The coordination of C4 photosynthesis and the CO2-concentrating mechanism in maize and Miscanthus x giganteus in response to transient changes in light quality.

Light quality coordinates a CO2-concentrating mechanism during C4 photosynthesis. Unequal absorption of photons between photosystems I and II, and between bundle-sheath and mesophyll cells, are likely to affect the efficiency of the CO2-concentrating mechanism in C4 plants. Under steady-state conditions, it is expected that the biochemical distribution of energy (ATP and NADPH) and photosynthetic metabolite concentrations will adjust to maintain the efficiency of C4 photosynthesis through the coordination of the C3 (Calvin-Benson-Bassham) and C4 (CO2 pump) cycles. However, under transient conditions, changes in light quality will likely alter the coordination of the C3 and C4 cycles, influencing rates of CO2 assimilation and decreasing the efficiency of the CO2-concentrating mechanism. To test these hypotheses, we measured leaf gas exchange, leaf discrimination, chlorophyll fluorescence, electrochromatic shift, photosynthetic metabolite pools, and chloroplast movement in maize (Zea mays) and Miscanthus × giganteus following transitional changes in light quality. In both species, the rate of net CO2 assimilation responded quickly to changes in light treatments, with lower rates of net CO2 assimilation under blue light compared with red, green, and blue light, red light, and green light. Under steady state, the efficiency of CO2-concentrating mechanisms was similar; however, transient changes affected the coordination of C3 and C4 cycles in M. giganteus but to a lesser extent in maize. The species differences in the ability to coordinate the activities of C3 and C4 cycles appear to be related to differences in the response of cyclic electron flux around photosystem I and potentially chloroplast rearrangement in response to changes in light quality.

The CO 2 -concentrating mechanism in C 4 plants reduces the carbon lost through the photorespiratory pathway by limiting the oxygenation of ribulose-1,5bisphosphate (RuBP) by the enzyme Rubisco (Brown and Smith, 1972;Sage, 1999). Through the compartmentalization of the C 4 cycle in the mesophyll cells and the C 3 cycle in the bundle-sheath cells (Hatch and Slack, 1966), C 4 plants suppress RuBP oxygenation by generating a high CO 2 partial pressure around Rubisco (Furbank and Hatch, 1987). To maintain high photosynthetic rates and efficient light energy utilization, the metabolic flux through the C 3 and C 4 cycles must be coordinated. However, coordination of the C 3 and C 4 cycles is likely disrupted due to rapid changes in environmental conditions, particularly changes in light availability (Evans et al., 2007;Tazoe et al., 2008).
Spatial and temporal variations in light environments, including both light quantity and quality, are expected to alter the coordination of the C 3 and C 4 cycles. For example, it has been suggested that the coordination of C 3 and C 4 cycles is altered by changes in light intensity (Henderson et al., 1992;Cousins et al., 2006;Tazoe et al., 2006Tazoe et al., , 2008Kromdijk et al., 2008Kromdijk et al., , 2010Pengelly et al., 2010). However, more recent publications indicate that some of the proposed light sensitivity of the CO 2 -concentrating mechanisms in C 4 plants can be attributed to oversimplifications of leaf models of carbon isotope discrimination (Δ 13 C), in particular, errors in estimates of bundle-sheath CO 2 partial pressure and omissions of respiratory fractionation (Ubierna et al., 2011(Ubierna et al., , 2013. Alternatively, there is little information on the effects of light quality on the coordination of C 3 and C 4 cycle activities and the subsequent impact on net rate of CO 2 assimilation (A net ).
In C 3 plants, A net is reduced under blue light compared with red or green light (Evans and Vogelmann, 2003;Loreto et al., 2009). This was attributed to differences in absorbance and wavelength-dependent differences in light penetration into leaves, where red and green light penetrate farther into leaves compared with blue light (Vogelmann and Evans, 2002;Evans and Vogelmann, 2003). Differences in light quality penetration into a leaf are likely to have profound impacts on C 4 photosynthesis, because the C 4 photosynthetic pathway requires the metabolic coordination of the mesophyll C 4 cycle and the bundle-sheath C 3 cycle. Indeed, Evans et al. (2007) observed a 50% reduction in the rate of CO 2 assimilation in Flaveria bidentis under blue light relative to white light at a light intensity of 350 mmol quanta m 22 s 21 . This was attributed to poor penetration of blue light into the bundle-sheath cells and subsequent insufficient production of ATP in the bundle-sheath cells to match the rates of mesophyll cell CO 2 pumping (Evans et al., 2007). Recently, Sun et al. (2012) observed similar low rates of steady-state CO 2 assimilation under blue light relative to red, green, and blue light (RGB), red light, and green light at a constant light intensity of 900 mmol quanta m 22 s 21 .
Because the light penetration into a leaf depends on light quality, with blue light penetrating the least, this potentially results in changes in the energy available for carboxylation reactions in the bundle-sheath (C 3 cycle) and mesophyll (C 4 cycle) cells. Changes in the balance of energy driving the C 3 and C 4 cycles can alter the efficiency of the CO 2 -concentrating mechanisms, often represented by leakiness (f), the fraction of CO 2 that is pumped into the bundle-sheath cells that subsequently leaks back out (Evans et al., 2007). Unfortunately, f cannot be measured directly, but it can be estimated through the combined measured and modeled values of Δ 13 C (Farquhar, 1983). Using measurements of Δ 13 C, it has been demonstrated that under steady-state conditions, changes in light quality do not affect f (Sun et al., 2012); however, it remains unknown if f is also constant during the transitions between different light qualities. In fact, sudden changes of light quality could temporally alter the coordination of the C 3 and C 4 cycles.
To understand the effects of light quality on C 4 photosynthesis and the coordination of the activities of C 3 and C 4 cycles, we measured transitional changes in leaf gas exchange and Δ 13 C under RGB and broadspectrum red, green, and blue light in the NADP-malic enzyme C 4 plants maize (Zea mays) and Miscanthus 3 giganteus. Leaf gas exchange and Δ 13 C measurements were used to estimate f using the complete model of C 4 leaf Δ 13 C (Farquhar, 1983;Farquhar and Cernusak, 2012). Additionally, we measured photosynthetic metabolite pools, Rubisco activation state, chloroplast movement, and rates of linear versus cyclic electron flow during rapid transitions from red to blue light and blue to red light. We hypothesized that the limited penetration of blue light into the leaf would result in insufficient production of ATP in the bundle-sheath cells to match the rate of mesophyll cell CO 2 pumping. We predicted that rapid changes in light quality would affect the coordination of the C 3 and C 4 cycles and cause an increase in f, but this would equilibrate as leaf metabolism reached a new steady-state condition.

