Exposed anthocyanic leaves of Prunus cerasifera are special shade leaves with high resistance to blue light but low resistance to red light against photoinhibition of photosynthesis

Abstract

Background and Aims

The photoprotective role of foliar anthocyanins has long been ambiguous: exacerbating, being indifferent to or ameliorating the photoinhibition of photosynthesis. The photoinhibitory light spectrum and failure to separate photo-resistance from repair, as well as the different methods used to quantify the photo-susceptibility of the photosystems, could lead to such a discrepancy.

Methods

We selected two congeneric deciduous shrubs, Prunus cerasifera with anthocyanic leaves and Prunus triloba with green leaves, grown under identical growth conditions in an open field. The photo-susceptibilities of photosystem II (PSII) and photosystem I (PSI) to red light and blue light, in the presence of lincomycin (to block the repair), of exposed leaves were quantified by a non-intrusive P700⁺ signal from PSI. Leaf absorption, pigments, gas exchange and Chl a fluorescence were also measured.

Key Results

The content of anthocyanins in red leaves (P. cerasifera) was >13 times greater than that in green leaves (P. triloba). With no difference in maximum quantum efficiency of PSII photochemistry (F_v/F_m) and apparent CO₂ quantum yield (AQY) in red light, anthocyanic leaves (P. cerasifera) showed some shade-acclimated suites, including lower Chl a/b ratio, lower photosynthesis rate, lower stomatal conductance and lower PSII/PSI ratio (on an arbitrary scale), compared with green leaves (P. triloba). In the absence of repair of PSII, anthocyanic leaves (P. cerasifera) showed a rate coefficient of PSII photoinactivation (k_i) that was 1.8 times higher than that of green leaves (P. triloba) under red light, but significantly lower (−18 %) under blue light. PSI of both types of leaves was not photoinactivated under blue or red light.

Conclusions

In the absence of repair, anthocyanic leaves exhibited an exacerbation of PSII photoinactivation under red light and a mitigation under blue light, which can partially reconcile the existing controversy in terms of the photoprotection by anthocyanins. Overall, the results demonstrate that appropriate methodology applied to test the photoprotection hypothesis of anthocyanins is critical.

INTRODUCTION

Anthocyanins are water-soluble pigments that, depending on their composition, the local pH, temperature, medium and co-pigmentation processes with other flavonoids, hydroxycinnamic acids or metallic cations, may appear red, purple, blue or black (Renner and Zohner, 2019; Alejo-Armijo et al., 2020). Anthocyanins are formed in a branch of the flavonoid pathway (Davies et al., 2022), and are found in fruits, flowers, leaves, stems and roots (Cooney et al., 2015). Anthocyanic (red) leaves (long-lasting or temporary) are common in nature (e.g. 10 % of the tree species of the temperate regions; Archetti et al., 2009), depending on species, leaf ontogeny (young, mature and senescent) and environmental conditions. Environmental factors that affect the formation of anthocyanins include growth irradiance (at an exposed location or understorey), UV-B radiation, temperature, drought, nutrient or biotic agents like fungi and herbivores (for reviews see Davies et al., 2018; Li et al., 2019; Tossi et al., 2021). Why some leaves appear red instead of green has attracted great attention for a long time (Pringsheim, 1879; Wheldale, 1916). Although there are many existing hypotheses [photoprotection, thermoregulation, scavenging of reactive oxygen species (ROS), osmotic regulation, communication to pollen and seed dispersal agents, etc.] to explain the functions of anthocyanins in leaves, there is no consensus in this regard (for review see Davies et al., 2022).

From the perspective of the photoprotection hypothesis, it is supposed that accumulation of anthocyanins in the vacuoles of (sub-)epidermal cells (Hughes et al., 2014; Landi et al., 2014) would confer an additional photoprotection capacity for the photosystems (Landi et al., 2015, 2021; Gould et al., 2018; Renner and Zohner, 2019, 2020). Because they can absorb incident irradiance (maxima in the 510–540 nm waveband) [e.g. for red Berberis leaves, up to 40 % of photosynthetically active radiation (PAR); Nichelmann and Bilger, 2017], anthocyanins can attenuate the excess absorbed light energy not used for photochemistry. Less excess absorbed light energy would reduce the excitation pressure on the photosystem reaction centres [primarily photosystem II (PSII), at favourable temperatures], and decrease the yield of triplet chlorophyll (³Chl) in the PSII reaction centres; otherwise, ³Chl can live long enough to interact efficiently with ground state O₂ (via the miss-associated recombination of P680⁺Q_A⁻; Mattila et al., 2023), generating toxic singlet O₂ and damaging the PSII reaction centre. This type of photoinhibition mechanism is called the ‘excess energy hypothesis’ (for review see Zavafer and Mancilla, 2021). Meanwhile, anthocyanins can presumably act as powerful antioxidants to scavenge ROS produced by excess absorbed light energy (Davies et al., 2022), though the light-attenuating property of anthocyanins seems to be more important than their antioxidative property in photoprotection (Zheng et al., 2021).

However, the current results on the effects of anthocyanins in the photoprotection of the photosystems are inconsistent and not well understood (reviewed by Manetas, 2006; Gould et al., 2018; Davies et al., 2022). The accumulation of anthocyanins has been shown to (1) exacerbate (Choinski and Johnson, 1993; Dodd et al., 1998; Woodall and Stewart, 1998; Esteban et al., 2008; Zeliou et al., 2009; Nikiforou et al., 2010, 2011; Juvany et al., 2012), (2) be marginal to (Burger and Edwards, 1996; Pietrini and Massacci, 1998; Lee et al., 2003; Karageorgou and Manetas, 2006; Henry et al., 2012; La Rocca et al., 2014; Fernández-Marín et al., 2015; Logan et al., 2015; Mlinarić et al., 2017; Mattila and Tyystjärvi, 2023) or (3) mitigate (Krol et al., 1995; Feild et al., 2001; Manetas et al., 2002, 2003; Oberbaueri and Starr, 2002; Pietrini et al., 2002; Gould, 2004; Hughes and Smith, 2007; Hughes, 2011; Nichelmann and Bilger, 2017; Tattini et al., 2017; Gould et al., 2018) net photoinactivation of PSII. The discrepancy can be largely ascribed to the selection strategies of green leaves as the control and the growth conditions, as well as the methodologies applied to quantify the PSII photo-susceptibility (Gould et al., 2018; Landi et al., 2020; Davies et al., 2022). For example, Manetas et al. (2002) found that young red leaves of Rosa sp. and Ricinus communis were more resistant to high light than mature green leaves. However, besides anthocyanins, other photoprotection strategies may also be involved that are different between anthocyanic and green leaves along the developmental trajectory (Tattini et al., 2014). It is, therefore, recommended that variegated leaves within a species at the same developmental phase and canopy position can be compared to reveal the photoprotection role of anthocyanins for field studies (Manetas, 2006). Nevertheless, phenotypes should be the same for a given genotype under identical growth and ontogeny conditions; thus, the intra-specific variation in photoprotection in that study (Manetas et al., 2002) can be related to unmeasured and/or unidentified abiotic or biotic factors other than anthocyanins.

