The Sclerophyllous Eucalyptus camaldulensis and Herbaceous Nicotiana tabacum Have Different Mechanisms to Maintain High Rates of Photosynthesis

It is believed that high levels of mesophyll conductance (gm) largely contribute to the high rates of photosynthesis in herbaceous C3 plants. However, some sclerophyllous C3 plants that display low levels of gm have high rates of photosynthesis, and the underlying mechanisms responsible for high photosynthetic rates in sclerophyllous C3 plants are unclear. In the present study, we examined photosynthetic characteristics in two high-photosynthesis plants (the sclerophyllous Eucalyptus camaldulensis and the herbaceous Nicotiana tabacum) using measurements of gas exchange and chlorophyll fluorescence. Under saturating light intensities, both species had similar rates of CO2 assimilation at 400 μmol mol−1 CO2 (A400). However, E. camaldulensis exhibited significantly lower gm and chloroplast CO2 concentration (Cc) than N. tabacum. A quantitative analysis revealed that, in E. camaldulensis, the gm limitation was the most constraining factor for photosynthesis. By comparison, in N. tabacum, the biochemical limitation was the strongest, followed by gm and gs limitations. In conjunction with a lower Cc, E. camaldulensis up-regulated the capacities of photorespiratory pathway and alternative electron flow. Furthermore, the rate of alternative electron flow was positively correlated with the rates of photorespiration and ATP supply from other flexible mechanisms, suggesting the important roles of photorespiratory pathway, and alternative electron flow in sustaining high rate of photosynthesis in E. camaldulensis. These results highlight the different mechanisms used to maintain high rates of photosynthesis in the sclerophyllous E. camaldulensis and the herbaceous N. tabacum.


