Variation in Photosynthetic Efficiency under Fluctuating Light between Rose Cultivars and its Potential for Improving Dynamic Photosynthesis

Photosynthetic efficiency under both steady-state and fluctuating light can significantly affect plant growth under naturally fluctuating light conditions. However, the difference in photosynthetic performance between different rose genotypes is little known. This study compared the photosynthetic performance under steady-state and fluctuating light in two modern rose cultivars (Rose hybrida), “Orange Reeva” and “Gelato”, and an old Chinese rose plant Rosa chinensis cultivar, “Slater’s crimson China”. The light and CO2 response curves indicated that they showed similar photosynthetic capacity under steady state. The light-saturated steady-state photosynthesis in these three rose genotypes was mainly limited by biochemistry (60%) rather than diffusional conductance. Under fluctuating light conditions (alternated between 100 and 1500 μmol photons m−2 m−1 every 5 min), stomatal conductance gradually decreased in these three rose genotypes, while mesophyll conductance (gm) was maintained stable in Orange Reeva and Gelato but decreased by 23% in R. chinensis, resulting in a stronger loss of CO2 assimilation under high-light phases in R. chinensis (25%) than in Orange Reeva and Gelato (13%). As a result, the variation in photosynthetic efficiency under fluctuating light among rose cultivars was tightly related to gm. These results highlight the importance of gm in dynamic photosynthesis and provide new traits for improving photosynthetic efficiency in rose cultivars.


Introduction
Plants use photosynthesis to convert light energy into stable chemical energy by photosynthetic electron transport and the Calvin-Benson cycle. Plants with high photosynthetic efficiency usually have relatively fast growth rate and high levels of biomass and productivity. The light-saturated photosynthetic capacity under steady state is thought to be the critical determinant of plant growth. For example, the higher steady-state photosynthetic capacity in C 4 plants facilitates their higher productivity than C 3 plants under optimal conditions [1,2]. Photosynthesis can be limited by CO 2 diffusional conductance and biochemical factors [3]. Stomatal conductance (g s ) and mesophyll conductance (g m ) together determine the CO 2 diffusion from air into chloroplast and thus influence chloroplast CO 2 concentration [4][5][6][7][8]. Biochemical factors represent the capacity for the Calvin-Benson cycle and photosynthetic electron flow. High values of g s and g m are the prerequisites of high CO 2 assimilation rate (A N ) in plants grown under high nitrogen condition and high light [6,8]. Generally, photosynthetic capacity in angiosperms is mainly limited by biochemical factors and g m rather than g s when measured under favorite conditions [5,9].

Photosynthetic Characteristics under Steady-State Differ Slightly between Rose Genotypes
The basal leaf functional traits of the three studied rose genotypes were measured and displayed in Table 1. Chlorophyll content (SPAD value) was significantly higher in Rosa hybrida cv. Orange Reeva and Gelato than in Rosa chinensis. Orange Reeva displayed the highest value of leaf mass per area (LMA), followed by Rosa chinensis and Gelato. Leaf N, K, P content in Orange Reeva and Gelato were significantly higher than those in Rosa chinensis. At a high light of 1500 µmol m −2 s −1 , values for steady state A N were 23.4, 21.7, and 20.7 µmol m −2 s −1 in Orange Reeva, Gelato, and Rosa chinensis, respectively. Concomitantly, no significant difference in g s was observed among these three rose genotypes, but Orange Reeva and Gelato had significantly higher g m than Rosa chinensis. Dark respiration rate (R d ) did not significantly differ among these rose genotypes, while the maximum rate of RuBP carboxylation (V cmax ) was significantly higher in Orange Reeva than Gelato and Rosa chinensis. Generally, the light response curves indicated that these three rose genotypes showed similar A N and g s at a given light intensity ( Figure 1). Therefore, the steady-state photosynthesis differed only slightly among different rose genotypes.
"Gelato" showed stronger photosynthetic performance under fluctuating light than the old germplasm Rosa chinensis. Therefore, the improved photosynthetic efficiency under fluctuating light partially contributes to the stronger growth potential of modern rose cultivars.

