Contrasting photosynthetic, stomatal, and mesophyll mechanisms drive common reductions in leaf water-use efficiency under blue light

Abstract

Forests experience complex light environments, but the detailed roles of photosynthetic, stomatal, mesophyll, and biochemical responses to dynamic blue light remain unclear in tree species. We measured the blue-light responses of leaf gas exchange, online isotope discrimination, photorespiration, and chlorophyll fluorescence in grey alder (Alnus incana) and holm oak (Quercus ilex), and investigated the underlying biochemical and physiological mechanisms. With increasing blue light, differing photosynthetic and stomatal responses consistently led to a decrease in intrinsic water-use efficiency (iWUE) in the two species. For alder, the decline in iWUE was primarily due to a reduced photosynthesis rate (A_n); for oak, although A_n also decreased, blue-light-stimulated stomatal opening played a major role. Although the reduction in A_n was linked to blue-light-induced photoprotective processes in alder, it was coordinated with mesophyll conductance (g_m) in both species. The maximum carboxylation rate of Rubisco and g_m imposed considerable photosynthetic limitations, especially at high blue-light levels. However, the component of g_m that responded to blue light and coordinated with A_n was the chloroplast membrane in alder whereas it was the cell wall and plasma membrane in oak. Our findings highlight species-specific physiological strategies in the response to blue light and underscore the importance of considering spectral composition when assessing carbon–water trade-offs in forest trees.

Exposing alder and oak to differing blue-to-red light ratios reveals species-specific responses in conductance and photosynthetic mechanisms, which show that reductions in water-use efficiency occur through different physiological pathways.

Introduction

Plants experience frequent variations in both light quantity and quality, which together shape their diverse light environments. Within the spectrum of photosynthetically active radiation (PAR) of incoming daylight, blue and red light typically exhibit greater variability than green light under different atmospheric conditions (Supplementary Fig. S1). Consequently, the effects of light quality on plants have been extensively studied in herbaceous crop settings, where optimizing the mix of blue and red light enhances biomass production and reduces water consumption (i.e. improves water-use efficiency, WUE) (Stamford et al., 2023); however, how tree species respond to light quality has generally been understudied (Pallozzi et al., 2013). The light spectral environment in forests is inherently complex, as it is influenced by factors such as the multi-layered canopy structure, dynamic light availability and shading through canopy gaps, and the occurrence of sun flecks (Parker et al., 2019; Durand et al., 2022; Forsström et al., 2023), in contrast to the more uniform light conditions in agricultural fields (Slattery et al., 2017). These natural variations in light quality serve as important signals, helping trees to adjust to short-term changes in their light environment to optimize performance (Kochetova et al., 2022; Lauria et al., 2024). Therefore, understanding how trees respond to the dynamic interplay of blue and red light is crucial for increasing our knowledge of tree functioning and improving management strategies for forest stands.

Both blue and red light are strongly absorbed by photosynthetic pigments, but red light is the more efficient in driving leaf photosynthesis (McCree, 1972). Blue light, however, is sensed by phototropins—the photoreceptors that trigger stomatal opening—with much greater efficiency than red light (Shimazaki et al., 2007; Inoue and Kinoshita, 2017). Red light can also induce stomatal opening via photosynthesis in the chloroplasts of the mesophyll and guard cell, but this requires sustained high-intensity illumination, and the underlying mechanisms are not well understood (Baroli et al., 2008; Busch, 2014). Our current understanding of photosynthesis and stomatal responses to blue and red light is largely based on measurements using transient light pulses, such as adding a pulse of weak blue light on a background of saturating red light (Grantz and Assmann, 1991; Shimazaki et al., 2007; Doi et al., 2015; Matthews et al., 2020). However, given that the synergistic effects of these two types of light on stomatal opening might not be observed under dynamic stimulation (Assmann, 1988; Vialet-Chabrand et al., 2021), the steady-state responses of photosynthesis and stomata to continuous illumination with blue light in the presence of background red light remain unclear. While several studies have been conducted to explore steady-state leaf gas exchange responses to blue light under discrete treatments (Loreto et al., 2009; Hogewoning et al., 2010; Savvides et al., 2011; Wang et al., 2016), only a few have involved investigation of the steady-state responses across a continuous gradient of blue-to-red ratios (Zhen and Bugbee, 2020; Izzo et al., 2021; Li et al., 2023). This hinders a comprehensive and quantitative understanding of the effect of blue light on photosynthesis and WUE.

Furthermore, the existing findings on the response of leaf gas exchange to blue light are contradictory. For instance, studies on the effects of an increasing blue-light fraction under constant mixed blue- and red-light intensity have pointed to both increases (Hogewoning et al., 2010; Wang et al., 2016; Izzo et al., 2021; Li et al., 2023) and decreases (Loreto et al., 2009; Zhen and Bugbee, 2020) in net photosynthesis rate (A_n) across different herbaceous C₃ plants. However, increases in stomatal conductance (g_s) or transpiration rate (E) with an elevated blue-light fraction were consistently observed in these studies, except for Loreto et al. (2009). These findings imply that a tight coupling between A_n and g_s might not be maintained under varying blue-light conditions, complicating the interpretation of the blue-light effects on WUE.

The coordination between A_n and g_s under varying blue-light conditions in C₃ and C₄ plants was investigated by Zhen and Bugbee (2020), who found that they were tightly coupled across different blue-light levels in C₄ species but not in C₃. Nevertheless, both the C₃ and C₄ species examined showed a consistent decreasing trend in WUE with increasing blue-light fractions. This suggests that the negative effect of blue light on WUE might arise from different mechanisms regulating the blue-light response of leaf photosynthesis in different species; however, the physiological mechanisms underlying this varied coordination between A_n and g_s have not been systematically explored.