INTRODUCTION
In C 3 plants, rates of photosynthesis differ widely among species. For individual leaves or whole plants, photosynthetic capacity mainly depends upon their biochemical composition and morphology. Generally, plants with high rates of CO 2 assimilation have higher levels of cytochrome f, ATP synthase, Rubisco, and other Calvin-Benson cycle enzymes (Evans, 1987;Terashima and Evans, 1988;Hikosaka, 1996;Hikosaka and Terashima, 1996;Yamori et al., 2010Yamori et al., , 2011. The rate of CO 2 assimilation is maximized in leaves that usually have high levels of stomatal conductance (g s ) and mesophyll conductance (g m ), which increase CO 2 diffusion into the chloroplasts (Yamori et al., 2010(Yamori et al., , 2011. Herbaceous plants, e.g., tobacco (Nicotiana tabacum), spinach (Spinacia oleracea), rice (Oryza sativa), and Triticum aestivum have commonly been used for studying the mechanism of photosynthetic acclimation and the rate-limiting step for CO 2 assimilation. In herbaceous C 3 plants, the ratelimiting step of photosynthesis depends on leaf N content and is mainly determined by N portioning between Rubisco and photosynthetic electron transport (Yamori et al., 2011). However, little is known about the coordination of photosynthetic electron flow and gas exchange in sclerophyllous plants that have high rates of photosynthesis. What is more, it is still not well understood whether the mechanisms responsible for high rates of photosynthesis vary among sclerophyllous and herbaceous C 3 plants.
Leaf anatomy plays an important role in determining photosynthetic capacity. The area of chloroplasts facing the intercellular space largely determines the light-saturated rate of photosynthesis (Oguchi et al., 2003(Oguchi et al., , 2005. High-photosynthesis herbaceous plants usually have thinner cell walls, leading to high values of g m . By comparison, leaves of sclerophyllous plants have thicker cell walls, with a high leaf dry mass per area (LMA) (Hassiotou et al., 2009). As an important morphological trait, LMA is inversely related to g m in sclerophyllous plants (Hassiotou et al., 2009). Therefore, sclerophyllous plants usually have low g m and slow rates of photosynthesis . For example, a sclerophyllous species Quercus guyavifolia has a low rate of the maximum photosynthesis being 13 µmol CO 2 m −2 s −1 (Huang et al., 2016). As we known, the sclerophyllous species Eucalyptus camaldulensis has been introduced for production of paper in China due to its high rate of photosynthesis. The high value of LMA for leaves of E. camaldulensis is assumed to result in low g m . Mesophyll conductance plays an important role in determining the rate of photosynthesis in C 3 plants (Flexas et al., 2008;Carriqui et al., 2015). A low g m value increases the resistance of CO 2 conductance to the chloroplasts, leading to a decline in the chloroplast CO 2 concentration (C c ) and, thus, restricted CO 2 assimilation Hanba et al., 2002;Flexas et al., 2012;Gago et al., 2013;Carriqui et al., 2015). Therefore, for the sclerophyllous E. camaldulensis, other mechanisms favoring CO 2 diffusion must be used to increase C c because it is essential for the maintenance of high CO 2 assimilation.
The net rate of CO 2 assimilation (A n ) is largely dependent upon the value of C c because the latter directly determines the affinity of Rubisco to CO 2 or O 2 (Farquhar et al., 1980;von Caemmerer, 2000). An increased C c increases the rate of RuBP carboxylation and, thus, results in a rise in the photosynthetic rate. As shown by the calculation of C c = C i − A n /g m , C c is mainly determined by three factors: intercellular CO 2 concentration (C i ), A n , and g m . During the steady-state phase under high light, A n and g m reach the steady-state values, the value of C i determines C c principally but is mainly influenced by g s . Stomata are the channels for gas exchange between leaf and atmosphere. During periods of drought or high temperatures, a decrease in g s leads to an inhibition of the Calvin-Benson cycle . In herbaceous plants of high values of g m , high rates of photosynthesis are usually accompanied by high values of g s (Yamori et al., 2010(Yamori et al., , 2011. However, it is unclear whether the high-photosynthesis sclerophyllous plant E. camaldulensis elevate g s to remedy the deficiency of g m . According to the C 3 photosynthesis model, photosynthesis can be limited by RuBP carboxylation and/or RuBP regeneration (Farquhar et al., 1980). When C c is higher than C trans (the chloroplast CO 2 concentration at which the transition from RuBP carboxylation limitation to RuBP regeneration limitation occurs), then CO 2 assimilation is limited by RuBP regeneration (Yamori et al., 2010(Yamori et al., , 2011. Once C c is lower than C trans , CO 2 assimilation tends to be limited by RuBP carboxylation. In herbaceous N. tabacum plants grown at high nitrogen concentration, the rate-limiting step of CO 2 assimilation is RuBP regeneration because C c is greater than C trans (Yamori et al., 2010(Yamori et al., , 2011. In the sclerophyllous E. camaldulensis, the high rate of photosynthesis and low g m can cause C c to be less than C trans . Consequently, the rate of CO 2 assimilation is probably limited by RuBP carboxylation in E. camaldulensis. If this occurs, then the reduced C c drives increased photorespiration (or RuBP oxygenation). The photorespiratory pathway is essential for photosynthesis at normal atmospheric CO 2 concentrations (Chastain and Ogren, 1989;Eisenhut et al., 2007), and impairment of that pathway decreases the rate of photosynthesis under such CO 2 conditions (Somerville and Ogren, 1980Takahashi et al., 2007).
Enhancement of the photorespiratory pathway leads to a considerably improved net photosynthetic rate in Arabidopsis thaliana (Timm et al., 2012(Timm et al., , 2015. Therefore, we might also speculate that E. camaldulensis enhances the capacity of that pathway to favor the Calvin-Benson cycle. Furthermore, if the photorespiratory pathway is up-regulated in plants of E. camaldulensis, then the stoichiometry of the ATP/NADPH energy demand by primary metabolism will increase. Therefore, such plants must utilize other flexible mechanisms to balance the ATP/NADPH ratio, e.g., cyclic electron flow (CEF) around photosystem I (PSI) or the water-water cycle (Makino et al., 2002;Walker et al., 2014). The WWC channels electrons obtained from splitting of water molecules at PSII. These electrons are transported to oxygen via the Cyt b 6 /f complex and PSI, resulting in the formation of a proton gradient across the thylakoid membranes (Asada, 1999(Asada, , 2000. However, little is known about how the WWC functions in the high-photosynthesis sclerophyllous plant E. camaldulensis.
N. tabacum is regarded as a model plant to study the mechanisms of photosynthetic regulation for herbaceous plants. However, it is not known how the sclerophyllous plant E. camaldulensis obtain a high rate of photosynthesis. Here, we compared g s , g m , CO 2 assimilation, photorespiration, and alternative electron flow between N. tabacum and E. camaldulensis. Our objective was to examine the potential differences in mechanisms underlying high rates of photosynthesis between herbaceous and sclerophyllous plants.