Photosynthetic Characteristics under Steady-State Differ Slightly between Rose Genotypes
The basal leaf functional traits of the three studied rose genotypes were measured and displayed in Table 1. Chlorophyll content (SPAD value) was significantly higher in Rosa hybrida cv. Orange Reeva and Gelato than in Rosa chinensis. Orange Reeva displayed the highest value of leaf mass per area (LMA), followed by Rosa chinensis and Gelato. Leaf N, K, P content in Orange Reeva and Gelato were significantly higher than those in Rosa chinensis. At a high light of 1500 μmol m −2 s −1 , values for steady state AN were 23.4, 21.7, and 20.7 μmol m −2 s −1 in Orange Reeva, Gelato, and Rosa chinensis, respectively. Concomitantly, no significant difference in gs was observed among these three rose genotypes, but Orange Reeva and Gelato had significantly higher gm than Rosa chinensis. Dark respiration rate (Rd) did not significantly differ among these rose genotypes, while the maximum rate of RuBP carboxylation (Vcmax) was significantly higher in Orange Reeva than Gelato and Rosa chinensis. Generally, the light response curves indicated that these three rose genotypes showed similar AN and gs at a given light intensity ( Figure 1). Therefore, the steadystate photosynthesis differed only slightly among different rose genotypes.  Based on the CO 2 response curves, A N differed very slightly between these three rose genotypes at C i below 300 µmol mol −1 (Figure 2A). However, when C i was higher than 300 µmol mol −1 , Rosa hybrida cv. Orange Reeva had significantly higher A N than Rosa hybrida cv. Gelato and Rosa chinensis (Figure 2A). Concomitantly, electron transport rate through PSII (J PSII ) was higher in Orange Reeva than the other two rose genotypes ( Figure 2B). At an atmospheric CO 2 concentration of 400 µmol mol −1 , A N just reached 40-50% of the maximum value, but J PSII reached approximately 80% of the maximum value  Figure 2A,B). Therefore, the major limitation imposed on A N at 1500 µmol m −2 s −1 and 400 µmol mol −1 CO 2 was Rubisco carboxylation rather than RuBP regeneration (i.e., electron transport rate). The quantitative analysis indicated that the relative limitation imposed on A N by biochemical capacity was approximately 0.6 in the three rose genotypes, the relative limitation of g s or g m was approximately 0.2 in them ( Figure 2C). Therefore, in the three studied rose genotypes, biochemistry was the major limitation of A N under atmospheric CO 2 concentration and high light, followed by diffusional conductance.

Modern Rose cultivars use Fluctuating Light more Efficiently Than the Old Rose Species
During the three low/high light cycles, Orange Reeva and Gelato had significantly higher AN in high-light phases than Rosa chinensis, while the value of AN in low-ligh