Gas Exchange and Photosynthetic Discrimination
The A net was significantly affected by changes in light treatment in both maize (Fig. 1A) and M. giganteus (Fig. 1B), with lower A net under blue light compared with RGB, red, and green light at 900 mmol quanta m 22 s 21 . In maize, it took 20 to 30 min for A net to stabilize when transitioned from either red or green light to blue light (Fig. 1A). However, in M. giganteus, A net stabilized within minutes under these light transitions (Fig. 1B). Both stomatal conductance (g s ) and the ratio of intercellular to ambient CO 2 partial pressure (P i /P a ) responded within minutes to changes in light quality in maize; however, g s was lower and P i /P a was higher under blue light relative to other light treatments ( Fig. 1, C and E). For M. giganteus, changing light treatment from RGB, red, and green to blue resulted in a gradual decrease in g s , whereas switching light from blue to red, green, and RGB caused a gradual increase in g s (Fig. 1D). Overall, the responses of g s to variation in light treatment in M. giganteus were not as rapid as in maize, and P i /P a was higher under blue light relative to other light treatments in M. giganteus (Fig. 1F).
Δ 13 C and f in maize ( Fig. 1, G and I) were less sensitive to changes in light treatments than in M. giganteus (Fig. 1,H and J). Values of Δ 13 C and f increased in M. giganteus during light transitions from blue to red, green, or RGB and decreased when switched from red or green to blue (Fig. 1, H and J). The Δ 13 C in maize was linearly related to P i /P a under all light qualities, consistent with a constant f ( Fig. 2A); however, the response in M. giganteus indicated a shift in f with light quality (Fig. 2B). In both species, Δ 13 C and f tended to converge to a constant value after 1 h of illumination in all light treatments (Figs. 1 and 2). Variation of Δ 13 C in maize was mainly related to changes in P i /P a , whereas in M. giganteus, shifts of Δ 13 C were associated with changes in both P i /P a and f (Fig. 2).
The allocation of absorbed excitation energy between the C 3 and C 4 cycles (x) was modeled by solving for the value of x that minimizes the difference between the modeled (see Eq. 3) and measured values of Δ 13 C, assuming that bundle-sheath conductance did not change with light quality and was equal to values estimated under RGB. To account for the measured 3.5‰ increase in D 13 C in M. giganteus, the x value had to increase 20%, with lower values under blue light compared with the other light treatments (Fig. 3). In maize, the required change in x needed to explain D 13 C was small compared with that in M. giganteus. The modeled bundle-sheath CO 2 partial pressure (P bs ) was lower and the Rubisco oxygenation-carboxylation ratio (v o /v c ) was higher under blue light compared with other light treatments in both maize and M. giganteus (Fig. 3, C-F).