Plant Materials and Growth Conditions
We compared the photosynthetic characteristics of N. tabacum and Eucalyptus camaldulensis Dehnh. The latter is a fast-growing species native to Australia that has been widely introduced into China for forest plantations. For this study, Eucalyptus samples were collected from plants grown in an open field at an elevation of 700 m in Dongchuan County, Kunming City, Yunnan Province, China. The monthly air temperature, total radiation and precipitation were displayed in Figure 1 (data were collected from 1961 to 1980). Seedlings of N. tabacum cv. k326 were cultivated in plastic pots in a phytotron at Kunming Institute of Botany, Yunnan, China. Growing conditions were 24/18 • C (day/night), 60% relative humidity, and an atmospheric CO 2 concentration maintained at 400 µmol mol −1 . The phytotron used sunlight as the source of illumination, and plants were exposed to approximately 95% of full sunlight (maximum at noon ≈ 1990 µmol photons m −2 s −1 ). Photosynthetic parameters were measured in June of 2014. Measurements were made using four mature leaves from four independent plants per species. Fully expanded mature leaves on 13-week-old plants of N. tabacum were used for photosynthetic measurements. For E. camaldulensis, mature leaves that flushed in the spring on 3-year-old plants were used for measurements.
The fluorescence parameters F m ′ and F s were evaluated as previously described (Baker and Rosenqvist, 2004), with F m ′ representing the maximum fluorescence after light-adaption and F s being the light-adapted steady-state fluorescence. The effective quantum yield of PSII was calculated as Genty et al., 1989). The maximum fluorescence after dark adaptation (F m ) was examined after 30 min of dark adaptation following measurement of the light response curve. Non-photochemical quenching was calculated as

Estimation of Photosynthetic Electron Flow
Using the data of chlorophyll fluorescence parameters, total photosynthetic electron flow through PSII is calculated as follows (Krall and Edwards, 1992): where PSII is the effective quantum yield of PSII and L abs represents leaf absorbance. We applied the constant of 0.5 based on the assumption that photons were equally distributed between PSI and PSII.
Using the basic equation of leaf carbon dioxide gas exchange and Rubisco specificity for carboxylation relative to oxygenation (von Caemmerer and Farquhar, 1981;Sharkey, 1988;Walker et al., 2016), the rate of Rubisco carboxylation (V c ) and that of Rubisco oxygenation (V o ) are calculated according to.
where A n represented the net rate of CO 2 assimilation, R d was the rate of mitochondrial respiration as measured after 30 min of dark adaptation, Γ * was the CO 2 compensation point in the absence of daytime respiration (Farquhar et al., 1980;Brooks and Farquhar, 1985), and C c was the chloroplast CO 2 concentration. The electron flow for photorespiratory carbon oxidation can be expressed as: The NADPH demands from CO 2 assimilation and photorespiration were calculated according to the models of Farquhar et al. (1980). Using the data from gas exchange measurements, we determined the rate of electron transport for NADPH required by carboxylation and oxygenation of RuBP (J g ) as follows (Zivcak et al., 2013;Walker et al., 2014) where C i was the intercellular CO 2 concentration. The alternative electron flow was calculated as follows (Makino et al., 2002;Zivcak et al., 2013;Huang et al., 2016):