Modern Rose Cultivars Use Fluctuating Light More Efficiently Than the Old Rose Species
During the three low/high light cycles, Orange Reeva and Gelato had significantly higher A N in high-light phases than Rosa chinensis, while the value of A N in low-light phases did not differ between them ( Figure 3A). Such difference in A N in high-light phases led to the higher carbon gain under fluctuating light in Orange Reeva and Gelato ( Figure 3B). During the 30 min fluctuating light treatment, g s gradually decreased with prolonged illumination under fluctuating light in all these three rose genotypes ( Figure 3C), and the average g s under fluctuating light was significantly higher in Orange Reeva and Gelato than Rosa chinensis ( Figure 3D). Upon transitioning to high light, g m gradually increased in the subsequent 5 min ( Figure 3E). No significant difference in g m was observed at low light, while Orange Reeva and Gelato had significantly higher g m at high-light phases than Rosa chinensis ( Figure 3F). When normalized to the initial values, Rosa chinensis displayed significant lower A N , g s , and g m under high-light phases than Orange Reeva and Gelato ( Figure 4). Therefore, the two modern Rose hybrida cultivars use fluctuating light more efficiently than the old rose genotype Rosa chinensis. Furthermore, tight relationships between A N and diffusional conductance (g s and g m ) were observed ( Figure 5), suggesting that the relatively lower photosynthetic efficiency under fluctuating light in Rosa chinensis was partially attributed to its lower g s and g m .
During fluctuating light treatment, C i did not significantly differ among these three rose genotypes ( Figure 6A). However, the C c values under high-light phases were significantly higher in Orange Reeva and Gelato than Rosa chinensis ( Figure 6B). Under steadystate photosynthesis at high light, these three rose genotypes had similar value of V cmax ( Figure 7A). After exposure to the three cycles of low/high light, V cmax could increase to the initial value after 5 min illumination at high light in Orange Reeva and Gelato but remarkedly decreased in Rosa chinensis ( Figure 7A), making the average V cmax under high light in Rosa chinensis was lower than the other two genotypes ( Figure 7B). By normalizing to the initial steady-state value, V cmax decreased to a much lower extent in Rosa chinensis when compared with Orange Reeva and Gelato ( Figure 7A,B). These results indicated that the difference in A N under fluctuating light between different rose genotypes was correlated to C c and V cmax rather than C i .
( Figure 4). Therefore, the two modern Rose hybrida cultivars use fluctuating light m efficiently than the old rose genotype Rosa chinensis. Furthermore, tight relationships tween AN and diffusional conductance (gs and gm) were observed ( Figure 5), suggest that the relatively lower photosynthetic efficiency under fluctuating light in Rosa chinen was partially attributed to its lower gs and gm.     During fluctuating light treatment, Ci did not significantly differ among these three rose genotypes ( Figure 6A). However, the Cc values under high-light phases were significantly higher in Orange Reeva and Gelato than Rosa chinensis ( Figure 6B). Under steadystate photosynthesis at high light, these three rose genotypes had similar value of Vcmax ( Figure 7A). After exposure to the three cycles of low/high light, Vcmax could increase to the initial value after 5 min illumination at high light in Orange Reeva and Gelato but remarkedly decreased in Rosa chinensis ( Figure 7A), making the average Vcmax under high light in Rosa chinensis was lower than the other two genotypes ( Figure 7B). By normalizing to the initial steady-state value, Vcmax decreased to a much lower extent in Rosa chinensis when compared with Orange Reeva and Gelato ( Figure 7A,B). These results indicated that the difference in AN under fluctuating light between different rose genotypes was correlated to Cc and Vcmax rather than Ci.

Discussion
In general, the major limiting factor of photosynthesis largely varied among di species or different genotypes of a given species. Alternating the relative limitati posed on photosynthesis at the leaf level can improve plant biomass and crop prod ity [19,[28][29][30]. The relative limitation of steady-state photosynthesis under saturatin has been investigated in many crops and groups [5,9]. However, leaves rarely co steady-state photosynthesis when exposed to natural sunlight [31][32][33]. While exp the major limitation under steady state is valuable for understanding photosynthet ulation, dynamic photosynthetic measurements provide insight into how crop lea spond to fluctuating light and has great potential in crop improvement [14,18,2 showed in Figure 1, the steady-state photosynthesis changed slightly among the thr cultivars. However, the dynamic photosynthetic efficiency under fluctuating ligh significantly higher in two modern rose cultivars Orange Reeva and Gelato when pared with the old rose plant Rosa chinensis (Figure 3), providing important new tr the modern rose cultivars. Therefore, improving dynamic photosynthesis under flu ing light is a potential target for increasing rose yield.

