developing soybean seeds


Author for correspondence: Ljudmilla Borisjuk Tel: +49 39482 5687 Fax: +49 39482 5500 Email: borysyuk@ipk-gatersleben.de


  • • Based on the topographical analysis of photosynthesis and oil storage, we propose in a companion paper that photosynthetic oxygen release plays a key role in the local energy state, storage metabolism and flux toward lipid biosynthesis in developing soybean seeds. To test this hypothesis, we combined topographical analysis of ATP gradients across tissues, microsensor quantifications of internal O2 levels, assays of energy balance, metabolite profiles and isotope-labelling studies.
  • • Seeds show a marked degree of oxygen starvation in vivo (minimum O2 levels 0.1 kPa, ≈ 1.3 µm), affecting ATP gradients, overall energy state, metabolite pools and storage activity.
  • • Despite the low light availability, photosynthesis supplies significant amounts of oxygen to the hypoxic seed tissue. This is followed by an increase in local ATP levels, most prominently within the lipid-synthesizing (inner) regions of the embryo. Concomitantly, partitioning of 14C-sucrose to lipids is increased, suggesting higher rates of lipid biosynthesis.
  • • It is concluded that both respiratory and biosynthetic fluxes are dynamically adjusted to photosynthetic oxygen supply.


The main progress towards our current understanding of regulatory networks governing lipid biosynthesis and its crosstalk with photosynthesis has come from studies on Arabidopsis and Brassica (White et al., 2000; Ruuska et al., 2002, 2004; Schwender & Ohlrogge, 2002; Beisson et al., 2003). However, the small seed size of these species hampered topographical approaches, and seeds or embryos were considered as homogeneous tissues. Previous studies on other species demonstrated that embryos possess marked developmental gradients with respect to cell structure, metabolite levels, energy status, storage activity, etc. (Borisjuk et al., 2003; Hills, 2004; Weber et al., 2005). Thus embryo tissue is highly heterogeneous, a fact that complicates the picture of regulatory networks in vivo. To overcome this limitation, soybean can be used as a model oilseed.

In the first part of our study (Borisjuk et al., 2005), the topography of photosynthetic activity and lipid deposition in developing soybean embryos was analysed. During development the interior the of embryo loses its photosynthetic capacity. Photosynthetically active chloroplasts differentiate into storage organelles, expressing metabolite transporters characteristic of heterotrophic metabolism. Storage processes are initiated, followed by the accumulation of both starch and lipids. This results in a marked separation of local photosynthetic and lipid biosynthetic activities within the embryo during late development. Under such conditions, direct interactions such as photosynthetic energy supply (ATP/NADPH) for lipid synthesis, as well as redox modulation of biosynthetic enzymes, are unlikely.

Photosynthetic oxygen release plays a fundamental role on a global scale, but has largely been overlooked with respect to seed metabolism. Recently the gain of photosynthetic capacity and oxygen release, respectively, was shown to be correlated with increasing energy levels in seeds of pea and Vicia faba (Borisjuk et al., 2003; Rolletschek et al., 2003). Concomitantly, fermentation activity of seeds decreased. Hence photosynthetic oxygen release may prevent internal hypoxia. This might also be of relevance for developing soybean embryos. There is indirect evidence for hypoxic conditions within soybean seeds (Shelp et al., 1995). Hypoxia would limit respiratory energy supply and, eventually, the energetically expensive synthesis of fatty acid (Vigeolas et al., 2003). According to our hypothesis, derived from topographical studies (Borisjuk et al., 2005), photosynthetically produced oxygen diffuses from the peripheral (photosynthetically active) regions towards the interior, allowing local increases in respiratory and biosynthetic fluxes. This would explain the light stimulation of lipid biosynthesis (for references see Borisjuk et al., 2005).

The overall motivation for this work was to investigate the effects of photosynthetic oxygen release on energy levels and lipid biosynthesis, and to determine the degree to which the potential oxygen depletion in vivo limits key aspects of seed metabolism in soybean. To this end, we combined topographical analysis of ATP gradients across tissues, with microsensor quantifications of O2 in transects through seeds, and assays of energy balance, metabolite profiles and labelling studies with 14C assimilates. These approaches were applied to experimental perturbations of oxygen inside intact, attached seeds. Our results show a marked degree of oxygen starvation inside developing seeds which contend with significant oxygen limitation to ATP gradients, energy status, metabolite pools and partitioning of sucrose to lipids. Moreover, data reveal that photosynthesis supplies significant amounts of oxygen to the seed tissue. It is concluded that both respiratory and biosynthetic fluxes are adjusted to photosynthetic oxygen supply.