Several factors have been frequently studied to understand the physiological mechanisms behind variations in photosynthesis, but less attention has been paid to their responses to blue light. Firstly, mesophyll conductance (g_m) is widely considered an important limiting factor for photosynthesis. The g_m responses have been investigated in only a few studies, and a decrease with an increasing blue-light fraction has been reported (Loreto et al., 2009; Pallozzi et al., 2013; Momayyezi and Guy, 2017). However, under varying blue-light conditions, no studies have elucidated whether the primary CO₂ resistance occurs at the cell wall and plasma membrane or through the cytosol to sites of carboxylation across the chloroplast membrane; how these g_m components coordinate with A_n; whether the decrease in g_m leads to a reduction in CO₂ concentration within the chloroplast stroma, potentially contributing to the decline in A_n; and the extent to which g_m imposes a limitation on A_n relative to Rubisco biochemistry and g_s. Secondly, chlorophyll fluorescence parameters, such as non-photochemical fluorescence quenching (NPQ) and effective quantum yield of PSII (Φ_PSII), are key indicators of photochemical energy conversion efficiency and the light-stress status of the photosynthetic apparatus, and these parameters have been shown to respond to the blue-light fraction (Yang et al., 2018; Hamdani et al., 2019; Gao et al., 2022). Thirdly, although it has yet to be observed directly, changes in CO₂ loss due to photorespiration (R_photo) could also play a crucial role in shaping the response pattern of A_n, as blue light influences chloroplast singlet-oxygen signalling and the production of reactive oxygen species (ROS) (Kangasjärvi et al., 2012). Despite this, we still lack a clear overall picture of how these key factors that drive leaf photosynthesis respond to changes in blue light.

In this study, we aimed to elucidate the underlying mechanisms of photosynthetic and stomatal responses to blue light. We exposed leaves of saplings of grey alder (Alnus incana) and holm oak (Quercus ilex) trees to different blue-light fractions under a constant incident light intensity, with the remaining fraction supplied as red light. Under each light condition, we measured leaf gas exchange, total g_m and its components, CO₂ concentrations in the intercellular air spaces, at the chloroplast surface, and in the chloroplast stroma, as well as photorespiration and chlorophyll fluorescence under steady-state conditions. We also analysed the photosynthetic limitations imposed by Rubisco biochemistry, g_m, and g_s. We selected the two tree species because they have contrasting autecological niches and leaf morphologies (Supplementary Fig. S2). Given the ecological and morphological differences between them, we hypothesized that whilst increasing blue light would reduce the WUE of both species, the underlying mechanisms would differ, driven by contrasting photosynthetic, stomatal, and mesophyll responses to blue-light fractions.

Materials and methods

Plant material, experimental conditions, and light-quality treatments

Five 4-year-old saplings of grey alder (Alnus incana) and of holm oak (Quercus ilex) were used in this study. The saplings were grown in 4 l pots with soil mixed with commercial slow-release NPK fertilizer (Osmocote Exact Standard 3–4 M, ICL). This provided sufficient volume and nutrient supply to support healthy growth throughout the experiment: no visible signs of root restriction or stress were observed during the study period. One month prior to the experiment, the saplings were transferred to a greenhouse at WSL, Birmensdorf, Switzerland, under natural light conditions supplemented with high-pressure sodium grow lights (HS 400W, Hortilux Schréder, The Hague, The Netherlands), providing a maximum PAR of ∼1000 μmol m⁻² s⁻¹ during daylight hours. The greenhouse temperature was regulated to be ≤30 °C. During the experimental period, the mean air temperature in the greenhouse was 22 °C and the mean relative humidity (RH) was 80%, corresponding to a mean air vapour pressure deficit (VPD) of 0.5 kPa. The saplings were watered every 2 d, ensuring that the trays beneath the pots always contained liquid water, to prevent soil water limitation.

Three experiments with different measurements were conducted in a climate chamber (Conviron PGR15, Controlled Environments Limited, Manitoba, Canada), as follows: Experiment 1, online stable isotope measurements; Experiment 2, photorespiration measurements; and Experiment 3, fluorescence measurements (Fig. 1). One day before the measurements, the selected saplings were transferred from the greenhouse to the climate chamber for acclimation. The environmental conditions in the chamber were regulated at 30 °C, 80% RH of (corresponding to an air VPD of 0.8 kPa), and PAR of 1000 μmol m⁻² s⁻¹. The temperature and VPD were consistent with the values in the leaf cuvette (3010-GWK1, Walz) used for Experiments 1 and 2, where the temperature was maintained at 30.0±0.0 °C and air VPD at 0.87±0.08 kPa (means ±SD) under each light-quality treatment. Such low VPD conditions prevent air-drying stress and unsaturation of leaf intercellular air spaces (Diao et al., 2024a). The leaf temperature was 30.2±0.3 °C. The CO₂ fraction of the air surrounding the leaf in the cuvette (c_a) was maintained at 400 μmol mol⁻¹ in Experiments 1 and 2. In Experiment 3, the leaf cuvette was not used and the ambient air inside the climate chamber was monitored using a CO₂ probe (GMP343, Vaisala, Helsinki, Finland). The CO₂ fraction inside the climate chamber during Experiment 3 was ∼420 μmol mol⁻¹.

Fig. 1.

Schematic diagram of the experiments. The different blue:red light ratios and corresponding photon flux densities (PFD) of spectra are shown on the left, where 0:10 is 0%B and 10:0 is 100%B. The set-up for each of the three experiments is shown on the right. LED, light emitting diode; IRGA, infrared gas analyser; IRIS, isotope ratio infrared spectrometer; PAM, pulse-amplitude-modulation. See Table 1 for explanations of the output variables.

For all three experiments, a light-emitting diode (LED) panel (RGBW-L084, Walz) was used as the light source. While air temperature and VPD were kept stable, either in the leaf cuvette or in the climate chamber, one entire leaf for grey alder or one foliated branch for holm oak of each sapling was subjected to seven different fractions of blue light: 0%B, 10%B, 30%B, 50%B, 70%B, 90%B, and 100%B. The remaining light fraction was red, for a total a photon flux density (PFD) of 1000 μmol m⁻² s⁻¹ for each light-quality treatment (Fig. 1).