Discussion
In general, the major limiting factor of photosynthesis largely varied among different species or different genotypes of a given species. Alternating the relative limitation imposed on photosynthesis at the leaf level can improve plant biomass and crop productivity [19,[28][29][30]. The relative limitation of steady-state photosynthesis under saturating light has been investigated in many crops and groups [5,9]. However, leaves rarely conduct steady-state photosynthesis when exposed to natural sunlight [31][32][33]. While exploring the major limitation under steady state is valuable for understanding photosynthetic regulation, dynamic photosynthetic measurements provide insight into how crop leaves respond to fluctuating light and has great potential in crop improvement [14,18,21]. As showed in Figure 1, the steady-state photosynthesis changed slightly among the three rose cultivars. However, the dynamic photosynthetic efficiency under fluctuating light was significantly higher in two modern rose cultivars Orange Reeva and Gelato when compared with the old rose plant Rosa chinensis (Figure 3), providing important new trait for the modern rose cultivars. Therefore, improving dynamic photosynthesis under fluctuating light is a potential target for increasing rose yield.

Steady-State Photosynthesis across Rose Germplasm Is Mainly Limited by Biochemical Capacity
Despite some uncertainties regarding the methods for g m estimation, the quantitative analysis indicated that the limitation to steady-state photosynthesis imposed by g m or g s in all three rose genotypes was approximately 20% ( Figure 2C). Therefore, increasing g s and g m might have minor roles in improving light-saturated photosynthesis under steady state in the breeding of rose cultivars. Concomitantly, the relative limitation imposed on A N by biochemistry was approximately 60% (Figure 2C), indicating that biochemical capacity was the major limitation imposed on photosynthesis at steady state in these three rose genotypes. This characteristics of photosynthetic limitation in rose plants were similar to herbaceous plants, such as rice [9] and tomato [27], but different from sclerophyllous angiosperms, such as evergreen Mediterranean oaks [10] and Rhododendron species [11].
At the atmospheric CO 2 concentration of 400 µmol mol −1 , photosynthetic electron transport reached 80-90% of the maximum value while A N just reached 40-50% of the maximum value ( Figure 2). Therefore, biochemical limitation was mainly attributed to Rubisco activity in vivo rather than regeneration of RuBP. On average, V cmax in the three studied rose genotypes was 108 µmol m −2 s −1 , which was low when compared to elite cultivars of wheat and rice [34,35]. V cmax estimated by A/C i curve is tightly determined by Rubisco content and efficiency, suggesting that rose genotypes grown under similar conditions of good nutrient might have relatively lower Rubisco content and/or efficiency than other high-yield C 3 crops. This difference in V cmax suggests that strategies proposed to improve Rubisco quantity and efficiency would have particular value in improving steady-state photosynthetic rate [36][37][38]. Therefore, increasing Rubisco content and activity through genetic manipulation might significantly increase yield potential in rose genotypes, which should be taken into consideration in molecular breeding of rose cultivars.

