Experimental and simulated CO2 responses of photosynthesis in leaves of Hippophae rhamnoides L. under different soil water conditions

CO2 concentrations and soil moisture conditions seriously affect tree growth and physiological mechanisms. CO2 responses of photosynthesis are an important part of plant physiology and ecology research. This study investigated the photosynthetic CO2 responses in the leaves of two-year-old Hippophae rhamnoides L. under eight soil water conditions in a semi-arid loess hilly region, and discussed the quantitative relationship between CO2 responses and soil moisture. CO2 response curves and parameters were fitted using a rectangular hyperbola model, non-rectangular hyperbola model, exponential equation, and modified rectangular hyperbola model. Results revealed that the relative soil water content (RWC) required to maintain a high photosynthetic rate (Pn) and carboxylation efficiency (CE) ranged from 42.8% to 83.2%. When RWC fell outside these ranges, the photosynthetic capacity (Pnmax), CE, and CO2 saturation point (CSP) decreased. CO2 response curves and three parameters, CE, CO2 compensation point (Γ), and photorespiration rate (Rp), were well fitted by the four models when RWC was appropriate. When RWC exceeded the optimal range, only the modified rectangular hyperbola model fitted the CO2 response curves and photosynthetic parameters better.


Introduction
Photosynthesis is a complex process affected by many factors in plants, including CO 2 and water, which have important effects [1,2]. CO 2 is the substrate of photosynthesis, and its atmospheric concentration is predicted to reach ~550 μmol·mol -1 by 2050 [3,4]. Global water shortages are aggravated by changes due to increasing CO 2 concentrations and the warming climate [5,6]. The increase in CO 2 can cause global climate change and directly affect the metabolism and growth of plants [7,8].Drought affects plant growth and development severely [9,10], as well as limits photosynthesis through carbon metabolism by restricting CO 2 diffusion [11,12]. However, plants display adaptability and resistance to water deficits [13,14]. Moreover, photosynthetic efficiency is higher within a certain water range, which varies according to the plant species and photosynthetic mechanism [15,16]. CO 2 responses are an important part of plant physiology and ecology research, the measurement and simulation of which are the main approaches for studying photosynthesis [17,18]. The photosynthesis CO 2 response model has played an important role in increasing our understanding of photosynthetic carbon uptake, which has thereby improved our understanding and predictions of plant photosynthetic physiology and its response to environmental changes and biogeochemical systems [19,20]. CO 2 response curves reflect the quantitative relationship between plant net photosynthetic rate (P n ) and CO 2 concentration and can be used to estimate photosynthetic parameters, including the CO 2 saturation point (CSP), photosynthetic capacity (P nmax ), compensation point (Γ), carboxylation efficiency (CE), and photorespiration rate (R p ) [21,22]. CO 2 responses have been fitted using biochemical models [23], empirical models [24,25], and optimized models [26,27], which are based on biochemical models. Biochemical models calculate two key model parameters, the maximum rate of carboxylation (V cmax ) and the maximum electron transport rate (J max ) [28,29]. Empirical models include the Michaelis-Menten model [30], rectangular hyperbola model [31], non-rectangular hyperbola model [32], and exponential equation [33], which have been applied in most crops [34,35]and some woody species [36,37]. Ye [38]thought the Michaelis-Menten and rectangular hyperbolic models were essentially the same. In recent years, some studies have proposed an improved rectangular hyperbolic model, namely, the modified rectangular hyperbola model [39,40]. This model has been applied to some plants, including some gramineous plants [41,42], herbs [43,44], and woody plants [45,46]. Results revealed that this new model could overcome the limitations of traditional models and accurately fit the CO 2 response curve and its characteristic parameters. Previous studies on photosynthesis CO 2 response models have focused on the estimation and optimization of key parameters in field crops [47,48]. However, the applicability of different models simulating the CO 2 response data of woody plants under adverse conditions, such as continuous drought, has rarely been reported.
Hippophae rhamnoides L. is a common afforestation species found in the arid and semi-arid regions of Northern China, which has a high economic value and plays an important role in ecological restoration and soil and water conservation. H. rhamnoidesL. is a non-leguminous and nitrogen-fixing species, deciduous shrubs, and is resistant to barren and dry conditions. In recent years, studies have focused on its growth [49], water consumption [50,51], and photosynthetic light response characteristics [52,53]. These studies have been conducted under water stress, while only a few studies related to the physiological characteristics of drought stress have been conducted [54,55]. However, continuous observations and the examination of photosynthesis CO 2 response in leaves of H. rhamnoides L. at many soil moisture gradients during the accelerated soil drought process have not been addressed. Therefore, the quantitative relationship between the photosynthetic CO 2 response process and soil moisture remains unclear.
In this study using potted seedlings of H. rhamnoides L., CO 2 response curves and parameters were evaluated and fitted with the rectangular hyperbola model, non-rectangular hyperbola model, exponential equation, and modified rectangular hyperbola model under different soil moisture conditions. The goals of this study were to define the quantitative relationship between photosynthetic CO 2 response processes and soil moisture, as well as explore the applicability of different CO 2 response models to fit CO 2 response processes and parameters. The findings of this study will provide an in-depth understanding of the photosynthetic physiology-ecological characteristics and cultivation of H. rhamnoides L. in the loess hilly-gully region of Northern China. Furthermore, the applicability of different CO 2 response models can be evaluated from these findings and used in future studies.

