Drought stress strengthens the link between chlorophyll fluorescence parameters and photosynthetic traits

Plant material and experimental design

Cucumbers (Cucumis sativus L.) were used as plant material, which was cultured in plastic seedling pots (12 × 8 × 10 cm) and cultivated in a growth chamber. Growth chamber temperature was 20−25 °C at day and 15−18 °C at night then light intensity was 1200 µmol m⁻² s⁻¹ with relative humidity (RH) at 75%. The potting soil was a composite culture substrate composed of wood chips, peat, pine bark, and sand. Six mature cucumbers were divided randomly into two treatments: drought and control, with three replicates per treatment. The drought treatment started on November 21, 2018. The soil moisture content (θ_g) of the drought was measured about 8  ± 2% by the weighing method on November 23, 2018, while the soil moisture content of the control was about 15  ± 1%. During the experiment, the plants of the drought treatment were not irrigated, while the plants of the control group were irrigated daily. The upmost, sunlit, dark green, fully unfolded and mainstem leaves were used to measure gas exchange and F_v/F_m, and adjacent leaves were used to measure laser-induced chlorophyll fluorescence and Chl_t.

Measurement of the CO₂ response curve and chlorophyll fluorescence

Typical A_n/C_i curves (light-saturated net CO₂ assimilation rate versus intercellular CO₂ concentrations) were measured using the Li-6800 portable photosynthesis system (LI-COR Inc., USA) from 8:00 to 11:30 after two days of drought treatment. The upmost fully unfolded, mainstem leaves were measured at leaf temperature of 25 °C, RH of 50–60%, and photosynthetic photon flux density (PPFD) of 1500 µmol m⁻² s⁻¹. The carbon dioxide concentration of the reference chamber was set as 400, 100, 50, 100, 400, 600, 800, 1,000 µmol mol⁻¹. A total of 54 A_n/C_i curves were taken (i.e., 3 samples per treatment × 7 times per sample × 2 treatments + 6 samples per treatment on the first day × 2 treatments = 54 curves). Before measuring the A_n/C_i curve, the leaves were adapted for 5 min at a CO₂ concentration of 400 µmol mol⁻¹. Measurements of the A_n/C_i curve were taken when gas exchange had equilibrated (taken to be when the coefficient of variation for the CO₂ partial pressure differential was below 1% between the sample and reference analyzers). This condition was typically achieved within 1–2 min after a stable CO₂ concentration had been reached.

F_v/F_m was measured using the Li-6800 fluorescence leaf chamber (LI-COR Inc., USA) connected to an LI-6800 portable photosynthesis system after dark treatment for one night. In the evening before the measurement, the upmost fully unfolded, mainstem leaves used to measure the dark-adapted fluorescence parameters were wrapped with tin foil. The rectangular flash was configured with a red target of 8,000 µmol m⁻² s⁻¹, a duration of 1,000 ms, the output rate of 100 Hz, and a margin of 5 points.

The laser-induced chlorophyll fluorescence system was composed of blue laser light source with a peak emission of 456 nm (Dslaser, China), USB4000 grating spectrometer, VIS-NIR band optical fiber (Ocean Optics, USA), Long-pass optical filter (AT600lp, Chroma Technology Corp, USA), and computer with software (Fig. 1). The light source output power was 40 mW, and the corresponding light source input voltage was 5.7 V. The spectrometer used in the experiment has a resolution of 1.5 nm, an integral time of 3.8 ms–10 s, and detector covers of 200–1,100 nm. The spectrometer was equipped with a USB port on the side, which was connected to the computer and directly powered by the computer. The linear array CCD detector (Toshiba, Japan) of the USB4000 spectrometer has a pixel count of 3648 (Li, Huang & Zhang, 2009). SMA905 fiber adapter (DingSuo Technologies, China) was used as a connector to match the VIS-NIR band optical fiber and USB4000 spectrometer. VIS-NIR band optical fiber has an optical fiber core diameter of 1,000 um, numerical aperture of 0.22, and divergent Angle of 25.4°. The included angle between light source and leaves is 45°, the fiber is perpendicular to the leaves with a distance of 4.5 cm. The long-pass optical filter with a transmission wavelength range greater than 600 nm, and transmittance greater than 90% to prevent the influence of reflected light on the fluorescence spectrum. The chlorophyll fluorescence of 650–850 nm was received by optical fiber then collected by the spectrometer. SpectraWiz software (StellarNet Inc., Tampa FL, USA) was used to set up to collect three spectra and take the average, the integrating time with 600 ms.