Linear Electron Flux and Proton Motive Force
Measurements of A net , g s , and P i /P a made with the LI-COR 6400 coupled to the electrochromatic shift (ECS) system were similar to those made with the LI-COR 6400 coupled to the tunable diode laser absorption spectroscope (TDLAS; compare Fig. 1 and Supplemental  Fig. S1). The efficiency of photosystem II (f PSII ) was higher under red light compared with blue light for both maize and M. giganteus (Supplemental Fig. S2D). However, the total ECS was lower under blue light compared with red light, and this difference was larger in maize compared with M. giganteus (Supplemental /LEF and f were measured on different leaves in separate gas-exchange systems; however, for comparison, measurements of A net , g s and P i /P a from both setups are presented in Supplemental Figure S1. , g s (mol m 22 s 21 ; C and D), P i /P a (unitless; E and F), D 13 C (‰; G and H), and f (I and J) in maize and M. giganteus with changes in light treatment in the light sequence RGB, blue, red, blue, green, blue, and RGB. The light intensity was set to 900 mmol quanta m 22 s 21 for all light treatments. The CO 2 partial pressure in the leaf chamber was maintained at 38.4 Pa, leaf temperature was 25˚C, and relative humidity was between 50% and 70% for all light treatments. Data are reported as means 6 SE (n = 4).

Pools of Photosynthetic Intermediates and Rubisco Activation State
There was more oxaloacetic acid (OAA) under blue light for both species (Supplemental Table S1); however, the change in concentration was slower in maize than in M. giganteus ( Fig. 5A; Supplemental Table S2). Pyruvate was constant under red light, but there was a significant change in content with time after the transition to blue light for both species (Supplemental Table S2), which took 30 min to reach a steady state in M. giganteus and only 10 min in maize (Fig. 5C). It took 30 min for RuBP to reach steady state under red light in maize but only 10 min in M. giganteus (Fig. 5D). Under blue light transitions, RuBP increased more in M. giganteus than in maize (Fig. 5D). The amount of triose phosphate (TP) was significantly lower under blue light compared with red light, whereas pyruvate was higher under blue light compared with red light for both maize and M. giganteus ( Fig. 5E; Supplemental Tables S1 and S2). Phosphoenolpyruvate (PEP) and phosphoglycerate (PGA) were not significantly different between light treatments ( Fig. 5F; Supplemental Table S2).
The ratios of C 3 metabolites (TP + RuBP + PGA) over C 4 metabolites (OAA + PEP + pyruvate) were significantly greater under red light compared with blue light treatment for both maize and M. giganteus ( Fig. 6A; Supplemental Table S2). There was no statistically significant difference in PGA/PEP, but TP/PGA was lower under blue light compared with red light for both maize and M. giganteus (Fig. 6, B and D; Supplemental Table S2). The ratio of RuBP to PGA reached steady state within 10 min in M. giganteus but took 30 min in maize (Fig. 6C). After the transition to blue light, RuBP/PGA increased within 2 min for maize but took 10 min in M. giganteus (Fig. 6C). The activation state of Rubisco was constant across light treatments for both species (Supplemental Table S2).

Chloroplast Movement
Light quality-driven movement and positioning of chloroplast can be monitored through changes in leaf light transmittance (Inoue and Shibata, 1973;Trojan and Gabrys, 1996;DeBlasio et al., 2003). In leaves of maize and M. giganteus, chloroplast movement was monitored through changes in leaf transmittance (DT) during 1-h exposures to RGB, blue, and red light under similar conditions used for the gas exchange and online discrimination measurements. In M. giganteus, there was little DT over the 60-min time period regardless of light treatment (Fig. 7). However, in maize, there was a relatively large response of DT under all light treatments compared with M. giganteus (Fig. 7).

