Diel Shifts in Carboxylation Pathway and Metabolite Dynamics in the CAM Bromeliad Aechmea ‘Maya’ in Response to Elevated CO₂

Abstract

Background and Aims

The deployment of temporally separated carboxylation pathways for net CO₂ uptake in CAM plants provides plasticity and thus uncertainty on how species with this photosynthetic pathway will respond to life in a higher-CO₂ world. The present study examined how long-term exposure to elevated CO₂ influences the relative contributions that C₃ and C₄ carboxylation make to net carbon gain and to establish how this impacts on the availability of carbohydrates for export and growth and on water use efficiency over the day/night cycle.

Methods

Integrated measurements of leaf gas exchange and diel metabolite dynamics (e.g. malate, soluble sugars, starch) were made in leaves of the CAM bromeliad Aechmea ‘Maya’ after exposure to 700 µmol mol⁻¹ CO₂ for 5 months.

Key Results

There was a 60 % increase in 24-h carbon gain under elevated CO₂ due to a stimulation of daytime C₃ and C₄ carboxylation in phases II and IV where water use efficiency was comparable with that measured at night. The extra CO₂ taken up under elevated CO₂ was largely accumulated as hexose sugars during phase IV and net daytime export of carbohydrate was abolished. Under elevated CO₂ there was no stimulation of dark carboxylation and nocturnal export and respiration appeared to be the stronger sinks for carbohydrate.

Conclusions

Despite the increased size of the soluble sugar storage pool under elevated CO₂, there was no change in the net allocation of carbohydrates between provision of substrates for CAM and export/respiration in A. ‘Maya’. The data imply the existence of discrete pools of carbohydrate that provide substrate for CAM or sugars for export/respiration. The 2-fold increase in water-use efficiency could be a major physiological advantage to growth under elevated CO₂ in this CAM bromeliad.

INTRODUCTION

The prospect of global changes in the earth's climate has stimulated extensive research into assessing the impacts of elevated CO₂ on different vegetation types. The pre-industrial atmospheric CO₂ concentration of approx. 280 ppm has risen steadily to a present-day concentration of 380 ppm (Körner, 2006). As predicted by carbon-balance models a doubling in this CO₂ concentration could be expected by the middle of this century due to the effects of deforestation, land-use changes and the burning of fossil fuels (Neftel et al., 1985; Schimel et al., 1996). The effects of CO₂ enrichment on plant growth and physiological performance are diverse and complex, depending on the carbon fixation pathway, the exposure duration and the accompanying environmental conditions (Körner, 2001; Bazzaz and Catovsky, 2002). According to the carbon fixation pathway, three functional groups can be distinguished, i.e. C₃, C₄ and CAM. Numerous reports agree that elevated CO₂ mainly increases net CO₂ uptake and growth of C₃ plants, reflecting the increased efficiency of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) for fixing CO₂ at higher CO₂:O₂ ratios (Isdo and Kimball, 1992; Bowes, 1993; Poorter and Navas, 2002). After a short-term exposure of 4–6 weeks, acclimation to increased CO₂ can occur, returning net CO₂ uptake and growth to previous lower rates (Sage et al., 1989; Stitt, 1991). The situation for C₄ plants is more uncertain and complex. In these plants Rubisco is localized in the chloroplasts of the bundle sheath cells where the internal CO₂ concentration is 3–6 times higher than in the atmosphere. The C₄ carbon-concentrating mechanism is a consequence of the prior uptake of CO₂ mediated via phosphoenolpyruvate carboxylase (PEPC) which results in an effective draw down of CO₂ in the mesophyll cells and produces C₄ acids that are subsequently decarboxylated in the bundle sheath cells (Bowes, 1993; von Caemmerer and Furbank, 2003). C₄ plants avoid photorespiration due to the saturation of Rubisco with CO₂ and, theoretically would not be expected to display greater rates of photosynthesis in a higher CO₂ environment. However, because of the typical reduction of stomatal conductance in plants exposed to elevated atmospheric CO₂ (Ainsworth and Long, 2005), yield improvements could be achieved in a higher CO₂ world through an amelioration of short-term drought stress via the conservation of soil moisture (Ghannoum et al., 2000; Leakey et al., 2004).