Varying the ratio of blue:red light intensity while maintaining a constant total PFD has been a common practice in previous studies (Hogewoning et al., 2010; Wang et al., 2016; Izzo et al., 2021; Li et al., 2023), and is necessary because it is impossible to change the intensity of one specific light component while keeping the other components and the total intensity constant. This methodology offers the advantage of stabilizing metabolic processes but has the drawback of skewing the activation of photosensory networks that contribute to developmental responses (Folta and Maruhnich, 2007). However, because developmental responses occur over a slow timescale of days to weeks, we opted to keep the total PFD constant in our study where we focused on leaf physiological processes over a timescale of tens of minutes. Therefore, it is worth noting that throughout the study an increasing proportion of blue-light PFD (%B) implies a reduction in the red-light fraction.

The PFD of blue and red light emitted by the LED panel was determined using a spectroradiometer (ASD FieldSpec 4; Malvern Panalytical, Malvern, UK) coupled to a remote cosine receptor (RCR; A124500, Malvern Panalytical). The surface of the RCR was positioned parallel to the LED panel surface at a distance of 3.6 cm, which was the same distance as between the leaf surface and the panel in all three experiments. To prevent interference from environmental light, the spectroradiometry measurements were conducted in darkness. First, the irradiance of the blue spectrum (peak at 455 nm) and the red spectrum (peak at 631 nm) were measured separately, with a wavelength resolution of 1 nm. Then, at each wavelength, the irradiant energy (W m⁻² nm⁻¹) was converted to quanta (μmol m⁻² s⁻¹ nm⁻¹) (Aphalo et al., 2012):

$$\mathrm{PF}{\mathrm{D}}_{\lambda }={\displaystyle \frac{I\times \lambda \times {10}^{-3}}{h\times c\times N_{\mathrm{A}}}}$$ (1)

where I is irradiance, λ is wavelength, h is Planck’s constant (6.63×10⁻³⁴ J s), c is the speed of light (2.998×10⁸ m s⁻¹), and N_A is Avogadro’s constant (6.022×10²³ mol⁻¹). Finally, the PFD of the blue or red light was calculated by the numerical integration of all wavelengths. Because there is no overlap between the blue and the red spectra, the sum of their PFDs is the PFD of the mixed blue and red lights (Fig. 1). The intensity settings for the blue and red LEDs required to achieve the seven target %B values were recorded and applied in all three experiments.

A list of the main abbreviations used throughout the text is given in Table 1.

Table 1.

List of main abbreviations used in the text

Symbol	Definition
A n	Net photosynthesis rate
%B	Fraction of photon flux density provided as blue light; the remaining fraction was provided by red light.
c a	CO2 mole fraction of the air surrounding the leaf
c c	CO2 mole fraction in the chloroplast stroma
c ca	CO2 mole fraction at the sites of carbonic anhydrase activity (chloroplast surface)
c i	CO2 mole fraction in the leaf intercellular air spaces
E	Transpiration rate
g cm	Mesophyll conductance to CO2 across the chloroplast membrane
g m13	Mesophyll conductance to CO2 estimated from 13C measurements, representing the total mesophyll conductance
g m18	Mesophyll conductance to CO2 estimated from 18O measurements, representing the cell wall and plasma membrane conductance
g s	Stomatal conductance to water vapour
iWUE	Intrinsic water-use efficiency
NPQ	Non-photochemical fluorescence quenching
PFD	Photon flux density
qP	Photochemical fluorescence quenching
R photo	Photorespiration rate
ΦPSII	Effective quantum yield of PSII
V cmax	Maximum carboxylation rate of Rubisco

Experiment 1: online stable isotope measurements

Online stable isotope measurements were conducted following the protocol described by Diao et al. (2024b). In brief, this approach is characterized by combining a leaf gas exchange system (GFS-3000, Walz) with isotope ratio infrared spectrometers for CO₂ (Delta Ray, ThermoFisher Scientific) and H₂O isotopologues (L2120-i, Picarro Inc., Santa Clara, CA, USA). The combined instruments were used to measure the air streams entering and leaving the leaf cuvette for both mole fractions and isotope compositions of CO₂ and H₂O at the same time. The carbon isotopic composition (δ¹³C) of CO₂ was expressed against Vienna Pee Dee Belemnite (VPDB), and the oxygen isotope composition (δ¹⁸O) of CO₂ and H₂O vapour were expressed against Vienna Standard Mean Ocean Water (VSMOW). A bypass humidity control system (NFRB0101, Walz) served as an additional measure to remove excess moisture in the cuvette and keep its humidity constant at the preselected value. The gas exchange system, the bypass humidity control system, and the sapling being measured were placed in the climate chamber. Under each %B level, the gas exchange and the isotope ratio data were recorded when the gas exchange rates were at steady state (∼30 min).

In addition to the leaf gas exchange parameters, Experiment 1 included estimations of g_m and CO₂ mole fractions inside the leaf mesophyll. The g_m estimated from the measurements of the ¹³C/¹²C fractionation (g_m13), representing the total g_m from the intercellular air spaces to the carboxylation sites inside the chloroplast stroma, was expressed as:

$$g_{\mathrm{m}13}={\displaystyle \frac{1+t}{1-t}\left({\displaystyle \frac{b-a_{\mathrm{m}}-{\displaystyle \frac{{\alpha }_{\mathrm{b}}}{{\alpha }_{\mathrm{e}}{\alpha }_{\mathrm{R}}}{e}^{\mathrm{\prime }}{\displaystyle \frac{R_{\mathrm{day}}}{A_{\mathrm{n}}}}}}{{\mathrm{\Delta }}^{13}{\mathrm{C}}_{\mathrm{i}}-{\mathrm{\Delta }}^{13}{\mathrm{C}}_{\mathrm{obs}}}}\right){\displaystyle \frac{A_{\mathrm{n}}}{c_{\mathrm{a}}}}}$$ (2)