Modern Rose cultivars have Stronger Dynamic Photosynthetic Efficiency Than the Old Rose Rosa chinensis
The loss of photosynthetic carbon gain under fluctuating light can significantly affect plant growth and biomass [15,19,31,39]. During fluctuating light treatment with low/high light cycles, the decline of A N under high light was observed in the three studied rose cultivars ( Figures 3A and 4A), which was similar to the phenomenon of Arabidopsis, rice, and tomato. Such loss of photosynthetic carbon gain in rose genotypes was particularly caused by the gradual decrease in g s under fluctuating light ( Figure 5). Previous studies indicated that improved induction speed of g s or increased g s under fluctuating light significantly increased photosynthetic efficiency and biomass in Arabidopsis thaliana and rice when grown under fluctuating light [14,15,19]. Similarly, the decline in g s is a common photosynthetic characteristic in rose genotypes when exposed to fluctuating light, indicating that increasing g s or altering the response of g s to change of light intensity is an attractive target for improving photosynthetic efficiency under fluctuating light in this crop.
In modern rose cultivars Orange Reeva and Gelato, the gradual decrease in g s , not the change of g m , accounted for the declines in A N under fluctuating light (Figure 4). By comparison, the decline in A N under fluctuating light in old rose cultivar Rosa chinensis was caused by the simultaneous decreases in g s and g m (Figure 4). Therefore, the underlying mechanisms for the decline in A N are different between different cultivars. Previous studies mainly focused on the effect of stomatal behavior on dynamic photosynthesis among different crop germplasms [16][17][18]40]. However, little attention is given to the behavior of g m under fluctuating light and its effect on photosynthetic carbon loss. Some recent studies reported that g m can exert a significant limitation of photosynthesis under fluctuating light [26,27]. Once light intensity abruptly increased, the induction speed of g m was rapider in Orange Reeva and Gelato than in Rosa chinensis. This different response of g m to fluctuating light led to significant higher C c and V cmax values in Orange Reeva and Gelato (Figures 6 and 7), which facilitated the higher efficiency of dynamic photosynthesis in them. Therefore, the response kinetics of g m significantly affect the photosynthetic efficiency under fluctuating light across rose germplasm. An improved kinetics of g m can favor photosynthesis under fluctuating light, which is an attractive strategy for the breeding of high-yield cultivars of other horticultural plants and crops.

Plant Materials and Growth Conditions
Two industrial Rosa hybrida cv. "Orange Reeva" and "Gelato" and an old Chinese rose plant Rosa chinensis cv. "Slater's crimson China" were used. These plants were cultivated in a greenhouse located in Kunming, Yunnan, China, with 50% full sunlight, day and night air temperatures of 35 and 20 • C, respectively, and relative air humidity of 45-60%. The maximum light intensity to which the leaves were exposed was approximately 1000 µmol photons m -2 s -1 . Plants were watered and fertilized (0.1% nutrient solution) every day. The uppermost mature leaves on the flower stems were chosen for measurements.

Gas Exchange and Chlorophyll Fluorescence Measurements
Gas exchange and chlorophyll fluorescence were measured simultaneously using an open gas exchange system (LI-6400XT; Li-Cor Biosciences, Lincoln, NE, USA) equipped with a leaf chamber fluorometer (Li-Cor Part No. 6400-40, enclosed leaf area: 2 cm 2 ) at leaf temperature of 25 • C, a relative humidity of approximately 60%, and air flow rate of 300 mmol min -1 . Irradiance was provided by a mixture of red (90%) and blue (10%) LEDs in the fluorometer. After fully induction at 1500 µmol photons m -2 s -1 , light response curves were measured under different light intensity (1500, 1000, 600, 300, 200, 100, 50 µmol photons m -2 s -1 ), and CO 2 response curves were measured at each CO 2 concentration (50, 100, 200, 300, 400, 600, 800, 1000 and 1500 µmol mol −1 ). In light and CO 2 response curves, photosynthetic parameters were logged after upon reaching steady-state conditions (at least 3 min). The maximum rates of RuBP carboxylation (V cmax ) and regeneration (J max ) were calculated using the A/C i curves [41]. Dynamic photosynthesis was measured under fluctuating light alternating between low light (100 µmol photons m -2 s -1 ; 5 min) and high light (1500 µmol photons m -2 s -1 ; 5 min). During three cycles of low/high light, photosynthetic parameters were logged every minute to calculate the kinetics of photosynthesis under fluctuating light.
Chlorophyll fluorescence parameters were determined using the multi-phase flash (MPF) protocol following recommended procedures [42]. The measuring light intensity and the maximum flash intensity were 1 and 8000 µmol m −2 s −1 , respectively. The flash intensity decreased by 60% during the second phase of the MPF and the durations of the three flash phases were 0.3 s, 0.7 s, and 0.4 s, respectively. The effective photochemistry quantum yield of photosystem II (ΦPSII) and total electron transport rate through PSII (J PSII ) were calculated using following equations [43,44]: where F s and F m are steady and maximum fluorescence under actinic light, respectively; PPFD is the light intensity, s is a unitless lumped calibration factor used to scale Φ PSII to J PSII [45], and a typical value of 0.45 was used in this study.