In comparison to C₃ and C₄ plants, relatively few studies have addressed the effects of elevated CO₂ on gas exchange characteristics in CAM plants which use both Rubisco and PEPC to take up CO₂. Different conclusions have been reached on how CAM species might be expected to respond to life in a higher CO₂ world, in terms of net carbon gain, growth and water use efficiency (WUE) (Drennan and Nobel, 2000). Specifically, a doubling of the atmospheric CO₂ concentration was found to have no effect on diel CO₂ uptake by the CAM species Kalanchoë daigremontiana (Osmond and Björkman, 1975; Holtum et al., 1983); night-time CO₂ uptake was decreased in Clusia uvitana and Portulacaria afra (Huerta and Ting, 1988; Winter et al., 1992) yet night-time CO₂ uptake was enhanced by elevated CO₂ in Agave vilmoriniana (Szarek et al., 1987) and a tropical orchid Mokara Yellow (Gouk et al., 1997; Li et al., 2002). In other CAM species, e.g. Agave deserti, Ananas comosus, Opuntia ficus-indica and Stenocereus queretaroensis, elevated CO₂ was shown to lead to an increase in both daytime and night-time CO₂ uptake (Nobel and Israel, 1994; Graham and Nobel, 1996; Nobel, 1996; Wang and Nobel, 1996; Zhu et al., 1999). Such contrasting reports on the impacts of elevated CO₂ on CAM plants may in part relate to the inherent plasticity afforded by the existence of temporally separated C₃ and C₄ carboxylation processes which serves to optimize carbon gain and WUE under a changing environment. Thus, a better understanding is required on how extended exposure to elevated CO₂ affects the relative contributions from PEPC- and Rubisco-mediated net CO₂ uptake over the day/night cycle.

Plasticity in the deployment of C₃ and C₄ carboxylation processes in CAM plants can be analysed by monitoring the shifts in net CO₂ uptake and metabolite levels that define the four phases of carbon supply and demand over a 24-h cycle (Osmond, 1978). Phase I represents the dark period when stomata are open and atmospheric and/or respiratory CO₂ is fixed in the cytosol by PEPC, resulting in malate formation. The CO₂ acceptor phosphoenolpyruvate (PEP) is formed by glycolysis of storage polysaccharides (e.g. starch) or soluble sugars (e.g. glucose, fructose, sucrose) assembled during the previous day. Phase II at dawn indicates the switch between C₄ and C₃ carboxylation but atmospheric CO₂ uptake still largely appears to be dominated by PEPC (Griffiths et al., 1990). Rubisco remains at low activation state until PEPC is inactivated by dephosphorylation (Maxwell et al., 1999). During the middle of the day (phase III) gas exchange is curtailed by stomatal closure as a result of high intercellular partial pressure of CO₂ generated by malate decarboxylation. The released CO₂ is fixed through Rubisco via the Calvin cycle into triose phosphates that have to be partitioned between the provision of carbohydrates for growth and as substrates for the production of PEP for phase I carboxylation. At the end of the light period, decarboxylation of malate has ceased causing CO₂ partial pressure to drop and stomates to reopen. During this phase IV, CO₂ is mainly fixed and assimilated via Rubisco with a major proportion of carbon fixed partitioned for export and growth (Borland et al., 1994). Over the latter part of this phase C₄ carboxylation by PEPC can be detected (Griffiths et al., 1990; Borland and Griffiths, 1996). The relative contribution that each phase makes to net carbon gain over a 24-h CAM cycle is determined by an interplay of circadian and metabolite control (Borland and Taybi, 2004). Control by a circadian clock sets up the diel phases of CAM to achieve appropriate synchronization of metabolic and transport processes. An additional layer of control over the CAM cycle is exerted by the day/night reciprocating accumulation/degradation of carbohydrates and malate which facilitates photosynthetic plasticity by entraining output from the clock to fluctuations in the environment (Borland et al., 1999; Borland and Taybi, 2004). By serving as an internal gauge of external environmental conditions, the reciprocating pools of organic acids and carbohydrates are central to the metabolic control of CAM. The day/night patterns and availability of carbohydrates in particular play a key role in setting the phases of CAM, by determining both the magnitude of dark CO₂ uptake and the amount of carbohydrate available for export and growth. Thus, it can be hypothesized that any change in carbohydrate or malate content elicited by growth under elevated CO₂ will result in a shift in the relative contributions that C₃ and C₄ carboxylation make to 24-h carbon gain and WUE.

