Skip to main content
NCBI home page
Plant Physiology logoLink to Plant Physiology
. 2004 Feb;134(2):801–812. doi: 10.1104/pp.103.032318

Adenylate Gradients and Ar:O2 Effects on Legume Nodules: I. Mathematical Models1

Hui Wei 1, Craig A Atkins 1, David B Layzell 1,*
PMCID: PMC344555  PMID: 14739350

Abstract

Mathematical models were developed to test the likelihood that large cytosolic adenylate concentration gradients exist across the bacteria-infected cells of legume nodules. Previous studies hypothesized that this may be the case to account for the unusually low adenylate energy charge (AEC; 0.65) measured in the plant fraction of metabolically active nodules (M.M. Kuzma, H. Winter, P. Storer, I. Oresnik, C.A. Atkins, D.B. Layzell [1999] Plant Physiol 119: 399–407). Simulations coupled leghemoglobin-facilitated O2 diffusion into the infected cell, through bacteroid nitrogenase activity, with the ATP demand for transport and ammonia assimilation in the plant fraction of ureide- and amide-producing nodules. Although large cytosolic adenylate gradients were predicted to exist in both nodule types, amide nodules were predicted to have steeper AEC gradients (0.82–0.52) than ureide nodules (0.82–0.61). The differences were attributed to an additional ATP demand for Asn synthesis in the amide nodule. Simulations for nodules transferred to an Ar:O2 atmosphere predicted a major reduction in the magnitude of adenylate gradients and an increase in the AEC of the plant fraction. Results were consistent with a number of experimental studies and were used to propose an experimental test of the models.

Carbon (C) metabolism and N2 fixation in legume nodules is limited by O2 supply. Using a nonaqueous technique to quantify adenylate pools in intracellular fractions, Kuzma et al. (1999) found that the adenylate energy charge (AEC = [ATP + 0.5 ADP]/[ATP + ADP + AMP]) was 0.65 ± 0.04 in the plant fraction of active, N2-fixing nodules, whereas bacteroids had an AEC of 0.76 ± 0.09. However, when nodule metabolism was more severely O2 limited by exposure to 10% (v/v) O2 for 3 min, AEC in the plant cell compartment was not affected but in the bacteroids decreased from 0.76 ± 0.09 to 0.56 ± 0.06. The decrease in bacteroid AEC accounted for the decrease in whole-nodule AEC (from 0.70–0.61, Kuzma et al., 1999) after low O2 treatment. Other treatments known to reduce infected cell pO2 (e.g. NO3– fertilization and stem girdling) also reduced the AEC of whole nodules (de Lima et al., 1994), presumably because of a decrease in bacteroid rather than plant AEC.

These data (de Lima et al., 1994; Kuzma et al., 1999) are consistent with the bacteroids, not the plant fraction, being the site of O2 limitation in legume nodules. This conclusion was supported by models of infected cells in nodules (Thumfort et al., 1994, 1999) that simulated diffusion and reactions involving O2. The models predicted steep gradients in O2 concentration from the cytosol adjacent to the intercellular space of infected cells where mitochondria and plastids are located to the center of the cell where symbiosomes containing bacteroids can be found. Therefore, mitochondria are likely to experience pO2 well in excess of the Km(O2) for their terminal oxidases (50–100 nm; Bergersen, 1994; Thumfort et al., 1994), and, as long as sufficient oxidizable substrate is available, ATP synthesis is unlikely to be limited by O2 supply.

Nevertheless, at 0.65 ± 0.04, AEC of the plant fraction is lower than that normally associated with hypoxic tissues (approximately 0.75 or higher; Pradet and Raymond, 1983). To explain this fact, Kuzma et al. (1999) proposed that a steep ATP gradient may be present across the infected cells, from the mitochondrial zone (MZ) near the intercellular spaces (and the site of ATP synthesis) to the cytosol-surrounding symbiosomes where the majority of plant-derived ATP would be used.

To develop a framework for an experimental test of this hypothesis, the present study modified the model of Thumfort et al. (1994) by incorporating a separate MZ adjacent to the intercellular gas spaces (as in Thumfort et al., 1999) and by including diffusion of adenylates between sites of ATP synthesis in the MZ and sites of ATP utilization (and ADP and AMP production) in the bacteroid zone (BZ), i.e. within the cytosol surrounding the symbiosomes. The main ATP demand in the cytosol would be associated with Gln synthetase (GS) activity for the assimilation of fixed nitrogen (N) and the ATP cost of transporting C4 acids across the peribacteroid membrane (PBM) to fuel bacteroid respiration. Separate models are developed for nodules that produce amides (Asn) or ureides (allantoin and allantoic acid) as predominant end products of assimilation of fixed N.

The simulations were also used to explore the effect of Ar:O2 exposure on ATP demand and the resulting cytosolic adenylate and AEC gradients across the cell. In Ar:O2, N2 fixation ceases, but nitrogenase activity continues, and we predicted that there should be less of an ATP gradient in Ar:O2-treated nodules. If true, this would help account for the results of de Lima et al. (1994) who observed that Ar:O2 treatment (1 h) had no effect on whole-nodule AEC, despite the fact that nodule permeability declines and infected cell O2 concentration decreases.

RESULTS

Developing the Models

The infected cell was assumed to be a three-dimensional (3D) rhombic dodecahedron, which was reduced to a one-dimensional (1D) representation and divided into 400 layers. Layers 400 to 365 were considered to be within the MZ adjacent to the gas space interface, whereas layers 364 to 1 (ureide nodules) or 364 to 181 (amide nodules) were identified as the BZ. In the amide nodules, layers 180 to 1 constituted a central vacuole zone (VZ). The O2 was assumed to diffuse from the gas space interface (i.e. layer 400) into the innermost layer of the BZ (i.e. layer 1 in ureide nodules or layer 181 in amide nodules).

Mathematical models of O2 and adenylate diffusion were developed in two phases. In the first, rates of ATP demand and synthesis were calculated in the plant cytosol within both the BZ and MZ of infected cell. This was done by coupling the C, N, and energy demands of 400 cell layers to the modeled O2 concentration gradients that are thought to exist between the intercellular spaces and the center of the cell. The second phase of the models used these values for ATP demand to generate gradients of cytosolic ATP, ADP, and AMP across the infected cell.

For ureide- and amide-based models, the O2 concentration gradient was anchored by setting an O2 concentration at the innermost layer to achieve a volume-weighted average fractional oxygenation of leghemoglobin (Lb; FOLAvg) of 0.4. The ATP gradient was anchored by setting the cytosolic ATP concentration at the innermost layer such that the AEC in the cytosol adjacent to the gas space (AEC400) was 0.82.

The amide model differed from the ureide model in two biochemical parameters: (a) A demand of three ATPs per N (versus one per N for ureide nodules) would be required for NH3 assimilation in cytosol of BZ, and (b) a value of 78 nm for O2 concentration at which 50% of Lb is oxygenated (KsLb; versus 48 nm for ureide model).

O2, LbO2 Gradient, and O2 Consumption in the Infected Cell

In the ureide model, innermost O2 concentration (O21) was fixed at 13 nm to generate a FOLAvg of 0.4. The resultant O2 concentration adjacent to the space (O2400) was 400 nm, and the volume-weighted O2 concentration (O2avg) was predicted to be 38 nm (Fig. 1A, solid line). As a consequence, the O2avg values were slightly higher than the average O2 concentration calculated from the average fractional oxygenation of Lb [FOL; i.e. O2avg(FOL) = 32 nm; O2avg(FOL) = FOL × KsLb/(1 – FOL), KsLb = 48 nm], an observation consistent with previous reports (Thumfort et al., 1994, 1999).

Figure 1.

Figure 1.