Light Quality and A net
After a 1-h acclimation, the steady-state A net was lower under blue light when compared with RGB, red light, and green light in both maize and M. giganteus (Fig. 1, A and B). The lower photosynthetic rate under blue light is consistent with previous findings for both C 3 (Evans and Vogelmann, 2003;Loreto et al., 2009) and C 4 (Evans et al., 2007;Sun et al., 2012) species. Low A net under steady-state blue light has been discussed previously by Sun et al. (2012); however, in this study, there were changes in f unrelated to changes in A net that can provide insight into the mechanisms controlling the coordination of C 4 photosynthesis. For example, transitions from blue light to any other light quality lead to a relatively fast recovery of A net in both species; however, f increased sharply in M. giganteus, but there was little change in maize (Fig. 1). Additionally, in maize, A net decreased significantly when shifted to blue light but without a change in f (Fig. 1), suggesting that f can be maintained regardless of changes in A net . It has been demonstrated that A net is typically determined by total available energy, whereas f is determined by x (Siebke et al., 1997;von Caemmerer, 2000;Ubierna et al., 2011;Pengelly et al., 2012). Over time, the coordination of both cycles was achieved under all light qualities; however, the rate at which this balance was achieved differed between maize and M. giganteus.
Blue light reduction in rates of C 4 photosynthesis has been described before by Sun et al. (2012); therefore, this article focuses on the dynamic coordination between the C 3 and C 4 cycles during the transition from one light quality to another. In brief, the decrease in photosynthesis under blue light is the result of reduced energy availability to drive ATP and NADPH production. This is demonstrated by the measured changes in metabolite pools and the decrease in f PSII under blue light. First, the analysis of photosynthetic metabolite concentrations demonstrates lower ATP and NADPH availability under blue light when A net is lower in both species. For example, pyruvate content under blue light was greater than under red light ( Fig. 5C; Supplemental Table S2), which implies insufficient energy supply under blue light to convert pyruvate to PEP in mesophyll cells von Caemmerer, 1988, 1989;Sun et al., 2012). Moreover, TP content was lower under blue light relative to red light ( Fig. 5E; Supplemental Table S2), whereas there were no differences in PGA content between red and blue light ( Fig. 5F; Supplemental Table S2). These results suggest a lack of reducing equivalent for the reduction of PGA to TP.
There was also a strong positive relationship between A net and f PSII (Fig. 4A) in response to light treatments, with lower f PSII under blue light (Supplemental Fig. S2). During C 4 photosynthesis, there is typically a linear dependence of f PSII and rates of CO 2 assimilation because the assimilation of CO 2 is the primary sink for NADPH generated by electron flux through PSII (Edwards and Baker, 1993). The lower f PSII under blue light observed in both species may result from (1) an imbalance in absorption between PSI and PSII within the mesophyll chloroplast, leading to an overreduction of the chloroplast electron transport chain (Evans, 1986);  (2) a change in the distribution of light energy absorption between the PSII-enriched mesophyll and the PSIIdepleted bundle-sheath chloroplast; (3) an increase in nonphotochemical quenching; and/or (4) an overall decrease in the absorption of energy by the chlorophyll under blue light due to strong absorption of blue light by carotenoids and nonphotosynthetic structures within the leaf. As discussed below, some of these explanations will not only affect the total available energy but also x and, therefore, the overall efficiency of the CO 2 -concentrating mechanism.
Coordination of the C 3 and C 4 Cycles in Response to Transient Changes in Light Quality Measurements of Δ 13 C indicated transitory changes in the coordination of the CO 2 -concentrating mechanism in response to changes in light quality for both maize (Fig. 1I) and M. giganteus (Fig. 1J). This response lasted about 1 h in M. giganteus but only a few minutes in maize (Figs. 1 and 2). As discussed above, and in more detail below, these differences between species may be associated with the ability to coordinate the production of ATP and NADPH under changing light environments. Transition from any light treatment to blue light caused a reduction in Δ 13 C, while the reverse transition from blue light to any other light quality resulted in increased Δ 13 C ( Figs. 1 and 2). A comparison of leaf-level models of Δ 13 C and C 4 photosynthesis was used to investigate the variation of f in response to rapid changes in light quality. This consisted of fitting measurements of Δ 13 C and gas exchange into theoretical models of C 4 photosynthesis and leaf CO 2 discrimination (Ubierna et al., 2013) and solving for several parameters of interest. These included x, P bs , and the rates of Rubisco oxygenation (v o ) versus carboxylation (v c ; Fig. 3). The C 4 photosynthesis model assumes steady-state conditions, which may be violated during the transition in light quality. However, there were no conditions where modeled parameters were outside physiologically feasible values. Additionally, this modeling assumed that the conductance of CO 2 between the bundle-sheath and mesophyll cells was constant. This may be an oversimplification; however, there is currently no evidence that bundle-sheath conductance (g bs ) changes in response to short-term manipulations in measurement conditions. Given these assumptions, these data illustrated that the transition from blue light to other light qualities resulted in a larger fraction of available energy allocated to the C 4 cycle, resulting in an increase in P bs , a decrease in v o /v c , and an increase in f (Fig. 3, A and B).
The change in f was larger in M. giganteus compared with maize. This suggests a faster coordination of the C 3 and C 4 cycles in maize compared with M. giganteus. Unfortunately, this modeling does not specifically identify the mechanism driving the change in energy allocation, but it does demonstrate that changes in light quality cause variations in the production of energy and reducing equivalents between the mesophyll and bundle-sheath cells.
Additionally, under blue light, there was a lower ratio of v H + /LEF (Fig. 4B). A lower v H + /LEF could be interpreted as lower activity of bundle-sheath cyclic electron flux; however, the ECS measurements integrate across the leaf cross section, and it is impossible to distinguish between the contributions of mesophyll and bundle-sheath cells. Nevertheless, it is reasonable that blue light would decrease cyclic electron flux preferentially in the bundle-sheath cells because of the limited penetration of blue light into the bundle-sheath cells (Evans and Vogelmann, 2003;Evans et al., 2007) and the insufficient blue light absorption by bundlesheath PSI (Evans, 1986). This would limit the Calvin-Benson-Bassham cycle, because approximately 50% of the ATP requirement for carbon assimilation is thought to be provided by bundle-sheath cyclic electron flux (Kanai and Edwards, 1999). In M. giganteus, there was a positive relationship of f and v H + /LEF, with higher f and higher v H + /LEF under red light compared with blue light. However, over time after the initial change in light quality, there was a relatively large decrease in f and only a slight increase in v H + /LEF for M. giganteus. Alternatively, in maize, there was no change in f between light treatments, but v H + /LEF increased from blue to red light (Fig. 4D). This indicates that changes in v H + /LEF and corresponding changes in ATP/NADPH production in maize may help maintain a balance in metabolic flux between the C 3 and C 4 cycles and low f. For example, in maize, increased cyclic electron flux under red light could drive additional phosphoenolpyruvate carboxylase (PEPC) refixation of leaked CO 2 from the bundle-sheath cells, maintaining a constant D 13 C and f. In M. giganteus, there was only a small shift in v H + /LEF between blue and red light compared with maize, suggesting that change in cyclic electron flux in M. giganteus was not the only mechanism that ultimately balances the energy production and coordination between the C 3 and C 4 cycles. These findings in M. giganteus are Figure 7. Transitional variation in DT in maize (black circles) and M. giganteus (white circles) when light treatment was switched in the sequence of RGB, blue, red, blue, and RGB. DT is defined as the difference between the initial averaged 5 min of transmittance of a given light quality and the transmittance at a given time point within the 60-min exposure. Data are reported as means 6 SE (n = 4 for maize; n = 3 for M. giganteus). similar to those found in F. bidentis, where the ratio of PSI to PSII quantum yields did not change in plants with higher f due to genetically reduced levels of Rubisco, suggesting that both linear and cyclic electron flux contribute to the production of ATP with increased f (Siebke et al., 1997) Chloroplast movement could also influence the metabolic and energy coordination between the C 3 and C 4 cycles. It has long been reported that blue light induces chloroplast movement to optimize photosynthetic activity and/or to prevent photodamage under strong light in both C 3 and C 4 species (Senn, 1908;Inoue and Shibata, 1974;Kagawa et al., 2001;Wada et al., 2003;Yamada et al., 2009;Luesse et al., 2010;Maai et al., 2011). In C 4 plants, blue light induced chloroplast movement in mesophyll but not in bundlesheath cells (Yamada et al., 2009). This differential response could result in changes to the energy allocation between the C 3 (bundle-sheath) and C 4 (mesophyll) cycles, different rates of linear versus cyclic electron flux, and changes in the metabolite flux between both compartments, all of which could affect the coordination and efficiency of C 4 photosynthesis. There was more chloroplast movement, estimated as DT, in response to the transition in light quality in maize compared with M. giganteus (Fig. 7). This corresponded to the transitional variation in f and x in response to light quality in M. giganteus but not in maize. Although the apparent chloroplast movement was relatively small, it does suggest that, in maize, it may help balance light energy distribution between the mesophyll and bundle-sheath cells, leading to a coordination of C 3 and C 4 cycle activities. In contrast, the lack of chloroplast movement in M. giganteus might be related to the longer time to coordinate the activities of C 3 and C 4 cycles in this species.
Although f responded differently to light transitions in maize and M. giganteus, there were no significant differences in the rapid response of the contents and ratios of metabolites between species (Figs. 5 and 6; Supplemental Tables S1 and S2). However, as demonstrated previously by Sun et al. (2012), there were strong correlations between A net and various metabolite concentrations (Figs. 5E and 6; Supplemental Tables S1 and S2). However, the lack of detectible transitional variation in photosynthetic metabolites may partially be attributed to the inability to measure metabolite fluxes between the mesophyll and bundle-sheath cells as well as an inability to determine mesophyll-or bundle-sheathspecific changes in metabolite concentrations.
Finally, the observed transitional variation in A net and f could be caused by light quality-associated differences in stomatal heterogeneity. Stomatal heterogeneity or patchiness has been observed in many species, particularly in response to environmental stress conditions (Terashima et al., 1988;Mott and Buckley, 1998). Stomatal heterogeneity could alter the estimation of Δ 13 C and P i /P a by affecting the relationship between assimilation rate and stomatal conductance (Lloyd et al., 1992). However, Lloyd et al. (1992) demonstrated that the effects of stomatal heterogeneity on the estimation of Δ 13 C are significant only under low light and low rates of stomatal conductance. We measured transitional variation in CO 2 assimilation rate and Δ 13 C at a constant photon density of 900 mmol m 22 s 21 for all light treatments. Furthermore, the rates of stomatal conductance were relatively high under all light treatments (Fig. 1, C and D). Therefore, the effects of stomatal heterogeneity on the variation in A net and Δ 13 C are likely minimal in both species.