In the present study, integrated measurements of leaf gas exchange and diel metabolite dynamics (e.g. malate, citrate, glucose, fructose, sucrose and starch) were made in leaves of the CAM bromeliad Aechmea ‘Maya’ after exposure to 700 µmol mol⁻¹ CO₂ for 5 months. Specific questions that were addressed included: (a) how does long-term exposure to elevated CO₂ influence the magnitude and duration of C₃ and C₄ carboxylation processes over the four phases of CAM; (b) how do shifts in the proportions of C₃ and C₄ carboxylation impact on leaf WUE over a 24-h light dark cycle; (c) how does carbohydrate availability influence the magnitude of CAM under elevated CO₂; and (d) how are carbohydrates partitioned between the potentially competing fates of dark carboxylation and export for growth under contrasting regimes of CO₂ supply.

MATERIALS AND METHODS

Plant material

Aechmea ‘Maya’ is a spineless cultivar resulting from a cross between A. tessmannii and A. fasciata. These aechmeas are CAM bromeliads and belong to the subfamily of the Bromelioideae (Londers et al., 2005; Benzing, 2000). Twelve-month-old vegetative A. ‘Maya’ plants with 27 (±5) leaves were equally divided between two greenhouse compartments (Leuven, Belgium). Control plants were grown under ambient atmospheric CO₂ concentrations (approx. 380 µmol mol⁻¹). The CO₂-treated plants were exposed to approx. 700 µmol mol⁻¹. This was achieved by automated continuous injection of pure CO₂ each time until the low set point (700 µmol mol⁻¹) was reached. All other environmental conditions in both compartments were identical. During day time a minimum temperature of 21 °C was maintained while at night a minimum of 19·5 °C was achieved. Between 0600 h and 2200 h artificial lighting was provided (PAR = 30 µmol m⁻² s⁻¹). Plants were watered once weekly with a conventional nutrient solution of 1 mS cm⁻¹.

Gas exchange measurements

After 5 months, leaf CO₂ and water vapour exchange data were collected on the youngest fully expanded leaf from plants under ambient and elevated CO₂. An LCi portable photosynthesis system (ADC BioScientific Ltd, UK) was used. The top part of the leaf was enclosed in a broad leaf chamber (6·25 cm²) and, to buffer short-term fluctuations in the CO₂ concentration, the incoming air was passed through a 25-L bottle. Gas exchange data were collected over a 24-h period using a 15-min interval (n = 5 plants). Total net gas exchange was calculated for specific periods during the 24-h time course by integrating specific areas under the CO₂ exchange curves (Griffiths et al., 1986).

Chemical analyses of metabolites

During a cycle of 24 h starting from 0600 h leaf samples were taken from the upper part of young fully expanded leaves (n = 10 plants) from plants under both CO₂ regimens every 3 h. These were immediately frozen in liquid nitrogen to arrest any enzyme activity.

Extraction of organic acids was performed as earlier described by Ceusters et al. (2008). Quantification of malic and citric acid was accomplished using high performance liquid chromatography (Waters 510, Waters, Milford, MA, USA) with detection at 210 nm (Waters 484) using an aminex HPX87-H (300 mm × 7·8 mm) resin-based column from Bio-Rad (Hercules, CA, USA).

Soluble sugars (glucose, fructose and sucrose) were extracted, subjected to enzymatic treatment (Enzytec, Scil Diagnostics GmbH, Germany) and analysed at 340 nm using a spectrophotometer (DU-65, Beckman, Fullerton, USA). Starch content was determined as glucose equivalents following digestion with amyloglucosidase (Enzytec, Scil Diagnostics GmbH, Germany). The procedures were conducted as described by Farrar (1993).

Carbon budgets

Carbon budgets were calculated and adapted from a previously described model (Borland, 1996; Borland and Dodd, 2002) which quantitatively describes: (a) the source of carbon in the leaf, i.e. C₃ (Rubisco-mediated CO₂ uptake) or C₄ (from breakdown of malate where 1 mol malate = 4 mol C); (b) the partitioning of this carbon between starch or soluble sugars (sucrose, glucose and fructose) with the excess going to daytime export where 1 mol Glc eq = 6 mol C; (c) the partitioning of carbohydrates between the generation of substrates for CAM, respiratory CO₂ or dark export. The budgets presented are net carbon flows, thus taking into account any respiratory losses of carbon during the light period. All numbers presented on budgets are mmol C m⁻² d⁻¹.

Data analysis

Where appropriate, the data were analysed using the statistical software package SAS Enterprise Guide 4·0. Before carrying out statistical tests normality of the data was checked by means of the Kolmogorov–Smirnoff statistic (P > 0·05). Means are compared by two-sample t-test (α = 0·05).