Besides the absence of a consensus on the photoprotection by anthocyanins, it is intriguing to find that exposed and mature anthocyanic leaves show some shade-acclimated properties. These properties include thinner laminae (Manetas et al., 2003), a lower proportion of palisade to spongy parenchyma (Manetas et al., 2003; Kyparissis et al., 2007; Tattini et al., 2014), a lower Chl a/b ratio (Manetas et al., 2003), lower levels of the xanthophyll cycle components and β-carotene (Manetas et al., 2003; Tattini et al., 2017), lower maximum net photosynthesis rates (Gould et al., 2002; Landi et al., 2020) and higher overexpression of genes promoting chlorophyll biosynthesis and light harvesting (Landi et al., 2020). This has led to a speculation that exposed anthocyanic leaves actually are more susceptible to strong irradiance because shade green leaves have a higher rate coefficient of photoinactivation than sun ones (Aro et al., 1993; Fan et al., 2016), which is against the hypothesis of photoprotection by anthocyanins. However, since the rate coefficient of photoinactivation is directly proportional to incident irradiance for green leaves (Tyystjärvi and Aro, 1996; Lee et al., 2001; He and Chow, 2003; Kato et al., 2003; Allakhverdiev and Murata, 2004; Yi et al., 2022), and anthocyanins attenuate irradiance received by photosynthetic pigments in a certain waveband, it is better to assess the PSII photo-susceptibility of anthocyanic leaves to red light, which is poorly absorbed by anthocyanins (Feild et al., 2001), provided absorption of red light by photosynthetic pigments between green and anthocyanic leaves is not so different.

Another caution in the investigation of photoprotection of anthocyanic leaves is the methodologies applied to quantify PSII photo-susceptibility. For example, many relevant studies measured the functional PSII content without blockage of repair during photoinhibition treatment, such that the measured susceptibility of PSII to photo-inactivation actually is a net effect of resistance and repair capacities (Aro et al., 1993; Tyystjärvi and Aro, 1996; He and Chow, 2003; Murata et al., 2012; Fan et al., 2016; Han et al., 2023; Mattila and Tyystjärvi, 2023). These two processes need to be disentangled because the excess energy hypothesis states that excess absorbed energy directly affects the resistance process (Adams et al., 2014), while the manganese-cluster hypothesis states that excess absorbed energy indirectly affects repair via ROS formation, without affecting resistance (Takahashi and Murata, 2008). Secondly, both hypotheses address the importance of the photoinhibitory light spectrum applied, as the manganese-cluster hypothesis gains support in the blue to the ultraviolet region, while the excess energy hypothesis gains support in the red region for green leaves (Zavafer, 2021). Thirdly, F_v/F_m (maximum quantum efficiency of PSII photochemistry), the widely applied metric representing the functional PSII content during photoinhibition treatment, has been questioned as some studies have shown no linear relationship between F_v/F_m and relative functional PSII content when photoinhibited across or within species (Losciale et al., 2008; Serôdio and Campbell, 2021), probably due to heterogeneous quenching of F_m after photoinhibition (Ruban and Horton, 1995; Malnoë et al., 2018) and/or other mechanisms (Pfündel, 1998; Serôdio and Campbell, 2021). Further, it was recently demonstrated that the F_v/F_m parameter may not be equated with the quantum efficiency of PSII photochemistry, while this parameter might still be used to monitor the functioning of PSII (Sipka et al., 2021).

In the present study, we selected two congeneric deciduous shrubs, purple leaf plum (Prunus cerasifera) and flowering plum (Prunus triloba), both belonging to Rosaceae. The exposed mature leaves of purple leaf plums are purple/red, while those of flowering plums are green. We also selected another two congeneric deciduous shrubs, hybrida vicary privet (Ligustrum × vicaryi) and purpus priver (Ligustrum quihoui), both belonging to Oleaceae. The exposed mature leaves of hybrid vicary privet appear yellow, while those of purpus privet appear green. The four species are widely distributed in South and North China, as famous foliage shrubs in landscaping practice. Leaf absorption, pigments, gas exchange and Chl a fluorescence were measured. Specifically, the PSII and PSI photo-susceptibilities to red light and blue light, in the absence of PSII repair, have been quantified by a P700⁺ kinetics area method we developed (Hu et al., 2013). The major aim of the present study is to test: (1) if PSII and PSI of anthocyanic leaves show less photo-susceptibilities to red/blue light than green leaves, and (2) whether yellow leaves function differently from anthocyanic leaves in terms of photoprotection and photosynthetic capacities.

MATERIALS AND METHODS

Plant materials

Shrubs of Prunus cerasifera, P. triloba, Ligustrum × vicaryi and L. quihoui were grown under natural sunlight in an open environment, in the China National Botanical Garden (39°59ʹN, 116°13ʹE, elevation 71 m) (Supplementary Data Fig. S1). During the experiment, the sampled shrubs were not artificially managed. Leaves were taken from three individuals for each shrub species. The sampling period was in August 2022 (growth season). The monthly mean maximum temperature and the monthly mean minimum temperature were 31 and 21°C, respectively. The average annual precipitation was 606 mm. The total precipitation in August 2022 was 78.8 mm (data from China Meteorological Administration).

Growth irradiance and soil nutrients in the field

All the growth irradiance measurements were made within 10 min at 1000 h on three days in a row, under clear weather conditions in August 2022. At least three measurements were made above each sampled leaf. The incident light spectrum was measured with a portable CCD array spectroradiometer with a spectral range of 200–1100 nm (AvaSpec-ULS2048 × 64, Avantes, The Netherlands). The integration time of the AvaSpec spectroradiometer was set to 10 ms and the spectral irradiance was measured in units of micromoles per square metre per second per nanometre. We followed the measurement and post-processing protocol of Hartikainen et al. (2020). Figure 1 shows the spectra of the growth light for the four shrubs.

Fig. 1.

The growth visible light environment (spectrum) for leaves of the four shrubs. PAR in the range of 400 700 nm was 583 ± 10 μmol m⁻² s⁻¹ for P. cerasifera, 615 ± 43 μmol m⁻² s⁻¹ for P. triloba, 727 ± 17 μmol m⁻² s⁻¹ for L. vicaryi and 723 ± 6 μmol m⁻² s⁻¹ for L. quihoui.

Soil samples were collected from the top 20 cm in August 2022. After removing dead branches, leaves and debris from the surface, soil samples were taken under each shrub species. The soil from each site was randomly sampled three times, fully mixed and air-dried for 7 d. Air-dried soil was sieved with 0.25-mm screens and stored in sealed plastic jars for analysis. The contents of total nitrogen (TN) were measured with an elemental analyser (Vario EL III CHNOS Elemental, AnalyzerElementar Analysensysteme, Germany) (Dong et al., 2016). Total phosphorus (TP) and total potassium (TK) were determined by the method of inductively coupled plasma optical emission spectrometry (ICP-OES) (iCAP 6300 ICP-OES Spectrometer, Thermo Scientific, USA) (Schnell et al., 2012). Each measurement had at least three replicates.

Measurements of specific leaf area

Specific leaf area (SLA, m² kg⁻¹) was calculated using the following formula: SLA = LA/DW, where LA is leaf area (m²) and DW is dry weight (kg). Leaf area was determined with a CanoScan LiDE 300 laser area meter, and leaf thickness (cm) was measured with a Vernier calliper. Then, leaves were dried in an oven at 80°C for ~72 h to constant weight, and leaf DW was recorded using an electronic balance with 0.1 mg accuracy. Leaf density (LD, g cm⁻³) is leaf dry mass per unit leaf volume, calculated as LD = 1/(SLA/leaf thickness) (Poorter et al., 2018).