CONCLUSION
The in vivo coordination of the C 3 and C 4 cycles is complex and difficult to evaluate, requiring in many cases the interpretation of indirect evidence. We used several tools, including measurements of gas exchange, Δ 13 C, metabolite pools, ECS, and chloroplast movement, to characterize C 4 photosynthesis in response to changes in light quality. In both maize and M. giganteus, there was a lower efficiency of PSII under blue light, suggesting that the reduction in A net under this condition was caused by insufficient production of NADPH through LEF. This was further supported by measurement of the pools of C 3 and C 4 metabolites. In maize, x was quickly optimized to minimize f; however, it took almost 1 h for M. giganteus to achieve steady-state values of f. Changes in linear versus cyclic electron flux in M. giganteus were slower to respond to light quality compared with those in maize. This may have led to an imbalance in the coordination of the C 3 and C 4 pathways. Additionally, the rapid rearrangement of chloroplasts under blue light in maize likely optimized light energy production between the mesophyll and bundlesheath cells. Together, these responses in maize may have helped coordinate the C 4 photosynthetic pathway and potentially increased PEPC refixation of CO 2 leaked from the bundle-sheath cells. In M. giganteus, there was a quick change in f PSII , but cyclic electron flux and chloroplast movement were not as responsive to changes in light quality. This potentially resulted in an imbalance of energy, leading to the overpumping of CO 2 and increased f. However, in both species, given enough time, the activities of C 3 and C 4 cycles appear to balance and maintain the efficiency of the C 4 -concentrating mechanism.