Materials and Methods

Plant growth

Soybean plants (Glycine max (L.) Merr.) were cultivated in the glasshouse under a light/dark regime of 16/8 h. Intact seeds at distinct developmental stages were harvested during the mid-light phase, snap-frozen in liquid N2 and stored at −80°C until analysed.

Measurement of internal O2 concentration

The O2 concentration inside seeds was measured using O2-sensitive optical microsensors in a procedure modified from that developed earlier for V. faba cotyledons (Rolletschek et al., 2002a). For soybean, the intact fresh seed was carefully moved into a fixed position and anchored firmly for subsequent micromanipulation. The microsensor was inserted into the seed using a micromanipulator. The site of sensor entry was sealed with silicone to prevent O2 movement along microchannels. Measurements were taken within 5 min. The O2 concentration is expressed in kPa (21 kPa = 100% of atmospheric saturation). After measurement, the seed was dissected at the measured transect to identify the exact position of the sensor.

Measurement of photosynthetic O2 release and light transmission

Photosynthetic oxygen release was measured as described in Borisjuk et al. (2005). Transmission of photosynthetically active radiation through pod wall, seed coat and embryo was measured using a quantum probe sensor (Li-Cor, Lincoln, NE, USA) and the illumination system of the microscope.

Imaging of local ATP concentration

Bioluminescence imaging was used to measure ATP distribution in cryosections of seeds. This technique allows the quantitative, histographical mapping of ATP and has been described in detail (Borisjuk et al., 2003). Briefly, cryosections were prepared from snap-frozen seeds and immersed in an enzyme solution linking ATP to the reaction of firefly luciferase. The enzyme emits photons with an intensity proportional to the content of ATP. The light emission was registered with a photon-counting system (Hamamatsu, Herrsching, Germany) linked to a microscope. ATP concentrations determined by bioluminescence are directly proportional to those determined by conventional high-performance liquid chromatography (HPLC) (correlation coefficient 0.85, P < 0.01; Borisjuk et al., 2003).

Metabolite analysis

Frozen plant material was rapidly weighed and immediately homogenized with an ice-cold mortar and pestle in liquid N2. Metabolic intermediates were extracted with trichloroacetic acid as described by Rolletschek et al. (2002a). Aliquots of snap-frozen extracts were used for subsequent metabolite analyses. Free amino acids were measured by HPLC (Rolletschek et al., 2002b); dissolved sugars were measured photometrically (Borisjuk et al., 2002); adenine nucleotides were analysed by HPLC with fluorescence detection (Rolletschek et al., 2005); and metabolites of glycolysis and the citrate cycle were measured by liquid chromatography coupled to mass spectrometry (Rolletschek et al., 2005).

Oxygen treatment

To study the effect of altered O2 supply on metabolite levels, seeds in an intact pod (attached to the mother plant) were aerated with gas mixtures (combined by a multi-gas controller type 647B, MKS, Muenchen, Germany) containing atmospheric O2 levels (21 kPa) and 42 and 2.1 kPa O2 (balanced by N2), respectively, at a flow rate of 3 l min−1. After a 6 h incubation, seeds were quickly removed, immediately frozen in liquid N2 and stored at −80°C until they were used for metabolite analysis. For the analysis of light effects, some intact pods were covered with aluminium foil. Seeds were sampled after 6 h and compared with neighbouring (uncovered) pods at the same developmental stage. The effect of changing external O2 levels on internal O2 concentrations was analysed by inserting the microsensor into intact seeds (≈ 500 µm) aerated with gas mixtures containing different O2 levels.