Calculations of g m , C c and V cmax
Based on the concurrent measurements of A N and J PSII , g m was calculated using the following equation [46]: where A N represents the net CO 2 assimilation rate; C i , intercellular CO 2 concentration; Γ * , CO 2 compensation point in the absence of daytime respiration [47,48], and a typical value of 40 µmol mol -1 was used in this study. R d , respiration rate in the dark and was considered to be half of the mitochondrial respiration rate as measured after dark adaptation for 10 min [5]. The chloroplast CO 2 concentration (C c ) was calculated using the values of A N , C i and g m [41,49]: The maximum rate of Rubisco carboxylation (V cmax ) was calculated as described by [48,50].

Quantitative Limitation Analysis of A N
Factors limiting steady-state photosynthesis in the studied species were also assessed. l s represents the relative photosynthetic limitation of g s ; l m represents the relative photosynthetic limitation of g m ; l b represents the relative photosynthetic limitation of biochemistry. The values of l s , l m and l b were calculated using the following equations [3]: where g tot was the total CO 2 diffusional conductance and was calculated as 1/g tot = 1/g s +1/g m [3], and ∂A N /∂C c was calculated according to the methods of [9,48].
where K c and K o are the Rubisco Michaelis-Menten constants for CO 2 and O 2 , respectively, and O is the oxygen concentration in the chloroplasts [48].

SPAD Index and Leaf Nutrient Content Measurements
The relative content of chlorophyll per unit leaf area (SPAD index) was measured using a SPAD-502 Plus (Minolta, Tokyo, Japan). After detached from plants, leaf area was measured using a LI-3000A (Li-Cor, Lincoln, NE, USA). Subsequently, these detached leaf samples were dried at 80 • C for 48 h, and dry weight was measured to calculate leaf mass per area (LMA). Finally, leaf N, P, K content was measured using a Vario MICRO Cube Elemental Analyzer (Elementar Analysensysteme GmbH, Langenselbold, Germany).

Statistical Analysis
Five independent leaves from five different plants were used for each measurement. One-way ANOVA was used to examine the significant differences between different rose cultivars (α = 0.05).
Average values ± SE (n = 5) are shown for leaf chlorophyll content (SPAD), leaf mass per area (LMA), leaf N content, leaf K content, leaf P content, net assimilation rate (A N ), stomatal conductance (g s ), mesophyll conductance (g m ), dark respiration rate (R d ), the maximum velocity of Rubisco carboxylation (V cmax ), and regeneration (J max ). Steady-state values of A N , g s and g m were measured at 1500 µmol photons m −2 s −1 as indicated in light response curves. V cmax and J max were calculated from CO 2 response curves. Different letters (a, b and c) indicate significant differences between different cultivars.

Conclusions
The results presented in this study highlight the main traits of the photosynthetic characteristics of rose cultivars under steady state and under fluctuating light. First, Rubisco activity is the major limiting factor of photosynthesis under steady state in rose cultivars, suggesting that increasing Rubisco activity might improve photosynthesis in this crop. Second, the decline in g s is an important reason for the loss of photosynthesis under fluctuating light in these three rose cultivars, pointing out that increasing g s is a potential target for improvement of photosynthetic efficiency under fluctuating light. Third, the rapid response kinetics of g m is a prerequisite of the high photosynthetic efficiency under fluctuating light in modern rose cultivars. Taking together, increasing Rubisco activity has large potential in improvement of photosynthetic efficiency in rose genotypes, which could be strengthened by improving the response kinetics of g s and g m under fluctuating light.