where a_m is the ¹³C/¹²C fractionation for CO₂ dissolution and diffusion from the intercellular air spaces to the sites of carboxylation in the chloroplasts (1.8‰), and b is the ¹³C/¹²C fractionation associated with carboxylation (30‰). e′ is e+e*, where e is the ¹³C/¹²C fractionation during day respiration (−3‰) and e* is an additional apparent fractionation accounting for the difference in substrate composition between freshly assimilated C and old C pools (Wingate et al., 2007). e* was zero in our case because the C turnover was assumed to be sufficiently effective at replacing the active respiratory pool with assimilates formed while the leaf was in the cuvette (Diao et al., 2024b). α_b and α_e are defined as 1+b and 1+e, respectively. R_day is the day respiration rate and is assumed to be half of the dark respiration rate (R_dark) (Adnew et al., 2020). The R_dark measurements for each replicate sapling were conducted under the same environmental conditions as the light measurements, with the leaf gas exchange system in complete darkness. α_R is 1+(R_day/A_n)(e/α_e) (Busch et al., 2020). The ternary correction term, t, was calculated according to Farquhar and Cernusak (2012). Δ¹³C_obs is the observed discrimination against ¹³C during CO₂ assimilation, which was calculated as presented by Evans et al. (1986). Δ¹³C_i is the modelled C isotope discrimination, which was calculated according to Busch et al. (2020). The CO₂ mole fraction inside the chloroplast stroma (c_c) was then calculated as:

$$c_{\mathrm{c}}=c_{\mathrm{i}}-{\displaystyle \frac{A_{\mathrm{n}}}{g_{\mathrm{m}13}}}$$ (3)

The CO₂ mole fraction at the sites of carbonic anhydrase activity (c_ca), which is thought to be mainly present on the chloroplast surface (Gillon and Yakir, 2000), was derived from δ¹⁸O measurements according to Barbour et al. (2016):

$$c_{\mathrm{ca}}=c_{\mathrm{i}}\left({\displaystyle \frac{{\delta }^{18}{\mathrm{O}}_{\mathrm{i}}-\overline{a}-{\alpha }_{\mathrm{ac}}{\delta }^{18}{\mathrm{O}}_{\mathrm{A}}}{{\delta }^{18}{\mathrm{O}}_{\mathrm{ce}}-\overline{a}-{\alpha }_{ac}{\delta }^{18}{\mathrm{O}}_{\mathrm{A}}}}\right)$$ (4)

where δ¹⁸O_i is the δ¹⁸O of CO₂ in the intercellular air spaces, with the ternary correction applied (Cernusak et al., 2004; Farquhar and Cernusak, 2012). δ¹⁸O_ce is the δ¹⁸O of CO₂ at the site of carbonic anhydrase activity. δ¹⁸O_A is the apparent δ¹⁸O of assimilated CO₂. α_ac is the ¹⁸O/¹⁶O fractionation factor for CO₂ diffusing across the boundary layer and stomata, defined as 1+ā. Detailed descriptions of the calculations of Δ¹³C_obs, Δ¹³C_i, δ¹⁸O_i, δ¹⁸O_ce, δ¹⁸O_A, and ā are presented by Diao et al. (2024b).

Finally, the g_m estimated from the measurements of the ¹⁸O/¹⁶O fractionation (g_m18), representing the g_m of the cell wall and plasma membrane, was calculated as:

$$g_{\mathrm{m}18}={\displaystyle \frac{A_{\mathrm{n}}}{c_{\mathrm{i}}-c_{\mathrm{ca}}}}$$ (5)

Subsequently, the g_m from the cytosol to the chloroplast (g_cm) was calculated from g_m13 (the total diffusional conductance to CO₂ in the mesophyll) and g_m18:

$$g_{\mathrm{cm}}={\displaystyle \frac{1}{{\displaystyle \frac{1}{g_{\mathrm{m}13}}-{\displaystyle \frac{1}{g_{\mathrm{m}18}}}}}}$$ (6)

The maximum carboxylation rate of Rubisco (V_cmax) was calculated according to Farquhar et al. (1980) :

$$V_{\mathrm{cmax}}={\displaystyle \frac{(A_{\mathrm{n}}+R_{\mathrm{day}})(c_{\mathrm{c}}+K_{\mathrm{m}})}{c_{\mathrm{c}}-{\mathrm{\Gamma }}^{*}}}$$ (7)

where Γ* is the CO₂ compensation point in the absence of day respiration and was calculated according to Sharkey et al. (2007). K_m is the effective CO₂ Michaelis–Menten constant for Rubisco at 21% O₂, which is expressed as K_c(1+O/K_o) where K_c and K_o are the Michaelis–Menten constants for CO₂ and O₂, respectively, and O is the O₂ partial pressure (210 mmol mol⁻¹). K_c and K_o were calculated using leaf temperature response functions described by Bernacchi et al. (2002).

Experiment 2: photorespiration measurements

Under each %B level, steady-state leaf gas exchange parameters were measured by sequentially supplying the gas exchange system with either ambient air (21% O₂) or with a mix of 2% O₂ and 98% N₂ (purity ≥99.999%; Linde Gas Schweiz AG, Dagmersellen, Switzerland). The gas selection was alternated using a three-way ball valve (SS-43GXS4, Swagelok, Solon, OH, USA). When the 2% O₂ was selected, the gas bottle flow rate was precisely adjusted using a pressure regulator (REG HBS 200-1-2, Air Liquide, Paris, France) to match the flow rate of the leaf gas exchange system (1000 μmol s⁻¹).

Photorespiration is substantial under normal air conditions because of the competitive inhibition of CO₂ fixation by O₂. Reducing the O₂ concentration to 2% suppresses the oxygenase activity of Rubisco, thereby minimizing photorespiration while maintaining photosynthesis (Sharkey, 1988). Hence, the photorespiration rate (R_photo) was estimated by subtracting A_n measured under ambient air conditions from the gross photosynthetic rate (A_g) measured at 2% O₂: R_photo=A_g–A_n. Leaf gas exchange data from Experiment 2 under ambient air conditions were pooled with data from Experiment 1 for further analyses (total n=10).