RESULTS

Leaf gas exchange

After exposure to 700 µmol mol⁻¹ for 5 months, diel gas exchange patterns were recorded and integrated CO₂ uptake was calculated for each of the different CAM phases over a 24-h period (Fig. 1 and Table 1). CO₂ enrichment significantly increased both the rate of CO₂ assimilation and total assimilation during the diel cycle (Fig. 1 and Table 1). More specifically, after a cycle of 24 h at elevated CO₂ 60 % more CO₂ was taken up compared with total uptake by plants kept at ambient CO₂ (Table 1). Night-time CO₂ uptake did not account for this substantial difference because neither the maximum rate of uptake (Fig. 1) nor the total amount of net CO₂ uptake (Table 1) during phase I were influenced by elevated CO₂. However, in the light (phases II and IV) net CO₂ uptake was significantly higher for plants grown at elevated CO₂ with a 3-h extended period of phase IV net CO₂ exchange compared with plants in ambient CO₂. The maximum rate of net CO₂ uptake during phase IV even exceeded uptake during the night under CO₂ enrichment (Fig. 1).

Fig. 1.

Net 24 h CO₂ uptake (μmol m⁻²s⁻¹), diel stomatal conductance (mmol m⁻² s⁻¹) and diel malic acid (mmol m⁻²) pattern for young fully expanded leaves of Aechmea ‘Maya’ at ambient (380 µmol mol⁻¹) and elevated (700 µmol mol⁻¹) CO₂. Gas exchange curves are representative of five replicate runs with s.e. <10 %. Malic acid data are means ± s.e. (n = 10 plants). The four different phases of CAM are indicated and the grey line shows a phase shift caused by CO₂ elevation.

Table 1.

Integrated net CO₂ uptake for the times of each of the four CAM phases (mmol CO₂ m⁻² phase⁻¹) by young fully expanded leaves of Aechmea ‘Maya’ under ambient (380 μmol mol⁻¹) and elevated (700 μmol mol⁻¹) CO₂ concentrations

CO2 (μmol mol−1)	Phase I 2200–0600	Phase II 0600–0900	Phase III 0900–1800*	Phase IV 1800–2200*	Total 24 h 1200–1200
380	52·5 ± 6·1	7·2 ± 0·6	−2·5 ± 4·7	10·5 ± 2·0	67·7 ± 6·9
700	55·4 ± 2·6	16·1 ± 1·3	−3·7 ± 2·5	42·5 ± 3·7	109·5 ± 10·5
P-value	>0·05	<0·05	>0·05	<0·05	<0·05

The data are means ± s.e. for five leaves, each from a separate plant and were compared for each phase between the two treatments by two-sample t-test (α = 0·05).

*Due to elevated CO₂ phase IV extended by starting 3 h earlier at 1500 h compared with controls; concomitantly phase III reduced by the same extent.

The diel courses of stomatal conductance (g_s) showed maximum values at night for plants under both CO₂ treatments (Fig. 1). However, in plants grown under elevated CO₂ lower values of g_s were observed simultaneously with a higher rate of CO₂ uptake in phase IV compared with phase I. Moreover, for an equal amount of CO₂ uptake during phase I, the values of g_s were 2-fold higher in ambient CO₂ compared with those under elevated CO₂ (Fig. 1), resulting in a nightly transpiration reduction of 50 % for plants under elevated CO₂ (Table 2). As a consequence WUE was 2-fold higher during the night under elevated CO₂. During the daytime phases II and IV, a 3- to 4-fold increase in WUE was recorded in plants under elevated CO_2. The resulting 24 h WUE of 15·3 mmol mol⁻¹ under elevated CO₂ was significantly higher than 7·4 mmol mol⁻¹ under ambient conditions (Table 2).

Table 2.

Mean photosynthetic water use efficiency (mmol CO₂ mol⁻¹ H₂O) for young fully expanded leaves of Aechmea ‘Maya’ under ambient (380 μmol mol⁻¹) and elevated (700 μmol mol⁻¹) CO₂ concentrations

CO2 (μmol mol−1)	Phase I 2200–0600	Phase II 0600–0900	Phase IV 1800–2200*	Total 24 h 1200–1200
380	13·1 ± 1·3	5·7 ± 0·8	6·6 ± 1·2	7·4 ± 1·2
700	24·7 ± 3·0	16·6 ± 1·0	24·6 ± 4·2	15·3 ± 2·8
P-value	<0·05	<0·05	<0·05	<0·05

Because no net CO₂ uptake occurred during phase III, WUE was not calculated for this period. The data are means ± s.e. for five leaves, each from a separate plant and were compared for each phase between the two treatments by two-sample t-test (α = 0·05).