Open in a new tab

Model predictions of the gradients in free O2 concentration (A), fractional oxygen of Lb (FOL; B), and O2 consumption in infected cells of ureide (solid lines) and amide (dotted lines) nodules under N2:O2 treatment (C), with values for the innermost O2 concentration (O2n = 1 or O2n = 181 for ureide and amide nodules, respectively) to give an average FOL (FOLAvg) of the cell = 0.4. The x axis shows the 400 layers along the diffusion path, from layer 400 (0-μm distance) to layer 1 (31.2 μm). The shaded areas show MZ (layers 400–365); BZ included layers 364 to 1 in the ureide-based nodule model and layers 364–181 in the amide model. Layers 180 to 1 in the amide model are occupied by central vacuole.

A steep O2 gradient was predicted to exist from the gas space interface to the center of the BZ (Fig. 1A). The decline in O2 concentration across the MZ, from 400 to 70 nm over a diffusion path of 2.8 μm, was much steeper than that in the BZ, where O2n declined from 70 to 13 nm over a path of 28.4 μm (Fig. 1A). A steep gradient in FOL was also predicted, declining from 0.90 to 0.22 across the infected cell (Fig. 1B).

The volume-specific respiration rates in the MZ (0.35–0.23 mol O2 m–3 s–1) were predicted to be considerably higher than those in the BZ (0.07–0.04 mol O2 m–3 s–1; Fig. 1C) as a result of the different kinetic constants for O2 uptake [i.e. Vmax and Km(O2)] were chosen for the two regions. The sharp discontinuity in respiration rate between the two zones (Fig. 1C) would be dampened if a thin, mixed zone of mitochondria and bacteroids was inserted between MZ and BZ, without altering the respiration rate in most of the layers in MZ and BZ. Total O2 consumption was 2.5 × 10–15 mol O2 s–1 cell–1 5.7 × 10–15 mol O2 s–1 cell–1 in the MZ and in the BZ.

The amide model predicted profiles for infected cell O2 concentration, FOL, and tissue respiration that were similar to those generated by the ureide model (Fig. 1, A–C, dotted lines). The innermost O2 concentration (O2181) was fixed at 32 nm to generate an FOLAvg of 0.4. The resultant O2 concentration adjacent to the space (O2400) was 600 nm. The volume-weighted O2 concentration (O2avg) was predicted to be 63 nm (Fig. 1A, dotted lines), and this value was slightly higher than the calculated O2avg(FOL) (52 nm) based on a KsLb value of 78 nm. In any layer between layer 400 and 181, the modeled O2 concentration was slightly higher in amide than in ureide nodules. By running a series of simulations, these differences were attributed to the higher value chosen for KsLb. Total O2 consumption was 2.7 × 10–15 mol O2 s–1 cell–1 in the MZ and 5.8 × 10–15 mol O2 s–1 cell–1 in the BZ.

Nitrogenase Activity and NH3 Production in Bacteroids

The models predicted a rate of NH3 production within the BZ (NH3n) that ranged from 0.030 to 0.014 mol NH3 m–3 s–1 in ureide nodules and from 0.033 to 0.024 mol NH3 m–3 s–1 in amide nodules (Fig. 2). Integrated over the entire infected cell, NH3 production was calculated to be 2.3 × 10–15 and 2.4 × 10–15 mol NH3 s–1 cell–1 in ureide and amide nodules, respectively.

Figure 2.

Figure 2.

Open in a new tab

Model predictions of nitrogenase activity in the production of NH3 in bacteroids of the infected cells of a ureide- (solid lines) and an amide- (dotted lines) based nodule under N2:O2 treatment. The nitrogenase activity in the MZ (shaded area) was zero.

Adenylate Demand in the Plant Fraction of the BZ

Large differences were observed in predicted ATP metabolic demand (PATPn) in the plant cytosol, with values for amide nodules being approximately twice that for ureide nodules (Fig. 3A), despite the fact that the models predicted similar rates of NH3 assimilation in the two nodule types (Fig. 2). These differences were attributed to a 3-fold higher ATP cost per NH3 assimilated for amide than ureide nodules.

Figure 3.

Figure 3.

Open in a new tab

Model predictions for ATP metabolic demand (PATP(metab)n; A and D), ATP demand by adenylate kinase (AK; PATP(AK)n; B and E), and the negative ratio of these two parameters (C and F) in the plant fraction of the infected cells of ureide- (solid lines) and amide- (dotted lines) producing nodules treated with the N2:O2 (A–C) or Ar:O2 (D–F). The shaded areas show the MZ (layers 365–400).

In the ureide nodule, total metabolic ATP demand Inline graphic in the plant fraction of BZ (ie, the sum of Inline graphic from the innermost layer of n = 3) for the entire cell was 5.4 × 10–15 mol ATP s–1 cell–1 where 42%, 54%, and 4% were used for NH3 assimilation (via GS), malate transport across the PBM to the bacteroids, and growth and maintenance respiration, respectively. In contrast, the amide model predicted a total metabolic ATP demand of 10.4 × 10–15 mol ATP s–1 cell–1, where 69%, 29%, and 2% were used for NH3 assimilation (via GS and Asn synthetase [AS]), malate transport across the PBM, and growth and maintenance respiration in the plant fraction of BZ, respectively.

Model estimates of metabolic ATP demand in each cell layer within the BZ were incorporated into a group of simultaneous equations describing the diffusion of cytosolic ATP, ADP, and AMP and the equilibrium between these pools as catalyzed by cytosolic AK. These simultaneous equations provided an estimate of the ATP “demand” [PATP(AK)n] associated with AK activity for each layer (n). Across most of the BZ, PATP(AK)n values for amide nodules were predicted to be more negative than in ureide nodules (Fig. 3B), denoting enhanced ATP regeneration and AMP production in the AK reaction (i.e. 2ADP → ATP + AMP). In the innermost part of the infected cell, the models predicted that ATP regenerated through AK activity was equivalent to 44% (ureide) or 49% (amide) of the metabolic ATP demand by those sites (Fig. 3C). Because AK converts two ADPs to one ATP and one AMP, these calculations indicate that in the innermost regions of the cell, 88% to 98% of the ADP produced through metabolic ATP demand would be substrate for AK.

It was noteworthy that the values of ATP “demand” by AK [PATP(AK)n] at layers 358 to 364 for amide nodules were predicted to be positive, implying that in these seven layers bordering the MZ, AK catalyzes the reaction in the direction of ATP + AMP → 2 ADP (Fig. 3B, dotted line). This was attributed to a model prediction that diffusion of AMP into these cell layers from the more centrally located layers of the BZ (see below) was sufficient to counter the activity of [PATP(metab)n] and reverse the AK reaction.

The total ATP “demand” by AK [Σinnermost (ie, the sum of PATP(AK)n from the innermost layer to n = 364)] in the plant fraction of the BZ for the entire cell was predicted to be 1.8 × 10–15 mol ATP s–1 cell–1 in the ureide nodule and 3.7 × 10–15 mol ATP s–1 cell–1 in the amide nodule. Dividing these values by the total metabolic ATP demand in the plant fraction of BZ (described previously), it was calculated that 33% (ureide) or 36% (amide) of the total ATP metabolic requirement was regenerated through AK activity.

Adenylate Demand in the MZ

The total metabolic ATP demand in the plant fraction of BZ for the entire cell was assumed to be met by oxidative phosphorylation within the MZ. The volume-specific metabolic ATP “demand” in layer “n” of the MZ [PATP(metab)n, moles ATP per meter cubed per second] was set as proportional to the contribution of that layer to the total O2 uptake and, therefore, was calculated to be –750 to –480 mmol ATP m–3 s–1 for ureide nodules and –1300 to –970 mmol ATP m–3 s–1 for amide nodules (Fig. 3A).

Estimates of metabolic ATP “demand” in each cell layer within the MZ were incorporated into a group of simultaneous equations describing the diffusion of cytosolic ATP, ADP, and AMP and the equilibrium between these pools as catalyzed by AK. These simultaneous equations provided an estimate of ATP “demand” by cytosolic AK [PATP(AK)n] for each layer “n” of MZ, which was predicted to be 220 to 180 mmol ATP m–3 s–1 for ureide nodules and 400 to 410 mmol ATP m–3 s–1 for amide nodules at layers 400 to 365 (Fig. 3B). The positive value of PATP(AK)n at layers 400 to 365 indicated that AK catalyzes the reaction in the direction of ATP + AMP → 2 ADP (Fig. 3B).