Experiment 3: fluorescence measurements

Fluorescence measurements were performed in the climate chamber using a portable chlorophyll fluorometer (PAM-2100) equipped with Special Fiberoptics 2010-F (both Walz). A leaf-clip holder (2030-B, Walz) was used to position the fibreoptics at a 60° angle to the leaf plane, at a distance of 7 mm. The target leaf area was defined by a steel ring with a diameter of 10 mm.

First, for each replicate sapling, a leaf was dark-adapted overnight and the minimum fluorescence yield (F_o) was then measured in the dark. This was followed by a saturation pulse (800 ms) from an actinic light via the fiberoptics of the fluorometer to determine the maximum fluorescence yield (F_m). The LED panel was then switched on. Under each %B level, the leaf was illuminated for 30 min for adaptation, after which the fluorescence yield (F_t) was measured followed immediately by application of a saturation pulse of the actinic light with a duration of 800 ms to determine the maximum fluorescence yield (F_m′). The F_t and F_m′ measurements were repeated five times in immediate succession for each leaf and averaged prior to further analysis. Finally, the LED panel was switched off and immediately afterwards the minimum fluorescence yield of the leaf (F_o′) was sampled five times and averaged prior to further analysis. The coefficient of photochemical fluorescence quenching (qP) was expressed as qP=(F_m′–F_t)/(F_m′–F_o′). The coefficient of non-photochemical fluorescence quenching (NPQ) was expressed as NPQ=(F_m–F_m′)/F_m′. The effective quantum yield of PSII (Φ_PSII) was expressed as Φ_PSII=(F_m′–F_t)/F_m′.

Analysis of photosynthetic limitation

The analysis of photosynthetic limitation was performed based on the mathematical framework of a differential method proposed by Liu et al. (2022). This method considers biochemical, stomatal, and mesophyll limitations as:

$$\mathrm{d}{A}_{\mathrm{calc}}=\mathrm{d}{A}_{\mathrm{biochem}}+\mathrm{d}{A}_{\mathrm{meso}}+\mathrm{d}{A}_{\mathrm{stom}}$$ (8)

where dA_calc is the difference between the maximum A_n across all the %B levels and the A_n at the given %B level, and dA_biochem, dA_meso, and dA_stom are respectively the biochemical, mesophyll, and stomatal components of dA_calc. Based on the fact that under low CO₂ concentration ranges, biochemical limitation operates mainly in the form of Rubisco limitation rather than RuBP-regeneration limitation (Deans et al., 2019; Acevedo-Siaca et al., 2020; Liu et al., 2022), dA_biochem was considered to be similar to the RuBP-carboxylation limitation in which V_cmax is the major limiting factor. dA_biochem, dA_meso, and dA_stom were calculated as follows:

$$\{\begin{array}{c}\mathrm{d}{A}_{\mathrm{biochem}}={\displaystyle \frac{\mathrm{\partial }{A}_{\mathrm{n}}}{\mathrm{\partial }{V}_{\mathrm{cmax}}}\mathrm{d}{V}_{\mathrm{cmax}}}\\ \mathrm{d}{A}_{\mathrm{meso}}={\displaystyle \frac{\mathrm{\partial }{A}_{\mathrm{n}}}{\mathrm{\partial }{g}_{\mathrm{m}13}}\mathrm{d}{g}_{\mathrm{m}13}}\\ \mathrm{d}{A}_{\mathrm{stom}}={\displaystyle \frac{\mathrm{\partial }{A}_{\mathrm{n}}}{\mathrm{\partial }{g}_{\mathrm{sc}}}\mathrm{d}{g}_{\mathrm{sc}}}\end{array}$$ (9)

where g_sc is the stomatal conductance to CO₂ (g_sc=g_s/1.6). dV_cmax, dg_m13, and dg_sc are the differences between the maximum values across all %B levels and the values at the given %B level for V_cmax, g_m13, and g_sc, respectively. The implicit differentiation in Equation 9 can be found in Liu et al. (2022). At each %B level, we further normalized dA_biochem, dA_meso, and dA_stom by multiplying them by dA_n/dA_calc, where dA_n is the difference between the maximum and current A_n.

The integrated relative limitations imposed by V_cmax (σ_biochem), g_m13 (σ_meso), and g_sc (σ_stom) across the %B levels were calculated as:

$${\sigma }_{\mathrm{biochem}/\mathrm{meso}/\mathrm{stom}}={\displaystyle \frac{\int \mathrm{d}{A}_{\mathrm{biochem}/\mathrm{meso}/\mathrm{stom}}\mathrm{d}\text{\%}\mathrm{B}}{\int \mathrm{d}{A}_{\mathrm{calc}}\mathrm{d}\text{\%}\mathrm{B}}}$$ (10)

Statistical analysis

All statistical analyses were conducted using R software v.4.4.1 (www.r-project.org). For each replicate of each species, the percent changes in the investigated variables with increasing %B were calculated relative to those under the 0% blue treatment. This normalization was performed to enhance the comparison of the treatment effect between species. One-sample Student’s t-tests were conducted to determine whether the raw measurements or the percent changes in the variables under each %B level were significantly different from that under 0%B. Two-sample t-tests were performed to test the significance of differences between the two species or between two different %B treatments. ANOVA was used to determine whether the response of a variable to increasing %B was significant. Pearson’s correlation analysis was used to investigate the relationships between A_n and g_s or g_m13 and between intrinsic water-use efficiency (iWUE; calculated using A_n/g_s) and Δ¹³C_obs.