Measurement of chlorophyll content

The leaves used to measure the ChlF spectrum were cut out and used to measure Chl_t. Starting from November 24, 2018, Chl_t was measured every other day. The control and drought were repeated three times (repeated six times on the first day), for a total of four measurements. A total of 30 chlorophyll content were collected (i.e., 3 samples per treatment × 3 times per sample × 2 treatments + 6 samples per treatment on the first day × 2 treatments = 30). Acetone and anhydrous ethanol were mixed into the extract at a volume of 1:1. The leaves were cut into filaments and weighed at 0.1 g in the test tube containing the mixture, which was placed in the dark place. After the material was completely white, the optical density at 663 nm and 645 nm was measured by spectrophotometer (MAPADA, China). The Chl_t in this study was expressed as follows: (1)${\displaystyle {\mathrm{Chl}}_{\mathrm{a}}=12.72\times {\mathrm{A}}_{663}-2.59\times {\mathrm{A}}_{645}\times 0.1}$ (2)${\displaystyle {\mathrm{Chl}}_{\mathrm{b}}=22.88\times {\mathrm{A}}_{663}-4.67\times {\mathrm{A}}_{645}\times 0.1}$ (3)${\displaystyle {\mathrm{Chl}}_{\mathrm{t}}={\mathrm{Chl}}_{\mathrm{a}}+{\mathrm{Chl}}_{\mathrm{b}}.}$ Where Chl_a is the chlorophyll A content (mg g⁻¹), and Chl_b isthe chlorophyll B content (mg g⁻¹), and Chl_t isthe total chlorophyll content (mg g⁻¹). The A₆₄₅ and A₆₆₃ are absorbances at wavelengths 645 and 663, respectively.

Statistical analysis

V_cmax and J_max were estimated by fitting the A_n/C_i curves using a spreadsheet-based software developed by Sharkey (Sharkey, 2016). The chlorophyll fluorescence collection program for the spectrometer was written based on the underlying program of the spectrometer with MATLAB software (Haiye, Haoyu & Lei, 2009). Dark current noise is removed from the chlorophyll fluorescence spectrum curve and the curve is smoothed by Savitzky-Golay filtering (Gorry, 1990).

We repeated measurements of the same six individuals. Repeated Measures ANOVA (RMANOVA) was used to test the effects of drought stress on photosynthetic traits and chlorophyll fluorescence parameters. The effects were considered to be significantly different if P < 0.05. Besides, a mixed-effect linear model was used to evaluate the effect of V_cmax, J_max, g_s, and Chl_t on A_n and chlorophyll fluorescence parameters. The individual plant was used as a random term. Similar method was used to test the relation between A_n and chlorophyll fluorescence parameters. All statistical analyses were performed using SPSS 25.0 (SPSS Inc., USA).