Diffusive Flux of Adenylates in Plant Fraction

Based on metabolic ATP demand and AK activity in maintaining adenylate equilibrium, the transfer of adenylates into layer n from layer n + 1 (moles per second) was solved at each layer (n). The predicted rates of adenylate transfer were converted into fluxes (nanomoles per meter squared per second) by dividing by the surface area of each layer. The resultant values, shown in Figure 4, A to C, provide a comparison of adenylate diffusion at various distances from the gas space interface (layer 400) to the innermost layer. The flux was positive for ATP but negative for ADP and AMP because ATP diffused from layer 400 to the innermost layer, whereas ADP and AMP diffused in the opposite direction.

Figure 4.

Figure 4.

Open in a new tab

Model predictions for the diffusive flux rates of ATP (A and D), ADP (B and E), and AMP (C and F) in the plant fraction of the infected cells of a ureide- (solid lines) and an amide- (dotted lines) producing nodule under N2:O2 (A–C) and Ar:O2. The shaded areas show the MZ (layers 365–400).

ATP flux between layers 399 and 364 of the MZ was predicted to increase from 39 to 716 nmol m–2 s–1 in the ureide nodule and from 69 to 1,330 nmol m–2 s–1 in the amide nodule (ATP flux was 0 at layer 400 according to the definition). In contrast, ATP flux between layers 364 and the innermost layer of BZ was predicted to decrease from 730 to 2 nmol m–2 s–1 in the ureide nodule and 1,350 to 4 nmol m–2 s–1 in the amide nodule. The highest ATP flux was predicted in layer 364, i.e. at the MZ:BZ interface for both ureide and amide nodules (Fig. 4A).

The fluxes of ADP and AMP showed a similar trend but in the opposite diffusion direction and are, therefore, negative in Figure 4, B and C. Similarly, the largest absolute value for ADP and AMP flux was observed at the MZ:BZ interface at layer 364 but with a lower magnitude than that predicted for ATP. At any layer within the infected cell, the sum of cytosolic ATP, ADP, and AMP fluxes equaled zero, reflecting the assumption of models that the total cytosolic adenylate concentration was constant throughout the infected cell.

Cytosolic Adenylate Gradients in Plant Fraction

The simultaneous equations also solved for concentrations of cytosolic ATP, ADP, and AMP in each layer of the infected cell. To ensure that the models predicted an AEC of 0.82 for AEC400, the innermost ATP (ATP1) had to be set to 1.04 mm in the ureide model and 0.79 mm in the amide model. The cytosolic ATP, ADP, and AMP gradients across the infected cells were predicted to be greater in amide than in ureide nodules (Fig. 5, A–C).

Figure 5.

Figure 5.

Open in a new tab

Model predictions for concentration gradients of cytosolic ATP (A and D), ADP (B and E), and AMP (C and F) in the plant cytosol of the infected cells of ureide- or amide-based nodules treated with N2:O2 (A–C) and Ar:O2 for 5 to 7 min (D–F). The shaded areas show the MZ (layers 365–400).

These cytosolic adenylate gradients resulted in predictions of large gradients in AEC (Fig. 6A) or ATP to ADP ratio (Fig. 6B) for both nodule types but much steeper gradients in amide than in ureide nodules. Although model parameters were chosen to give an AEC of 0.82 (ATP:ADP = 3.1) at the gas space interface (n = 400), the AEC predicted to exist in the cytosol at the innermost layer was 0.62 (ATP:ADP = 1.4) in the ureide nodule and 0.52 (ATP:ADP = 1.0) in the amide nodule. The average AEC (AECAvg) of the entire plant fraction of the cell (BZ and MZ) was predicted to be 0.61 in the amide nodules and 0.70 in the ureide nodules. The latter values were similar to the experimentally determined value for soybean (Glycine max) nodules (0.65 ± 0.04; Kuzma et al., 1999).

Figure 6.

Figure 6.

Open in a new tab

Model predictions of AEC (A and C) and ATP to ADP ratio (B and D) in the plant fraction of the infected cells of ureide- or amide-based nodules treated with N2:O2 (A and B) and Ar:O2 for 5 to 7 min (C and D). The shaded areas show the MZ (layers 365–400). The volume-weighted average AEC (AECAvg) associated with each form of assimilation are shown with arrows on the figures.

Effect of Ar:O2 Treatment on Predictions of Models

The models were also used to simulate the effect of exposing ureide- or amide-forming nodules to an Ar:O2 atmosphere, a treatment that prevents N2 fixation and NH3 assimilation without immediate effects on nitrogenase activity or nodule carbohydrate metabolism. Extended exposure to Ar:O2 is known to decrease the nodule's permeability to O2 diffusion (Hunt and Layzell, 1993), but the simulations carried out here were focused on the physiological status of the nodule after a few minutes of Ar:O2 exposure (5–7 min), when the rate of NH3 assimilation was reduced to zero but before the diffusion barrier responds to reduce the incoming flux of O2 (King and Layzell, 1991).

Cessation of NH3 production in bacteroids would reduce ATP demand for NH3 assimilation in the plant fraction. Therefore, the models predicted much lower ATP demand [PATP(metab)n] in Ar:O2-treated (Fig. 3D) than in N2:O2-treated nodules (Fig. 3A). As a consequence, the AK activity indicated by its ATP demand [PATP(AK)n] was also predicted to be lower in Ar:O2-treated nodules (Fig. 3E).

As a result of lower adenylate demand, diffusive fluxes of adenylate were predicted to be much lower in Ar:O2 (Fig. 4, D–F) than in N2:O2 (Fig. 4, A–C). Although parameters of models were chosen to give an AEC of 0.82 (ATP:ADP = 3.1) in layer 400, the simulation showed that the cytosolic adenylate gradients predicted during steady-rate N2 fixation in N2:O2 were significantly reduced after the switch to Ar:O2 (compare Fig. 5, D to F with A to C). The average AEC in the plant fraction of an infected cell was 0.77 in both ureide and amide nodules treated with Ar:O2 (Fig. 6C). Thus, the ureide model predicted a rise of average plant AEC from 0.70 in N2:O2 to 0.77, whereas the amide model predicted a rise from 0.61 in N2:O2 to 0.77.

DISCUSSION

Cytosolic Adenylate Gradients in Legume Nodules

The models predicted that substantial gradients in adenylate concentration could occur in the plant cytosol of the infected cell, even in the absence of membrane barriers. Both adenylate and AEC gradients would be reduced dramatically after short-term exposure of nodules to an Ar:O2 atmosphere, a treatment that stops N2 fixation and NH3 assimilation but not nitrogenase activity.

The predictions from these models are the first of their kind, to our knowledge, and may account for a number of physiological observations that to date have defied easy explanation. For example, Oresnik and Layzell (1994) have reported low values for AEC (i.e. less than 0.8), even in metabolically active, N2-fixing nodules. Using a nonaqueous fractionation technique, Kuzma et al. (1999) showed values for the plant AEC of soybean nodules as low as 0.65 ± 0.04, but when nodules were exposed to a severe O2 limitation, the plant AEC was unaffected, whereas bacteroid AEC declined sharply from 0.76 to 0.56. These findings indicated that bacteroids were the site of O2 limitation but did not provide an explanation for the low AEC in the plant compartment. A low AEC value of 0.70 was also observed by Sa and Israel (1991) in plant fraction of soybean nodules, whereas the AEC values in bacteroids and whole nodules were measured to be 0.84 and 0.75, respectively. Results of the current study readily account for these earlier observations by predicting steep AEC gradients across the infected cell. Even when the AEC at the interface with the gas-filled spaces was set at 0.82, the average AEC within the plant compartment of the infected cell was predicted to be 0.70, a value consistent with that measured in soybean nodules (0.70, Sa and Israel, 1991; 0.65 ± 0.04, Kuzma et al., 1999).