Measurements of absorptance spectra of leaves

Absorptance spectra of leaves were measured with the aid of a fibre spectroradiometer (AvaSpec-ULS2048 × 64, Avantes, The Netherlands) attached to an integrating sphere, using a barium sulphate block as the reference, and a bromine tungsten lamp (WC-150, Wence Photoelectricity Technology, Shanghai, China) as the light source. Before making these measurements, leaves were washed with distilled water and blotted gently with a paper towel. Leaf reflectance and transmittance are defined as the proportion of incident light that is reflected from and transmitted through the leaf, respectively. Absorptance was calculated as absorptance = 1 − reflectance − transmittance. Absorptance spectra were determined from 400 to 800 nm at a scanning interval of 0.536 nm. The integrated time of the AvaSpec spectroradiometer was set to 10 ms and measured in the irradiance mode (Davis et al., 2011). Figure 2 shows the absorptance spectra of the four shrubs.

Fig. 2.

(A) Leaf absorptance spectra of P. cerasifera (anthocyanic leaves) and P. triloba (green leaves). (B) Leaf absorptance spectra of L. vicaryi (yellow leaves) and L. quihoui (green leaves). Average absorptance (400–700 nm) was 0.90 ± 0.02 for P. cerasifera leaves, 0.84 ± 0.003 for P. triloba leaves, 0.34 ± 0.05 for L. vicaryi leaves and 0.91 ± 0.01 for L. quihoui leaves. Mean ± standard error (n ≥ 3) are shown. Shadow areas show standard error.

Measurements of chlorophyll, anthocyanins and other pigments

Leaves were frozen in a mortar by adding a small volume of liquid nitrogen. Pigment extraction was performed in dim light by grinding the frozen samples in the mortar with 100 % purified acetone in the presence of a small amount of CaCO₃. The extract was centrifuged at 5000 g for 10 min at 2°C and the supernatant was further cleared by passing through a 0.45-μm filter. Carotenoids were measured spectrophotometrically, using a UV-1800PC spectrophotometer (AOE Instruments, Shanghai, China) (Ritchie, 2008). The absorbance of the extracts at 470, 474 and 485 nm was determined. Carotenoid, carotene and phylloxanthin contents were assayed according to Attaran Dowom et al. (2022).

Chlorophyll content was measured according to Porra et al. (1989). Leaf discs (~0.25 g per sample, three replicates, fresh weight) were cut from sampled leaves and ground in quartz sand. Ground material was resuspended in 50 mL of 80 % (v/v) acetone and centrifuged at 1776 g for 10 min. The supernatant was transferred to a spectrophotometer (UV-2550, Shimadzu Corporation, Kyoto, Japan) to measure the absorbance at 646.6 and 663.6 nm for calculation of chlorophyll a and b contents.

For determination of anthocyanins, 95 % ethanol leaf extracts were acidified with 1.5 % HCl and the absorbance at 525 nm was measured after extraction in the dark for 24 h. The anthocyanin content was calculated according to the method of Lee and Wicker (1991). For determination of flavonoids, leaves were dried at 70°C for 24 h until constant weight and then crushed. Crushed leaf powder (1 g) was mixed with anhydrous methanol solution in a ratio of 1 g powder to 6 mL methanol, and sonicated at 60°C in a 240-W ultrasonic bath (KQ-300DE, Kunshan Ultrasonic Instruments, China). Following extraction for 40 min, the suspension was centrifuged at 10 000 g for 5 min. The colour development reaction was performed by adding 0.2 mL of 5 % NaNO₂ solution, mixing well and leaving for 5 min. Then 10 % Al(NO₃)₃ solution (0.2 mL) was added, mixed well and left for 5 min. Finally, 4 % NaOH solution (1 mL) was added, diluted with methanol to a constant volume of 25 mL, and the absorbance at 352 nm was measured (Nabavi et al., 2008). The total flavonoid content was expressed as rutin (C₂₇H₃₀O₁₆) equivalent. To make a rutin standard curve, 0, 0.08, 0.2, 0.4, 0.6 or 0.8 mg rutin was added to 0.2 mL of 5 % NaNO₂, 0.2 mL of 10 % Al(NO₃)₃ and 1 mL of 4 % NaOH, and diluted with methanol to a constant volume of 25 mL. The absorbance at 352 nm for each solution was measured to establish a standard curve.

Measurements of apparent quantum yield, maximum net photosynthesis rate, dark respiration rate and stomatal conductance of leaves

Apparent quantum yield, maximum net photosynthesis rate and stomatal conductance were measured in situ with an infrared gas analyser (Li-6400; Li-Cor, Lincoln, NE, USA) maintaining a CO₂ partial pressure of 400 μmol mol⁻¹. Leaf temperature was kept at 27°C and the relative humidity in the leaf cuvette was 50–60 % during measurements. The Li-6400 default red–blue light source (100 % red) was used.

The apparent quantum yield [AQY, mol CO₂ (mol photons)⁻¹] (Maxwell et al., 1994) was measured as the slope of the linear relationship between photosynthesis rates and low irradiances (30, 60, 90, 120 μmol m⁻² s⁻¹ or 30, 60, 90, 120, 150 μmol m⁻² s⁻¹ in red light). Dark respiration rate (R_dark, μmol CO₂ m⁻² s⁻¹) was the gas exchange rate at 0 μmol m⁻² s⁻¹. After measurements of quantum yield and dark respiration, leaves were exposed to 1200 μmol m⁻² s⁻¹ for 10 min, before maximum net photosynthesis rate (P_n,max, μmol CO₂ m⁻² s⁻¹) and stomatal conductance (g_s, mol m⁻² s⁻¹) were measured. All measurements were conducted from 0900 to 1200 h.

PSII functionality determined by redox kinetics of P700

The functional fraction of PSII was determined by the method of Kou et al. (2012). The flash-induced redox kinetics of P700⁺, the oxidized special chlorophyll pair in PSI, was measured. A dual wavelength (810/870 nm) unit (ED-P700DW) attached to a Chl a fluorometer (PAM 101, Walz, Effeltrich, Germany) was used. To obtain P700⁺ redox changes due to a brief flash superimposed on continuous far-red light, a steady state was induced by illumination with far-red light (50 μmol photons m^–2 s^–1, peak wavelength 729 nm, 102-FR, Walz, Effeltrich, Germany). Then a single-turnover, saturating xenon flash (XST 103, full width at half height = 6 μs) was applied to the adaxial side of the leaf disc. The start of data acquisition (time constant = 95 μs), the triggering of the flash, and the repetition rate were controlled by a pulse/delay generator (Model 8224, Quantum Composers, USA). The analogue output of the fluorometer was digitized and then acquired using a program (Chow and Hope, 2004). Flashes were given at 0.2 Hz; four to six consecutive signals were averaged automatically. The maximum signal immediately after the flash, representing the total amount of photo-oxidizable P700⁺, was used to normalize the trace. After the flash, electrons were delivered from PSII to PSI, tending to reduce P700⁺, while the background far-red light brought the P700⁺ concentration back to the steady state in far-red light. The area between the dipping curve and the horizontal line representing the steady-state [P700⁺] in continuous far-red light is here called the P700⁺ kinetics area; it is a simple empirical measure of the electrons delivered by PSII, directly proportional to the oxygen yield per single-turnover flash in non-photorespiratory conditions (Kou et al., 2012; Hu et al., 2013) and hence represents relative functional PSII content. The P700⁺ kinetics area was measured ~1 min after the end of high-light pretreatment to diminish the effect of energy-dependent quenching of excitation energy while minimizing any substantial recovery in prolonged darkness (Hu et al., 2013). Infiltration with methyl viologen, which is known to intercept electrons on the acceptor side of PSI, did not reduce the P700⁺ kinetics area (Jia et al., 2014), suggesting that, under the conditions of measurement, no electrons would return from the acceptor side of around PSI in a cyclic electron flux. One advantage of using the P700⁺ kinetics area is that it is a whole-tissue measurement, unlike chlorophyll fluorescence, which is detected from an unspecified depth in the leaf tissue that varies according to the extent of photoinactivation (Oguchi et al., 2011). Another advantage is that the measurement is fast and non-intrusive; therefore, any recovery of photoactivated PSII in prolonged darkness can be avoided. Figure 3 shows the decrease in the P700⁺ kinetics area with progressive photoinactivation of the PSII population.