Labelling experiments

For monitoring of in vivo fluxes, developing seeds in an intact pod (attached to the mother plant) were labelled with U-14C-sucrose (Amersham-Buchler, Freiburg, Germany) and treated with different gas mixtures, as described above. Using a microsyringe, 3 µl labelled 14C-sucrose (7.4 MBq ml−1) were injected into the adaxial region of the embryo. The site of needle entry was sealed with silicone to prevent gas exchange. After 8 h incubation the seeds were removed and immediately frozen in liquid N2. After weighing, seeds were extracted as described by Bligh & Dyer (1959), giving a chloroform-soluble fraction (lipids), a water/methanol-soluble fraction and the remaining pellet. Radioactivity was determined by liquid scintillation counting. Counts were corrected for background and quenching by external standards.


Light supply to the embryo is increasingly restricted toward the interior

Low light transmittance of tissues surrounding the embryo is regarded as a critical factor for seed photosynthesis. To test this, we analysed the amount of available light (Fig. 1a). Nearly 80% of incident light was absorbed by the pod wall, and another 10–12% by the seed coat. Only 8% of incident light was available at the embryo surface, corresponding to ≈ 160 µmol photons m−2 s−1 (assuming maximal light intensities of ≈ 2000 µmol photons m−2 s−1 in full sunlight in summer). Transmission values for pod wall and seed coat differed insignificantly for the period analysed (main storage phase between 80 and 400 mg seed f. wt). Within the embryo, light supply declined nearly exponentially, and was minimal in adaxial regions.

Figure 1.

Light availability and its utilization by soybean (Glycine max) seed. (a) Profile of light transmittance through the seed tissue (p, pod; sc, seed coat; e, embryo). (b) Tissue O2 level (circles, left axis) and photosynthetic O2 release rate (bars, right axis) in nmol g−1 f. wt seed and min−1 as means ± SD, in response to light intensity.

Soybean embryos are photosynthetically active and produce O2 under light (Sugimoto et al., 1987). To analyse how the low light supply affects the photosynthetic activity of the embryo as well as its internal O2 balance, whole seeds (seed coat + embryo) were exposed to stepwise increasing light intensities (Fig. 1b). Raising light intensity from 60 to 320 µmol photons m−2 s−1 resulted in a several-fold increase in photosynthetic activity measured as O2 release rate. Concomitantly, the mean O2 concentrations within the embryo tissue also increased (Fig. 1b). Light intensities exceeding 450 µmol photons m−2 s−1 did not result in a further boost of tissue O2 levels (data not shown). It is concluded that, despite the low light availability, in vivo seed photosynthesis supplies significant amounts of oxygen to seed tissue and elevate internal levels.

Photosynthetic oxygen release is balanced in vivo by adjusted oxygen consumption

To analyse the in vivo O2 concentration in soybean seeds and how photosynthesis affects the O2 balance, we measured the steady state O2 levels along transects in seeds exposed to either light or darkness. During the early to mid-storage phase (Fig. 2, left and middle panels), embryos were characterized by very low internal O2 concentrations with minimum levels of 0.1 kPa when measured in darkness (corresponding to ≈ 0.5% atmospheric saturation). When seeds were illuminated (≈ 160 µmol photons m−2 s−1), O2 levels rose strongly to 11–16 kPa. The concentration profile showed slightly decreasing levels towards the interior, especially in embryos of the mid-storage stages. Remarkably, within ≈ 20 min of light-adaptation, O2 levels decreased to values observed under darkness. Assuming nearly constant photosynthetic O2 release rates at the given light intensity, this decrease in O2 levels can only be explained by elevated O2 demands (seed respiration). To further confirm this suggestion, we exposed darkened seeds to elevated external O2 concentrations. Increases in external levels, up to ≈ 63 kPa (threefold atmospheric O2 saturation), did not affect the internal level (data not shown). More marked elevations of external O2 concentration eventually raised its levels in embryo. Rising O2 supply was thus apparently balanced by increasing O2 consumption. External levels greater than 63 kPa were needed before internal O2 demand could be exceeded. Consequently, respiration inside developing seeds was considered to be nonO2-saturated at normal ambient levels, that is, it was O2-limited in vivo.

Figure 2.

Representative O2 maps within developing soybean (Glycine max) seeds. O2 was measured by microsensors along the x axis (penetration depth in µm) within seeds at 80 mg (early storage stage); 240 mg (mid-storage); and 400 mg (late storage) of seed f. wt. The O2 concentration is given as percentage of atmospheric saturation (21 kPa = 100%). Measurements were done in dark-adapted; light-nonadapted (3 min after illumination); and light-adapted (20 min after illumination) seeds (black, red and white circles, respectively).

