The pot experiment was carried out from Januray in Wenshan Miao Xiang P. notoginseng Technology Park (23°05′N, 104°03′E), Yunnan, China. The healthy two-year-old rhizome of P. notoginseng were cultivated in plastic pots (30 cm × 25 cm × 25 cm), with each containing 3 rootstocks. Total photon exposure per day in screened growth house for seven treatments was equivalent to 29.8%, 11.5%, 9.6%, 5.0%, 3.6%, 1.4% and 0.2% of that in the full sunlight (FL), respectively. Figure S1 shows the diurnal variation of photosynthetic photon flux density (PPFD) under seven light treatments, respectively. 210 pots were used for each light intensity regimes, and a total of 1470 pots were arranged (n = 7). Polyoxin and agricultural streptomycin were used to control pests and diseases. In September, the youngest fully expanded functional leaf on each treatment at the maximum nutritional period from pot planting was used for the determination of photosynthetic performance, photosynthesis-related pigments, leaves anatomical characteristics and antioxidant defense system analysis.

Chlorophyll (Chl) was extracted as described by Pérez-Patricio et al. (2018). A LI-3000 leaf-area meter (Li-Cor, USA) was used to determine leaf area. 0.5 g of fresh leaves were immersed in a 15 mL extraction mixture [99% acetone was mixed with ethanol (2:1 v/v)]. 3 h of standing in the dark were followed by a 10 min centrifugation at 3000 g. Absorbance readings were performed at wavelengths of 665 nm and 649 nm. Chl a and b content were calculated based on the method of Gu et al. (2016). Total Chl content was the sum of Chl a and b.

Gas exchange measurements were performed between 09:00 and 11:00 on fully expanded function leaves using an LI-6400XT portable photosynthesis system equipped with a 6400-40 leaf chamber (LI-Cor, UAS). Leaf temperature was maintained at 25°C in the chamber. PPFD was 500 μmol·m^-2·s^-1 and CO₂ concentration was adjusted to 400 mmol·mol^-1 with a mixture. After equilibration to a steady state, net photosynthesis rate (P _n), stomatal conductance (Cond), transpiration rate (Tr), and intercellular CO₂ concentration (C _i) were recorded.

Dual-PAM 100 chlorophyll (Chl) fluorometer (Walz, Germany) was used to determine PSI and PSII Chl fluorescence parameters at 25°C. Seven plants were dark-adapted for 20 min, and both PSI and PSII parameter were monitored to record Chl fluorescence and P700 state. Then leaves were light-adapted at 172 μmol·m^-2·s^-1 for 20 min. Subsequently, PSI and PSII parameters were determined after 120 s exposure to each light intensity (0, 36, 94, 132, 172, 272, 421, and 611 μmol·m^-2·s^-1; PPFD, photosynthetic photon flux density). The chlorophyll fluorescence parameters were calculated as follows (Genty et al., 1989; Oxborough and Baker, 1997; Hendrickson et al., 2004): F _v/F _m = (F _m - F _o)/F _m; Y(II) = (F _m`-F <sub>s</sub>)/F <sub>m</sub>`; Y(NO) = F _s/F _m; NPQ = (F _m - F _m`)/F <sub>m</sub>`; 1 – qP = (F _s - F _o`)/(F <sub>m</sub>` - F _o`); Y(NPQ) = F <sub>s</sub>/F <sub>m</sub>` - F _s/F _m. F _o and F _o` were the minimum fluorescence after dark- and light- adaptation, respectively; F <sub>m</sub> and F <sub>m</sub>` were the maximum fluorescence after dark- and light-adaptation, respectively; and F _s was the dark-adapted steady-state fluorescence. F _v/F _m was the maximum quantum yield of photosystem II. Y(II) was the effective quantum yield of PSII photochemistry. Y(NO) and Y(NPQ) were the yield of non-regulated and regulated energy dissipation of PSII, respectively. NPQ was the non-photochemical quenching in PSII. 1-qP was the redox poise of the primary electron acceptor of PSII.