*Due to elevated CO₂ phase IV extended by starting 3 h earlier at 1500 h compared with controls.

Metabolite dynamics

Under both CO₂ concentrations the overall day/night changes in malic acid content were similar (Fig. 1; P > 0·05). However, malic acid started to accumulate 3 h earlier in the afternoon, at 1500 h, in plants grown under elevated CO₂ compared with plants in ambient CO₂ where malic acid did not start to accumulate until 1800 h. As the beginning of malic acid accumulation indicates the start of phase IV (Osmond, 1978), the earlier suggested phase extension from this phase with 3 h and an associated reduction of phase III is hereby confirmed. Lowest values of malic acid were reached at the end of the respective phase III (19 ± 5 mmol m⁻²), while the highest values were recorded in the morning, at the end of phase II (94 ± 11 mmol m⁻²) (Fig. 1). Citric acid levels did not show a varying diel trend or significant differences concerning CO₂ concentration (P > 0·05) and varied around 40 ± 11 mmol m⁻² throughout the whole cycle (results not shown).

The diel cycle of starch content under both CO₂ treatments varied reciprocally to that of malic acid and was not significantly altered by CO₂ enrichment (Fig. 2; P > 0·05). In the morning during phase II the mean minimum starch concentration was 12 ± 5 mmol Glc eq m⁻². Around 1800 h the maximal starch content was achieved for leaves from plants under both ambient and enhanced CO₂ with an average value of 43 ± 9 mmol Glc eq m⁻². Starch degradation commenced during phase IV and continued during phase I (Fig. 2).

Fig. 2.

Diel pattern of starch (mmol Glc eq m⁻²) for young fully expanded leaves of Aechmea ‘Maya’ at ambient (380 µmol mol⁻¹) and elevated (700 µmol mol⁻¹) CO₂. The four different phases of CAM are indicated and the grey line shows a phase shift caused by CO₂ elevation. Data are means ± s.e. (n = 10 plants).

Sucrose showed a diel pattern of accumulation and depletion analogous to starch and reciprocal to malic acid and remained unaffected by CO₂ enhancement (Fig. 3; P > 0·05). The maximum concentrations of sucrose (13 ± 2 mmol m⁻²) were reached around 1800 h and nocturnal breakdown of sucrose during phase I yielded 9 ± 1 mmol m⁻² (Fig. 3). Hexose sugars (glucose and fructose) fluctuated around 8 ± 2 mmol m⁻² and they did not participate in the diel carbohydrate flux to sustain CAM under ambient CO₂ conditions (Fig. 3). However, under elevated CO₂, substantial accumulation of glucose and fructose was noted over the latter part of the day. Analogous to starch the lowest concentrations for both glucose and fructose were present in the morning at the end of phase II. At the start of the extended phase IV under elevated CO₂, concentrations of hexose sugars started to rise to achieve significantly different (P < 0·05) maximum values from those in plants grown under ambient CO₂ (i.e. 18 ± 5 mmol m⁻² for glucose and 31 ± 5 mmol m⁻² for fructose at the end of phase IV). In contrast to starch, however, only limited net nocturnal degradation of both glucose and fructose took place during phase I under elevated CO₂ (Fig. 3; P > 0·05) and as such did not appear to contribute to the nocturnal provision of PEP.

Fig. 3.

Diel patterns of sucrose, glucose and fructose (mmol m⁻²) for young fully expanded leaves of Aechmea ‘Maya’ at ambient (380 µmol mol⁻¹) and elevated (700 µmol mol⁻¹) CO₂. The four different phases of CAM are indicated and the grey line shows a phase shift caused by CO₂ elevation. Data are means ± s.e. (n = 10 plants).