Estimations of Mesophyll Conductance and Chloroplast CO 2 Concentration
Values for mesophyll conductance (g m ) were estimated through a combination analysis of gas exchange and chlorophyll fluorescence, and according to the following equation Loreto et al., 1992;Warren and Dreyer, 2006;Yamori et al., 2010Yamori et al., , 2011: Using the estimated g m , we calculated the chloroplast CO 2 concentration (C c ) according to the following equation (Long and Bernacchi, 2003;Warren and Dreyer, 2006;Yamori et al., 2010Yamori et al., , 2011: The response of net CO 2 assimilation rate to CO 2 concentration was examined at 2000 µmol photons m −2 s −1 and 25 • C. Before A/C i measurement, leaves were light adapted at 2000 µmol photons m −2 s −1 and 400 µmol mol −1 CO 2 concentration for at least 20 min to obtain the maximum values of g s and A n . Afterwards, the CO 2 concentrations were set to 50 µmol mol −1 and increased stepwise. For E. camaldulensis, CO 2 concentrations were set to 0, 50, 100, 150, 200, 300, 400, 600, 800, 1000, and 1200 µmol mol −1 . The CO 2 concentrations in A/C i measurement in N. tabacum were set to 0, 50, 100, 150, 200, 300, 400, 600, 800, 1000, 1200, 1600, 2000 µmol mol −1 . Each stepwise measurement was completed within 2-3 min. Using A/C i curves, we calculated the maximum rates of RuBP regeneration (J max ) and RuBP carboxylation (V cmax ) according to the method of Long and Bernacchi (2003). To identify the limiting step of CO 2 assimilation, we determined the chloroplast CO 2 concentration at which the transition from RuBP carboxylation to RuBP regeneration occurred (C trans ) as follows (Yamori et al., 2010(Yamori et al., , 2011: where K c (µmol mol −1 ) and K o (mmol mol −1 ) were the Michaelis constants for CO 2 and O 2 , respectively (Farquhar et al., 1980), and were assumed to be 406.8 µmol mol −1 and 277 mmol mol −1 , respectively, at 25 • C (Long and Bernacchi, 2003); O was the partial pressure of O 2 and was assumed to be 210 (Farquhar et al., 1980); J max was the maximum rate of RuBP regeneration; and V cmax was the maximum rate of RuBP carboxylation. The rate-limiting step for CO 2 assimilation was then determined by comparing the values of C c and C trans .

Modeling ATP Supplied via Flexible Mechanisms
The total amount of ATP demand from Rubisco carboxylation and oxygenation was obtained with the following formula (Walker et al., 2014): Assuming that the stoichiometry of ATP/NADPH produced by LEF (electron transport from PSII to NADP + ) is 1.29 (Sacksteder et al., 2000;Seelert et al., 2000;Walker et al., 2014), the amount of ATP produced by LEF was calculated as: v ATP(LEF) = 1/2 × J g × 1.29 Rates of ATP supply from other flexible mechanisms were determined by subtracting the amount of ATP produced by LEF from v ATP according to: Quantitative Limitation Analysis of A n Photosynthetic limitations in E. camaldulensis and N. tabacum were assessed according to the method of Grassi and Magnani (2005) and Carriqui et al. (2015). The values for stomatal (l s ), mesophyll conductance (l mc ), and biochemical (l b ) limitations represented measures of the relative importance of stomatal diffusion, mesophyll diffusion, and photosynthetic biochemistry in setting the observed value of A n . Relative photosynthetic limitations were calculated as follows (Grassi and Magnani, 2005;Carriqui et al., 2015): where g tot was total conductance to CO 2 between the leaf surface and carboxylation sites (calculated as 1/g tot = 1/ g s + 1/ g m ).

Statistical Analysis
All results were displayed as mean values of four independent measurements. We used one-way ANOVA and SPSS 16.0 software (SPSS Inc., Chicago, IL, USA) to examine differences between the two species. Those differences were considered significant at P < 0.05.

A/C i Curves and the Rate-Limiting
Step of CO 2 Assimilation The A/C i curves indicated that the maximum rate of photosynthesis was higher in N. tabacum than in E. camaldulensis (Figure 2A). Values for the maximum rate of RuBP regeneration (J max ) and RuBP carboxylation (V cmax ) were significantly higher in N. tabacum (Figure 2B). Both species showed similar ratios of J max /V cmax (Figure 1B) as well as the same value for A 400 at 400 µmol mol −1 CO 2 . Because the CO 2 compensation point in the absence of daytime respiration (Γ * ) has an important impact on g m , C c , C trans , and limitations on the stomata (l s ), mesophyll conductance (l mc ), and biochemical functions (l b ), we conducted a sensitivity analysis and examined the rate-limiting step for CO 2 assimilation to Γ * (range from 30 to 40 µmol mol −1 ). For E. camaldulensis, l mc was the most important constraining factor for photosynthesis, followed by l s and l b (Figure 2C). By contrast, biochemical limitations were the most significant in N. tabacum, followed by l mc and l s (Figure 2C). Furthermore, irrespective of Γ * , the value of C c at 400 µmol mol −1 CO 2 was significantly lower than C trans in E. camaldulensis (Table 1), suggesting that the rate-limiting step of A 400 tended to be RuBP carboxylation. By comparison, in N. tabacum the value of C c at 400 µmol mol −1 CO 2 was significantly higher than C trans (Table 1), indicating A 400 tended to be limited by RuBP regeneration.