Plants
Maize (Zea mays var. Trucker's Favorite; Victory Seed Company) seeds and Miscanthus 3 giganteus rhizomes were planted in 6-L pots. The plants were grown in a greenhouse with day temperature of 25°C to 28°C, night temperature of 20°C to 25°C, and daylength of 14 h. Illumination in the greenhouse was a combination of sunlight and supplementary light provided by 400-W high-pressure sodium lamps. Plants were watered daily and fertilized weekly with Peters 20-20-20 (J.R. Peters) and a slow-release fertilizer (17-3-6 monopotassium phosphate). Four-week-old maize and 3-month-old M. giganteus were used for the measurements.
Gas Exchange and Online Leaf Δ 13 C Leaf gas exchange and online Δ 13 C were measured on the uppermost fully expanded maize and M. giganteus leaves using the LI-COR 6400xt gasexchange analyzer with the opaque conifer chamber and the RGB light source (LI-COR 6400-18; LI-COR Biosciences) coupled to a TDLAS (TGA 100A; Campbell Scientific). The RGB light source generates a broad-spectrum red, green, and blue light with peak wavelengths of 635, 522, and 460 nm and bandwidths of 16, 35, and 24 nm, respectively. The 12 CO 2 and 13 CO 2 partial pressures in the LI-COR reference and sample cells were measured by the TDLAS concurrently with a CO 2 -free tank and two standard tanks (Liquid Technology) with known 12 CO 2 and 13 CO 2 partial pressures. The TDLAS was sampled from each of the five sites at a flow rate of 150 mL min 21 and a frequency of 40 s per site. Only the last 10 s at each site was used for the calculation of 12 CO 2 and 13 CO 2 partial pressures. The partial pressures of 12 CO 2 and 13 CO 2 in the reference and sample lines were calibrated using a gain and offset calculated from the two calibration tanks (Bowling et al., 2003;Ubierna et al., 2011;Sun et al., 2012). The leaves were acclimated to 900 mmol m 22 s 21 RGB for 1 h in the opaque conifer chamber. The light treatment was then shifted in the order RGB, blue, red, blue, green, blue, and RGB, with an interval of 1 h per light treatment. The irradiance of 900 mmol m 22 s 21 was the maximum light intensity all light treatments could achieve with the LI-COR 6400-18 RGB light source. The CO 2 partial pressure in the leaf chamber was maintained at 38.4 Pa, a leaf temperature of 25°C, and a relative humidity between 50% and 70% for all gas-exchange measurements. The simultaneous gas exchange and Δ 13 C measurements were conducted on four individual plants of maize and M. giganteus.
D 13 C was calculated as : where d e and d o represent the isotopic composition of CO 2 entering and leaving the leaf chamber, respectively; j is calculated as: where P e and P o are the partial pressures of CO 2 in the dry air entering and leaving the leaf chamber, respectively.