Carbon budgets

To integrate gas exchange and carbohydrate measurements leaf carbon budgets were calculated for plants under ambient and elevated CO₂ (Fig. 4). Doubling of the CO₂ concentration resulted in an increase of daily carbon input by 35 %, mainly due to enhanced atmospheric CO₂ uptake during the transitional phases II and IV. In ambient conditions, CO₂ uptake in phases II and IV was almost entirely mediated via PEPC activity (Table 3). Under elevated CO₂ a 2- to 4-fold increase in CO₂ acquisition in phases II and IV, respectively (Table 1) was mediated by stimulation in both C₄ and C₃ carboxylation. As such, 60 % of CO₂ uptake during the translational phases was mediated by PEPC while 40 % was taken up through direct assimilation by Rubisco in a high CO₂ environment (Table 3). For both CO₂ treatments the major sink for the acquired carbon was the daytime accumulation of starch. Under ambient CO₂, a further 23 % of fixed carbon was allocated to the accumulation of soluble sugars, allowing 8 % of total acquired carbon to be exported. Under elevated CO₂ the net flux of carbon into soluble sugars doubled over that in ambient conditions. Net flux to starch accumulation remained the same under elevated and ambient CO₂ but daytime net export of sugars was abolished under elevated CO₂ (Fig. 4). Starch degradation was the major source of carbon skeletons for the synthesis of PEP for C₄ carboxylation under both CO₂ regimes (Fig. 2). Additionally, limited breakdown of soluble sugars, especially sucrose occurred during the period of malate formation (Figs 1 and 3). In total, 80 % of carbohydrate degradation was used to fuel PEPC uptake resulting in 20 % of carbon available for export/respiration during the period of malate formation under ambient CO₂. Under elevated CO₂, 65 % of previously accumulated carbohydrates were used to provide substrate for PEPC-mediated CO₂ uptake, 20 % of carbohydrates were exported/respired and 15 % was retained in the form of fructose and glucose (Figs. 3 and 4). Overall, the total net export of carbon during the diel cycle was similar under both CO₂ treatments.

Fig. 4.

Diel net carbon budgets for young fully expanded leaves of Aechmea ‘Maya’ at ambient (380 µmol mol⁻¹) (left) and elevated (700 µmol mol⁻¹) (right) CO₂. Fluxes shown above the horizontal line mainly coincide with degradation of malic acid and concomitant accumulation of carbohydrate reserves while those under the horizontal line are characterized by the reverse process. Gas exchange data were calculated as the integrated mean from five replicate leaves and metabolites were sampled from ten replicate leaves on individual plants. Export includes growth and maintenance and all units are mmol C m⁻² d⁻¹. * accumulated hexose carbon which is not degraded or exported overnight.

Table 3.

Relative contribution of PEPC- and Rubisco-mediated CO₂ uptake during translational phases II and IV for young fully expanded leaves of Aechmea ‘Maya’ under ambient (380 μmol mol⁻¹) and elevated (700 μmol mol⁻¹) CO₂ concentrations

	% C4 carboxylation		% C3 carboxylation
CO2 (μmol mol−1)	Phase II 0600–0900	Phase IV 1800–2200*	Phase II 0600–0900	Phase IV 1800–2200*
380	85	100	15	0
700	56	60	44	40

Percentage of C₄ carboxylation was quantified by comparison of malic acid build up and total net CO₂ uptake in each period whereby 1 mol malic acid corresponds to 1 mol CO₂. The remaining CO₂ uptake was attributed to C₃ carboxylation through Rubisco activity.

*Due to elevated CO₂ phase IV extended by starting 3 h earlier at 1500 h compared with controls.

DISCUSSION

Impact of elevated CO₂ on C₃ and C₄ carboxylation processes

Extended exposure of A. ‘Maya’ to elevated CO₂ resulted in a 60 % increase in 24-h net carbon gain, but the stimulation was due entirely to an increase in the direct uptake of CO₂ during the day and nocturnal CO₂ uptake remained unchanged. Although Ting (1994) previously postulated that exposure to elevated CO₂ will result in a minimal increase in dark CO₂ uptake due to carbon saturation of PEPC at current atmospheric CO₂ concentrations, CAM species like Hylocereus undatus (Raveh et al., 1995), Opuntia ficus-indica (Cui et al., 1993) and Ananas comosus (Zhu et al., 1999) have shown significant increases in CO₂ uptake during phase I in response to CO₂ enrichment. The present results for A. ‘Maya’ might be explained by an average stomatal conductance that was approx. 50 % lower under CO₂ enrichment, which would imply carbon saturation of PEPC at ambient CO₂ levels. It is also possible that dark CO₂ uptake could have been limited by vacuolar storage capacity for malic acid given that the leaves of A. ‘Maya’ are less succulent than many other CAM species, e.g. pineapple (Popp et al., 2003). Another potential limiting factor for dark CO₂ uptake could be carbonic anhydrase (CA) activity. In the C₄ plant Flaveria bidentis it has been reported that leaf CA activity is in excess and does not limit CO₂ assimilation under normal conditions (Cousins et al., 2006) but in several other C₄ species it would appear that CA activity is only just sufficient to support photosynthetic rates, especially in monocots (Hatch and Burnell, 1990; Gillon and Yakir, 2000). The situation in CAM plants with regards to CA activity and the allocation of CA between cytosol (to support PEPC-mediated CO₂ uptake) and chloroplast (to support Rubisco-mediated CO₂ uptake) remains to be resolved.