Results of this study may also account for the observation of de Lima et al. (1994) that extended exposure to an Ar:O2 atmosphere differs from all other treatments known to reduce infected cell O2 concentration (e.g. low O2 concentration and NO3– treatment and carbohydrate deprivation) in that it was not correlated with a decrease in whole-nodule AEC. The predictions of models from the current study indicate that, unlike these other treatments, Ar:O2 exposure would dramatically reduce the adenylate gradient across the infected cell and, in effect, increase AEC in the plant compartment. This may offset a decrease in bacteroid AEC that would be expected when the O2 permeability of these nodules decreases and the infected cell becomes severely limited by O2 supply.

Validity of the Predicted Rates of Respiration and Nitrogenase Activity

To test the validity of models, the predicted rates of respiration and NH3 assimilation in the infected cell were converted and then compared with measured values.

The models predicted a rate of mitochondrial respiration of 2.5 × 10–15 to 2.7 × 10–15 mol O2 s–1 cell–1 in ureide and amide nodules. Assuming 12,472 mitochondria per cell and 0.25 pg of protein per mitochondrion (Table IV of Millar et al., 1995), the mitochondrial respiration rate was equivalent to 48 to 51 nmol O2 min–1 mg–1 protein, which was comparable with reported data for isolated nodule mitochondria (35–140 nmol O2 min–1 mg–1 protein for ureide-producing soybean nodules, Millar et al., 1995; 25 nmol O2 min–1 mg–1 protein for amide-producing lupin nodules, Tomaszewska et al., 1991). The predicted rate of NH3 production was 2.3 × 10–15 to 2.4 × 10–15 mol NH3 s–1 cell–1 in ureide and amide nodules in N2:O2 (apparent nitrogenase activity). When this value was scaled to a whole nodule using an electron allocation coefficient of 0.67, the total nitrogenase activity was estimated to be equivalent to 310 and 340 μmol H2 g–1 dry weight (nod) h–1 for ureide and amide nodules (nod), respectively. These values were similar to those observed in intact, undisturbed nodules (i.e. 200–350 μmol H2 g–1 dry weight [nod] h–1 for ureide nodules, Hunt and Layzell, 1993; 367 ± 31 μmol H2 g–1 dry weight [nod] h–1 for amide-forming lupin nodules, Diaz del Castillo et al., 1992). The above conversions were based on: (a) for ureide nodules, 1.8 × 105 cells and 9.5 × 10–3 g dry weight per soybean nodule (Bergersen, 1994); and (b) for amide nodules, 1.1 × 105 cells per nodule (calculated from a nodule volume of 2.2 × 10–8 m–3, Fernandez-Pascual et al., 1992; assuming 55% of nodule volume is occupied by CZ, Oresnik and Layzell, 1994; infected cell volume of this work) and 5.7 × 10–3 g dry weight per lupin nodule (Fernandez-Pascual et al., 1992).

Validity of 1D Models Compared with 3D Models

The method used in this study was largely based on 1D models (Thumfort et al., 1994; Thumfort, 1996) that retained the essential features of a polyhedron cell with respect to O2 diffusion from gas-filled intercellular spaces. To address the concern that 1D models of O2 diffusion may be oversimplified and, therefore, their predictions compromised, Thumfort (1996) and Thumfort et al. (2000) constructed full 3D models of O2 diffusion and compared their predictions with those of the 1D models. The predicted values for the two simulations differed by less than 15%. Thus, the 1D modeling approach proved to be a simple yet powerful and valid tool for simulating O2 diffusion into the infected cell of a legume nodule.

For simplicity, the 1D models developed in this study did not consider O2 diffusion into and out of the adjacent uninfected cell. Based on a two-cell model prediction, the amount of O2 entering the infected cell from an adjoining uninfected cell accounted for only 15% of the total amount of O2 entering when average FOL was from 0.4 to 0.5 (Thumfort et al., 1999).

Assumptions for Anchoring the Gradients of O2 and Adenylates

The O2 concentration gradient in the models was anchored by setting an O2 concentration at the innermost layer to achieve an average FOL of 0.4, a value measured by nodule oximetry and thought to exist in active, undisturbed ureide and amide nodules (Table I, item 1). Increasing the average FOL from 0.4 to 0.6 would significantly increase O2400 from 400 to 7,600 nm (Table II, item 1, a and b); this range of O2 concentrations was still in agreement with that in a two-cell model of Thumfort et al. (1999). The resultant plant AEC gradients would become steeper (Table II, item 1, a and b). Increasing the average FOL from 0.4 to 0.6 had a similar impact on amide nodules (data not shown).

Table I.

Biochemical and physiological parameters used in the models

Parameter (abbreviation) Value and Units References
1. FOL in ureide- and amide-forming nodules 0.4 King and Layzell (1991); Kuzma et al. (1993); Shimada et al. (1997)
2. Diffusion coefficient for Lb and LbO2 in cytosol 5.9 × 10-11 m2 s-1 Bergersen (1993)
3. Diffusion coefficient for O2 in cells 7.4 × 10-10 m2 s-1 Denison (1992)
4. Diffusion coefficient for adenylate in cytosol 1.10 × 10-10 m2 s-1 Kushmerick and Podolsky (1969)
5. Tortuosity factor for diffusion of Lb and adenylate 1.7 Satterfield and Sherwood (1963)
6. P:O ratio for bacteroid oxidative phosphorylation 2 Wittenberg et al. (1974); p. 4065)
7. Ks(O2) for Lb oxygenation (KsLb) 4.8 × 10-5 mol m-3 in ureide nodules; 7.8 × 10-5 mol m-3 in amide nodules Gibson et al. (1989)
8. Apparent equilibrium constant of adenylate kinase in plant fraction (KAK) 0.9 (unitless) Calculated from Kuzma et al. (1999)
9. Total cytosolic adenylate concentration (Aden) 2.29 mol m-3 Calculated from Oresnik and Layzell (1994) and Kuzma et al. (1999)
10. Vmax for respiration 0.40 mol m-3 s-1 in MZ; 0.096 mol m-3 s-1 in BZ Bergersen (1994); Thumfort et al. (1999)
11. Ks(O2) for respiration 5.3 × 10-5 mol m-3 in MZ; 2.0 × 10-5 mol m-3 in BZ Bergersen (1994); Kuzma et al. (1993); Bergersen and Turner (1993)
12. Total cytosolic Lb concentration 3.9 mol m-3 Bergersen (1993)
Open in a new tab

Table II.

Effect of changing input parameters for the models on the predicted values for O2 concentrations at the gas space interface (O2400) and innermost layer (O2min), average O2 concentration of cell (O2Avg), FOL at the gas space interface (FOL400), average FOL of cell (FOLAvg), O2 respiration in mitochondria and its percentage in the whole cell's respiration, cytosolic AEC at the gas space interface (AEC400) and innermost layer (AECmin), and average cytosolic AEC of cell (AECAvg)

All simulations assume an AEC400 of 0.82.