P700 redox state was calculated by the saturation pulse (600 ms, 10000 μmol·m^-2·s^-1) method (Klughammer and Schreiber, 2008). The P 700 + signals (P) may vary between a minimal (P700 fully reduced) and a maximal level (P700 fully oxidized); the maximum photo-oxidation P 700 + (P _m) and P _m` were ascertained the application of a saturation pulse after pre-illumination with far-red light and actinic light, respectively (Huang et al., 2010; Yamori et al., 2016; Takagi et al., 2017). The chlorophyll fluorescence parameters were determined by Klughammer and Schreiber (2008) method: Y(I) = (P <sub>m</sub>` - P)/P _m; Y(ND) = P/P _m; Y(NA) = (P _m - P _m`)/P _m. Y(I) was the effective quantum yield of PSII; Y(ND) and Y(NA) were the donor side and acceptor side limitation of PSI, respectively.

Photosynthetic electron flows through PSI and PSII were analyzed according to the method described by Huang et al. (2012a); Huang et al. (2017); Huang et al. (2019): ETRII = Y(II) × PPFD × 0.84 × 0.5; ETRI = Y(I) × PPFD × 0.84 × 0.5. ETRI was the electron transport rate of PSI; ETRII was the electron transport rate of PSII. Furthermore, the electron transport rate of cyclic electron flow around PSI was estimated as ETRI - ETRII; the quantum yield of cyclic electron flow around PSI was estimated as Y(I) – Y(II), or expressed as Y(I)/Y(II) (Miyake et al., 2005; Fan et al., 2016; Sagun et al., 2019).

Fast Chl fluorescence measurements were conducted by a pulse-amplitude modulation (PAM) fluorometer (PAM-2500, Walz, Germany). After a dark adaptation for 4 h, Chl fluorescence transient curves (OJIP transients) were inducted by a red light (652 nm) of 3000 μmol·m^-2·s^-1 by the PAM-2500 through an array of light-emitting diodes. Cha a fluorescence emission inducted by the strong light pulses was measured and digitized between 10 μs and 320 ms (Kanutsky curve; Kautsky and Hirsch, 1931). Meanwhile, four characteristic levels of fluorescence yield can be distinguished in a plot with logarithmic time scale: F _o, I ₁, I ₂ and F _m (alternatively also denoted O, J, I and P; Schreiber et al., 1986; Schreiber et al., 1989),. The F _o - I ₁ (or O-J) phase of the transient directly reflects the closure of PSII reaction centers by charge separation (Q_A-reduction). The initial rate of increase of this phase is proportional to the applied light intensity (photochemical phase). At a given light intensity, the initial rate provides a relative measure of the optical absorption cross-section of PSII. The I ₁- I ₂ - F _m (or J-I-P) phases of the transient reflect the reduction of the rest of the electron transport chain defined mainly by the reduction of the plastoquinone pool and the acceptor side of PSI; the rate of which is limited by dark reactions (thermal phase) (Schreiber and Klughammer, 2021). The point of time corresponding to 300 μs on the OJIP kinetic curves was defined as the “K” characteristic points (Eggenberg et al., 1995; Strasser et al., 2000; Strasser et al., 2004). The OJIP transients were analyzed using JIP-test, and the JIP-test is a multiparametric analysis of the OJIP transients, which is based on the theory of energy fluxes in bio-membranes (Strasser, 1981; Strasser and Strasser, 1995). From OJIP transient, the extracted parameters (F _{20 µs}, F _{300 µs}, F _{2 ms}, F _{30 ms} etc.) led to the calculation and derivation of a range of new parameters according to previous authors ( Table S1 ; Yusuf et al., 2010).

After photosynthetic parameters measurement, leaf sections of 1.00 × 1.00 cm were also cut from the middle of fully expanded function leaves (avoiding midribs). Leaves were cleaned by sterilizing water and stored in the FAA fixative. Leaf tissues were dyed by hematoxylin staining method and fixed with paraffin before observed (Xiong et al., 2017; Chang et al., 2023). The tissue sections were observed under electron microscope and analyzed through separately quantifying variables in the visible field using Case Viewer software.