Although there was no difference in the amount of CO₂ taken up at night in A. ‘Maya’ under ambient or elevated CO₂, there did appear to be a stimulation of net C₄ carboxylation during phase II and, particularly during late afternoon, phase IV under elevated CO₂, even although stomatal conductance was lower than that measured in plants grown under ambient CO₂. On the assumption that PEPC was carbon saturated, there are two possible explanations for the CO₂-induced increase in daytime net C₄ carboxylation: (1) given reports that PEPC may be limited by availability of PEP under ambient conditions, exposure to elevated CO₂ may have resulted in an increased flux into the production of PEP during the day; or (2) since the contribution from PEPC to net carbon balance was assessed in terms of net malate accumulation, it is possible that exposure to elevated CO₂ affected malate turnover. Thus, PEPC activity in phase II may have been similar under ambient and elevated CO₂, but malate breakdown may have been delayed/curtailed at the start of the day in plants exposed to elevated CO₂. Although the factors that regulate the timing and rate of malate decarboxylation in CAM plants remain unclear, exposure to elevated CO₂ resulted in an increased rate of malate decarboxylation during phase III so that stomata re-opened and phase IV commenced some 3 h earlier than in plants under ambient CO₂. Thus, there was a significant increase in direct C₃-mediated net carbon gain over the latter part of the day in plants grown under elevated CO₂, despite the lower stomatal conductance compared with plants in ambient conditions. The typically thick, succulent leaves of CAM plants are believed to impose significant diffusional limitations to CO₂ transfer from the atmosphere to Rubisco in the chloroplast, with the consequence that Rubisco may be severely CO₂ limited during the latter part of the photoperiod when malate decarboxylation has reached completion (Maxwell et al., 1997). Thus, as shown here, it seems likely that direct daytime C₃ photosynthesis will show substantial and sustained stimulation in CAM plants growing in a higher CO₂ world.

In total, approx. 50 % of the 24 h net CO₂ uptake occurred during the day in A. ‘Maya’ grown under elevated CO₂, representing a larger proportion of net carbon gain than previously reported for CAM plants with a comparable daytime contribution of approx. 22 % under ambient conditions. Under CO₂ enrichment, Agave deserti (Nobel and Hartsock, 1986), Ananas comosus (Zhu et al., 1999) and Hylocereus undatus (Raveh et al., 1995) still showed a dominating night-time CO₂ uptake of approx. 65 % of total diel uptake. Despite the relative increase in daytime to night-time CO₂ uptake and the concomitant stimulation in C₃ carboxylation in A. ‘Maya’, diel WUE increased from 7·4 to 15·3 mmol mol⁻¹ under CO₂ enrichment. Comparable findings have been obtained for the CAM plants Ananas comosus and Opuntia ficus-indica (Cui et al., 1993; Zhu et al., 1999). By restricting most transpirational water loss to the cooler nocturnal period, high WUE values are usually attributed to PEPC-mediated CO₂ uptake during phase I of the CAM cycle (Black, 1973; Lüttge, 2004), a statement which holds under ambient conditions for A. ‘Maya’. However, under CO₂ enrichment high values of WUE (25 mmol mol⁻¹) were established during both phases I and IV. Compared with plants under ambient CO₂, the gain in WUE during phase I under elevated CO₂ was achieved by a 50 % reduction of stomatal conductance without a concomitant reduction in CO₂ uptake. However, during the daytime phases both increased Rubisco, and PEPC-mediated CO₂ uptake resulted in high WUE values, confirming the statement of Eller and Ferrari (1997) that high WUE is not exclusively associated with performing C₄ carboxylation at night.