Fig. 3.

Time course of redox changes in the P700⁺ signal in P. cerasifera leaves when a single-turnover, saturating flash was superimposed on continuous background far-red light. Weak continuous far-red light resulted in the photo-oxidation of ~95 % of the total photo-oxidizable P700⁺. The single-turnover saturating flash photo-oxidized the remainder of the P700, giving the spike (set to 1.0). Subsequent to the flash, electrons arrived from PSII to P700, but the background far-red light brought the [P700⁺] back to the steady-state level. Each signal trace is an average of four scans. The P700⁺ kinetics area, bounded by a signal trace and the horizontal line corresponding to the steady-state oxidation level of P700, decreased as the photoinhibition pretreatment time increased from 0 to 2 h using light at 660 nm and irradiance 800 μmol m⁻² s⁻¹.

The P700⁺ kinetics area can also indicate the ratio of PSII to PSI reaction centres on an arbitrary scale (Albrecht-Borth et al., 2013). The PSII/PSI ratio on this arbitrary scale is the ratio of the absolute P700⁺ kinetics area to the absolute P700⁺_max signal. The absolute P700⁺ kinetics area was obtained by integration of the area between the dipping curve and the horizontal line representing the steady-state [P700⁺] in continuous far-red light, with an integral time interval of 1 ms.

Photoinhibition treatments

Leaf discs (1.33 cm²) were first floated on a 3-mm lincomycin solution overnight in darkness to allow sufficient uptake of the inhibitor of chloroplast-encoded protein synthesis by the 70S ribosome. Leaf discs were floated with the abaxial side facing air and the adaxial side in contact with the lincomycin solution in a clear Petri dish; photoinhibitory light was applied vertically up onto the adaxial side. Illumination at 800 μmol m^–2 s ^–1 using red LEDs (660 nm, with half peak width = 10 nm) and blue LEDs 600 μmol m^–2 s ^–1 (450 nm, with half peak width = 10 nm) were applied for up to 2.5 or 4 h to obtain the time course of photoinactivation of PSII, which yielded the rate coefficient of photoinactivation k_i. The function y = exp(−k_i × t) was applied to obtain k_i, where y is the relative functional PSII and t is the photoinhibition time (h); k_i was obtained by curve fitting with the software Origin 7 (Microcal Software, Northhampton, MA, USA), allowing k_i to vary from an initial estimate until a stable value was obtained after several iterations. The maximum P700⁺ signals before and after photoinhibition treatment were also recorded to evaluate whether PSI was photoinhibited (Hu et al., 2013; Oguchi et al., 2021). During exposure to light, the leaf discs were kept at 25°C in an air-conditioned laboratory, using a fan to increase boundary layer conductance (Miyata et al., 2015).

Measurement of maximum quantum efficiency of PSII photochemistry (F_v/F_m)

The F_v/F_m ratio was measured by a chlorophyll fluorometer (Maxi-Imaging-PAM). Imagingwin 2.40b software (camera: Kappa DX4-285) was used to analyse the fluorescence image data. Before measurement, the sampled leaves were kept in the dark for 30 min. The fluorescence F_o, which represents minimized fluorescence yield when PSII primary electron acceptors (Q_A) are open, was evaluated by using a modulated light with a low intensity (1 μmol m⁻² s⁻¹). The maximal fluorescence yield, F_m, was induced by a short, saturating pulse of white light (5000 μmol m⁻² s⁻¹, 0.8 s duration), when PSII traps are closed. We calculated F_v/F_m using the equation: F_v/F_m = (F_m − F_o)/F_m (Baker, 2008). Since the actinic light was applied by an array of blue LEDs, other fluorescence parameters [e.g. the PSII photochemical yield in the light-adapted state, Y(II)] cannot be directly compared between anthocyanic leaves and green leaves under the same incident irradiance, because it is not known how much irradiance has been absorbed by anthocyanins (see Discussion and Supplementary Data Fig. S2).

Statistical analysis

Differences in the various measured parameters between the four shrub species in this paper were analysed with one-way ANOVA tests by using the SPSS 11.0 statistical package (SPSS, Chicago, IL, USA). The observed levels of significance (P) are included in all tables. Where necessary, data were transformed to meet the assumptions of ANOVA.

RESULTS

The investigated individuals of the four shrub species had no significant difference in soil nutrients (Table 1). The sampled leaves were at the exterior of the canopy [the ratio of red light irradiance (655–665 nm) to far-red light irradiance (725–735 nm), calculated based on Turnbull and Yates (1993) (R/FR ratio) was >1.15; Smith, 1994]. The exposed anthocyanic leaves (P. cerasifera) and green leaves (P. triloba) had similar growth irradiance (Fig. 1). The exposed yellow leaves (L. vicaryi) and green leaves (L. quihoui) also had similar growth irradiance, but slightly higher than P. cerasifera and P. triloba. The maximum quantum efficiency of PSII, F_v/F_m, was not different among P. cerasifera, P. triloba and L. quihoui, and ranged between 0.83 and 0.86 (Table 1), which means no stress due to light was experienced by the plants (Baker, 2008). However, yellow leaves (L. vicaryi) had a F_v/F_m value as low as 0.34, suggesting dysfunction of PSII photochemistry.

Table 1.

Mean and standard error of PAR (400–700 nm), R/FR ratio, soil TN, soil TP, soil TK, soil pH and F_v/F_m of four shrub species

Variable	P. cerasifera	P. triloba	L. vicaryi	L. quihoui
PAR (400–700 nm)	582.7 ± 8.5b	614.6 ± 35.2b	727.2 ± 14.2a	723.3 ± 5.3a
R/FR ratio	1.46 ± 0.01a	1.45 ± 0.17a	1.44 ± 0.01a	1.44 ± 0.01a
Soil TN (g kg−1)	1.71 ± 0.04a	1.60 ± 0.07a	1.69 ± 0.10a	1.60 ± 0.06a
Soil TP (g kg−1)	0.75 ± 0.00a	0.73 ± 0.01a	0.77 ± 0.02a	0.75 ± 0.01
Soil TK (g kg−1)	19.22 ± 0.11ab	18.71 ± 0.06b	20.08 ± 0.65a	19.53 ± 0.41ab
Soil pH	7.15 ± 0.04a	7.37 ± 0.07a	7.30 ± 0.02a	7.13 ± 0.05a
F v/Fm	0.86 ± 0.01a	0.83 ± 0.02a	0.37 ± 0.06b	0.84 ± 0.01a

Values labelled with different letters are significantly different from each other (P < 0.05).