Item No. and Parameter Changed
Predicted Values
O2400 O2min O2Avg FOL400 FOLAvg Mito Respirationa AEC400 AECmin AECAvg
nm mol O2 s-1 cell-1 × 10-15
1. Ureide nodule model
    (a) Standard parameters and predictionsb 400 13 39 0.90 0.40 2.5 (31%) 0.82 0.62 0.70
    (b) Increased average FOL to 0.6 7,600 26 205 0.99 0.60 3.5 (32%) 0.82 0.56 0.66
    (c) Decreased P:O ratio from 2 to 1.3 400 13 39 0.90 0.40 2.5 (31%) 0.82 0.66 0.72
2. Amide nodule model
    (a) Standard parameters and predictionsb 600 32 63 0.88 0.40 2.9 (32%) 0.82 0.52 0.61
    (b) Decreased P:O ratio from 2 to 1.3 600 32 63 0.88 0.40 2.9 (32%) 0.82 0.60 0.67
    (c) Increased KsLb from 78 to 127 nm 1,300 46 103 0.91 0.40 3.2 (33%) 0.82 0.48 0.58
    (d) Decreased KsLb from 78 to 48 nm 210 21 36 0.81 0.40 2.3 (31%) 0.82 0.57 0.64
    (e) Increased Aden from 2.29 to 3.29 mm 600 32 63 0.88 0.40 2.9 (32%) 0.82 0.61 0.67
    (f) Decreased Aden from 2.29 to 1.29 mm 600 32 63 0.88 0.40 2.9 (32%) 0.82 0.29 0.44
Open in a new tab
a

The value in parentheses (31%-33%) indicates the percentage of O2 respiration in mitochondria accounting for O2 respiration of whole cell, which is lower than that predicted value of 56% in Bergersen's model (Table 5 of Bergersen, 1996). The latter value may reflect an overestimation of mitochondrial respiration (Thumfort, 1996). b Standard parameters and predictions as presented in Figures 1 to 6.

The cytosolic ATP gradient in the models was anchored by choosing the cytosolic ATP concentration at the innermost layer such that the AEC400 was 0.82, a value similar to that measured in fully aerobic nodule cortex tissue (Oresnik and Layzell, 1994). To test the effect of this assumption, AEC at the intercellular space was set at a range of 0.72 to 0.92. This had no effect on the magnitude of the predicted gradient in AEC; an effect was only observed on the absolute values for AEC across the infected cell.

Validity of Physiological Parameters

The diffusive path for O2 was defined in both the cytosol and symbiosome of the infected cell, whereas the diffusive paths for LbO2 and plant ATP, ADP, and AMP were restricted to the cytosol. Diffusion coefficients (D) of LbO2, O2, and ATP in cell (Table I, items 2–4) used in the simulation were expected to have an impact on the predicted features of the models. The reported D values for LbO2, O2, and ATP in water were 12.5 × 10–11, 17 × 10–10, and 2.70 × 10–10 m2 s–1 (Kushmerick and Podolsky, 1969; Bergersen, 1993), respectively. The calculated DWater:DCell ratio for these three molecules was 2.11, 2.30, and 2.45, respectively. Assuming the effective radius of a diffusing molecule does not vary between pure water and the cytosol, DWater:DCell would reflect the relative viscosity of cytosol, resulting in similar values for different diffusing molecules (Kushmerick and Podolsky, 1969). Because D of LbO2 (Table I, item 2) was measured in the cytosol of an infected cell of a nodule, the similarity in DWater:DCell (2.11–2.45) for LbO2, O2, and ATP permits confidence in the D values used in the models. Simulations in which D values were varied by ±30% only slightly changed the predictions in both ureide and amide models (data not shown).

The models used a P:O ratio of 2 (Table I, item 6) to couple bacteroid O2 consumption rate to the production of ATP in bacteroids of BZ, which was then used to generate the rate of nitrogenase activity. Wittenberg et al. (1974) suggested that the P:O ratio in bacteroids might vary in the range of 2 to 1 with an average of 1.3. Simulations with a P:O of 1.3 would reduce the rate of nitrogenase activity to 1.4 × 10–15 and 1.5 × 10–15 mol NH3 s–1 cell–1 (apparent nitrogenase activity) in ureide and amide nodules, respectively. The equivalent rates of total nitrogenase activity would be 190 to 200 μmol H2 g–1 dry weight (nod) h–1, values at the lower limit but still comparable with those reported for nodules in vivo (Hunt and Layzell, 1993). Adopting a P:O of 1.3 the predicted cytosolic adenylate gradients were reduced in magnitude but were still generated (Table II, items 1c and 2b).

It is noteworthy that the P to O ratio was not used to calculate the ATP production in mitochondria in MZ (layers 400–365). Instead, the metabolic ATP “demand” (i.e. production) in each layer of MZ was assumed to be proportional to the contribution of that layer to the total O2 uptake within the MZ.

Impact of Changes in Vacuole Volume, KsLb, KAK, and Aden and in the Amide Model

The amide model assumed a large central vacuole in the infected cells, accounting for 12% of whole cell volume (i.e. layers 180–181 of the cell). Increasing VZ volume to 20% shortened the diffusion pathway in the BZ but did not significantly reduce the adenylate gradients (data not shown). Pugh et al. (1995) reported that the fraction of VZ volume was increased from 20% to 74% in infected cells of white clover (Trifolium repens) nodules by waterlogging. A larger central vacuole may make O2 in the intercellular space more accessible to the bacteroids, thus increasing the nodule's tolerance to waterlogging. A simulation of an infected cell in an amide nodule with a fraction of VZ as high as 74% essentially abolished both the O2 and adenylate gradients (data not shown).

A KsLb value (the O2 concentration at which 50% of Lb is oxygenated) of 78 nm (Table I, item 7) was used for the amide model. This value was originally determined for lupin Lb II and was between the reported KsLb for soybean Lb (48 nm) and pea (Pisum sativum) Lb IV (127 nm, Kawashima et al., 2001). Changing KsLb over this range (i.e. 48–127 nm) still predicted a substantial gradient of plant AEC (Table II, item 2, a, c, and d).

The apparent equilibrium constant of cytosolic AK (KAK = 0.90; Table I, item 8) was based on measurements from ureide-producing nodules. The reported values of KAK in plant tissues were in the range of 0.3 to 1.5 (Igamberdiev and Kleczkowski, 2001). For the amide model, increasing KAK to 1.5 would slightly change the ATP gradient to 1.65 mm in layer 400 and 0.86 mm in layer 181. In contrast, decreasing KAK to 0.3 would predict a steeper ATP gradient ranging from 1.55 mm in layer 400 to 0.64 mm in layer 181.

The value for total cytosolic adenylate concentration (Aden, 2.29 mm; Table I, item 9) was based on measurements from ureide-producing nodules. For amide model, increasing the value for Aden would reduce the magnitude of adenylate gradients (Table II, item 2, a and e). In contrast, decreasing the value for Aden would cause the adenylate gradients to become steeper (Table II, item 2, a and f), consistent with the infected cell requiring steeper adenylate gradients to meet the ATP demand in the inner part of cell.

Ammonia Assimilation Pathways in Nodules and Implications for the Models

The central assumptions about assimilation of fixed N in the present models was that NH4+ was the sole N solute transferred from bacteroids to the plant cytosol and that it was incorporated exclusively into the amide group of Gln through cytosolic GS. Further, the ultimate products of N2 fixation were solely ureides in one and Asn in the other type of nodule. There have been a number of proposals that amino compounds may be exported from bacteroids (for review, see Day et al., 2001), and recently, evidence has been presented (Waters et al., 1998) for a significant flux of fixed-N as Ala from soybean bacteroids. Even if this were to be the case in vivo, Asn and ureide synthesis depend on the amide group of Gln, and Ala-N would need to be transferred, most likely through a cytosolic pool of NH4+, to the amide; thus, the ATP demand by GS would be unchanged. 15N2 labeling of cowpea (Vigna unguiculata) nodules (a ureide-based symbiosis) indicates a flow of N to Ala, but the labeling kinetics are not consistent with this N being a precursor for the amide group of Gln (Atkins and Thumfort, 2002).

If the amidotransferases of either purine (phosphoribosyl pyro-phosphate or 5′-phosphoribosyl-N-formylglycinamide amidotransferases) or AS could utilize NH4+ directly in nodules, and then ATP demand by GS would be reduced. Although Vance (2000) reported that AS from alfalfa (Medicago sativa) nodules could use NH4+ in vitro, unreasonably high concentrations were needed, and similar observations were made for de novo purine synthesis by cowpea nodule extracts (Atkins et al., 1982). Thus, it is reasonable to assume that the amide-N of Gln, formed by cytosolic GS, is the substrate in both cases. However, the sites at which Gln-N is used are different in the two types of nodule. In the case of Asn synthesis, AS is soluble in the cytosol of the infected cell (Shelp and Atkins, 1984) and, like GS, is assumed to be functional in the BZ. Purines, on the other hand, are formed in both plastids and mitochondria in the MZ (Smith and Atkins, 2002).