Leaf was homogenized on ice with a mortar and pestle in a 0.1 M potassium phosphate buffer (pH 7.0). The homogenate was centrifuged at 12000 g for 15 min at 4°C. The supernatant was used immediately for enzyme assays (Wang et al., 2009). The activity of superoxide dismutase (SOD) was measured according to a method using xanthine, xanthine oxidase, and cytochrome c (Giannopolitis and Ries, 1977). The activity of peroxidase (POD) was assayed according to the method described by Zhang et al. (2005), using pyrogallol as a substrate. Catalase (CAT) activity was assayed according to the method described by Aebi (1984), by measuring the decrease at 240 nm for 1 min, due to H₂O₂ consumption.

SPSS 20.0 software (Chicago, IL, USA) was used to statistical analysis. The variables were means ± standard deviation (SD) (n = 7). Significant differences are indicated by letters (One-way ANOVA; P < 0.05). Graphing was made by SigmaPlot 10.0 (Systat Software Inc, San Jose) and GraphPad Prism 8.0 (GraphPad Inc, USA) software.

Leaves were significantly smaller and yellowish in P. notoginseng under high light; moderate-light-grown leaves were dark-green ( Figure 1A ). The content of Chl a, Chl b, total Chl increased first and then decreased with the increase of growth irradiance ( Figures 1B–D ). The maximum values of photosynthetic pigments were recorded in 5.0% FL-grown P. notoginseng ( Figure 1 ; as reflected by Chl a, Chl b, total Chl content). Chl a, Chl b, total Chl contents were lowest in P. notoginseng under 29.8% FL ( Figures 1B–D ).

P _n and Cond were significantly enhanced in 11.5% FL-grown plants compared with other treatments ( Figures 2A, B ). Compared with 11.5% FL-grown P. notoginseng, P _n were decreased 36.55% and 65.17% in 29.8% FL- and 0.2% FL-grown plants, respectively ( Figure 2A ). The maximum and minimum values of C _i were recorded in 0.2% FL- and 9.6% FL-grown plants, respectively ( Figure 2C ). The minimum values of P _n, Cond, and Tr were obtained in P. notoginseng under 0.2% FL condition ( Figures 2A, B, D ).

The thickness of the upper epidermis, palisade tissue, and lower epidermis were declined with increasing growth irradiance ( Table 1 , Figure S2 ). 29.8% FL-grown leaves were dramatically increased in the thickness of the upper epidermis, palisade tissue, and spongy tissue ( Table 1 ). The thickness of the lower epidermis was greatest in P. notoginseng grown under 29.8% and 11.5% FL condition ( Table 1 ). These differences were not significant for the upper epidermis thickness in the range 3.6% to 11.5% FL ( Table 1 ). The palisade/spongy increased first and then decreased with the increase of growth irradiance, and the maximum values of palisade/spongy were recorded in 5.0% FL-grown plants ( Table 1 ).

Growth irradiance significantly influenced PSI and PSII activity in the leaf ( Figure 3 ). The minimum values of F _v/F _m were showed in P. notoginseng grown under long-term low light (1.4% FL, 0.2% FL) ( Figure 3B ), and P _m in high-light-grown plants were lower (29.8% FL, 11.5% FL) ( Figure 3A ). The difference between moderate- and low-light-grown plants in P _m was only marginal ( Figure 3A ), but P _m was highest in P. notoginseng grown under 5.0% FL ( Figure 3A ).

ETRI, ETRII and ETRI - ETRII were raised with increasing PPFD ( Figure 4 ). ETRI and ETRII were significantly greater in 29.8% FL- and 9.6% FL-grown plants compared with other individuals ( Figures 4A, B ). ETRI and ETRII were significantly reduced in low-light-grown plants (0.2% FL; Figures 4A, B ). When PPFD was lower than 200 μmol·m^-2·s^-1, the maximum values of ETRI - ETRII were obtained in 0.2% FL and 29.8% FL P. notoginseng ( Figure 4C ). When plants were exposed to higher PPFD, the maximum values of ETRI - ETRII were recorded in 29.8% FL individuals, but the ETRI - ETRII were declined in low-light-grown P. notoginseng (0.2% FL, Figure 4C ).