Light Response Changes in CO 2 Assimilation and Photosynthetic Electron Flow
The light response curves demonstrated that the response of A n to incident light was similar between E. camaldulensis and N. tabacum ( Figure 3A). However, the maximum value of g s was much higher in E. camaldulensis (Figure 3B), while C i was slightly higher in that species (Figure 3C). When Γ * was assumed to be 40 µmol mol −1 , then C c under saturating light was much lower in E. camaldulensis (Figure 3D). At 2000 µmol photons m −2 s −1 in light response curves, the value of C c was 124 µmol mol −1 in E. camaldulensis and 226 µmol mol −1 in N. tabacum (Figure 3D). This large difference of C c between E. camaldulensis and N. tabacum was principally caused by the contrast in g m .
Under all light intensities, E. camaldulensis had significantly higher effective quantum yield of PSII ( PSII ) compared to N. tabacum, especially under high light ( Figure 4A). Concomitantly, NPQ values were lower in E. camaldulensis than N. tabacum ( Figure 4B). According to the data of PSII , E. camaldulensis had significantly higher values of total electron  flow through PSII (J T ) than N. tabacum ( Figure 5A). Meanwhile, the ratios of the rate of Rubisco carboxylation (V c ) to that of Rubisco oxygenation (V o ) under saturating light intensities were much higher in E. camaldulensis ( Figure 5B) Plotting the V o /V c ratio and C c indicated that the difference in the V o /V c ratio between E. camaldulensis and N. tabacum was mainly determined by the difference in C c (Figure 6), which in turn caused by the change in g m .

Alternative Electron Flow and ATP Synthesis from Flexible Mechanisms
Under high light, E. camaldulensis had significantly increased alternative electron flow, as indicated by the higher values of J a and the J a /J g ratio (Figure 7). To examine the relationship between the alternative electron flow and photorespiration, we evaluated possible associations between J a and Je(PCO) under light intensities higher than 300 µmol photons m −2 s −1 . Interestingly, J a was positively and linearly correlated with Je(PCO) (Figure 7A, P < 0.0001). Similarly, the J a /J g ratio was positively and linearly correlated with V o /V c ratio ( Figure 7B, P < 0.0001). According to photosynthesis model, an increase in the V o /V c ratio requires a higher ATP/NADPH energy demand from other flexible mechanisms such as cyclic electron flow and alternative electron flow. Increased alternative electron flow  promotes the formation of a proton gradient across the thylakoid membrane ( pH), which can be used for activating NPQ and ATP synthesis. We found that under high light the rate of ATP supplied from alternative electron sinks [v ATP(Flex) ] were much higher in E. camaldulensis than in N. tabacum (Figure 8A), and the rate of alternative electron flow was positively correlated to v ATP(Flex) (Figure 8B, P < 0.0001). Furthermore, NPQ values under high light were significantly lower in E. camaldulensis than in N. tabacum (Figure 4B), suggesting that the main role for alternative electron flow in E. camaldulensis is not to activate NPQ but to provide extra ATP. The greater capacity of the photorespiratory pathway in E. camaldulensis plants is sustained by the enhanced alternative electron flow.