Results

Responses of A_n, g_s, R_photo, and iWUE to increasing blue light

Across the %B gradient, the mean values of A_n, g_s, and R_photo were significantly higher for grey alder than for holm oak (hereafter abbreviated to alder and oak; P<0.001; Fig. 2A–C). A_n of alder decreased significantly with increasing %B (P<0.001), with a 43% reduction at 100%B compared with its value at 0%B (Fig. 2D). In contrast, A_n of oak increased significantly at both 10%B (P<0.001) and 30%B (P=0.002) compared with 0%B, followed by a decreasing trend with further increases in %B. Only at ≥90%B was A_n of oak significantly lower than at 0%B (P<0.001), with a 16% reduction observed at 100%B.

Fig. 2.

Responses of photosynthesis, stomatal conductance, and photorespiration to increasing blue-light fraction for grey alder and holm oak. Plants in Experiments 1 and 2 were exposed to 0–100% blue light, with the remaining fraction being red light and the total photon flux density being constant (Fig. 1). (A) Net photosynthesis rate (A_n), (B) stomatal conductance (g_s), and (C) photorespiration rate (R_photo). (D–F) Percentage changes relative to the 0% blue-light treatment for (D) A_n, (E) g_s, and (F) R_photo. Data are means (±SD), for n=10 (A_n, g_s) or n=5 (R_photo).

With increasing %B, g_s of alder first increased to a peak at 30%B and then decreased, although these changes were not significant (Fig. 2B, E). In contrast, g_s of oak increased significantly with increasing %B (P=0.02), albeit starting from a much lower value compared with alder. Specifically, g_s of oak had increased by 33% at 70%B, but further increases in %B had no significant effect.

The response of R_photo to %B was not significant for alder (Fig. 2C, F), and although it decreased significantly overall with increasing %B for oak (P=0.01), the values under individual %B levels were not significantly different from those at 0%B, indicating a limited response in both species.

The different responses of A_n and g_s to increasing %B between the two species resulted in contrasting A_n–g_s relationships (Fig. 3A). For alder, A_n and g_s were strongly coupled from 30%B to 100%B, as evidenced by a significant positive correlation (r=0.99, P<0.001; Fig. 3A). In contrast, increases in both A_n and g_s for oak were observed only from 0%B to 10%B. From 0%B to 30%B for alder and from 10%B to 70%B for oak, an increase in g_s was accompanied by a decrease in A_n. At 70%B to 100%B, A_n and g_s became completely decoupled in oak, as A_n declined without any significant changes in g_s.

Fig. 3.

Relationship between changes in photosynthesis and stomatal conductance, and the response of intrinsic water-use efficiency to increasing blue-light fraction for grey alder and holm oak. Plants in Experiment 1 were exposed to 0–100% blue light, with the remaining fraction being red light and the total photon flux density being constant (Fig. 1). (A) Relative percentage changes in net photosynthesis rate (A_n) versus stomatal conductance (g_s) and (B) relative percentage changes in intrinsic water-use efficiency (iWUE) at different fractions of blue light. The fractions are indicated by the numbers in (A). In both cases, changes are relative to the value under the 0% blue-light treatment, the value of which was set as zero. Data are means (±SD), n=10.

Despite the contrasting A_n–g_s patterns in response to %B changes, iWUE (calculated using A_n/g_s) decreased significantly with increasing %B for both species and followed the same pattern (P<0.001; Fig. 3B). At 100%B, iWUE had declined by 37% relative to 0%B for alder and by 36% for oak.

Responses of g_m and the CO₂ mole fraction inside the leaf to increasing blue light

Across the %B levels, the values of total mesophyll conductance to CO₂ (g_m13), the cell-wall and plasma-membrane conductance (g_m18), and the chloroplast-membrane conductance (g_cm) were significantly higher on average for alder than for oak (P<0.05), except for g_cm at 70%B to 100%B (Fig. 4A–C). For alder, g_m13 decreased significantly with increasing %B (P<0.001), with a 46% reduction at 100%B relative to 0%B (Fig. 4D). For oak, g_m13 increased at 10%B, then showed a decreasing trend with further increases in %B, but the responses were not significant.

Fig. 4.

Responses of mesophyll conductance and its components to increasing blue-light fraction for grey alder and holm oak. Plants in Experiment 1 were exposed to 0–100% blue light, with the remaining fraction being red light and the total photon flux density being constant (Fig. 1). (A) Total mesophyll conductance, as estimated from ¹³C measurements (g_m13). (B) Mesophyll cell wall and plasma membrane conductance, as estimated from ¹⁸O measurements (g_m18). (C) Chloroplast membrane conductance, as calculated from g_m13 and g_m18. (D–F) Percentage changes relative to the 0% blue-light treatment for (D) g_m13, (E) g_m18, and (F) g_cm. Data are means (±SD), n=5 biological replicates.

The response of g_m18 to %B was not significant for alder (Fig. 4B, E). For oak, although g_m18 decreased significantly with increasing %B (P=0.01), the values under individual %B levels were not significantly different from those at 0%B, indicating a limited overall response for this species as well. As a result of the variations in g_m13 and g_m18, the calculated g_cm showed a significant decreasing trend with increasing %B for alder (P<0.001), with a reduction of 67% at 100%B (Fig. 4C, F); no significant response was observed for oak.

In response to changing %B levels, A_n showed a significant correlation with g_m13 in both alder and oak (Fig. 5A). However, A_n was significantly correlated with g_m18 only in oak (Fig. 5B) and with g_cm only in alder (Fig. 5C).

Fig. 5.

Relationships between changes in photosynthesis rate and changes in the different components of mesophyll conductance for grey alder and holm oak. Plants in Experiment 1 were exposed to 0–100% blue light, with the remaining fraction being red light and the total photon flux density being constant (Fig. 1). All percentage changes were calculated relative to the 0% blue-light treatment. Changes in net photosynthesis rate (A_n) are plotted against (A) changes in total mesophyll conductance (g_m13), (B) changes in mesophyll cell wall and plasma membrane conductance (g_m18), and (C) chloroplast membrane conductance (g_cm). See Table 1 for details of the conductance variables. Data are means (±SD), n=5 biological replicates. Linear relationships that are significant (P<0.05) according to ordinary least-squares linear regression are shown.