ATP is used in the MZ for purine synthesis, but this is unlikely to significantly affect the simulated adenylate gradients generated by the ureide model. Plastids are distributed together with mitochondria adjacent to the intercellular space. These organelles are assumed to be distributed evenly across the MZ and that in each layer, the ATP demand in plastids would be met by ATP produced by the mitochondria located in the same layer. Therefore, the extra ATP use in plastids would not affect the adenylate gradients generated by diffusion from mitochondria toward the inner cytosol.

ATP Gradients and the Regulation of Biochemical Activities in Infected Cells

It is known that cytosolic ATP level can regulate the activity of specific glycolytic enzymes, including phosphofructokinase, pyruvate kinase, hexokinase, enolase, and phosphoenolpyruvate carboxylase in plants (Plaxton, 1996). For example, in chickpea (Cicer arietinum) nodules, cytosolic phosphofructokinase was disaggregated in the presence of high ATP, resulting in a lower affinity for its substrate, Fru-6-phosphate (Lee and Copeland, 1996). Large gradients across an infected cell would create the potential for very dynamic and complex interactions in metabolic regulation.

ATP “demand” by AK [PATP(AK)n] was negative in most layers of BZ (Fig. 3B), indicating the AK reaction was in the direction 2ADP → ATP + AMP, thereby regenerating 33% to 36% of the total ATP metabolic demand in the plant fraction. Such activity of AK could be a strategy for the cell to reduce the ATP diffusive flux from mitochondria to the inner cytosol. In contrast, the predicted ATP “demand” by AK [PATP(AK)n] in MZ was positive, indicating that the AK reaction was in the opposite direction, thereby helping provide ADP for mitochondrial oxidative phosphorylation.

The pyrophosphate (PPi) cleavage of ATP by AS raises one further possibility that might alter the ATP dynamics assumed in the models. If PPi is substrate for an H+-pyrophosphatase (PPase) in the PBM, then its activity could generate sufficient proton motive force for malate transport, eliminating the need for ATP to drive an alternate H+ pump. However, in PBM preparations from soybean, there is no evidence for a PBM-localized PPase (D. Day, personal communication). Thus, the models assume that PPi generated by AS is cleaved by a PPase such that there are no consequences for the adenylate dynamics of the BZ.

A Possible Experimental Test of the Models

The simulations with Ar:O2 indicate a rationale to test the models experimentally. Both amide and ureide models predict that the cytosolic adenylate gradients will decrease in Ar:O2, resulting in large increases in the average AEC of the plant fraction as soon as the N pools are depleted. This could be tested experimentally using a nonaqueous fractionation technique (Kuzma et al., 1999) to separate and analyze adenylate pools in plant and bacteroid fractions before and after Ar:O2 exposure. If the predictions of the models are realistic, brief (<8 min), exposure to Ar:O2 should increase AEC in the plant fraction with minimal effect on bacteroid AEC. Then, as the nodule's O2 gaseous permeability is decreased, there should be a large decline in AEC of the bacteroid fraction but only a minor effect on AEC in the plant fraction, resulting in little change from the AEC of whole nodules treated with air.

MATERIALS AND METHODS

Cell and Tissue Geometry

The central, bacteria-infected zone of nodules is assumed to consist only of tightly packed infected cells (the actual ratio of infected cell:uninfected cell volume = 10:1; Dakora and Atkins, 1990; Thumfort et al., 1999), each in the shape of a rhombic dodecahedron (Fig. 7A) with a side length of 33 μm and a volume of 1.1 × 10–13 m3 (Dakora and Atkins, 1990). The edges of each plane of the cell were modified by removal of an “arc” representing one-third of a circle so as to create a network of cylindrical gas-filled spaces (total volume = 0.98% of cell) surrounding each cell. This resulted in a gas space interface of 1719 μm2 cell–1 and a ratio of gas space interface:infected cell volume of 0.0156 μm2 μm–3 (Dakora, 1989). These parameters were similar to those used previously (Thumfort et al., 1994, 1999, 2000).

Figure 7.

Figure 7.

Open in a new tab

Geometry and surface area profile of an infected cell modeled as a rhombic dodecahedron. A, Diagrammatic representation of the infected cell as a rhombic dodecahedron. The gas space is modeled by truncating the 24 edges of the dodecahedron to create a network of cylindrical (in cross section) tubes. The O2 diffuses from these spaces into the cell. The shaded cut-out region depicts the shape and volume of the symmetrical units that were modeled. B, Surface area profile derived for the rhombic dodecahedron cell described above. The shaded area represents the MZ (layers 400–365), whereas the BZ occupies layers 364 to 1 in ureide nodules and layers 364 to 181 in amide nodules, and the VZ occupies layers 180 to 1.

To model the adenylate diffusion within the infected cell, the method of Thumfort et al. (1994, 2000) and Thumfort (1996) was used to derive a 1D representation, i.e. the “surface area versus distance profile” of a 3D cell assumed to be a rhombic dodecahedron (Fig. 7B). The infected cell was divided into two or three zones: a MZ near the gas-filled intercellular spaces that contained mitochondria, plastids, and plant cytosol, and a BZ that contained bacteroids and cytosol and that occupied the rest of cell volume in ureide nodules (Thumfort et al., 1999) or all but the innermost 12% of cell volume in amide nodules that was assumed to be the VZ (Pugh et al., 1995). The cytosolic fraction occupied 15% of the volume of both MZ and BZ.

The 31.2-μm distance from the gas space interface to the innermost point in the cell was divided into 400 layers, each with an identical depth (0.078 μm) and with the relative surface area shown in Figure 7B. Layers 400 to 365, occupying a depth of 2.8 μm, were considered to be within the MZ adjacent to the gas space interface, whereas layers 364 to 1 (ureide nodule) or 364 to 181 (amide nodule) were identified as the BZ. In the amide nodules, layers 180 to 1 constituted a central VZ. This partitioning resulted in the volume ratio of MZ:BZ = 8.5%:91.5% in the infected cell of ureide nodules (Bergersen, 1994; Thumfort et al., 1999) or the volume ratio of MZ:BZ:VZ = 8.5%:79.5%:12% in that of amide nodules.

A five-step process (Fig. 8) was used to model cytosolic adenylate gradients in the infected cells of ureide and amide nodules. A detailed description of the methods is described in the supplemental material or from the authors.

Figure 8.

Figure 8.

Open in a new tab

Scheme showing a five step process to simulate the O2 gradient, the NH3 production and malate consumption in bacteroids, the cytosolic ATP demand, and the production and diffusion of adenylates throughout the mitochondrial and BZs in the infected cell of a nodule. The numbers in the figure refer to the equation numbers associated with the calculation of the models, as described in the supplemental material that are available from the authors. In each layer within the BZ (layer 364 to innermost), O2 is consumed by bacteroids to produce ATP, which is used for growth and maintenance respiration (G&M), N2 fixation, and H2 evolution. In the plant cytosol of the BZ, net ATP consumption occurs through ammonium assimilation, growth and maintenance respiration (G&M), and H+ pumping that is coupled to malate (Mal) transport into bacteroids. The ammonia assimilation costs for amide nodules was assumed to have an additional cost associated with the assimilation of ASN (dashed line). In both ureide and amide nodules, the ATP, ADP, and AMP pools were equilibrated by an AK in simultaneous equations that included the diffusion of the adenylates to or from adjacent layers. The production of ATP used in the plant fraction was assumed to occur within the MZ (layers 400–365) of the infected cell.