The content of anthocyanins in red leaves (P. cerasifera) was >13 times greater than those in yellow and green leaves (P < 0.001; Table 2). Contents of Chl a on a leaf area basis were not different between anthocyanic leaves (P. cerasifera) and green leaves (P. triloba) (P > 0.05), while the Chl b content of P. triloba was lower than that of P. cerasifera (P < 0.05), leading to a higher Chl a/b ratio of green leaves of P. triloba. Yellow leaves of L. vicaryi had very low Chl a and Chl b contents (P < 0.001), compared with the other three types of leaves. Contents of flavonoids, carotenoid, carotene and phylloxanthin (per unit leaf fresh weight) were not significantly different between the two Prunus species (P > 0.05). By contrast, yellow leaves of L. vicaryi had very low contents of these pigments, except flavonoids, which showed no difference from green leaves of L. quihoui (P > 0.05).

Table 2.

Mean and standard error of Chl a, Chl b, Chl a/b, anthocyanins, flavonoids, carotenoids, carotene and phylloxanthin contents of four shrub species

Variable	P. cerasifera	P. triloba	L. vicaryi	L. quihoui
Chl a (μmol m−2)	231.6 ± 15.5b	232.4 ± 15.5b	6.76 ± 3.37c	335.6 ± 7.2a
Chl b (μmol m−2)	75.88 ± 2.48b	67.34 ± 6.03c	2.50 ± 1.04d	114.17 ± 2.30a
Chl a/b	3.04 ± 0.10b	3.45 ± 0.28a	2.99 ± 0.30b	2.94 ± 0.003b
Anthocyanins (mg g−1 FW)	3.39 ± 0.18a	0.26 ± 0.02b	0.03 ± 0.01b	0.17 ± 0.01b
Flavonoids (mg g−1 DW)	1.45 ± 0.60a	0.97 ± 0.13ab	0.32 ± 0.12b	0.40 ± 0.06ab
Carotenoids (mg g−1 FW)	0.50 ± 0.01a	0.55 ± 0.03a	0.02 ± 0.01c	0.32 ± 0.004b
Carotene (mg g−1 FW)	0.33 ± 0.01a	0.37 ± 0.02a	0.02 ± 0.01c	0.17 ± 0.01b
Phylloxanthin (mg g−1 FW)	0.16 ± 0.02ab	0.18 ± 0.01a	0.01 ± 0.001c	0.14 ± 0.003b

Values labelled with different letters are significantly different from each other (P < 0.05).

FW, fresh weight; DW, dry weight.

The mean leaf absorptance spectra of the four studied species are shown in Fig. 2. Leaf absorbance at 680 nm correlated with chlorophyll content across the four studied species (Supplementary Data Fig. S3). Owing to very high anthocyanin content, red leaves (P. cerasifera) absorbed more light of blue to orange wavelengths (400–610 nm) than green leaves (P. triloba), especially in the green light region (500–600 nm), consistent with previous studies (Feild et al., 2001). It is well known that the maximum absorption wavelength of cyanidian-3-glucoside, which is the most common anthocyanin species in red leaves, is ~520–540 nm (Feild et al., 2001; Mejía-Giraldo et al., 2016; Gitelson et al., 2017). For yellow leaves (L. vicaryi), due to the low content of chlorophyll, anthocyanins, carotenoid, carotene and phylloxanthin, much less visible light (400–700 nm) was absorbed compared with green leaves (L. quihoui). The absorption in the blue light region (400–500 nm) for yellow leaves was mainly due to flavonoids, which show a maximum specific absorption coefficient at 410 nm in the visible range in situ (Gitelson et al., 2017).

Respiration rate in the dark (R_dark) of P. cerasifera was slightly lower than that of P. triloba, but the difference was not significant (Table 3). As the irradiance increased in the light-limiting range, the net photosynthetic rate P_n increased linearly (Fig. 4). The slope of each straight line gives the AQY of carbon assimilation, in units of mol CO₂ (mol incident photons)⁻¹. The AQY was not different between the two Prunus species (P > 0.05) (Table 3). The light compensation point (LCP) of P. cerasifera was higher than that of P. triloba (Table 3), yet the difference was not statistically significant due to the high variation of the results. The parameters P_n,_max, g_s and PSII/PSI ratio (on an arbitrary scale) of anthocyanic leaves (P. cerasifera) were significantly lower than those of green leaves (P. triloba) (all P < 0.01; Table 3). As for the yellow leaves (L. vicaryi), the values of all photosynthetic parameters were the lowest among the four studied species, while its LCP was the highest.

Table 3.

Mean and standard error of R_dark, AQY, LCP, P_n,max, PSII/PSI ratio, g_s and SLA, leaf thickness and leaf density of four shrub species

Variable	P. cerasifera	P. triloba	L. vicaryi	L. quihoui
R dark (μmol−2 s−1)	−1.47 ± 0.43ab	−1.49 ± 0.06ab	−0.75 ± 0.10a	−1.62 ± 0.13b
AQY mol CO2 (mol photons)−1	0.035 ± 0.01a	0.032 ± 0.01a	0.0089 ± 0.01b	0.036 ± 0.01a
LCP (μmol m−2 s −1)	16.05 ± 2.28b	11.80 ± 0.57b	51.33 ± 10.56a	18.02 ± 6.80b
P n,max (μmol CO2 m−2 s−1)	6.91 ± 0.77b	11.29 ± 1.19a	1.90 ± 0.67c	10.92 ± 0.47a
PSII/PSI ratio (arbitrary scale)	24.52 ± 2.38b	36.07 ± 2.40a	---	38.73 ± 1.00a
g s (mol m−2 s−1)	0.09 ± 0.01ab	0.13 ± 0.02a	0.04 ± 0.01a	0.08 ± 0.01b
SLA (m2 kg−1 DW)	14.55 ± 0.46b	19.10 ± 3.19a	13.22 ± 0.56b	9.83 ± 1.23 c
Leaf thickness (cm)	0.014 ± 0.001c	0.013 ± 0.001c	0.017 ± 0.001b	0.026 ± 0.001a
Leaf density (g DW cm−3)	0.49 ± 0.01a	0.44 ± 0.06a	0.45 ± 0.01a	0.40 ± 0.04a

R _dark, leaf respiration rate in the dark; g_s, stomatal conductance; AQY, apparent CO₂ quantum yield, measured under red light (660 nm, 800 μmol m⁻² s⁻¹); PSII/PSI ratio, PSII content/PSI content ratio on an arbitrary scale, based on the P700⁺ kinetics area method (Albrecht-Borth et al., 2013); P_n,max, maximum net photosynthetic rate under 1200 μmol m⁻² s⁻¹ light; –, could not be measured due to very low signal.

Fig. 4.

Net photosynthetic rate P_n as a function of low irradiance of leaves of the four shrubs. Data are mean ± standard error (n ≥ 3). P. cerasifera (anthocyanic leaves), P. triloba (green leaves), L. vicaryi (yellow leaves) and L. quihoui leaves (green leaves). Measurement of P_n was under red light (660 nm, 800 μmol m⁻² s⁻¹), n ≥ 3.