Leakiness
Following the recommendation of Ubierna et al. (2013), CO 2 f was estimated by rearranging the equation proposed by Farquhar (1983) and Farquhar and Cernusak (2012): Cð1 2 tÞP a 2 aðP a 2 P i Þ 2 ð1 þ tÞP i b 4 ð1 þ tÞ½b 3 P bs 2 sðP bs 2 P i Þ þ aðP a 2 P i Þ 2 P a D 13 Cð1 2 tÞ ð3Þ where P a , P i , and P bs are CO 2 partial pressures in the atmosphere, the intercellular air spaces, and the bundle-sheath cells, respectively. The term a is the weighted fractionation across the boundary layer and stomata: where a b (2.9‰) and a s (4.4‰) are the fractionation across the boundary layer and stomata, respectively. P L is CO 2 partial pressure at the leaf surface. The terms b 3 and b 4 are defined as (Farquhar, 1983): where R d is the nonphotorespiratory CO 2 release in the light (assumed to equal measured rates of dark respiration, 1.80 and 1.27 mmol m 22 s 21 for maize and M. giganteus, respectively); R m is the mesophyll cell dark respiration rate, which is assumed to be half of the R d (von Caemmerer, 2000); v c and v o are as already defined and v p is the PEP carboxylation rate (for details on the modeling of v c , v o , v p , and P bs , see Ubierna et al. [2011]); b9 3 is fractionation by Rubisco, 30‰ (Roeske and O'Leary, 1984); b9 4 represents the net effect of CO 2 dissolution, hydration, and PEPC activity, which at 25°C has a value of 25.7‰ (Farquhar, 1983;Henderson et al., 1992); f is fractionation during photorespiration, 11.6‰ ; and e9 is fractionation associated with dark respiration, which is calculated as (Wingate et al., 2007): where e is fractionation during decarboxylation, assumed to equal 26‰ (Ghashghaie et al., 2001;Sun et al., 2010); d 13 C measurement (231‰) is the carbon isotope signature of CO 2 used for the online discrimination measurement, which was measured by the TDLAS; and d 13 C growth is the carbon isotope composition of CO 2 of the growth conditions, which is assumed to be 28‰. The ternary effects term t is defined as (Farquhar and Cernusak, 2012): where E is the transpiration rate, g ac is the conductance to diffusion of air in CO 2 , and a ac ð¼ 1 þ aÞ is the fractionation factor for the isotopologs of CO 2 diffusing in air. g bs was estimated from the least-square difference between predicted and observed leaf discrimination using data obtained from the first 1-h RGB measurements (Kromdijk et al., 2010;Ubierna et al., 2011). This resulted in g bs values of 0.0046 mmol m 22 s 21 Pa 21 for maize and 0.01 mmol m 22 s 21 Pa 21 for M. giganteus, which are within the range of values previously suggested for this parameter of 0.005 to 0.02 mmol m 22 s 21 Pa 21 (He and Edwards, 1996;von Caemmerer and Furbank, 2003).

Modeling Energy Partitioning between the C3 and C4 Cycles (x)
The light-limited C 4 photosynthesis model (von Caemmerer, 2000), which describes the relationships between total electron flux (J t ) and A net , was used to model how x could explain the changes in D 13 C observed in Figure 1. J t can be partitioned between the C 4 cycle (J m ) and the C 3 cycle (J s ) by x, which is the portion of ATP required for J m , and 1 2 x is the requirement for J s . The C 4 photosynthetic pathway has a theoretical minimum energetic cost of five ATPs, three for the C 3 cycle and two required to regenerate PEP from pyruvate in the mesophyll cells, so it is generally assumed that x = 0.4 (von Caemmerer, 2000). Therefore, during the first 1 h of RGB, x was assumed to be 0.4; however, for other light treatments, x was solved to minimize the least-square difference between modeled and measured Δ 13 C. Currently, there is no evidence that g bs changes in response to variation in measurement conditions. Therefore, during this modeling, it was assumed that g bs (0.0046 mmol m 22 s 21 Pa 21 for maize and 0.01 mmol m 22 s 21 Pa 21 for M. giganteus) and dark-type respiration (1.80 and 1.27 mmol m 22 s 21 for maize and M. giganteus, respectively) were constant for all light treatments and mesophyll conductance to CO 2 was infinite. Additionally, the modeled relationship between light, A net , and J t was used to calculate P bs and v o /v c (Fig. 3), as described (Ubierna et al., 2011;Supplemental Equations S1). To demonstrate the sensitivity of x, P bs , and v o /v c to g bs , these parameters were modeled with g bs of 0.0069 and 0.0023 mmol m 22 s 21 Pa 21 for maize and 0.015 and 0.005 mmol m 22 s 21 Pa 21 for M. giganteus (Fig. 3, shaded areas).

Chlorophyll a Fluorescence and ECS
Measurements were performed with a LI-COR 6400 gas-exchange analyzer coupled to a nonfocusing optics spectrophotometer/chlorophyll fluorometer (Avenson et al., 2004(Avenson et al., , 2005Kiirats et al., 2010). In brief, 0.79 cm 2 of leaf was clamped into a custom-built leaf chamber that allowed simultaneous measurements of net CO 2 exchange, chlorophyll fluorescence, and leaf optical properties. The leaf chamber was maintained at 38.4 Pa of CO 2 and 900 mmol photon m 22 s 21 . Leaves were acclimated for 1 h under blue actinic light (460 nm), and then the actinic light was switched to red (637 nm) for 1 h and then to blue for 1 h. Measurements of chlorophyll a fluorescence and 520-nm absorbance in the dark were taken every 10 min per light treatment. The measuring and actinic light were the same wavelength for the measurements of chlorophyll a fluorescence.
The f PSII derived from fluorescence measurements (Genty et al., 1989) was used to estimate the LEF as f PSII 3 Abs leaf 3 fraction PSII (Donahue et al., 1997;Avenson et al., 2005), where Abs leaf and fraction PSII are leaf absorbance and the fraction of light energy capture by PSII as compared with total photosystems, respectively. Leaf absorbance was determined from measurements of reflectance and transmittance as described previously by Sun et al. (2012), and the fraction PSII was assumed to be 0.5. Absorbance changes at 520 nm from rapid light-to-dark transitions were used to estimate ECS. The total proton motive force was defined as the amplitude of ECS decay from steady state to a quasistable minimum after a 300-ms dark period (Avenson et al., 2004(Avenson et al., , 2005. The relative rate of v H + was estimated by ECS decay kinetics (Avenson et al., 2004(Avenson et al., , 2005Livingston et al., 2010).