The response of leaf photosynthetic traits and ChlF ratio to drought

The drought stress caused a significant reduction in A_n, V_cmax, J_max, g_s, and Chl_t compared with control (P < 0.05). The averages of A_n were 1.0 ± 0.1 µmol m⁻² s⁻¹ and 2.0 ± 0.2 µmol m⁻² s⁻¹ in drought and control (Fig. 2A). The averages of V_cmax and J_max were 92.0 ± 5.0 µmol m⁻² s⁻¹ and 117.0 ± 5.0 µmol m⁻² s⁻¹ in drought, respectively (Figs. 3A and 3C). For control, the averages of V_cmax and J_max were 105.9 ± 4.3 µmol m⁻² s⁻¹ and 141.0 ± 5.2 µmol m⁻² s⁻¹, respectively (Figs. 3A and 3C). Compared to control, the averages of V_cmax and J_max decreased 13.1% and 17.1%, while the g_s and Chl_t in drought (0.1 ± 0.01 mol m⁻² s⁻¹ and 1.6 ± 0.1 mg g⁻¹) were reduced by 27.1% and 21.5% compared with control (0.1 ± 0.01 mol m⁻² s⁻¹ and 2.1 ± 0.1 mg g⁻¹) (Figs. 3E and 3G). Compared to the control plants, the averages of F_v/F_m decreased 6.8% (0.74 ± 0.01 to 0.69 ± 0.13), while LD₆₈₅/LD₇₄₀ increased 10.7% (1.1 ± 0.4 to 1.2 ± 0.3).

Control factors for A_n and ChlF ratio

Significant positive correlation was found between A_n and V_cmax in the control (R² = 0.15, P = 0.03) and drought stress (R² = 0.45, P < 0.001). There was a significant positive correlation between J_max and A_n in the control (R² = 0.12, P = 0.04) and under drought stress (R² = 0.60, P < 0.001). Besides the significant positive correlation (R² = 0.11, P = 0.05) between g_s and A_n in the control, a significant positive correlation (R² = 0.48, P < 0.001) in the drought stress was observed. A poor correlation (R² = 0.07, P = 0.84) between A_n and Chl_t was observed in the control, whereas the correlation was positive (R² = 0.56, P < 0.001) in the drought stress (Figs. 4A–4D).

A significant positive correlation between F_v/ F_m and V_cmax was found in the drought (R² = 0.13, P = 0.03). Meanwhile, a marginally positive correlation between F_v/F_m and J_max was observed in the control (R² = 0.09, P = 0.09), and a significant positive correlation was observed in the drought stress (R² = 0.28, P = 0.03). In addition, a significant positive correlation was found between F_v/F_m and g_s in the control (R² = 0.28, P = 0.003) while the correlation was poor under drought stress (R² = 0.01, P = 0.22). No significant correlation (R² = 0.03, P = 0.27) was observed between F_v/F_m and Chl_t in the control group, whereas a significant positive correlation was found between F_v/F_m and Chl_t under drought stress (R² = 0.29, P = 0.02) (Figs. 4E–4H).

A significant negative correlation was found between LD₆₈₅/LD₇₄₀ and V_cmax in drought stress (R² = 0.18, P = 0.009), while there was no significant correlation between LD₆₈₅/LD₇₄₀ and V_cmax in control (R² = 0.03, P = 0.50). There was no significant correlation between LD₆₈₅/LD₇₄₀ and J_max in control (R² = 0.03, P = 0.69), while a significant negative correlation was observed between LD₆₈₅/LD₇₄₀ and J_max in drought stress (R² = 0.13, P = 0.04). No significant correlation was observed between LD₆₈₅/LD₇₄₀ and g_s in control (P = 0.13) and drought stress (P = 0.09). There was no significant correlation between the LD₆₈₅/LD₇₄₀ and Chl_t in the control (R² = 0.06, P = 0.68), while a significant negative correlation between LD₆₈₅/LD₇₄₀ and Chl_t in drought stress (R² = 0.25, P = 0.03) (Figs. 4I–4L).

Correlation between A_n and ChlF ratio

There was a marginally positive correlation between A_n and F_v/F_m in the control (R² = 0.08, P = 0.09). However, the correlation between A_n and F_v/F_m was significant positive in drought stress (R² = 0.28, P = 0.003). Similarly, there was no significant correlation between A_n and LD₆₈₅/LD₇₄₀ in the control (R² = 0.02, P = 0.45), while a significant negative correlation was found between A_n and LD₆₈₅/LD₇₄₀ in drought stress (R² = 0.17, P = 0.02) (Figs. 5A–5B).