Across the %B levels, the mean c_i/c_a, c_ca/c_a, and c_c/c_a ratios were significantly higher for alder than for oak (P<0.001; Fig. 6A–C) and significant increasing trends in all three in response to increases in %B were observed for both species (P<0.01), although the percent changes in c_c/c_a of alder were not significantly different from zero (Fig. 6F). The percent changes in all three ratios were higher overall for oak than for alder, especially for c_c/c_a (Fig. 6D–F).

Fig. 6.

Responses of the relative CO₂ mole fractions within the leaf to increasing blue-light fraction for grey alder and holm oak. Plants in Experiment 1 were exposed to 0–100% blue light, with the remaining fraction being red light and the total photon flux density being constant (Fig. 1). (A) The ratio of intercellular to ambient CO₂ (c_i/c_a), (B) the ratio of chloroplast-surface to ambient CO₂ (c_ca/c_a), and (C) the ratio of chloroplast stroma to ambient CO₂ (c_c/c_a). See Table 1 for further details. (D–F) Percentage changes relative to the 0% blue-light treatment for (D) c_i/c_a, (E) c_ca/c_a, (F) and c_c/c_a. Data are means (±SD), for n=10 (c_i/c_a) or n=5 (c_ca/c_a, c_c/c_a).

Responses of V_cmax, NPQ, and Φ_PSII to increasing blue light

V _cmax significantly decreased with increasing %B in both alder (P<0.001) and oak (P=0.003) (Fig. 7A), with reductions of 44% and 46%, respectively, at 100%B compared with 0%B (Fig. 7D). For alder, NPQ showed a significant increasing trend with increasing %B (P<0.001), with a 35% increase at 100%B compared with 0%B (Fig. 7B, E). In contrast, NPQ of oak declined significantly by 50%, from 0%B to 50%B, followed by a significant increase from 50%B to 90%B (P<0.01); the reduction in NPQ at 100%B was 45% relative to 0%B. For both species, qP showed no significant response to increasing %B (Supplementary Fig. S3).

Fig. 7.

Responses of maximum carboxylation rate of Rubisco, non-photochemical fluorescence quenching, and effective quantum yield of PSII to increasing blue-light fraction for grey alder and holm oak. Plants in Experiments 1 and 3 were exposed to 0–100% blue light, with the remaining fraction being red light and the total photon flux density being constant (Fig. 1). (A) Maximum carboxylation rate of Rubisco (V_cmax), (B) non-photochemical fluorescence quenching (NPQ), and (C) effective quantum yield of photosystem PSII (Φ_PSII). (D–F) Percentage changes relative to the 0% blue-light treatment for (D) V_cmax, (E) NPQ, and (F) Φ_PSII. Data are means (±SD), n=5 biological replicates.

The responses of Φ_PSII to %B showed opposite patterns to those observed for NPQ for both species (Fig. 7C, F). Specifically, Φ_PSII of alder decreased significantly with increasing %B (P=0.002), with a 10% reduction at 100%B compared with 0%B. For oak, Φ_PSII increased significantly by 31% at 50%B (P<0.001), but further increases in %B had no significant effect.

Photosynthetic limitations imposed by biochemistry, mesophyll conductance, and stomatal conductance

Biochemical (dA_biochem), mesophyll (dA_meso), and stomatal (dA_stom) limitations on photosynthesis in alder and oak were quantitatively estimated at different levels of blue light according to Equation 9. In both species, the calculated values of dA_biochem and dA_meso were consistently higher than those of dA_stom, with dA_biochem being higher than dA_meso, (Fig. 8A, B). In addition, dA_biochem and dA_meso were generally higher at higher levels of %B. However, for oak, dA_stom was higher than both dA_biochem and dA_meso at 0%B. For the integrated limitation across all %B levels, σ_biochem, σ_meso, and σ_stom accounted for 55%, 37%, and 7%, respectively, of the total photosynthetic limitation in alder; and 60%, 27% and 13% in oak (Fig. 8C).

Fig. 8.

Biochemical, mesophyll, and stomatal limitations on the photosynthetic response to increasing blue-light fraction for grey alder and holm oak. The biochemical (dA_biochem), mesophyll (dA_meso), and stomatal (dA_stom) limitations were calculated according to Equation 9. (A) Grey alder. (B) Holm oak. At each %B level, dA_biochem, dA_meso, and dA_stom were further normalized by multiplying them by dA_n/dA_calc, where dA_n is the difference between the maximum and current A_n. (C) The integrated relative limitations by V_cmax (maximum carboxylation rate of Rubisco; σ_biochem), g_m13 (total mesophyll conductance; σ_meso), and g_sc (stomatal conductance to CO₂; σ_stom) across the different %B levels. All data are means (±SD), n=5 biological replicates.

Discussion

Sensitivities of photosynthesis and stomatal conductance to the blue-to-red light ratio differ between the species

Trees experience complex changes in light environment, with the blue fraction of PAR typically higher in the early morning (Chiang et al., 2019; Kotilainen et al., 2020) and the blue/red ratio greater within the canopy compared with above or below it under clear-sky conditions (Hertel et al., 2011; Dengel et al., 2015; Durand et al., 2021b). Sun flecks usually have a lower blue/red ratio compared with surrounding shade (Durand et al., 2021a; Sellaro et al., 2025). However, despite this, the response patterns and mechanisms of leaf gas exchange to blue- and red-light compositions remain largely underexplored in tree species. Our study revealed contrasting photosynthetic and stomatal sensitivities to increasing blue-light fraction between alder and oak, probably reflecting species-specific adaptations to their respective environments. In alder, increasing the blue-light fraction did not enhance A_n, whilst oak exhibited a notable increase under low doses of blue light (10–30%) (Fig. 2A, D). Similarly, blue light stimulated stomatal opening in oak across all the tested %B levels, with saturation occurring at higher levels, whereas alder showed a stimulation only at lower doses (∼30%) (Fig. 2B, E).