Specific leaf area ranged from 9.83 ± 1.23 (L. quihoui) to 19.10 ± 3.19 m² kg⁻¹ (P. triloba) (Table 3). However, the SLA of anthocyanic leaves (P. cerasifera) was lower than that of green leaves (P. triloba) (P < 0.05), attributed to both higher leaf thickness and higher leaf density (Table 3). Yellow leaves (L. vicaryi) had higher SLA than green leaves (L. quihoui), mainly due to lower leaf thickness, not leaf density.

The P700⁺ kinetics area is a rapid, empirical, whole-tissue and non-intrusive measurement of the fraction of functional PSII (Kou et al., 2012; Fan et al., 2016). As the PSII population became progressively photoinactivated in the presence of lincomycin (Mattila and Tyystjärvi, 2023), the P700⁺ kinetics area decreased (Fig. 3). However, this method requires that the maximum photo-oxidizable P700⁺ signal does not decline after high-light treatment. As shown in Table 4, indeed, the relative value of P700⁺ maximum signal after 2 h of photoinhibition treatment by either red light or blue light did not decrease significantly for the two Prunus species and L. quihoui, indicating no photodamage to PSI at favourable temperatures.

Table 4.

Mean and standard error of the mean for k_i (red) and k_i (blue) of four shrub species

Variable	P. cerasifera	P. triloba	L. vicaryi	L. quihoui
k i (red) (h−1)	0.51 ± 0.02a	0.28 ± 0.01b	–	0.32 ± 0.04b
k i (blue) (h−1)	0.23 ± 0.02b	0.28 ± 0.03a	–	0.20 ± 0.01b
Relative P700+ maximum signal (red)	0.98 ± 0.04a	1.00 ± 0.01a	–	0.93 ± 0.04a
Relative P700+ maximum signal (blue)	1.00 ± 0.04a	0.98 ± 0.02a	–	0.99 ± 0.02a

Values labelled with different letters are significantly different from each other (P < 0.05).

k _i (red), rate coefficient of photoinactivation under red light (660 nm, 800 μmol m⁻² s⁻¹); k_i (blue), rate coefficient of photoinactivation under blue light (450 nm, 600 μmol m⁻² s⁻¹).

Relative value of P700⁺ maximum signal is the ratio of the P700⁺ maximum value after 2 h of photoinhibition to that before photoinhibition treatment.

–, could not be measured due to very low signal.

We investigated the time course of photoinactivation of PSII in the absence of repair by floating leaf discs (which had been allowed to take up lincomycin overnight) with the adaxial side in contact with a lincomycin solution (allowing stomata on the abaxial side to exchange gas freely), with illumination from the adaxial side by photoinhibitory light. Figure 5 shows that the time course of net photoinactivation of PSII in the presence of lincomycin followed first-order kinetics (Tyystjärvi and Aro, 1996), yielding the rate coefficient of photoinactivation k_i. The k_i of anthocyanic leaves (P. cerasifera) was 1.8 times higher than that of green leaves (P. triloba) under red light (P < 0.01) (Table 4). On the contrary, the k_i of anthocyanic leaves under blue light was 18 % lower than that of green leaves (P < 0.05). The P700⁺ signal of yellow leaves (L. vicaryi) was very low, hence neither the P700⁺ kinetics area value nor k_i could be deduced.

Fig. 5.

Time course of PSII photoinactivation in leaves of three shrubs. Illumination was performed using a red LED (660 nm, half-peak width = 10 nm) at 800 μmol m⁻² s⁻¹ for up to 2.5 h (A) and a blue LED 600 μmol m⁻² s⁻¹ (450 nm, half-peak width = 10 nm) at 600 μmol m⁻² s⁻¹ for up to 4 h (B). Prunus cerasifera leaves (anthocyanic leaves), P. triloba leaves (green leaves) and L. quihoui leaves (green leaves). Leaf discs were pre-floated on a 3-mm lincomycin solution in darkness overnight. Illumination was directed at the adaxial side, which was in contact with lincomycin solution. Curve fitting to yield k_i of the three species. y = exp(−k_i × t). n ≥ 3.

DISCUSSION

Exposed anthocyanic P. cerasifera leaves are special shade leaves with high resistance to blue light but low resistance to red light

Under similar growth conditions (with little or no difference in soil nutrients, light quantity and quality; Table 1), exposed anthocyanic and mature leaves of P. cerasifera showed some shade-acclimated properties (Table 3), including lower Chl a/b (Table 2, indicating larger antenna size; Björkman, 1981), lower P_n,max, lower g_s and lower PSII/PSI ratio (on an arbitrary scale), compared with green leaves of P. triloba. These findings are consistent with previous reports (Manetas et al., 2003; Kyparissis et al., 2007; Tattini et al., 2014; Gould et al., 2018; Landi et al., 2021). Specifically, it was found that exposed anthocyanic leaves of P. cerasifera had a lower PSII/PSI ratio than green leaves of its congeneric species. Many studies have shown that under conditions of the same irradiance spectrum, low-light-grown leaves have fewer PSII units with a larger light-harvesting antenna, relative to PSI, while high-light-grown leaves have more PSII units each with a smaller effective light-harvesting antenna size, relative to PSI (Anderson et al., 1988; Melis, 1991; Fan et al., 2007). Such light-regulation of the stoichiometry of the two photosystems aims to optimize the absorbed light energy partitioning between the two photosystems, leading to a maximum photochemical quantum yield under growth irradiance (Chow et al., 1990). This shade syndrome is linked to blue-light-mediated upregulation of relevant genes (Landi et al., 2021). On the other hand, Zeliou et al. (2022) reported that anthocyanic leaves of seven Mediterranean shrubs had higher PSII/PSI ratios (estimated by 77K fluorescence, as excited by a 490-nm light) with lower Chl a/b ratios, compared with green leaves. However, whereas our method gives the ratio of PSII to PSI reaction centres (albeit on an arbitrary scale), the results obtained by Zeliou et al. (2022) depend not only on the contents of reaction centres, but also on the antenna size of each photosystem. For example, if PSII complexes possess a large antenna due to the low-light growth environment, the Chl fluorescence emission at 685 and 695 nm would be high even if the PSII reaction centre content is low; consequently, the ratio of F685 (emitted by PSII) to F735 (emitted by PSI) would be high.

Since many reports have shown that shade-green leaves have a higher rate coefficient of PSII photoinactivation than that of sun ones (e.g. Aro et al., 1993; Flexas et al., 2001; Fan et al., 2016), we supposed that shade-like anthocyanic leaves should be more susceptible to photoinactivation of PSII than green leaves, provided the photoinhibitory light colour is red in order to bypass attenuation by anthocyanins (Feild et al., 2001), and provided the absorption of red light by photosynthetic pigments is not much different between the red and green leaves (Fig. 2). As expected, in the presence of lincomycin (which blocks the repair process), anthocyanic leaves showed 1.8 times greater k_i than green leaves. According to the excess energy hypothesis (Zavafer et al., 2015), such higher PSII susceptibility of anthocyanic (shade-like) leaves is probably linked to larger antenna size (indicated by a lower Chl a/b ratio Table 2) (Park et al., 1995), higher unregulated constitutive excitation dissipation as heat in reaction centres (Miyata et al., 2015), lower light-regulated excitation dissipation as heat in the antenna (Ruban and Wilson, 2021), lower photochemical utilization of energy (P_n,max; Table 3) (Melis, 1999), and lower sink strength (Adams et al., 2013).