Study area
The experimental site was located in the Tuqiaogou watersheds (37°36′58″N, 110°02′55″E) of Yukou Town, Fangshan County, Shanxi Province, China, a portion of the gully-hilly area of the Loess Plateau in the middle reaches of the Yellow River. This area has a sub-arid, warm temperate, continental monsoon climate. The average annual precipitation is 525.0 mm with more than 70% of the precipitation concentrated between July and September. The annual potential evaporation is 1839.7 mm with the greatest amount of evaporation occurring between April and June. The annual frost-free period lasts 140 d. The soil is classified as medium loessial soil, and the soil texture is uniform with a pH value ranging from 8.0 to 8.4. Vegetation consists mainly of trees, shrubs, lianas, and subshrubs. Tree species are predominantly Robinia pseudoacacia, Ulmus pumila, Platycladus orientalis, and Syringa oblata. Shrubs are mainly Rosa xanthina and Ulmus macrocarpa. Herbs consist of Compositae and Gramineae, of which the Compositae belong to the Artemisia genus. Most of the forest land consists of sparse woodland with poor stand stability.

Materials and water treatments
Two-year-old H. rhamnoides L. were used as the experimental materials and were selected carefully to ensure consistency in their height, diameter, and growth. Plants were investigated and marked one by one before transplantation. In March 2018, seedlings were transplanted in containers (50 cm in height, 35 cm in diameter) that had drainage holes in the bottom. A total of six basins with one plant in each pot were used. The relative soil water content (RWC) and photosynthetic CO 2 responses in the leaves of H. rhamnoides L. were determined in August. Three strong plants were selected and watered to saturation, and the initial RWC was obtained; the first CO 2 response was also determined. Then, soil moisture gradients were obtained every two days through the natural water consumption method after artificially supplying water. The soil mass water content (MWC, %) was measured by the stoving method. The RWC was considered as the ratio of MWC to the field water capacity (FC, %). The potting soil FC was 24.3%, according to the cutting ring method, and the soil bulk density was 1.26 ± 0.13 g·cm -3 . Eight RWC gradients were obtained and found to be 91.7%, 83.2%, 71.5%, 54.6%, 42.8%, 31.9%, 26.1%, and 21.4%. The experiment was monitored under a canopy with a plastic film covering the top on rainy days to prevent rain from interfering with the RWC.