Embryos of the late storage phase (Fig. 2, right panel) had similarly low O2 levels when measured in darkness. However, when seeds were illuminated, O2 level increased to ≈ 6.3 kPa. After light adaptation, the O2 level remained at a higher level compared with the early stages. Thus during the late-storage phase, day/night variations of O2 levels were evident, possibly indicating a lower adaptive increase in respiratory O2 demand.

In summary, our measurements show that O2 levels in embryos were quite low, and became dramatically elevated under illumination because of photosynthesis. Furthermore, tissue O2 demand was adjusted to O2 delivery, that is, increased when additional O2 (via photosynthetic release) becomes available. As a result, O2 levels within the embryo tissue were more-or-less stable after an period of adaptation. Such balancing of O2 consumption to its supply was observed during the entire development period, but declined towards maturation.

Responses of energy state of seeds to both light and oxygen supply

Tissue O2 measurements indicated an adaptive increase in respiratory activity (or O2 demand) in response to light supply. We tested whether this affected the energy state of seeds. As shown in Fig. 3a, the amount of single adenine nucleotides changed only slightly in response to light supply, and changed significantly only for ADP. However, the ATP : ADP ratio of light-adapted embryos increased markedly, indicating a higher energy state at light vs darkness (Fig. 3b). In principle this increase might be attributed to both photosynthetic and mitochondrial energy production. To elucidate the mitochondrial input, we tested the effect of O2 treatment on seeds under darkness. Subambient (2.1 kPa) O2 levels resulted in a drastic increase in AMP and a decrease in ATP (Fig. 3c), whereas superambient (42 kPa) O2 levels reduced the level of ADP. Elevating the external O2 levels resulted in marked increases in the ATP : ADP ratio (Fig. 3d). In this way, increasing O2 supply mimicked light supply. This indicates that (1) ATP production in developing seeds is limited in vivo by low internal O2 levels (hypoxia); and (2) increasing O2 supply (either experimentally or via photosynthesis) is coupled to higher respiratory activity, resulting in an elevated energy state. Importantly, this suggests that the increasing energy state of seeds following illumination cannot be restricted to the energetic function of photosynthesis, and is essentially caused by respiration (stimulated via photosynthetic O2 release).

Figure 3.

Effect of light and changing external O2 levels on concentration of adenine nucleotides and ATP : ADP ratios in soybean (Glycine max) seeds (mid-storage stage). (a,b) Intact seeds were kept in either darkness (black bars) or light (white bars). (c,d) Intact seeds were aerated with gas mixtures containing 2.1, 21 and 42 kPa oxygen (black, grey and white bars, respectively). Data are given as means ± SD. *, Significant differences vs control (either dark or 21 kPa treatment), t-test P < 0.05.

Oxygen supply affects ATP gradients within the embryo

To see how ATP is distributed under in vivo conditions, and how this pattern is affected by O2 supply, we investigated ATP distribution using bioluminescence imaging (Borisjuk et al., 2003). When darkened seeds were exposed to ambient air, ATP was highest within the embryo's interior (Fig. 4a), where high starch and lipid biosynthetic activity occur (Borisjuk et al., 2005). There was a near-linear decrease in ATP levels towards the outer regions, as shown by the regression curve (red in Fig. 4b). When seeds were exposed to twofold atmospheric O2 levels (42 kPa), ATP distribution remained constant (data not shown) as the overall concentration did (Fig. 3c). However, exposing seeds to only 10% of the ambient O2 level (2.1 kPa) caused the loss of ATP enrichment within the interior region of the embryo (Fig. 4a). The ATP gradient across embryo tissue nearly disappeared (Fig. 4b). These results demonstrate that the in vivo O2 supply is an essential factor in maintaining the high-energy state, especially within the interior region of embryos where lipid deposition occurs. A limited O2 supply resulted in a loss of ATP gradients within cotyledons.

Figure 4.

ATP distribution pattern in soybean (Glycine max) seeds. (a) Distribution under in vivo conditions (21 kPa) and at decreased external O2 levels (2.1 kPa). Concentration of ATP is given in a colour scale. (b) Local concentration of ATP within the same tissue region, represented as a graph along the α axis (a).