On the other hand, considering that blue light can be absorbed by anthocyanins, it is expected that the difference in k_i based on incident blue irradiance between anthocyanic and green leaves should be less than that based on red light according to the excess energy hypothesis. Indeed, the k_i of anthocyanic leaves under blue light was significantly lower (−18 %) than that of green leaves (Table 4), supporting the photoprotection hypothesis of anthocyanins as an attenuator of light. Further, such higher PSII sensitivity to red light and lower sensitivity to blue light can partially explain the existing contradictory reports on the PSII photoprotection by anthocyanins because previous studies applied photoinhibitory light sources with different spectra. Besides, different growth light spectrums, no separation of the resistance and repair processes, and using F_v/F_m as the functional PSII content during photoinhibition treatment can also lead to such ambiguity. Therefore, our results extend the warning by Gould et al. (2018) that the responses of red and green leaves to light stress need to be evaluated in the context of not only the spectrum used, but also other methodological factors.

We noticed that anthocyanic leaves also showed some sun-acclimated properties, which include little or no difference in total chlorophyll, carotenoid, carotene and phylloxanthin contents, no difference in dark respiration rate and light compensation point between the two types of exposed leaves. Further, SLA of anthocyanic leaves was significantly lower than that of green leaves of P. triloba. Specific leaf area is a critical functional trait representing leaf light interception capacity, shade leaves tending to increase their SLA to intercept incident irradiance relative to sun leaves within (Evans and Poorter, 2001) or across species (Hoffmann and Franco, 2003) (Table 3). Specific leaf area contains two components, namely leaf thickness and leaf density (Poorter, 2009); both contributed to the lower SLA in the present study. This result is consistent with some studies (Steyn et al., 2002; Close et al., 2004; Cirillo et al., 2021), but not others (Manetas et al., 2003), possibly due to different leaf ontogenies in the latter case. Under similar growth irradiance, a decrease of SLA is a recognized strategy for resistance to environmental stresses such as drought (Wright and Westoby, 2001), low temperature (Close et al., 2004), low nutrients (Turner, 1994) or low palatability (Walters and Reich, 1999). This may come about by arresting cell expansion through reduced turgor, or accumulating osmotic substances, palatability-reducing compounds and ROS-scavenging substances, as a consequence of more limited sink strength under environmental stresses (Poorter, 2009). It is worth noting that, although the selection of green leaves of P. triloba as the control seems reasonable because many traits were not different between anthocyanic leaves of P. cerasifera and green leaves of P. triloba, the use of genetic lines with preferably near-isogenic mutants or transgenics would be a better choice for such comparisons (Gould et al., 2018).

A closer examination of anthocyanins as an attenuator of light in photoprotection

We estimated the efficiency of light attenuation of anthocyanins on PSII photoprotection at 450 nm. The k_i of an anthocyanic leaf (P. cerasifera) was 0.51 h⁻¹, while the k_i of a green leaf (P. triloba) was 0.28 h⁻¹ in red light (not absorbed by anthocyanins) (Table 4). In blue light, assuming anthocyanins have no effect on PSII photoprotection, the k_i of an anthocyanic leaf (P. cerasifera) should be close to 0.51 h⁻¹, because the blue light-induced k_i of green leaf (P. triloba) was the same (0.28 h⁻¹) as that of red light-induced k_i [assuming the low content of anthocyanins (0.26 mg/g) in the green leaf has negligible impact on k_i]. In fact, the ratio of blue light-induced k_i to red light-induced k_i at the same irradiance is not much different among Arabidopsis NPQ mutant strains (Fig. 2B in Sarvikas et al., 2006). A much lower k_i (0.23 h⁻¹) in P. cerasifera means that the presence of anthocyanins could protect PSII better by ~54 % at 450 nm, which can be the combined effect of light attenuation and other functions by anthocyanins. The light-attenuating effect could be roughly estimated by the relative absorbance of anthocyanins to chlorophyll based on their contents and corresponding extinction coefficients (for simplicity, the sieve effect and detour effect were not considered), and can be compared with the relative change of k_i, as overwhelming experimental evidence by independent groups has shown that k_i is directly proportional to absorbed visible irradiance (Allakhverdiev and Murata, 2004; Tyystjärvi and Aro, 1996; Lee et al., 2001; He and Chow, 2003; Kato et al., 2003; Tyystjärvi, 2013) or irradiance with a specific wavelength (Sarvikas et al., 2006) for green leaves. The extinction coefficient of cyanidin-3-glucoside is 38 000 L g⁻¹ cm⁻¹ at 525 nm (Feild et al., 2001); thus, its extinction coefficient at 450 nm is ~12 920 L g⁻¹ cm⁻¹ (Longo and Vasapollo, 2006). The extinction coefficient of chlorophyll based on measured leaf optical spectra (Supplementary Data Fig. S3) can be extrapolated to ~48 200 L g⁻¹ cm⁻¹ at 450 nm, according to the absorbance spectrum of chlorophyll in ethanol solution (Fig. 3 in Pápista et al., 2002) at 450 nm. Therefore, the ratio of absorbance of anthocyanins to chlorophyll in red leaves (P. cerasifera) was 3.39 × 12 920/(1.90 × 48 200) ≈ 0.48, where 1.90 (mg g⁻¹) is the total chlorophyll content per unit fresh weight.

Next, we constructed a simplified two-layer anthocyanic leaf model, the first layer from the adaxial side being the anthocyanin layer and the layer below being the chlorophyll layer, as supported by anatomical evidence (Manetas et al., 2003; Nichelmann and Bilger, 2017). Based on the Lambert–Beer law and the measured leaf optical spectrum (Fig. 2, Supplementary Data Fig. S4), it was deduced that the total absorbance of an anthocyanic leaf (P. cerasifera) was 1.39 (A_total), composed of an anthocyanic component (A_Anc = 0.45) and a chlorophyll component (A_chl = 0.94). Suppose a unit incident light firstly penetrates the anthocyanin layer, ~63 % has been absorbed and the remaining 37 % comes out of the anthocyanin layer and reaches the chlorophyll layer. Then, another 33 % has been absorbed by the chlorophyll layer, and only 4 % of the incident light comes out of the system, which matches well with our light transmission measurement (I_t = 0.04). The calculated 63 % of light attenuated by anthocyanins was also consistent with previous determinations (Nichelmann and Bilger, 2017). Thus, an attenuation of 63 % of incident light bringing about 54 % better photoprotection of PSII suggests that, at 450 nm under the current experimental conditions, anthocyanins protect PSII mainly via the light-attenuating mechanism, not the ROS-scavenging one, consistent with the conclusion of Zheng et al. (2021). The slight difference (9 %) between 63 % attenuation and 54 % PSII photoprotection effects is probably due to the fact that we did not consider the photoprotection by the low content of anthocyanins in the green leaf, or the difference in flavonoid content between the two types of leaves. Further, the rough estimation does not consider the sieve effect and detour effect of light penetration within the leaf, although these two effects execute opposite impacts on light absorbance (Terashima et al., 2009). Actually, from Supplementary Data Fig. S3, there existed a good linear relationship between absorbance and chlorophyll content across the four studied species, suggesting that the net effects of the sieve effect and detour effect of light penetration impacts on absorbance of the four studied species are similar. Nevertheless, the two-layer anthocyanic leaf model is still a simplified model that does not consider the heterogeneous distribution of chlorophyll and anthocyanin pigments across the leaf section.