The minimum values of Y(I) were shown in the 0.2% FL individuals ( Figure 5A ), and Y(ND) in low-light-grown individuals was greatest ( Figure 5B ). The opposite of Y(ND), Y(NA) was increased when PPFD is lower than 272 μmol·m^-2·s^-1 in plants grown under moderate shading environments ( Figure 5C ). There was no significant difference in Y(NA) when PPFD is more than 272 μmol·m^-2·s^-1. Compared with PSI, the lowest values of Y(II) were always observed in low-light-grown P. notoginseng ( Figure 5D ), and Y(NPQ) was highest in 0.2% FL plants ( Figure 5E ). Y(NO) was rapidly increased when PPFD is higher than 272 μmol·m^-2·s^-1 ( Figure 5F ), and the Y(NO) were increased in low-light-grown plants ( Figure 5F ). NPQ and 1-qP increased with increasing PPFD ( Figure 6 ). NPQ was increased in P. notoginseng were exposed to high light (29.8% FL, 11.5% FL; Figure 6A ), and 1-qP in 0.2% FL plants were highest ( Figure 6B ).

The quantum yield of cyclic electron flow around PSI [Y(I)/Y(II)] increased with increasing PPFD ( Figure 7A ). Y(I)/Y(II) was activated earlier when PPFD was higher than 36 μmol·m^-2·s^-1 in P. notoginseng under light stress (29.8% FL, 0.2% FL; Figure 7A ). Y(I)/Y(II) was inversely correlated with Y(II) ( Figures 5D , 7B ), and the greatest values were shown in 0.2% FL individuals ( Figure 7B ). As showed in Figure 8 , Y(NPQ), NPQ and Y(ND) were positively correlated with ETRI - ETRII ( Figure 8 ). Y(NPQ), NPQ and Y(ND) were greatest in the 0.2% FL individuals when ETRI - ETRII is lower ( Figure 8 ). Y(NPQ), NPQ and Y(ND) were increased in the high-light-grown plants when ETRI - ETRII was greater ( Figure 8 ).

POD activity was greater in P. notoginseng grown under 29.8%, 11.5%, and 9.6% FL condition ( Figure 9A , P < 0.05). The POD activity was declined with decreasing growth irradiance ( Figure 9A ), and the minimum values of POD activity was obtained in 0.2% FL-grown P. notoginseng ( Figure 9A ). CAT activity was significantly increased in high-light-grown plant (29.8% FL, 11.5% FL; Figure 9B ). CAT activity was lowest in 5.0% FL-grown plants ( Figure 9B ). SOD activity was reduced with decreasing grown irradiance in the range 29.8% to 9.6% FL ( Figure 9C ). SOD activity was significantly decreased in 3.6% FL-grown plants compared with 5.0%, 1.4% and 0.2% FL treatments ( Figure 9C , P < 0.05).

The OJIP kinetic curve showed an “S”-shaped in all light regimes ( Figure 10A ). The lower fluorescence values were shown in high-light-grown individuals, F _o≌F _{20 μs} (O phase) was greater in the 9.6% FL individuals, and the maximum values of F _M=F _P=F _{300 ms} (P phase) were recorded in the 5.0% FL individuals ( Figure 10A ). W _k was lower in moderate-light-grown plants (9.6% FL, 5.0% FL, 3.6% FL; Figure 11B ), and the maximum values of W _k were recorded in 0.2% FL individuals ( Figure 11B ).

In the JIP-test parameters, change in M _o, V _J and ψ _o can reflect activity of PSII acceptor sides (Force et al., 2003). Changes of M _o and V _J are similar ( Figures 10B , 11A ), and M _o and V _J were greater in low-light-grown plants (0.2% FL, Figures 10B , 11A ). ψ _o was significantly lower in 0.2% FL plants than in other light regimes plants ( Figure 10B ). Compared with F _v/F _m, PI_ABS could more sensitively reflect the activity of PSII acceptor sides (Crafts-Brandner and Salvucci, 2002). The minimum values of PI_ABS were surveyed in 0.2% FL individuals ( Figure 11C ), and there were not significantly different in other light regimes ( Figure 11C ). DI_o/RC and ABS/RC were highest in the 9.6% FL plants ( Figure 10B ), and ET_o/RC were higher in low-light-grown individuals (0.2% FL; Figure 10B ). ABS/RC and TR_o/RC were increased when the growth irradiance is lower than 5.0% FL ( Figure 10B ).