DISCUSSION
In herbaceous C 3 crop species such as N. tabacum, spinach, rice and wheat, high stomatal and mesophyll conductance are essential for their strong photosynthetic capacities (Yamori et al., 2010(Yamori et al., , 2011. However, for sclerophyllous Eucalyptus camaldulensis, we found that its high rate of CO 2 assimilation Values are means ± SE (n = 4).
( Figure 3A) was accompanied with a low g m (≈0.12 molm −2 s −1 ) ( Table 1). Thus, high levels of mesophyll conductance do not appear to be a common mechanism for high rates of photosynthesis in C 3 plants. Surprisingly, in the sclerophyllous E. camaldulensis, the high rates of photosynthesis occurred at the low levels of C c (Figure 3D), which directly increased the V o /V c ratio ( Figure 5B). What is more, E. camaldulensis showed increased capacities of photorespiratory pathway ( Figure 5C) and electron flow to alternative sinks (Figure 7). Enhancement of photorespiratory pathway increased the release of CO 2 in mitochondria. This photorespired CO 2 can be trapped and reassimilated by chloroplast, thereby boosting photosynthesis (Busch et al., 2013). Concomitantly, the increased photorespiratory pathway in E. camaldulensis needs more extra ATP supply from alternative electron sinks rather than electron transfer from water to NADP + . Interestingly, the increased alternative electron flow provided essential extra ATP to balance the energy budget and sustain the high rate of photorespiratory pathway, then increasing the rate of photosynthetic CO 2 assimilation. These results highlight that the sclerophyllous E. camaldulensis enhanced the capacities of photorespiratory pathway and alternative electron flow to sustain a high rate of photosynthesis.

Quantitative Limitation Analysis of A n
In N. tabacum leaves, l b was the most important constraining factor for photosynthesis, followed by l mc and l s , being consistent with high levels of stomatal and mesophyll conductance. By comparison, in E. camaldulensis, l mc had the greatest influence while l s was the least limiting factor. This was consistent with the higher g s and lower g m measured for that species. These results confirmed that, the major limiting factor for photosynthesis differs intrinsically between high-photosynthesis herbaceous and sclerophyllous plants. Diffusional limitations of CO 2 are the main reason for lower photosynthetic rates in ferns than in angiosperms (Gago et al., 2013;Carriqui et al., 2015). Mesophyll conductance is the most constraining factor for photosynthesis in ferns (Carriqui et al., 2015). These findings suggested that, not only in low-photosynthesis species (e.g., ferns) but also in some high-photosynthesis sclerophyllous species such as E. camaldulensis, mesophyll conductance is the primary limiting factor for photosynthesis. The difference of g m between E. camaldulensis and N. tabacum is probably determined by their leaf anatomy (Terashima et al., 2006). The thicker cell walls (B) The value of V o /V c as a function of J a /J g for leaves of Eucalyptus camaldulensis and tobacco. Data were obtained from light response curves (light intensities higher than 300 µmol photons m −2 s −1 ) measured at 25 • C and 400 µmol mol −1 CO 2 . Values are means ± SE (n = 4). The coefficient of correlation (r) and significance of correlation (P) are 0.98 and <0.0001 for Figure 6A, respectively. The coefficient of correlation (r) and significance of correlation (P) are 0.98 and <0.0001 for Figure 6B, respectively.
lead to a lower g m in E. camaldulensis when compared with N. tabacum.

The Rate-Limiting Step of CO 2 Assimilation
The rate of CO 2 assimilation can be limited by RuBP carboxylation and/or RuBP regeneration in C 3 plants (Farquhar et al., 1980;Yamori et al., 2010Yamori et al., , 2011. The specific ratelimiting step of CO 2 assimilation is determined by the relative values of C c and C trans . When the value of C c is higher than C trans , the CO 2 assimilation rate is mainly limited by RuBP regeneration. Otherwise, the CO 2 assimilation rate is limited by RuBP carboxylation when C c is lower than C trans . In N. tabacum leaves, the value of C c under saturating light was higher than C trans , and thus the rate-limiting step of CO 2 assimilation tended to be RuBP regeneration ( Table 1). On the contrary, the value of C c under saturating light was lower than C trans in E. camaldulensis, and thus the rate of CO 2 assimilation was limited by RuBP carboxylation (Table 1). According to the calculation of C c = C i − A n /g m , the value of C c can be affected by three parameters C i , A n , and g m . Here, because A n and C i values under saturating light were similar in E. camaldulensis and N. tabacum (Figures 3A,C), Because the value of C c can be largely determined by g m , the difference in the main rate-limiting step of CO 2 assimilation between N. tabacum and E. camaldulensis was mainly caused by the contrasts in their mesophyll conductance.
In the herbaceous N. tabacum, which has thin, flexible leaves, g m can reach 0.5 mol m −2 s −1 when plants are exposed to high nitrogen concentrations (Yamori et al., 2011). By comparison, sclerophyllous plants have relatively lower g m values that range between 0.09 and 0.25 mol m −2 s −1 (Lloyd et al., 1992;Hassiotou et al., 2009). Conductance in the mesophyll can be influenced by leaf anatomical traits such as the cell surface area and the chloroplast surface area that is exposed to intercellular air spaces (Evans et al., 1994;Oguchi et al., 2005;Terashima et al., 2011), chloroplast rearrangements (Tholen et al., 2008), cell wall thickness (Terashima et al., 2006(Terashima et al., , 2011Flexas et al., 2012), and LMA (Flexas et al., 2008;Hassiotou et al., 2009). The value of LMA for in E. camaldulensis is approximately 169.5 g m −2 (Suganuma et al., 2006), which is much higher than N. tabacum (approximately 25 g m −2 ) (Yamori et al., 2010). The large differences of leaf anatomical traits between E. camaldulensis and N. tabacum might result in the diversity of g m .