CO 2 response determination
Three strong, mature leaves were selected and marked in a central test plant. CO 2 responses under different soil moisture conditions were measured using a CIRAS-2 (PP Systems, Hitehin, UK) portable photosynthesis system. The light saturation point for H. rhamnoides L. was 1400 μmol·m -2 ·s -1 [52,54]. Measurements were obtained under each soil moisture condition on separate days. The time of measurements occurred from 08:30 to 11:00 h in completely clear weather to reduce the effects of outside light fluctuations. Measurements were obtained three times for each leaf, and the average value was calculated and used for the analyses. The atmospheric temperature ranged from 24°C to 26°C, and the relative humidity was approximately 60% ± 4.0%. The CO 2 concentration in the leaf chamber was controlled and regulated from 0 to 1400 μmol·mol -1 by a small cylinder with high CO 2 concentrations. The CO 2 concentration gradients were 1400, 1200, 1000, 800, 600, 400, 200, 180, 150, 120, 90, 60, 30, and 0 μmol·mol -1 . The duration of the measurement lasted 120 s at each CO 2 concentration, and the apparatus automatically recorded the photosynthetic physiological parameters, including the P n (μmol·m -2 ·s -1 ) and intercellular CO 2 concentration (C i , μmol·mol -1 ).

Data analysis
CO 2 response curves were drawn with C i as the horizontal axis and P n as the vertical axis. According to the measured data point trends, CSP (μmol·mol -1 ), P nmax (μmol·m -2 ·s -1 ), and Γ (μmol·mol -1 ) were estimated and regarded as measured values. CE Γ (mol·m -2 ·s -1 ) at Γ, the intrinsic carboxylation efficiency (CE 0 , mol·m -2 ·s -1 ) at 0 Γ, the absolute value (CE Γ0 , mol·m -2 ·s -1 ) of the slope of the line from C i = 0 to C i = Γ in the CO 2 response curve, and R p (μmol·m -2 ·s -1 ) were calculated using the traditional linear regression method and used as the measured values to compare to the fitted values of the four models.
Statistical analyses were performed using Microsoft Excel 2003(Microsoft Corp., Redmond,Wash.). Significant differences were analyzed by a one-way ANOVA and Duncan's posthoc test. Nonlinear regression was analyzed using SPSS v18.0 (SPSS Inc., Chicago, Illinois). Data were expressed as the mean ± standard deviation (S.D.), and significance was interpreted as p < 0.05. The CO 2 response curve was fitted using the rectangular hyperbola model, non-rectangular hyperbola model, exponential equation, and modified rectangular hyperbola model (described below).

Rectangular hyperbolic model
The rectangular hyperbolic model is expressed as follows [31]: where P n is the net photosynthesis rate, C i is the intercellular CO 2 concentration, α is the slope of the CO 2 response curve when C i = 0 (namely, the initial slope of the CO 2 response curve and the initial CE ), P nmax is the photosynthetic capacity, and R p is the photorespiration rate.
CE Γ , CE 0 , and CE Γ0 are expressed as follows: where the line y = P nmax intersects the linear equation when C i is below 200 μmol mol -1 , and the value of the intersected point on the x-axis is CSP [56].

Non-rectangular hyperbola model
The non-rectangular hyperbola model is expressed as follows [32]: where k is the curved angle of the non-rectangular hyperbola; the definitions of other parameters are the same as above.
CE Γ , CE 0 , and CE Γ0 are expressed as follows: Γ is expressed as follows: where the line y = P nmax intersects the linear equation when C i is below 200 μmol·mol -1 , and the value of the intersected point on the x-axis is CSP [38].

Exponential equation
The exponential equation is expressed as follows [33]: (11) where the definitions of P n , C i , P nmax , α, and R p are the same as above.
CE Γ , CE 0 , and CE Γ0 are expressed as follows:

R CE
Γ is expressed as follows: where the line y = P nmax intersects with the linear equation at Ci ≤ 200 μmol·mol -1 , and the value of the intersected point on the x-axis is CSP [57].

Modified rectangular hyperbola model
The modified rectangular hyperbola model is expressed as follows [39]: where b and c are coefficients; the definitions of other parameters are the same as above. CE Γ , CE 0 , and CE Γ0 are expressed as follows: CSP and P nmax are expressed as follows:

Results
Photosynthetic CO2 response Soil moisture significantly affected the photosynthetic CO 2 response of H. rhamnoides L. (Fig.  1). Under different soil moisture conditions, P n increased rapidly as C i increased, when C i was below ~200 μmol·mol -1 . P n increased slowly as C i increased, and the maximum P nmax appeared at CSP. When C i reached CSP, the CO 2 response was considerably different under different soil water conditions, specifically when RWC ranged from 42.8% to 83.2%，P n of each CO 2 response curve changed slightly as C i increased after C i reached CSP. When RWC was out of the above ranges, P n decreased considerably after C i reached CSP，P n in each curve at the highest C i was significantly smaller than its P nmax under the same soil moisture conditions (p < 0.05) ( Table 1) Clearly, CO 2saturated inhibition had occurred. Furthermore, the CO 2 responses to soil moisture had an obvious RWC threshold. The overall level of P n in each CO 2 response curve increased initially, then decreased as RWC decreased. The P n level was the highest when RWC was 71.5%; thus, an increase or decrease in RWC led to a decrease in the overall P n level. CSP and P nmax were high and P n did not decrease at high CO 2 concentrations when RWC ranged from 42.8% to 83.2%; thus, these RWC ranges were suitable for photosynthesis in the leaves of H. rhamnoides L.. Note: Different letters indicate significant differences between values in the same row (p < 0.05); the same letter indicates no significant differences.  The simulated effects of the four models fitting the CO 2 response data were notably different under different soil moisture conditions ( Fig. 2; Table 2). CO 2 responses curves and photosynthetic characteristic parameters(CE 0、 CE Γ 、CE Γ , Γ, and R p ) were well simulated by the four models, and the determination coefficients were all > 0.991 when RWC ranged from 42.8% to 83.2% (Table 2). Moreover, within the above RWC range, P nmax and CSP fitted by the modified rectangular hyperbola model were comparable with the measured value. The P nmax values fit by the other three models were significantly higher than their observed values, while the fit CSP values were significantly lower than their observed values. When RWC was outside the range of 42.8%-83.2%, only the modified rectangular hyperbola model fit the CO 2 responses (R 2 > 0.99) and characteristic parameters well (Figue 2D; Table 2), the other three kinds of models produced large deviation to fit the CO 2 response process and its characteristic parameters(Figues 2A,B,C; Table 2).

Effects of soil moisture on CO 2 response curves and photosynthetic parameters
Water is a major limiting factor in the recovery and restoration of vegetation found in the loess, hilly-gully regions of China. RWC seriously affected light-response curves and photosynthetic parameters, which also profoundly affected CO 2 response curves and photosynthetic parameters in the leaves of H. rhamnoides L. The classical form of a P n -C i curve can be summarized in three stages [58,59]. First, an approximately linear segment is observed when C i ≤ 200 μmol·mol -1 . Thus, P n increases rapidly as C i increases, namely, during the ribulose bisphosphate (RuBP) restriction phase. The slope of the straight line is the mesophyll conductance, CE, which reflects the assimilative capacity of plant responses to low CO 2 [60,61]. Second, the curved segment is observed when C i is ~200 μmol·mol -1 to CSP, and P n increases slowly as C i increases, gradually entering the restriction stage of RuBP regeneration [62]. Third, an almost linear segment when C i > CSP, P n changes insignificantly as C i increases, moving into the restriction stage of triose-phosphate utilization (TPU). P n at this stage is P nmax , which reflects photosynthetic electron transport and photophosphorylation activity [63].
The form of the P n -C i curve changes when plants encounter stressful conditions, such as drought. Bernacchi et al. [64]considered that numerous factors would influence the curve of P n -C i which included physiological changes (e.g. Vcmax,Jmax or Rd) and environmental changes (e.g. drought, temperature and/or atmospheric CO 2 concentration).However, the quantitative relationship between this change and soil moisture has remained unclear. This study demonstrated that the photosynthetic P n -C i curve of H. rhamnoides L. exhibited a classical form, with P nmax , CE, CSP, and R p being high and Γ being low within a suitable RWC range (i.e., 42.8%-83.2%); P n levels were highest when RWC was 71.3%( Fig. 1; Table 1). Three photosynthetic parameters, P nmax , CE, and CSP, declined dramatically when soil moisture was beyond this range. H. rhamnoides L. exhibited wide photosynthetic adaptability to soil moisture compared to the suitable RWC ranges of Robinia pseudoacacia L. (50.0%-81.6%), Platycladus orientalis L. (5.3%-75.0%) [65], Syringa oblata Lindl. (58.8%-76.6%) [66], and Ziziphus jujube (46.0%-80.5%) [67].
The common method for obtaining CE is the traditional linear regressive method, whereby CE is the slope of the straight line of the P n -C i curve at a low CO 2 concentration (C i ≤ 200 μmol mol -1 ) [35,68]. Many studies have shown that the CE values of different plants vary greatly [69,70] [70]. Although Hu et al. [44]showed that soil moisture greatly affects the CE values of plants, the quantitative relationship between CE and soil moisture has remained unclear. According to a previous study, the P n -C i curve of photosynthesis does not have a strictly linear relationship at a low CO 2 concentration [43].