The plasticity index of P _m was much greater than that of F _v/F _m among the photosystem activity variables ( Figure 12 ); The higher plasticity index values of ETRI, ETRII, Y(II) and Y(I) were shown among photosynthetic electron transport and light energy distribution ( Figure 12 ). The plasticity indices of M _o were largest among PSII receptor side parameters ( Figure 12 ). Noteworthy, the plasticity indices of P _m, ETRII, ETRI, Y(II) and Y(I) exceeded 0.5, and the lowest plasticity indices values of F _v/F _m, Y(NPQ), ET_o/RC and W _k ( Figure 12 ).

Photosynthetic capacity is at least in part determined by leaf anatomy and P _n is limited by the rate of CO₂ diffusion from the atmosphere to the chloroplast (Gratani and Bombelli, 2000). The reduction of palisade tissue thickness increases the density of chloroplast distribution and enchants light-receiving area and light capture capability, thus improving photosynthetic capacity in shade -tolerant species (e.g., Phoebe bournei, Cyclobalanopsis gilva, Zelkova serrata, Cinnamomum camphora; Xue, 2020). Thicker upper epidermis protects mesophyll tissue from damage in high-light-grown Acer rybrum (Goulet and Pierre, 1986). The thickness of palisade tissue was declined with increasing growth irradiance, and 29.8% FL-grown leaves were dramatically increased in the thickness of the upper epidermis ( Table 1 , Figure S2 ). These results imply that P. notoginseng leaves made favorable adaption to high and low light, respectively. Correspondingly, the increase of upper epidermis, palisade tissue, and lower epidermis would reduce liquid phase diffusion of CO₂ in mesophyll cells ( Table 1 ), this might partly explain the fact that a significant decline in P _n was observed in the high-light-grown plants ( Figure 2 ), as has also been observed in Zhang et al. (2020). Meanwhile, low-light-grown leaves were declined in Tr and Cond, and C _i and P _n had opposite trends ( Figure 2 ). These results imply that the decline of photosynthetic rate in low-light-grown P. notoginseng was mainly caused by non-stomatal limitation factors, and this is consistent with the results reported by Rylski and Spigelman (1986). Thus, light-driven changes in P _n are in part explained by leaf anatomy.

It has commonly accepted that the primary sites of photoinhibition are PSI and PSII (Gerganova et al., 2016). The PSI and PSII photoinhibition is characterized by a significant decrease in P _m and F _v/F _m, respectively (Demmig-Adams and Adams, 1992). PSII activity is inhibited under high light, but PSI activity remains stable, and this has been confirmed in Solanum lycopersicum and Arabidopsis thaliana (Gerganova et al., 2019; Chen et al., 2020). F _v/F _m was greatly reduced in 1.4% FL- and 0.2% FL-grown plants ( Figure 3B ), but PSI activity was relatively increased in low-light-grown plants ( Figure 3A ). This is inconsistent with the results reported that inhibition of the activity of PSII under strong light is referred to as photoinhibition (Murata et al., 2007). This may be due to the different light demands of the study species (as reflected by P. notoginseng is a typically shade-tolerant species). These results imply that the degree of PSII photoinhibition is significantly affected by long-term low light stress, as confirmed in the shade-tolerant species P. henryi treated by short-term low light (Huang et al., 2016b). Meanwhile, the degree of inhibition of P _n under 0.2% FL was greater than that of 29.8% FL ( Figure 2A ), it implied that P. notoginseng are more sensitive to long-term low light compared to high light. Furthermore, compared with F _v/F _m, PI_ABS could more sensitively reflect the activity of PSII (Crafts-Brandner and Salvucci, 2002; Li et al., 2009b). PI_ABS in 0.2% FL plants was significantly lowest than other counterparts ( Figure 11C ). Obviously, PSII was more sensitive to low light stress compared with PSI. Therefore, long-term low light stress exacerbates the photoinhibition to PSII in the shade-tolerant species.