Oxygen availability affects steady state levels of metabolites

In addition to changes in the energy state of seeds, O2 supply might affect their metabolic activity. To further elucidate this aspect, we analysed the differential responses of intermediates of glycolysis and the citric acid cycle, as well as related sugars and free amino acids, to changes in O2 supply (Fig. 5; Table 1). The positioning of subfigures is based on their proximity in a generalized metabolic context (for details see Borisjuk et al., 2004; Dwyer et al., 2004). Neither individual soluble sugars (glucose, fructose, sucrose) nor the total sugar pool changed significantly in response to shifts in external O2 supply. When exogenous O2 supply was lowest (2.1 kPa), the steady state levels of most glycolytic intermediates (glucose-1-phosphate, glucose-6-phosphate, fructose-6-phosphate, fructose-1,6-bisphosphate and pyruvate) were highest. The concomitant increases in both lactate and alanine may indicate accelerated lactate/alanine fermentation under these low-O2 conditions. High pool sizes for succinate and gamma-amino butyric acid (GABA) were further indications of hypoxic metabolism (in addition to the lower energy state, Fig. 3c). A slight but significant increase was also obvious for ADP-glucose (the direct precursor for starch biosynthesis). When exogenous O2 supply was increased to twice ambient levels (42 kPa), a nearly opposite pattern was observed: lowest levels of most glycolytic intermediates, and even significant decreases for both lactate and GABA. This suggests decreasing fermentation activity or, conversely, hypoxic metabolism under in vivo conditions (21 kPa). Importantly, the ratio of fructose-6-phosphate to fructose-1,6-bisphosphate increased about twofold in superambient O2 levels. This further indicates a shift from fermentative to aerobic metabolism. Metabolite levels that increased most markedly in response to O2 supply were Gln, Asp and, notably, acetyl-coenzyme A, the latter being the precursor for plastidic fatty acid synthesis. The Gln : Glu ratio increased progressively with rising O2 supply, whereas the Asn : Asp ratio was highest at the subambient O2 treatment (2.1 kPa).

Figure 5.

Simplified schematic representation of primary metabolites in soybean (Glycine max) seeds and their response to changing O2 supply. Intermediates of glycolysis and the citrate cycle are shown, as well as branch paths to related sugars and free amino acids. Arrows indicate probable direction of predominant carbon flow. Vertical bars show the level of each metabolite in seeds after a 6 h treatment with 2.1, 21 and 42 kPa oxygen (black, grey and white bars, respectively). Data are given in relative units (mean ± SD). *, Significant differences vs control treatment (21 kPa), t-test P < 0.05.

Table 1.  Primary metabolites and their response to changing O2 supply
 2.1 kPa O221 kPa O242 kPa O2
  • *

    Data in µmol g−1 f. wt.

  • Data in relative units.

  • Data in nmol g−1 f. wt.

  • Data are given as mean ± SD.