Rapid Freeze Clamping of Leaves and Photosynthetic Metabolite Measurements
Leaves similar in age and position to those used for the transitional gas exchange and Δ 13 C measurements were rapidly freeze clamped in situ using a custom-built rapid kill system attached to a LI-COR 6400 gas-exchange analyzer (Badger et al., 1984;Hendrickson et al., 2008). Leaves were acclimated to either red or blue light for 1 h at a light intensity of 900 mmol quanta m 22 s 21 , a sample CO 2 partial pressure of 38.4 Pa, and a leaf temperature of 25°C. Leaves were then rapidly freeze clamped between two liquid nitrogen-cooled copper rods at 2, 10, 30, and 60 min after light was switched from red to blue or blue to red. The frozen leaf discs were stored in a 280°C deep freezer prior to extraction for photosynthetic metabolites and Rubisco activation state assays (Sun et al., 2012). Leaf discs (approximately 4.5 cm 2 ) were ground to a fine powder in liquid nitrogen in a precooled pestle and mortar and extracted with 1 M HClO 4 (Leegood and von Caemmerer, 1988). OAA, pyruvate, and PEP were measured consecutively in an enzymatic assay by monitoring changes in absorbance at 340 nm (Leegood and Furbank, 1984). TP, PGA, and RuBP were also assayed spectrophotometrically by coupled enzyme as described previously (He et al., 1997) using Rubisco purified from tobacco (Nicotiana tabacum).

Rubisco Activation State
Rubisco activity was determined by the incorporation of 14 CO 2 into acidstable products at 25°C (Salvucci and Anderson, 1987). The initial activity was determined by adding 20 mL of the leaf extract to the reaction mixture (100 mM Tricine-NaOH, pH 8.0, 10 mM MgCl 2 , 10 mM NaH 14 CO 3 [0.2 mCi mmol 21 ], and 0.4 mM RuBP) and quenching the reaction after 30 s with 100 mL of acid (1 N HCl/4 N formic acid). Total activity was measured after incubating 20 mL of the same leaf extract in the reaction mixture for 3 min at 25°C minus RuBP to carbamylate all Rubisco catalytic sites. Subsequently, the reaction was initiated with RuBP and quenched after 30 s. Rubisco activation state was determined by the ratio of initial to total Rubisco activity.

Chloroplast Movement
Chloroplast movement was determined from the DT during exposure to different wavelengths of light (DeBlasio et al., 2003). Leaves were clamped to an integrating sphere (Labsphere) in a custom-built apparatus that included a multispectral light-emitting diode light source (Cree) that could be toggled between preset light qualities. Red, RGB, and blue light intensity was adjusted so that incident light had a similar intensity to that used during gas-exchange measurements. A photometer (LI-COR Biosciences) was attached to the integrating sphere, and light intensity was averaged over 5-min intervals using a data logger (Campbell Scientific). Leaves were exposed for 1 h each to the following light regime: RGB, blue, red, blue, and RGB. Reference traces were measured under the same regime without a leaf and used to calculate DT according to: where I t and I r are the light intensities measured under a given light treatment and time point for the leaf and reference runs, respectively. I t0 and I r0 are the averaged light intensities from the leaf and reference runs measured during the initial 5 min of each light quality treatment. Therefore, DT is defined as the difference between the initial averaged 5 min of transmittance of a given light quality and the transmittance at a given time point within the 60-min exposure.

Statistical Analyses
Three-way ANOVA was conducted to assess the effects of species (maize and M. giganteus), time (2, 10, 30, and 60 min), and light treatment (red and blue) on the contents of OAA, pyruvate, PEP, RuBP, TP, and PGA, the ratios of C 3 to C 4 metabolites (TP to PGA, RuBP to PGA, and PGA to PEP), and Rubisco activation state. Statistical analyses were carried out using SAS version 9.0 (SAS Institute). Average values are reported as arithmetic means 6 SE.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Comparison in gas exchange measurements between gas analyzers coupled to a tunable diode laser and an electrochromatic shift spectrophotometer.
Supplemental Figure S2. Transitional variation in gas exchange parameters, chlorophyll a fluorescence, and electrochromatic shift.
Supplemental Equations S1. Calculation of total electron transport rate and CO 2 partial pressure inside the bundle-sheath cells.