PSI activity is slow to recover from photoinhibition compared with the recovery of PSII activity (Zhang and Scheller, 2001; Zhou et al., 2019). PSI photoinhibition mainly occurs in plants grown under high light and chilling temperatures condition (Zhang and Scheller, 2001), as has been recorded in the shade-tolerant plants P. rubra, P. henryi and Nephrolepis falciformis (Huang et al., 2015; Huang et al., 2017; Huang et al., 2018b). P _m in 29.8% FL plants was greatly reduced by 51.57% in relative to 0.2% FL counterparts ( Figure 3A ), and PSI activity is significantly reduced in high-light-grown plants. The excess electrons on PSI acceptor side induce the formation of superoxide anion radicals and the reduction of the iron-sulfur center in PSI, which leads to photoinhibition to PSI (Sonoike, 2011). Y(NA) in 29.8% FL individuals was significantly higher than 0.2% FL individuals ( Figure 5C ), implying that the occurrence of PSI photoinhibition in high-light-grown P. notoginseng might is due to the excess accumulation of superoxide anion radicals on the PSI acceptor side as has been proposed by Kim et al. (2005). PSI is sensitive in high-light-grown P. notoginseng. On the other hand, the degree of PSI photoinhibition is greater than that of PSII photoinhibition in high-light-grown individuals ( Figure 3 ), and the plasticity index of P _m was larger than that of F _v/F _m ( Figure 12 ). PSI photoinhibition is the basis for the sensitivity of shade-tolerant plants P. rubra to high light condition (Huang et al., 2015). Thus, PSI photoinhibition might be a vital reason for explaining why the shade-tolerant plants P. notoginseng cannot grow under high light.

On the condition of excess light, the utilization and dissipation of light are increased to protect PSII and PSI against photoinhibition (Zhang et al., 2015; Bascuñán-Godoy et al., 2018). Higher NPQ dissipates excess energy as heat in order to prevent damage to PSII of high-light-grown A. thaliana and Chromera velia (Belgio et al., 2018; Howard et al., 2019). 29.8% FL-grown plants possessed a high NPQ ( Figure 6 ). These results imply that excess light energy could be effectively dissipated in the form of heat photochemistry in high-light-grown plants. Thus, high-light-grown plants show greater photochemical efficiency and photoprotective capacity, contributed by higher Y(II) and NPQ ( Figures 5D , 6A , 8 ), while the NPQ of shade plants is more sensitive to changes in high light. This is consistent with the results reported by Ishida et al. (2014) that a larger proportion of Y(II) and Y(NPQ) has been observed in high-light-grown O. sativa. Moreover, the utilization of excess light is increased by increasing electron transport and photochemistry in high-light-grown (Genty and Harbinson, 1996). Y(I), Y(II), ETRI, ETRII and NPQ were increased in the 29.80% FL individuals ( Figures 4 , 5 , 6A ); and the plasticity indices of ETRII, ETRI, Y(II) and Y(I) all exceeded 0.5 ( Figure 12 ). These results imply that excess light energy could be effectively dissipated in the form of heat or photochemistry in high-light-grown plants. However, excess light energy could not be effectively dissipated in time, which accumulates ROS (Zhou et al., 2019). Plants up-regulate the antioxidant enzyme system to scavengethe ROS under stress (Li et al., 2009). The activities of SOD, POD and CAT showed different degrees of changes in high-light-grown P. notoginseng ( Figure 9 ). This is consistent with the results reported by Zhang et al. (2022) that the activation of SOD and POD could avoid photooxidative damage in Pyropia haitanensis grown under high light condition. Overall, high-light-grown P. notoginseng had stronger capability of scavenging ROS and non-photochemical quenching. Moreover, light capture capability was decreased by inhabiting Chl content (as reflected by Chl a, Chl b, and total Chl content) in 29.80% FL-grown P. notoginseng ( Figures 1B-D ), as has been confirmed by Sato et al. (2015) in A. thaliana grown under high light stress. The degree of PSI photoinhibition is higher than that of PSII photoinhibition in high-light-grown P. notoginseng ( Figure 3 ). PSI photoinhibition in P. notoginseng grown under high light condition was primarily caused by the excess electron transport from PSII to PSI (Huang et al., 2015). PSI activity is protected against photodamage in pgr5 mutants of A. thaliana upon moderate PSII photoinhibition, due to the depression of electron flow from PSII to PSI (Tikkanen et al., 2014). Moderate photoinhibition of PSII is a protective response (Huang et al., 2016a; Huang et al., 2018a). F _v/F _m, Ψ _o, W _K and V _J were relatively stable when P. notoginseng were exposed to high light ( Figures 3B , 10B , 11A, B ), as has been confirmed by Thachle et al. (2007) in Graptophyllum reticulatum. These results imply that moderate photoinhibition of PSII occurs in high-light-grown P. notoginseng. Therefore, the enhanced photosynthetic electron transport and moderate PSII photoinhibition of P. notoginseng under high light condition were presented as photoprotection strategies.