Combination of the stomatal limitation, non-stomatal limitations, and chlorophyll content regulated leaf photosynthesis under drought stress

As expected, drought stress significantly decreased the leaf photosynthesis of cucumber (Fig. 2A). Although drought stress is known to reduce leaf photosynthesis, the processes responsible for the key limitations are still a matter of debate (Chaves et al., 2002; Flexas & Medrano, 2002; Pinheiro & Chaves, 2011). Here, our study found that the combination of the stomatal limitation, non-stomatal limitations, and chlorophyll content regulated the decrease of leaf photosynthesis under drought stress (Figs. 4A–4D). Increasing evidence shows that leaf photosynthesis under drought stress is not limited by a single process (Zhou et al., 2015). It has been demonstrated that stomatal closure reduces photosynthesis and transpiration while improving water use efficiency due to acclimation under drought stress (Feller, 2016; Lamaoui et al., 2018). Non-stomatal limitations were defined as the sum of the contributions of mesophyll conductance and leaf biochemistry, which directly reflect the biochemical process of photosynthesis (Grassi & Magnani, 2005). The decrease in V_cmax and J_max may result from a decrease in the amount of active Rubisco and an inadequate supply of ATP or NADPH or to a low enzymatic activity during the photosynthetic carbon reduction cycle (Campos et al., 2014; Flexas et al., 2004; Peña Rojas, Aranda & Fleck, 2004). Recent studies found that stomatal and non-stomatal limitations to photosynthesis are coordinated on similar timescales, and suggested that non-stomatal limitations should be included in predict the model of photosynthesis response to drought (Drake et al., 2017; Salmon et al., 2020). A study of three Mediterranean species has found that the decrease of A_n was simultaneous regulated by stomatal and biochemical limitations during drought stress (Varone et al., 2012). In addition, Chl_t is the main pigment that absorbs photosynthetically active radiation and can indirectly reflect the integrity of the photosynthetic device (Streit et al., 2005). Chl_t can be used as a functional trait to evaluate drought stress (Pilon et al., 2018; Pilon et al., 2014). Therefore, our observations suggest that the combination of the stomatal limitation, non-stomatal limitation, and chlorophyll content should be taken into consideration in process-based models for simulating photosynthesis in terrestrial ecosystems under drought stress.

J_max and Chl_t governed ChlF ratio under drought stress

Leaf F_v/F_m is a crucial chlorophyll fluorescence parameter for evaluating the health or integrity of the internal apparatus under drought stress (Krause & Weis, 1991; Urban, Aarrouf & Bidel, 2017). Here, we found that F_v/F_m was significantly decreased under drought stress (Fig. 2B), which revealed that the PSII may be damaged under drought stress, and the primary reaction of photosynthesis may be inhibited (Lichtenthaler & Rinderle, 1988). The fluorescence parameters of the leaves are changed in two ways under stress conditions. The minimal fluorescence (F_o) increases due to obstruction of the electron flow through PSII, and plastoquinone acceptor (QA⁻) cannot be completely oxidized during stress. Simultaneously, the reduction of F_m during stress may be affected by decreased activity of the water-splitting enzyme complex and perhaps a concomitant cyclic electron transport within or around PSII (Porcar-Castell et al., 2014). Therefore, F_v/F_m will decrease under drought stress. Our finding was consistent with a previous studying, in which drought stress inhibited the photochemical activity of PSII and decreased leaf F_v/F_m (Meng et al., 2016). Meanwhile, our study showed that F_v/ F_m was largely related to J_max and Chl_t under drought stress (Figs. 4F and 4H). It has been proposed that J_max is decreased by drought stress, which prevents the electron from rapidly transferring back, and hinders the whole photochemical process (Baker & Rosenqvist, 2004; Khatri & Rathore, 2019). Similarly, the decrease in Chl_t will weaken the photochemical process, which demonstrates the dependence of the light absorption and fluorescence emission on the concentration of chlorophyll molecules in the chloroplast (Nyachiro et al., 2001). Thus, the significant linear relationship between F_v/F_m and J_max and Chl_t observed in drought stress jointly indicates the importance of J_max and Chl_t in governing chlorophyll fluorescence.