Soluble sugars*
 Glucose 4.8 1.6  6.0 2.1  8.1 1.5
 Fructose 6.5 3.0  7.1 2.3  9.1 1.6
 Sucrose32.6 4.9 34.9 1.7 35.610.2
Free amino acids*
 Ala 5.19 1.18  1.97 0.48  1.77 0.35
 Arg 1.95 1.03  1.70 0.31  2.27 0.82
 Asp 0.35 0.06  1.41 0.12  1.50 0.36
 Asn15.8 4.8  7.5 1.4 11.7 2.5
 Cys 0.16 0.03  0.23 0.04  0.33 0.09
 Glu 7.12 0.60  5.59 0.40  4.26 0.51
 Gln 0.35 0.18  0.68 0.17  0.90 0.31
 Gly 0.38 0.10  0.27 0.02  0.27 0.03
 Gaba 0.08 0.02  0.07 0.01  0.03 0.02
 Ile 0.11 0.06  0.08 0.02  0.20 0.08
 Leu 0.15 0.04  0.18 0.02  0.14 0.02
 Lys 0.18 0.07  0.11 0.01  0.16 0.04
 Met 0.06 0.02  0.05 0.01  0.13 0.03
 Phe 0.09 0.04  0.05 0.01  0.07 0.01
 Ser 1.17 0.44  1.01 0.13  0.92 0.18
 Thr 1.30 0.14  1.15 0.14  1.54 0.20
 Tyr 0.09 0.04  0.02 0.00  0.04 0.01
 Val 0.36 0.17  0.28 0.05  0.53 0.17
Nucleotide sugars
 UDP-glucose 1.02 0.09  1.45 0.06  1.37 0.06
 ADP-glucose16.5 1.2 13.6 1.2 14.2 2.0
 AMP44.4 8.1  6.2 1.7  6.6 1.1
 ADP53.7 3.0 45.8 4.3 35.2 3.1
Glycolytic intermediates
 Glucose-1-P10.57 1.41  5.73 0.65  5.43 0.24
 Glucose-6-P 8.82 0.51  3.76 0.12  2.62 0.11
 Fructose-6-P32.81 3.05 12.72 0.95  9.20 0.50
 Fructose-1,6-bisP 1.65 0.48  0.87 0.23  1.40 0.25
 3-PGA 1.08 0.41  2.02 0.18  0.89 0.09
 PEP 2.50 0.28  3.09 0.19  2.28 0.19
 Pyruvate 0.36 0.06  0.27 0.02  0.10 0.01
 Acetyl-CoA 0.24 0.06  0.62 0.08  0.80 0.06
Organic acids
 Lactate 3.76 0.59  1.86 0.36  1.09 0.11
 Succinate28.39 2.65 15.36 1.40 16.26 0.80
 Malate 8.28 0.56  7.42 0.41  7.30 0.47
 Citrate10.60 0.63  9.95 0.56  9.74 0.41
 Isocitrate 3.61 1.12  6.14 1.65  2.61 0.66
 Shikimate 0.82 0.10  0.82 0.06  0.77 0.07
 Chorismate 0.29 0.04  0.37 0.03  0.32 0.04

Elevated oxygen supply shifts partitioning of labelled sucrose into lipids

Metabolite patterns in Fig. 5 suggest changing fluxes towards lipid but not starch biosynthesis in response to elevated O2 supply. To follow metabolic fluxes, we analysed the partitioning of sucrose in isotope tracer experiments. Presuming the label will be evenly distributed over the incubation time, 14C-sucrose was injected into the adaxial region of the embryo of intact, attached seeds. These were subsequently exposed to either ambient or superambient O2 levels (21 and 42 kPa, respectively). The results in Fig. 6 show that the lipid fraction contained nearly 40% more 14C label when atmospheric O2 levels had been doubled. The label within the pellet fraction consisting mainly of starch (≈ 85%) and protein + cell wall (15%) was not significantly changed. Thus changing fluxes from endogenous sources of unlabelled sugars (e.g. altered starch synthesis/breakdown) are very unlikely.

Figure 6.

Incorporation of 14C-labelled sucrose into different fractions in soybean (Glycine max) seeds as affected by elevated external O2. Radiolabelled sucrose (3 µl) was delivered via air-sealed minute injection into seeds attached to the plant. After incubation at ambient (21 kPa, grey bars) and superambient (42 kPa, white bars) O2 levels, the seed was isolated and analysed. All data are means ± SD. *, Significant differences vs control treatment (21 kPa), t-test P < 0.05.

Higher labelling of the lipid fraction clearly indicates a higher lipid biosynthetic activity in response to elevated O2 supply. This is also consistent with increases in the steady state pool size of acetyl-coenzyme A.


This study tests the hypothesis that the light stimulation of lipid biosynthesis in soybean seeds (Willms et al., 1999) is mediated via photosynthetic O2 supply. Key evidence is provided that photosynthetic O2 release significantly elevates internal O2 levels within seeds. Under conditions where endogenous O2 concentrations are limiting, this O2 release stimulates respiratory O2 demand followed by increases in the overall energy state of seeds, especially in regions where lipid storage occurs. Moreover, metabolic activity is affected and becomes adjusted by increasing the partitioning of assimilates towards lipid biosynthesis.