Step 1. O2 Diffusion and Consumption across the Infected Cell

The modeling of O2 diffusion (Fig. 8, item 1) into the cells was similar to that described by Thumfort et al. (1994). However, the infected cells were assumed to have two metabolically active zones (MZ and BZ), each with separate kinetic constants for O2 uptake [i.e. Vmax and Km(O2); Table I, items 10 and 11]. The ureide and amide models were also assigned different half-saturation constants for Lb oxygenation (Table I, item 7).

Equations were derived that described Lb equilibrium with free O2, the diffusion of O2 and LbO2 between layers in the infected cell, and the respiratory consumption of O2 by the bacteroids or mitochondria within each layer (Eq. 1–4 in Thumfort et al., 1994). These equations were combined to derive a quadratic equation that permitted the calculation of O2 concentration in layer n + 1 (O2n+1, moles per meter cubed) from parameters obtained for layers n and n – 1.

The model was built in Excel 98 (Microsoft, Redmond, WA) by first selecting an O2 concentration for the innermost layer (O2n, n = 1 for ureide nodules, n = 181 for amide nodules, moles per meter cubed; layers 180 to 1 in amide nodules were occupied by vacuole) and then calculating the corresponding oxygenated Lb concentration and O2 consumption rate for that layer. The quadratic equation was used to calculate the O2 concentration in the layer adjacent to the innermost layer (i.e. O2n+1) that the O2 demands of the innermost layer were provided through layer (n + 1). Values were then derived for the oxygenated Lb concentration, the O2 consumption rate, and the required transfer of O2 from layer n + 2. This process was repeated for all layers from n + 2 to 400.

In each layer, the FOL was calculated as the LbO2 concentration divided by the total cytosolic Lb concentration (Table I, item 12). Model simulations were carried out for a range of innermost O2 concentrations (O2n = 5–30 nm O2), and for each, values were calculated for the volume-weighted average FOL (FOLAvg).

Step 2. Bacteroid Metabolism and NH3 Production

Given that the bacteroid metabolism is limited by O2 (Kuzma et al., 1999), the uptake of O2 by bacteroids within each layer in the BZ was used to calculate the rate of ATP production (Fig. 8, item 3) in the bacteroids assuming a P:O ratio of 2.0 (Table I, item 6). After assigning a portion of the ATP production to bacteroid growth and maintenance (Fig. 8, item 4), the remaining ATP production in each layer was used to support nitrogenase activity (Fig. 8, item 5). In air (N2:O2 = 79:21 [v/v]), it was assumed that nine ATPs were required to reduce 0.5 N2 gas to one NH3 and 0.75 H2 (i.e. electron allocation coefficient of 0.67). However, under an Ar:O2 (79:21 [v/v]) atmosphere, all of the ATP demand by nitrogenase was assumed to involve H2 gas production, and no NH3 was produced (Layzell and Atkins, 1997).

Step 3. ATP Demand for NH3 Assimilation and Malate Transport in Plant Cytosol of BZ

The NH3 produced by the bacteroids was assumed to diffuse into the plant cytosol, where it was assimilated into amino acids by GS and Glu synthase, in ureide nodules or by GS, Glu synthase, and AS in amide nodules. The ATP requirements for GS (one ATP/NH3) and AS (two ATP/NH3) were assumed to be synthesized by the mitochondria in the MZ, and diffuse through the cytoplasm to the layers in which NH3 was produced. The ATP demand along the diffusion pathway included that for the transport of malate into the symbiosomes (Fig. 8, item 6) and the plant growth and maintenance (Fig. 8, item 7), in addition to the cost of NH3 assimilation.

The ATP demand for malate transport across the symbiosome membrane (Fig. 8, item 6) was calculated assuming one ATP per malate transported. To calculate malate transported into the symbiosomes, values were generated to account for: (a) the C requirement for bacterial growth (relative growth rate = 6.49 × 10–7 g dry weight new g–1 dry weight s–1; Bouma et al., 1997), and (b) the requirement for reducing power (12 e–/malate) to satisfy the demands for bacteroid O2 consumption (4e–/O2) and nitrogenase activity (4.5 e–/NH3) calculated above.

Step 4. Adenylate Diffusion Gradients in Plant Cytosol of BZ

Estimates of the ATP demand within the plant fraction of each layer [PATP(metab)n] of the BZ were used to generate the diffusion gradients of adenylates across the cells. The total cytosolic adenylate pool (Aden; Table I, item 9) was assumed to be constant for each layer, and within each layer (n), an AK was assumed to maintain a balance among the adenylate pools according to the equation of ATPn × AMPn = KAK × (ADPn)2, where KAK is the apparent equilibrium constant of AK (Table I, item 8). The ATP demand for AK activity [PATP(AK)n] was generally negative because the AK was needed primarily to build the AMP pool, which involves the production of ATP from 2ADP.

The calculations of the cytosolic adenylate gradient were begun by first choosing a value for the cytosolic ATP concentration for the innermost layer (range of 0.01–2.29 mol m–3, determined according to the value of Aden; Table I, item 9). Then, a series of equations were used to calculate the diffusion of cytosolic ATP, ADP, and AMP (Fig. 8, item 11, a–c) between adjacent layers while ensuring that the adenylate pools were in equilibrium via the AK and assuming each layer met its respective demand for ATP.

Step 5. Continuity of Adenylate Diffusion Gradients in Plant Cytosol of MZ

Calculation of the cytosolic ATP gradients across the MZ assumed that this zone was responsible for providing the ATP needs of the plant fraction in the entire BZ. In the MZ, as in the BZ, cytosolic AK was assumed to maintain a balance among the adenylate pools.

The contribution of each MZ layer to the total ATP demand of plant fraction of BZ was assumed to be proportional to the contribution of that layer to the total O2 uptake within the MZ. This permitted the calculation of the volume-specific metabolic ATP “demand” in each layer of the MZ [PATP(metab)n; Fig. 8, item 15].

Although formation of ureides relies on purine synthesis, which itself requires ATP (Smith and Atkins, 2002), this demand was considered to be met within the plastids and mitochondria and, therefore, to have no impact on the demand for ATP within the cytosol.

Given these values, a series of equations were used to calculate the diffusion of cytosolic ATP, ADP, and AMP between adjacent layers while ensuring that the adenylate pools were in equilibrium. Finally, the AEC [AEC = ([ATP] + 0.5 × [ADP])/([ATP] + [ADP] + [AMP])] was calculated for each layer, and these values were used to calculate the volume-weighted average AEC for the plant cytosol of entire cell.

For both ureide and amide models, the cytosolic ATP concentration for the innermost layer (initialized in Step 4) was set at a value that resulted in an AEC at the gas space (AEC400) of 0.82, a value representing fully aerobic tissue (Oresnik and Layzell, 1994).

Acknowledgments

The conceptual basis for this modeling exercise was developed in collaboration with Dr. Peter Thumfort (Department of Biology, Massachusetts Institute of Technology, Cambridge).

Article, publication date, and citation information can be found at http://www.plantphysiol.org/cgi/doi/10.1104/pp.103.032318.

1

This work was supported by the National Science and Engineering Research Council of Canada (grants to D.B.L.).