Therefore, further efforts on the action spectrum of PSII photoinactivation of anthocyanic leaves are warranted, in order to completely assess the hypothesis of photoprotection by anthocyanins. It is worth noting that the absorption spectrum of anthocyanins (Gould et al., 2018) is similar to the action spectrum of PSII photoinactivation of green leaves (Takahashi et al., 2010). Unfortunately, the yellow-peak action spectrum as measured by a prism is problematic because it did not take the optical path effect as induced by the incident light angle into account (Zavafer et al., 2015), and is very different from other reported action spectra (Jones and Kok, 1966; Jung et al., 1990; Santabarbara et al., 2001; Hakala et al., 2005; Sarvikas et al., 2006; He et al., 2015; Schreiber and Klughammer, 2013; Karim et al., 2015). Detailed studies on the repair process associated with ROS scavenging by anthocyanins are also needed. In fact, enhanced repair ability of anthocyanic leaves could be inferred from a higher recovery rate of F_v/F_m of young anthocyanic leaves of Quercus coccifera after photoinhibition treatment, compared with its mature green counterpart (Manetas et al., 2003).

For both anthocyanic and green leaves, PSI, on the contrary, remained very stable in red or blue light in the presence of lincomycin (Table 4). This result is in accordance with previous studies showing that PSI is well protected from strong irradiance at room temperature (Terashima et al., 1994; Barth et al., 2001; Tikkanen and Aro, 2014). The effective protection mechanism(s) of PSI may come from the strong quenching capacity of P700⁺ (Shubin et al., 2008), alleviation of excitation pressure of the PSI reaction centre from the photoinactivated PSII (Tikkanen and Aro, 2014), alternative electron fluxes around PSI to avoid over-reduction on the PSI acceptor side (Kono et al., 2014), and the RISE mechanism at the Cyt b₆/f site, a form of photosynthetic control in which an over-reduction of the plastoquinone pool slows both linear and cyclic electron flow (Shimakawa et al., 2018).

Exposed yellow leaves of L. vicaryi act as blue light attenuator with only a trace of photosynthetic capacity

There are many studies on the physiological functions of either anthocyanins or flavonoids during leaf senescence in autumn (for review, see Renner and Zohner, 2019; Agati et al., 2021; Mattila and Tyystjärvi, 2023). However, whether yellow leaves in their growth season function similarly to anthocyanic leaves in terms of photoprotection and photosynthetic capacity remains largely unknown (Davies et al., 2022). The yellow colour of L. vicaryi leaves, mainly due to the accumulation of flavonoids (e.g. aurones; Romani et al., 2000), showed a striking difference in leaf absorption spectrum from the other three types of leaves (Fig. 2). The leaf absorption spectrum is in good accordance with very low concentrations of photosynthetic pigments (Chl a, Chl b) and other pigments (anthocyanins, carotenoid, carotene, phylloxanthin), except flavonoids (Table 2), compared with the green leaves of L. quihoui. Such extremely low photosynthetic pigment contents led to very low P_n,max, g_s and AQY and very high LCP compared with its green leaf counterpart, consistent with the observation of defective chloroplasts in the mesophyll cells of the golden-yellow leaf of hybrid paper mulberry (Wang et al., 2021). All the results demonstrate that carbon assimilation is not the major function of yellow leaves of L. vicaryi. This raises a question about the purpose(s) of L. vicaryi keeping these yellow leaves in the growth season instead of letting them fall as during senescence in autumn.

The leaves of L. vicaryi from the canopy interior appear green (Supplementary Data Fig. S1). We measured the k_i of green leaves of L. vicaryi and L. quihoui in the canopy at a fixed distance from the exposed leaves (Supplementary Data Table S1). The total incident PAR was higher for L. vicaryi, particularly in the green waveband (500–600 nm), but not different in the blue waveband (400–500 nm) (Supplementary Data Fig. S5), matching well with the absorption spectrum of an exposed leaf (Fig. 2). However, the incident PAR at the red waveband (600–700 nm) in L. vicaryi was slightly higher than that in L. quihoui, not corresponding well to the absorption spectrum of an exposed leaf (Fig. 2), possibly due to the stronger detour effect of red light in exposed leaves compared with the sampled interior leaves. Interestingly, the values of k_i of leaves from interior of the L. vicaryi canopy under blue and red photoinhibitory lights were significantly lower than the respective values of L. quihoui. The results indicate that one possible function of preserving exposed yellow leaves of L. vicaryi in the growth season is to generate a moderately high light environment with abundant green and red light component for leaves in the canopy interior to enhance their photoprotection capacity (and possibly the photosynthesis rate), in order to compensate for the negligible carbon assimilation of exposed leaves. This also supports the physiological significance of green light in leaf photoprotection and/or photosynthesis capacity at the leaf level or canopy level, as addressed by Terashima et al. (2009). Future efforts on comparisons of pigment analysis and photosynthesis between the two types of green leaves, and/or leaf removal experiments, are warranted to evaluate the photosynthetic function zonation hypothesis between yellow exposed and green canopy interior leaves of L. vicaryi.

Conclusions

Although the effect of accumulation of anthocyanins on the photoprotection of photosystems has been studied extensively, there is a paucity of data on the PSII as well as PSI photo-resistance of exposed anthocyanic leaves in the absence of a repair process, and under light of a specific spectrum. We found that exposed anthocyanic P. cerasifera leaves are special shade leaves with high resistance to blue light but low resistance to red light. Such high PSII sensitivity to red light and low sensitivity to blue light of anthocyanic leaves relative to green leaves can partially explain the existing contradictory reports on PSII photoprotection by anthocyanins. For both anthocyanic and green leaves, on the contrary, PSI showed great resistance to both red and blue light at a favourable temperature. Further efforts on the action spectrum of PSII photoinactivation of anthocyanic leaves as well as the quantification of the repair process are warranted. We also found that exposed yellow leaves of L. vicaryi in the growth season act as a pure blue light attenuator with only a trace of photosynthetic capacity, demonstrating that carbon assimilation is not the major function of yellow leaves of L. vicaryi. We hence presented a ‘photosynthetic function zonation’ hypothesis that in the growth season the exposed yellow leaves can generate a moderately high light environment (with abundant green light and red component) for leaves in the canopy interior to enhance their photoprotection/photosynthetic capacities.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following.

Table S1: mean and standard error of PAR, R/B ratio and R/FR ratio inside the canopy and k_i and k_i of leaves of the two shrub species in the interior of the canopy. Figure S1: growth environment of four shrub species. Figure S2: light response curves of PSII photochemical yield in the light-adapted state Y(II), regulated light-induced non-photochemical quenching Y(NPQ) and constitutive non-regulated non-photochemical quenching Y(NO). Figure S3: linear equation fitted to light absorbance at 680 nm as a function of Chl a + b content of the leaves of four shrub species. Figure S4: two anatomical models for anthocyanic leaf. Figure S5: growth light environment of L. vicaryi and L. quihoui leaves at a fixed distance of 15 cm from the outside of the canopy.

FUNDING

This research was funded by the National Natural Science Foundation of China (No. 32271652, No. 32201258 to X.P.W., No. 32271832 to D.Y. F.), the Major Program for Basic Research Project of Yunnan Province (No. 202101BC070002 to X.P.W.) and an Australian Research Council grant (No. DP 0664719 to W.S.C.).