The enhanced absorption and utilization of light energy is a predominated strategy for plants to adapt to low light (Lei et al., 1996; Ruberti et al., 2012), and this has been confirmed in the shade-tolerant species Paeonia veitchii, Paeonia intermedia and Paeonia anomala grown under low light (Wan et al., 2020). ABS/RC, TR_o/RC, 1-qP, and M _o were enhanced in 0.2% FL-grown P. notoginseng ( Figures 6B , 10B ). The capture and absorption of light energy were improved by the increased active reaction centers per unit area in P. notoginseng grown under low light. Additionally, antenna sizes are increased by enhancing Chl b and LHCII levels in low-light-grown A. thaliana, resulting in higher light capture capability (Sato et al., 2015). The previous observation is consistent with present results that the maximum values of Chl b content were recorded in 5.0% FL-grown P. notoginseng ( Figure 1C ). These results imply that light capture capability is enhanced by increasing antenna size in P. notoginseng grown under low-light stress.

It has commonly accepted that the state transition is a photoprotective mechanism that improves the utilization of plant light energy by balancing the excitation energy of PSI and PSII (Bailey and Grossman, 2008; Khuong et al., 2019). In the present study, the maximum values of 1-qP were recorded in 0.2% FL plants ( Figure 6B ). The maintenance of state 1 of P. notoginseng at 0.2% FL may be due to the strong PSII excitation, resulting in high excitation pressure on PSII (Tikkanen et al., 2006). These results imply that PSII reaction centers are inactivated in plants grown under low light, as has been confirmed by Chen and Xu (2006). However, the imbalance between the absorption and utilization of light energy could cause a damage to photosynthetic apparatus (Zavafer et al., 2019; Kodru et al., 2020). Y(II), Y(I), NPQ, φD_o and F _v/F _m were decreased in the 0.2% FL individuals, but Y(NO) was increased ( Figures 3B , 5A,D,F , 6A , 10B ), suggesting that excess light energy could not be effectively dissipated in the form of thermal in low-light-grown individuals, and it probably lead to the reduction in PSII activity and the damage to PSII. On the other hand, plants would use light energy through photosynthetic electron transport to protect photosynthetic apparatus, and this has been confirmed in the light-demanding species Shorea leprosula and Cerasus cerasoides grown under light stress (Scholes et al., 1996; Yang et al., 2019b). ETRI, ETRII, ETRI - ETRII, ET_o/RC and F _v/F _m were reduced in low-light-grown P. notoginseng (0.2% FL or 1.4% FL; Figures 3B , 4 , 10B ). Low-light-grown P. notoginseng cannot increase the utilization of light energy by enhancing electron transport. The decline in PSII activity result in the inhibition to electron transport in low-light-grown P. notoginseng ( Figures 3B , 4 ). This is consistent with the results reported by Huang et al. (2018a) that the decline in electron transport under low light is induced by a decline in PSII activity in P. notoginseng. The imbalance between PSI and PSII leads to reduced electron transport (Wen et al., 2005; Sonoike, 2011; Oguchi et al., 2021). The previous observation is consistent with present results that the lower value of ETRI, ETRII and ψ _o was observed in the 0.2% FL individuals ( Figures 4A, B , 10B ).