CO 2 response curves and photosynthetic parameters fitted by different models
The major use of different CO 2 response models lies in the equations used to fit the CO 2 response and its characteristic parameters to extract physiologically meaningful variables; these parameters are used to describe physiological responses of leaves to different treatments [64,73]. For example, CE Γ at the CO 2 compensation point, CE 0 , and the absolute value of the slope of the line between C i = 0 and C i = Γ on the CE Γ0 curve can be fitted, and they have clear physiological meaning and unique values. However, the applicability and simulated accuracy of the empirical models are limited by their asymptotic form with no extreme values [38,39] (Ye & Gao 2009, Ye 2010. In some studies [43,45,46], P nmax was much larger than the measured value, while CSP was far less than the measured value. In particular, the CO 2 response curves could not be fitted under stressful conditions. The same problem was noted in this study. Although the modified rectangular hyperbola model proposed in recent years can fit and analyze various forms of CO 2 response curves more accurately [77,41] , overcoming the limitations of other models to a certain extent, there are few reports regarding its application in plants under different soil moisture conditions. This study indicated that when the soil moisture was within a suitable RWC range, the CO 2 response curves and characteristic parameters were well fitted by the four models (R 2 > 0.99, Fig. 2; Table 2), where the non-rectangular hyperbola model and modified rectangular hyperbola model fit the data better than the other two models (Fig. 2 B,D). When soil moisture was too high or too low, the modified rectangular hyperbola model was better than the other three models fitting the CO 2 response process and its characteristic parameters in the leaves of H.rhamnoides L. (Fig. 2 D). This result is consistent with the findings of Jiao & Wei [45]and Lv et al. [46]. This study demonstrated that the simulation results of the photosynthetic-CO 2 response model were closely related to soil moisture content.

Conclusions
Research on the effects of soil moisture on the physiological mechanisms related to photosynthetic responses is garnering attention toward CO 2 response curves and photosynthetic parameters in trees. This study indicated that soil moisture content affected the CO 2 response processes in the leaves of H.rhamnoides L. The photosynthetic P n -C i curve exhibited a classical form, with P nmax , CE, CSP, and R p being high, while Γ was low when the RWC ranged from 42.8% to 83.2%. H.rhamnoides L. exhibited high photosynthetic efficiency in this soil moisture range, and the P n levels were highest when RWC was 71.5%. Three photosynthetic parameters, P nmax , CE, and CSP, declined dramatically when soil moisture was outside the aforementioned range. Thus, the suitable RWC for P. sibirica L. ranged from 46.5% to 81.6%, and the most suitable RWC was ~66.8%.
The CE (i.e CE 0, CE Γ ,CE Γ0 )values of H. rhamnoides L. were significantly different under different soil moisture conditions. For example,the H. rhamnoides L. CE Γ0 ranged from 0.0260 to 0.0564, with a comparatively higher level > 0.047 in the RWC range of 42.8%-83.2%; the maximum (0.0564) appeared when RWC was71.5%. CE of H. rhamnoides L. decreased markedly when the soil moisture was too high or too low. When soil moisture was within the suitable RWC range, the CO 2 response curves and characteristic parameters were well fitted by the four models (R 2 > 0.99). The non-rectangular hyperbola model and modified rectangular hyperbola model fitted better than the other two models (R 2 > 0.998). However, when soil moisture exceeded the suitable RWC range, only the modified rectangular hyperbola model fit the CO 2 response curves and photosynthetic parameters well. Compared to the other three models, the modified rectangular hyperbola model demonstrated extensive applicability for fitting photosynthetic CO 2 responses under different soil moisture conditions.