Photorespiration
We found that the capacity of the photorespiratory pathway was greater for E. camaldulensis than for N. tabacum (Figures 5B,C). Rubisco is a dual functional enzyme that catalyzes the carboxylation of RuBP, but also oxygenates RuBP in photorespiration. A lower C c in E. camaldulensis would increase the rate of RuBP oxygenation by Rubisco. Photorespiration begins with RuBP oxygenation that generates glycolate-2phosphate and glycerate-3-phosphateare. To maintain a steady-state high rate of photosynthesis in E. camaldulensis, the RuBP pool must remain stable. In chloroplast, the steady-state of RuBP pool depends on two different RuBP regeneration pathways: the Calvin-Benson cycle and the photorespiratory pathway. Impairment of photorespiratory pathway induced a gradient decrease in photosynthetic rate at atmospheric CO 2 concentration (Takahashi et al., 2007). Importantly, during steady-state phases, the rates of RuBP oxygenation and RuBP regeneration through the photorespiratory pathway must be balanced. Under high light, E. camaldulensis accelerate the photorespiratory pathway to favor the regeneration of RuBP via glycerate-3-phosphate, thereby preventing the RuBP pool from shrinking.
A reduction in C c also accelerates the production of photorespiratory intermediates, e.g., glycine and glycerate, which inhibit the Calvin-Benson cycle (Chastain and Ogren, 1989;Eisenhut et al., 2007;Timm et al., 2012Timm et al., , 2015. In Arabidopsis thaliana plants with elevated glycine decarboxylase activity, the rapid acceleration of the photorespiratory pathway lowers the accumulation of photorespiratory metabolites including those which impair Rubisco activation and possibly the activity of other enzymes of the Calvin-Benson cycle, increasing the performance of the Calvin-Benson cycle (Timm et al., 2012(Timm et al., , 2015. Furthermore, N. tabacum plants grown under high light and high nitrogen concentration up-regulate photorespiratory pathway to maintain high rates of photosynthesis (Huang et al., 2014(Huang et al., , 2016. To overcome those detrimental effects, the photorespiratory pathway should be enhanced in E. camaldulensis. In addition, although CO 2 is released in the mitochondria in the photorespiratory pathway, C 3 plants can trap photorespired CO 2 within individual mesophyll cell. This causes chloroplast CO 2 concentrations to rise and, ultimately, improves the rate of photosynthesis in C 3 plants (Busch et al., 2013). Taking together, the enhanced photorespiratory capacity strongly contributes to the high rate of CO 2 assimilation in E. camaldulensis.