The OJIP kinetic curve reflects the degree of damage to PSII under light stress (Kumar et al., 2020; Lysenko et al., 2021). The appearance of the K-phase in OJIP is related to the injury of PSII donor side, particularly the OEC (Oxygen-evolving complex) (Zhang et al., 2016; Kumar et al., 2020). However, evidence is accumulating that K-phase is observed when plants are exposed to environmental stress, and K-phase are more pronounced in short-term stressed plants compared with long-term stressed individuals (Pagliano et al., 2006; Tóth et al., 2007). The appearance of the K-phase and the high value of W _k was obtained in P. notoginseng grown under long-term 0.2% FL condition ( Figures 10 , 11B ; P < 0.05), and this has been confirmed in Rosa hybrida grown under long-term drought stress (Pinior et al., 2005). These results indicate that electron transport is inhibited from electron donor of PSII to the reaction center in low-light-grown individuals, which in turn lead to the OEC injury of PSII donor side. M _o, Ψ _o, V _J and φE _o mainly reflects changes in PSII acceptor side (Ayyaz et al., 2020; Kumar et al., 2020; Khan et al., 2021). V _J and M _o were increased, andΨ _o was decreased in 0.2% FL-grown P. notoginseng compared with other counterparts ( Figures 10B , 11A ), implying that PSII reaction center is closed, a large amount of oxidized Q_A is accumulated and the electron transport after Q_A is inhibited, consequently resulting in a damage to the acceptor side of the PSII. Nevertheless, the increase in V _J and W _k reflects the degree of damage to the acceptor side and the donor side of PSII, respectively (Lu and Zhang, 2000). A similar effect has been observed in Glycine max and Zea mays grown under environmental stress (Li et al., 2009a; Li et al., 2009b). V _J and W _k were significantly increased in 0.2% FL compared with other counterparts, but the increase of V _J was larger than that of W _k ( Figures 11A, B ). Anyways, PSII acceptor side is more readily damaged than the donor side in P. notoginseng grown under low light condition.

Y(I)/Y(II) was activated earlier when PPFD was higher than 36 μmol·m^-2·s^-1 in when P. notoginseng were exposed to high light and low light condition (29.8% FL, 0.2% FL; Figure 7A ), but ETRI - ETRII in 29.8% FL plants was consistently higher than in 0.2% FL plants ( Figure 4C ). These results imply that ΔpH and ATP might be enhanced in high-light-grown P. notoginseng compared with the counterparts as has been suggested by Miller et al. (2020). In addition, high ΔpH not only decelerates the damage to PSII by protecting the OEC, but also protect PSI by regulating electron transport from PSII to PSI (Takahashi et al., 2009; Tikkanen et al., 2015). Similarly, cyclic electron flow around PSI plays an essential role in photoprotection for P. henryi, C. cerasoides and Phaeodactylum tricornutum under high-light (Huang et al., 2017; Yang et al., 2019b; Zhou et al., 2020; Sun et al., 2021). ETRI - ETRII, NPQ, ETRI and ETRII were increased, P _m was substantially reduced in the 29.8% FL plants ( Figures 3A , 4C , 6A ), and Y(NPQ), NPQ and Y(ND) have a positive correlation with ETRI - ETRII ( Figure 8 ), suggesting that cyclic electron flow around PSI protects PSI and PII from damage by enhancing thermal dissipation capacity and regulating P700⁺ redox state and electron transport in high-light-grown individuals.

Cyclic electron flow around PSI also shows photoprotection in plants exposed to low light (Laisk et al., 2005; Huang et al., 2011; Huang et al., 2012a; Huang et al., 2012b; Huang et al., 2019; Flannery et al., 2021). The maximum values of Y(NPQ), NPQ and Y(ND) were recorded in 0.2% FL-grown plants when ETRI - ETRII is lower ( Figure 8 ). High Y(NPQ), NPQ and Y(ND) depend on cyclic electron flow around PSI to produce ΔpH in low-light-grown plants (Munekage et al., 2004). ETRI - ETRII was reduced in the 0.2% FL plants when PPFD is above the value of 272 μmol·m^-2·s^-1 ( Figure 4C ), indicating that cyclic electron flow around PSI could not build up a sufficient ΔpH to protect PSII from photodamage in low-light-grown P. notoginseng. Severe photoinhibition to PSII would limit the transport of electrons from PSII to PSI, which in turn prevents damage to PSI (Huang et al., 2015). PSII activity and ETRII were drastically decreased when plants were exposed to low light (1.4% FL & 0.2% FL; Figures 3B , 4B ), but P _m was relatively stable ( Figure 3A ). The results obtained herein suggest that severe photoinhibition to PSII protects PSI from photodamage in low-light grown P. notoginseng. Overall, cyclic electron flow around PSI cannot completely protect PSII from damage under low light stress, but can prevent PSI photodamage.