Under internal hypoxia, O2 demand of seeds is adjusted to photosynthesis

Oxygen concentration profiles, as well as increases in the overall energy state in response to O2 supply, clearly indicate that developing soybean seeds experience marked hypoxia. The extent and duration of this hypoxia are marked. Our data provide evidence for earlier assumptions based on rather indirect biochemical data (Shelp et al., 1995; Gambhir et al., 1997). It further explains the findings of Sinclair et al. (1987), who showed that the seed growth rate of soybean is dependent on external O2 levels. Internal hypoxia might be attributed to highly gas-impermeable seed coats (Sinclair, 1988) as well as high respiratory activity, and appears to be a common feature of developing seeds (Geigenberger, 2003; Rolletschek et al., 2003). The importance of photosynthetic O2 release for internal O2 balance is highlighted in Figs 1, 2. O2 concentrations transiently increased to high levels (> 12 kPa) when soybean embryos were illuminated. This increase corresponds to the photosynthetic (16O–) O2 release of illuminated soybean seeds (Willms et al., 1999). The transient increase occurred to a lesser extent in later stages, which probably corresponds to decreasing photosynthetic activity during development (Sugimoto et al., 1987). Remarkably, O2 levels within embryos decreased again to values observed under darkness after relatively short periods (15–20 min). This adaptive change can be explained only by significant increases in respiratory O2 demand. Dynamic adjustments of the respiratory chain might be caused via phytochrome signalling (Escobar et al., 2004) and/or increased availability of substrates (e.g. O2). We strongly suggest that the latter is the decisive factor under conditions where O2 limits respiration (hypoxia), and it is reasonable to assume that an elevated O2 supply promotes respiration. This conclusion also corresponds to findings by Gale (1974), who has shown that respiration in soybean seeds is positively related to external O2 concentrations up to 21% (v/v). We could further show that even more marked elevations of up to 3 × atmospheric levels did not raise internal O2 levels, but were apparently fully balanced by increasing O2 consumption. Concomitantly, the energy state of seeds increased. The dynamic adjustment of respiration finally leads to nearly constant steady state O2 concentrations within the embryo. This might be different in later developmental stages where mean O2 levels were significantly increased in light vs darkness. A shifted balance between respiratory O2 demand and photosynthetic O2 release would explain this observation. Both processes declined during development (Miller et al., 1983; Sugimoto et al., 1987).

The dynamic adjustment of respiration also explains major problems concerning in vitro studies using excised embryos with removed seed coats. First, O2 entry into the embryo is significantly increased which, in turn, promotes respiration as well as overall energy state (Shelp et al., 1995). Thereby, hypoxic constraints, important in vivo, are abolished. Second, CO2 levels which are very high in vivo (up to 2000-fold atmospheric level, Goffman et al., 2004) drop. Respiration, previously possibly arrested because of high CO2, is promoted. Third, rising CO2 : O2 ratios may considerably increase photorespiration, which does not play a role in vivo (Sriram et al., 2004). Consequently, conclusions drawn from in vitro studies are seriously limited.

Energy levels in lipid-storing regions are dependent on photosynthetic oxygen release

Both illumination of seeds and exposition to elevated O2 concentrations resulted in an increase in their energy state (ATP : ADP ratio, Fig. 3). This suggests that the increase following illumination cannot be restricted to the energetic role of photosynthesis, but is essentially coupled to respiratory activation via photosynthetic O2 delivery. This conclusion is further evidenced by the observed ATP distribution pattern (Fig. 4a). ATP levels were relatively low in peripheral (abaxial) regions of the embryo where highest photosynthetic activities are found (Borisjuk et al., 2005). However, the highest ATP levels are found in the inner (adaxial) regions of the embryo where lipid (and starch) biosynthesis preferentially occurred. Consequently, ATP distribution pattern is not coupled with photosynthesis, but coincides with storage activity. Similar findings come from studies with seeds of broad bean (V. faba), barley (Hordeum vulgare) and maize (Zea mays). In barley, the highest ATP levels are generated within highly active starch-synthesizing regions of caryopses (Rolletschek et al., 2004). In V. faba, high ATP is coupled with the establishment of photosynthetic capacity during early growth stages (Borisjuk et al., 2003). However, during the main storage phase, the highest ATP levels are found in abaxial regions where major storage proteins are accumulated. In maize, high ATP coincides spatially with both starch and lipid deposition (Rolletschek et al., 2005). The results of these studies support the suggestion that high local energy levels are indicative of the metabolic status of the tissue, and may represent an important feature of expanding, actively storing cells.