Alternative Electron Flow
We were surprised to learn that, under saturating illumination, alternative electron flow was enhanced in E. camaldulensis when compared with N. tabacum. The water-water cycle, nitrate reduction and malate shunt are potential candidates responsible for this increase in alternative electron flow (Yi et al., 2014). When plants are illuminated at atmospheric CO 2 and O 2 , most of this alternative flow accounts for the electron flux to oxygen and oxidized ascorbic acid (Miyake and Yokota, 2000;Makino et al., 2002). Thus, we concluded that the WWC activity was greater in E. camaldulensis. During the early phase of photosynthetic induction in rice, the WWC first generates a ∆pH across the thylakoid membranes to form NPQ and supply ATP for carbon assimilation (Neubauer and Yamamoto, 1992;Miyake and Yokota, 2000;Makino et al., 2002). However, when photosynthesis reaches a steady-state rate, the WWC no longer maintains a high NPQ but instead provides additional ATP for primary metabolism in rice leaves (Makino et al., 2002). Our light response curves indicated that NPQ values under high light were significantly lower in E. camaldulensis (Figure 4B). Because the NPQ activation relies on the acidification of thylakoid lumen, and which also depresses electron transport through Cyt b 6 /f complex via "photosynthetic control, " this result suggested the higher levels of lumen acidification in leaves of N. tabacum when illuminated at high light. Therefore, the steady-state rate of WWC under intense illumination in E. camaldulensis mainly contributed to additional ATP synthesis rather than lumen acidification, which is consistent with previous studies on the role of the WWC (Makino et al., 2002;Huang et al., 2016).
In electron flow from PSII to NADP + , the stoichiometry of the ATP/NADPH ratio is thought to be 1.29 (Sacksteder et al., 2000;Seelert et al., 2000). By comparison, each Rubisco oxygenation consumes 3.5 ATP and 2 NADH equivalents in total. Therefore, the occurrence of photorespiratory pathway needs a higher ATP/NADPH ratio than 1.29 to maintain primary metabolism (Edwards and Walker, 1983;Walker et al., 2014). Under high light and ambient CO 2 , the ATP/NADPH ratio required by CO 2 assimilation, photorespiration, and nitrite assimilation is approximately 1.6 (Walker et al., 2014). Furthermore, the ATP/NADPH energy demand for primary metabolism will rise as photorespiration increases. As a result, the rate of ATP supplied from other flexible pathways must be higher in E. camaldulensis due to its higher rate of photorespiratory pathway. Cyclic electron flow and the WWC are the main flexible pathways that contribute to extra ATP synthesis under high light and atmospheric CO 2 concentrations (Makino et al., 2002;Walker et al., 2014;Huang et al., 2015Huang et al., , 2016. Here, we found that the greater rate of photorespiration was accompanied by higher alternative electron flow (Figure 7), and the rate of ATP supplied from other flexible mechanisms was positively correlated to the rate of electron flow to alternative sinks ( Figure 8B). Therefore, it appears that E. camaldulensis enhances the alternative electron flow to balance the ATP/NADPH energy demand for high rates of photorespiration.

CONCLUSIONS
In order to illustrate the potential different mechanisms underlying the high rates of photosynthesis in sclerophyllous and herbaceous C 3 plants. Gas exchange and chlorophyll fluorescence were measured in E. camaldulensis (sclerophyllous) and N. tabacum (herbaceous). Although E. camaldulensis and N. tabacum had similar A n under saturating light, the value of g m differed largely between N. tabacum and E. camaldulensis. In N. tabacum, a higher g m increased the value of C c , resulting in the rate-limiting step of CO 2 assimilation tended to be RuBP regeneration. On the contrary, RuBP carboxylation was the main rate-limiting step of CO 2 assimilation in E. camaldulensis because CO 2 diffusion to the chloroplasts was restricted by a lower g m . Therefore, the rate-limiting step of CO 2 assimilation appears to be more related to g m rather than g s in high-photosynthesis species. The lower C c aggravated RuBP oxygenation in E. camaldulensis. Meanwhile, increased flux through the photorespiratory pathway minimizes the accumulation of photorespiratory metabolites, benefiting photosynthetic CO 2 fixation in the Calvin-Benson cycle in E. camaldulensis. In order to balance the ATP/NADPH energy demand for high rates of photorespiration, E. camaldulensis up-regulated alternative electron flow to provide extra ATP. Thus, coordination of photorespiratory pathway and alternative electron flow is crucial for the high rates of CO 2 assimilation in E. camaldulensis. These results highlight the different mechanisms responsible for high rates of photosynthesis in the sclerophyllous plant E. camaldulensis and the herbaceous plant N. tabacum.

AUTHOR CONTRIBUTIONS
WH and YT conceived and designed research. WH conducted experiments. WH, GY, and WY analyzed data. WH wrote the manuscript.