References

  1. Atkins CA, Ritchie A, Rowe PB, McCairns E, Sauer D (1982) De novo purine synthesis in nitrogen-fixing nodules of cowpea (Vigna unguiculata [L.] Walp.) and soybean (Glycine max [L.] merr.). Plant Physiol 70: 55–60 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Atkins CA, Thumfort PP (2002) Assimilation of fixed-N in a ureide-forming symbiosis. In TM Finan, MR O'Brian, DB Layzell, JK Vessey, W Newton, eds, Nitrogen Fixation: Global Perspectives. CAB International, Oxon, UK, pp 256–257
  3. Bergersen FJ (1993) Diffusion of oxyleghemoglobin. Ann Bot 72: 577–582 [Google Scholar]
  4. Bergersen FJ (1994) Distribution of O2 within infected cells of soybean root nodules: a new simulation. Protoplasma 183: 49–61 [Google Scholar]
  5. Bergersen FJ, Turner GL (1993) Effects of concentrations of substrates supplied to N2-fixing soybean bacteroids in flow chamber reactions. Proc R Soc Lond Biol 251: 95–102 [Google Scholar]
  6. Bouma TJ, Nielsen KL, Eissenstat DM, Lynch JP (1997) Soil CO2 concentration does not affect growth or root respiration in bean or citrus. Plant Cell Environ 20: 1495–1505 [Google Scholar]
  7. Dakora FD (1989) Studies of N2 fixation, N assimilation and nodule development in grain legumes: effects of O2. PhD thesis. The University of Western Australia, Nedlands
  8. Dakora FD, Atkins CA (1990) Morphological and structural adaptation of nodules of cowpea to functioning under sub- and supra-ambient oxygen pressure. Planta 182: 572–582 [DOI] [PubMed] [Google Scholar]
  9. Day DA, Poole PS, Tyerman SD, Rosendahl L (2001) Ammonia and amino acid transport across symbiotic membranes in nitrogen-fixing legume nodules. Cell Mol Life Sci 58: 61–71 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. de Lima ML, Oresnik IJ, Fernando SM, Hunt S, Smith R, Turpin DH, Layzell DB (1994) The relationship between nodule adenylates and the regulation of nitrogenase activity by O2 in soybean. Physiol Plant 91: 687–695 [Google Scholar]
  11. Denison RF (1992) Mathematical modelling of oxygen diffusion and respiration in legume root nodules. Plant Physiol 98: 901–907 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Diaz del Castillo L, Hunt S, Layzell DB (1992) O2 regulation and O2 limitation of nitrogenase activity in root nodules of pea and lupin. Physiol Plant 86: 269–278 [Google Scholar]
  13. Fernandez-Pascual M, de Lorenzo C, Pozuelo JM, de Felipe MR (1992) Alteration induced by four herbicides on lupine nodule cortex structure, protein metabolism and some senescence-related enzymes. J Plant Physiol 140: 385–390 [Google Scholar]
  14. Gibson QH, Wittenberg JB, Wittenberg BA, Bogusz D, Appleby CA (1989) The kinetics of ligand binding to plant hemoglobins. J Biol Chem 264: 100–107 [PubMed] [Google Scholar]
  15. Hunt S, Layzell DB (1993) Gas exchange of legume nodules and the regulation of nitrogenase activity. Annu Rev Plant Physiol Plant Mol Biol 44: 483–511 [Google Scholar]
  16. Igamberdiev AU, Kleczkowski LA (2001) Implications of adenylate kinase-governed equilibrium of adenylates on contents of free magnesium in plant cells and compartments. Biochem J 360: 225–231 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kawashima K, Suganuma N, Tamaoki M, Kouchi H (2001) Two types of pea leghemoglobin genes showing different O2-binding affinities and distinct patterns of spatial expression in nodules. Plant Physiol 125: 641–651 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. King BJ, Layzell DB (1991) Effect of increases in O2 concentration during the argon-induced decline in nitrogenase activity in root nodules of soybean. Plant Physiol 96: 376–381 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kushmerick MJ, Podolsky RJ (1969) Ionic mobility in muscle cells. Science 166: 1297–1298 [DOI] [PubMed] [Google Scholar]
  20. Kuzma MM, Hunt S, Layzell DB (1993) Role of oxygen in the limitation and inhibition of nitrogenase activity and respiration rate in individual soybean nodules. Plant Physiol 101: 161–169 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kuzma MM, Winter H, Storer P, Oresnik I, Atkins CA, Layzell DB (1999) The site of oxygen limitation in soybean nodules. Plant Physiol 119: 399–407 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Layzell DB, Atkins CA (1997) The physiology and biochemistry of legume N2 fixation. In DT Dennis, DB Layzell, DD Lefebvre, DH Turpin eds, Plant Metabolism. Addison Wesley Longman, Essex, UK, pp 495–505
  23. Lee HS, Copeland L (1996) Phosphofructokinase from the host fraction of chickpea nodules. Physiol Plant 96: 607–614 [Google Scholar]
  24. Millar AH, Day DA, Bergersen FJ (1995) Microaerobic respiration and oxidative phosphorylation by soybean nodule mitochondria: implications for nitrogen fixation. Plant Cell Environ 18: 715–726 [Google Scholar]
  25. Oresnik IJ, Layzell DB (1994) Composition and distribution of adenylates in soybean (Glycine max L.) nodule tissue. Plant Physiol 104: 217–225 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Plaxton WC (1996) The organization and regulation of plant glycolysis. Annu Rev Plant Physiol Plant Mol Biol 47: 185–214 [DOI] [PubMed] [Google Scholar]
  27. Pradet A, Raymond P (1983) Adenylate nucleotide ratios and adenylate energy charge in energy metabolism. Annu Rev Plant Physiol 34: 199–224 [Google Scholar]
  28. Pugh R, Witty JF, Mytton LR, Minchin FR (1995) The effect of waterlogging on nitrogen-fixation and nodule morphology in soil-grown white clover (Trifolium repens L.). J Exp Bot 46: 285–290 [Google Scholar]
  29. Sa TM, Israel DW (1991) Energy status and functioning of phosphorus-deficient soybean nodules. Plant Physiol 97: 928–935 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Satterfield CN, Sherwood TK (1963) The Role of Diffusion in Catalysis. Addison-Wesley, Reading, MA, p 14
  31. Shelp BJ, Atkins CA (1984) Subcellular location of enzymes of ammonia assimilation and asparagine synthesis in root nodules of Lupinus albus L. Plant Sci Lett 36: 225–230 [Google Scholar]
  32. Shimada S, Rousseau R, Denison RF (1997) Wavelength options for monitoring leghaemoglobin oxygenation gradients in intact legume root nodules. J Exp Bot 48: 1251–1258 [Google Scholar]
  33. Smith PMC, Atkins CA (2002) Purine biosynthesis: big in cell division, even bigger in nitrogen assimilation. Plant Physiol 128: 793–802 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Thumfort PP (1996) Nitrogen fixation and oxygen in legumes: the mathematical modelling of O2 diffusion into infected cells. PhD thesis. The University of western Australia, Nedlands
  35. Thumfort PP, Atkins CA, Layzell DB (1994) A re-evaluation of the role of the infected cell in the control of O2 diffusion in legume nodules. Plant Physiol 105: 1321–1333 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Thumfort PP, Layzell DB, Atkins CA (1999) Diffusion and reaction of oxygen in the central tissue of ureide-producing legume nodules. Plant Cell Environ 22: 1351–1363 [Google Scholar]
  37. Thumfort PP, Layzell DB, Atkins CA (2000) A simplified approach for modeling diffusion into cells. J Theor Biol 204: 47–65 [DOI] [PubMed] [Google Scholar]
  38. Tomaszewska B, Jarmuszkiewicz W, Schramm RW (1991) The role of malate in the metabolism of mitochondria and symbiosomes from lupin root nodules. Plant Physiol Biochem 29: 489–496 [Google Scholar]
  39. Vance CP (2000) Amide biosynthesis in root nodules of temperate legumes. In EW Triplett, ed. Prokaryotic Nitrogen Fixation: A Model System for Analysis of a Biological Process. Horizon Scientific Press, Wymondham, UK, pp 589–607
  40. Waters JK, Hughes BL, Purcell LC, Gerhardt KO, Mawhinney TP, Emerich DW (1998) Alanine, not ammonia, is excreted from N2-fixing soybean nodule bacteroids. Proc Nat Acad Sci USA 95: 12038–12042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wittenberg JB, Bergersen FJ, Appleby CA, Turner GL (1974) Facilitated Oxygen diffusion. The role of leghemoglobin in nitrogen fixation by bacteroids isolated from soybean root nodules. J Biol Chem 249: 4057–4066 [PubMed] [Google Scholar]

Articles from Plant Physiology are provided here courtesy of Oxford University Press

RESOURCES