Under O2 shortage (as induced experimentally by exposing seeds to reduced O2 concentrations), ATP gradients across soybean cotyledons nearly disappeared (Fig. 4b). Consequently energy levels drop, especially in lipid-synthesizing regions which potentially possess an intensely energy-demanding metabolism. This illustrates the importance of photosynthetic O2 supply for the maintenance of high local energy levels as a prerequisite for storage activity.

Light-stimulated increase in lipid storage is mediated by photosynthetic O2 supply

To test if the above-mentioned changes also affect the overall metabolism of developing soybean seeds, and especially their storage activity, we analysed metabolite pattern and assimilate partitioning in response to changing O2 supply (Figs 5, 6). The steady state levels of primary metabolites showed characteristic changes when the O2 supply was decreased: (i) an increase in most glycolytic intermediates; and (ii) an increase in fermentation products (lactate, alanine, GABA). Together with decreasing energy levels (Fig. 3), such changes are indicative of partially restricted respiration, elevated glycolytic fluxes and fermentative metabolism (Armstrong et al., 1994; Drew, 1997; Geigenberger, 2003). Furthermore, they argue for a demand-driven regulation of the glycolytic pathway, as was suggested recently by Fernie et al. (2004). Reducing the O2 supply to seeds did not decrease the level of ADP-glucose, which represents the direct precursor for starch biosynthesis. Instead there was a small but significant increase, indicating that starch biosynthesis was not negatively affected. Increasing the O2 supply to seeds resulted in nearly inverse changes, as described before. Remarkably, the level of both lactate and GABA (indicative of fermentation) decreased compared with those obtained under ambient O2 levels. This points to metabolic limitations at ambient O2 supply. Similarly important was the finding that acetyl-CoA (the precursor for lipid biosynthesis) progressively increased with O2 supply. Increases in O2 levels also changed the pool size of some free amino acids. This may be related to changes in assimilate supply (phloem unloading was shown to be energy- and O2-dependent; Thorne, 1982) or, alternatively, to changing rates of protein biosynthesis.

Using isotope tracer experiments, we have shown that not only metabolite profiles but also metabolic fluxes responded to varying O2 supply. Partitioning of (14C-labelled) sucrose towards lipids was significantly increased by ≈ 40% when seeds were exposed to 2 × atmospheric O2 levels (Fig. 6). This also corresponded to increased levels of acetyl-CoA. However, it should be noted that total metabolite levels were measured, but not those of plastids (which would be more appropriate). As already mentioned, the energy state of soybean seeds increased likewise (Fig. 3). There are several studies showing that rates of lipid synthesis are strongly energy-dependent, and that lipid synthesis requires more ATP than the synthesis of other storage products (reviewed by Neuhaus & Emes, 2000). Our data are consistent with recent findings of Vigeolas et al. (2003) that lipid but not starch synthesis in rapeseed correlates with the energy state of seeds. Based on theoretical considerations, lipids possess the highest chemical potential energy compared with either protein or starch (Vertregt & deVries, 1987). This fact may well explain the strict energy dependence. The experimental approach used here was to inject labelled substrates into otherwise intact seeds. The observed changes in carbon partitioning therefore indicate a changing sink activity – the use of assimilates for storage product synthesis. Under in vivo conditions, fluxes towards lipids as well as proteins may be additionally affected by changes in the O2-dependent assimilate supply (Thorne, 1982; Saravitz & Raper, 1995).

In summary, we have shown that illumination of seeds causes photosynthetic release of significant amounts of O2. Under internal hypoxia, this O2 is instantly used for respiration which, in turn, elevates the energy supply. Finally, this affects metabolic fluxes and increases assimilate partitioning towards lipids. It is concluded that the light stimulation of lipid biosynthesis in developing soybean seeds can be substantially attributed to photosynthetic O2 release. Energy metabolism, as well as lipid biosynthetic fluxes, is adjusted to O2 availability and photosynthetic O2 release, respectively.


We are grateful to Dr M. Hajirezaei for help with 14C-isotope studies, and K. Blaschek for excellent technical assistance. We acknowledge funding by the Deutsche Forschungsgemeinschaft (Project number RO 2411/2-1/2-2).