Impact of elevated CO₂ on carbohydrate partitioning

In accordance with gas exchange analyses the diel pattern of accumulation/degradation of malate was hardly affected by elevated CO₂. Similarly, CO₂ enrichment did not alter the dynamics of the reciprocating carbohydrates. Stoichiometric considerations indicated that under both CO₂ regimes starch appeared to be the predominant substrate for nocturnal carboxylation while sucrose also accumulated during daytime. Comparable results were obtained under ambient CO₂ for pineapple by Antony et al. (2008). These observations confirm that bromelioids accumulate both sugar and starch to fuel the dark reactions of CAM (Christopher and Holtum, 1998; Popp et al., 2003). Soluble hexose sugars (glucose and fructose) did not fluctuate over the diel cycle under ambient CO₂ in A. ‘Maya’. However, under elevated CO₂ glucose and fructose started to accumulate to levels 3- and 5-fold higher, respectively, compared with controls during phase IV, utilizing the carbon gained during the period of malate decarboxylation and abolishing net daytime export of sugars. Since there was no net sugar accumulation during the day under ambient CO₂, 8 % of the carbon gained over the photoperiod could be exported for growth and maintenance. During the subsequent period of malate formation, approx. 80 % of accumulated carbohydrates were degraded to provide substrate for PEPC-mediated CO₂ uptake, while a further 20 % of carbohydrates were depleted and used for export/respiration. Under elevated CO₂ a hexose (glucose and fructose) pool of 32 ± 5 mmol Glc eq m⁻² was available at the start of phase I and thus constituted a pool of carbon skeletons almost equivalent to those available from starch and which could potentially be degraded to support night-time CO₂ uptake and malate formation. However, in contrast to starch, only limited net nocturnal degradation of either glucose or fructose took place, thereby withholding 25 % of accumulated sugar carbon from the competing fates of PEPC-mediated CO₂ uptake and export for growth. Carbohydrate availability is considered to be a major limiting factor for the magnitude of dark CO₂ uptake in CAM plants (Borland and Dodd, 2002). Moreover, the restoration of nocturnal malate accumulation in a starch-deficient mutant of Mesembryanthemum crystallinum by feeding leaves with glucose or sucrose has implied flexibility of the CAM pathway in terms of utilizing different carbohydrate sources to generate substrate for nocturnal carboxylation (Cushman et al., 2008). However, the results obtained here indicate that for A. ‘Maya’ under elevated CO₂, PEPC was not carbohydrate limited or else the glucose and fructose reserves, presumably stored in the vacuole, were unavailable for the generation of PEP. Previous considerations of ¹³C signatures of different carbohydrate species have implied the existence of discrete pools of storage carbohydrates destined for CAM or export/respiration (Borland and Dodd, 2002). The data obtained here for A. ‘Maya’ are consistent with this suggestion and could provide a means by which the potentially competing fates of CAM and export/respiration for growth could be regulated independently, thereby facilitating photosynthetic plasticity in a changing environment. As a consequence of hexose accumulation under elevated CO₂, the total net export of carbon for growth and maintenance during the diel cycle in A. ‘Maya’ was similar in both regimes of CO₂ supply, although a shift occurred from combined daytime and night-time export to exclusive night-time export of carbohydrates under elevated CO₂.

In conclusion, elevation of CO₂ was found to exert a pronounced influence on the magnitude and duration of both C₃ and C₄ carboxylation processes during the day while nocturnal PEPC-mediated CO₂ uptake remained unaffected. This impacted on the synchronization of both carboxylation processes by altering the balance of almost exclusive C₄ carboxylation under ambient CO₂ to approx. 60 % PEPC- and approx. 40 % Rubisco-mediated uptake over the day under elevated CO₂. Nevertheless, WUE doubled under elevated CO₂, demonstrating that WUE is not always related to specific CAM phases. The extra gain of carbon under high CO₂ was largely accumulated as hexose sugars during phase IV, abolishing any daytime net export of carbon. Around 25 % of the sugars that accumulated under elevated CO₂ were not degraded or exported overnight, indicating that factors other than carbohydrate pool size might limit dark carboxylation. Consequently, total net export of carbon for growth and maintenance was similar under the contrasting CO₂ regimes. The data indicate that whilst some CAM species may not show enhanced biomass production in a higher CO₂ world, productivity could be maintained with reduced inputs of water. Such findings indicate a major ecophysiological advantage for CAM plants under a changing climate and suggest that CAM species should also be considered in an agronomic context as potential sources of biomass production on arid, marginal lands.