In our study, the LD₆₈₅/LD₇₄₀ based on spectral analysis was significantly increased under drought stress (Fig. 2C). This finding was similar to the study of Meng et al. (2016), in which the fluorescence intensity ratio increased when the PSII was damaged. Moreover, LD₆₈₅/LD₇₄₀ was largely regulated by J_max and Chl_t under drought stress (Figs. 4J and 4L). The previous study found that changes in the chlorophyll content resulted in changes of more than 90% for the F₆₉₀/F₇₃₅ ratio (Csintalan, Tuba & Lichtenthaler, 1998). There was a significant negative correlation between the LD₆₈₅/LD₇₄₀ and Chl_t under drought stress (Fig. 4L). The chlorophyll absorption spectrum overlaps with the chlorophyll fluorescence emission spectrum in the red band, which results in the LD₆₈₅ being decreased by re-absorption in the case of higher chlorophyll content. The effect of re-absorption on the red band is stronger than that of the far-red band, and therefore the LD₆₈₅/LD₇₄₀ will decrease (Buschmann, 2007). The LD₆₈₅/LD₇₄₀ represents an ideal tool for evaluating the change of Chl_t and reflects the photochemical activity of PSII indirectly under drought stress (Figs. 4J and 4L). Spectral analysis has been used to directly assess ecosystem functioning under climate change. For instance, Gameiro et al. (2016) found a significant linear relationship between the fluorescence intensity ratio and leaf water content of Arabidopsis. Norikane et al. (2003) successfully monitor the growth of tomato under drought stress based on spectral analysis. Here, synchronous observation of LD₆₈₅/LD₇₄₀ and F_v/ F_m based on spectral analysis and fluorescence kinetics suggest that LD₆₈₅/LD₇₄₀ can be used as an indicator for detection of plant stress.

Drought stress strengthens the relationship between net CO₂ assimilation rate (A_n) and ChlF ratio

Our study reported a significant relationship between the A_n and ChlF ratio under drought stress, while no significant correlation was found in the control (Figs. 5A and 5B). The strengthening relationship between A_n and ChlF ratio may be ascribed to variations of J_max and Chl_t under drought stress. On the one hand, the reduction of F_v/F_m and J_max under drought indicated that the photosynthetic electron transport was damaged. Drought stress damages the reaction center of PSII and inhibits the electron transfer process of photosynthesis, which reduces the light energy conversion efficiency of PSII (Brestic et al., 1995; Cornic & Fresneau, 2002; Longenberger et al., 2009). On the other hand, drought stress deforms the leaf chloroplast layer structure and reduces chlorophyll content (Batra, Sharma & Kumari, 2014). So, J_max and Chl_t became limiting factors for the A_n and ChlF ratio under drought stress (Figs. 4B, 4D, 4F, 4H, 4J and 4L). The strengthening relationship between A_n and ChlF ratio has been observed in previous studies (Murchie & Lawson, 2013; Su et al., 2015). For example, Wang et al. (2018) found a significant linear relationship between A_n and F_v/F_m in the soybean experiment under drought conditions. Batra, Sharma & Kumari (2014) found similar results by studying the ChlF ratio characteristics of mung beans under drought conditions. Therefore, ChlF ratio based on spectral analysis and fluorescence kinetics was a better indicator of the photosynthetic capacity under drought stress.