These divergent responses suggest that oak possesses a higher photosynthetic tolerance to elevated blue-light doses, effectively benefiting from blue light through increased A_n and g_s. Our findings also suggest that oak has more conservative photosynthetic and stomatal response patterns to fluctuating blue and red light than alder, allowing it to maintain stable leaf gas exchange even under large light changes, such as sun flecks (Way and Pearcy, 2012). Thus, oak demonstrates an ecological adaptation to maximize light utilization across dynamic spectral conditions in its native habitats. This aligns with the general expectation that the leaf morphology of oak is shaped by the climate of its habitat, which is characterized by warm, dry conditions and high light intensity (Supplementary Fig. S2). The small, coriaceous, and thick leaves of oak with leaf mass per area (LMA) of 14.1±1.7 mg cm⁻², in contrast to the large, soft, and thin leaves of alder with LMA of 5.5±0.5 mg cm⁻² (Supplementary Fig. S2), provide an advantage in terms of protection against light stress and photodamage caused by high-energy blue light (Martín-Sánchez et al., 2024). Since both species are light-demanding, alder would thrive better in light habitats with lower %B, such as forest understory gaps; however, when exposed to high %B levels, where the negative effects on its g_s are more pronounced, it might still maintain a better water balance than oak.

Differences in internal light distribution between thick and thin leaves might have contributed to the species-specific responses observed in this study. Previous work has shown that both red and blue light are predominantly absorbed by the upper layers of the palisade mesophyll, with red light penetrating deeper into the leaf than blue (Vogelmann and Han, 2000; Vogelmann and Evans, 2002). Thinner leaves, such as those of alder, allow more blue light to penetrate compared with thicker leaves (Brodersen and Vogelmann, 2010), which might increase their sensitivity to blue-light exposure, resulting in the reductions in photosynthesis that we observed even at the lowest level of %B (Fig. 2A, D). Moreover, light quality has been shown to influence chloroplast distribution within mesophyll cells, for example through chloroplast movement induced by blue light (Wada, 2013). Patterns of chlorophyll distribution also differ with leaf thickness, with thicker leaves typically exhibiting more stratified chlorophyll profiles than thinner leaves (Borsuk and Brodersen, 2019). More studies are needed to fully understand how the dynamics of chloroplast distribution in response to changes in leaf internal light environments drive photosynthesis across different leaf morphologies.

Notably, in oak the increase in A_n at low %B and the strong sensitivity of g_s to low %B followed by saturation of the response at higher levels (Fig. 2D, E) matched patterns previously observed in herbaceous C₃ species including cucumber (Cucumis sativus), sunflower (Helianthus annuus), and kale (Brassica napus var. pabularia) (Savvides et al., 2011; Zhen and Bugbee, 2020). In contrast, alder exhibited a consistent decline in A_n and a rise-and-fall pattern in g_s, resembling the blue-light response of the C₄ species Sorghum bicolor (Zhen and Bugbee, 2020). These findings underscore the fact that contrasting stomatal responses to blue light are not strictly associated with differences in the C₃ and C₄ photosynthetic pathways, as they also exist between these two C₃ tree species.

Contrasting coupling of photosynthesis and stomatal conductance under varying blue-light fractions

The contrasting patterns of the blue-light responses of A_n and g_s led to different A_n–g_s coupling behaviours between the two species, with alder exhibiting tight coupling, while oak showed clear decoupling (Fig. 3A). Interestingly, of these patterns were accompanied by a linear increase in c_i/c_a with increasing %B (Fig. 6A, D), suggesting that intercellular CO₂ is unlikely to be the primary driver of A_n and g_s coordination under shifting blue-to-red ratios, contrary to previous assumptions (Roelfsema et al., 2002). The decoupled A_n–g_s observed in oak might reflect a temporal mismatch between photosynthetic and stomatal responses to blue light due to differences in photoreceptor sensitivity or downstream signalling pathways (Lawson et al., 2010; McAusland et al., 2016).

Reduced water-use efficiency under blue light is caused by different mechanisms in the two species

The contrasting blue-light response patterns of A_n and g_s in alder and oak surprisingly converged on a common outcome, namely a reduction in iWUE with increasing %B (Fig. 3B). This trend was independently supported by the Δ¹³C_obs measurements, which showed a significant negative relationship with iWUE in both species (Supplementary Fig. S4), consistent with the well-established link between Δ¹³C_obs and iWUE via c_i/c_a (Farquhar et al., 1989). Notably, WUE, calculated using A_n/E, matched the decrease in iWUE, calculated using A_n/g_s, owing to constant VPD across treatments, which maintained tight coupling between g_s and E (Supplementary Fig. S5).

Although both species exhibited declining iWUE and WUE under increasing %B, the underlying mechanisms differed. For alder, the decrease was mainly due to a reduction in A_n, whereas for oak it was additionally caused by a simultaneous increase in g_s and E (Fig. 2A, B). On top of that, the factors limiting A_n in each species were different. In alder, increasing %B induced a rise in NPQ and a decline in Φ_PSII (Fig. 7), indicating that greater light energy was diverted away from photochemical processes to non-photochemical processes to prevent overexcitation, resulting in reduced efficiency in photochemical energy conversion during photosynthesis (Murchie and Lawson, 2013). This energy reallocation probably reduced carbon assimilation as the fraction of blue light increased. In contrast, in oak NPQ decreased while Φ_PSII increased from 0%B to 50%B, corresponding to the initial increase in A_n (Fig. 2A, D). As suggested by Evans et al. (2017), this could reflect greater Φ_PSII in deeper mesophyll layers, leading to an overall greater sum of Φ_PSII at higher blue-light fractions. The subsequent decline in A_n in oak at even higher %B, despite stable NPQ and Φ_PSII (Fig. 7), suggests that it was not associated with reduced photosynthetic efficiency due to excess blue-light stress. Additionally, for both species, R_photo (Fig. 2C, F) and qP (Supplementary Fig. S3) did not show significant responses to increasing %B. This suggests that Rubisco oxygenation and the proportion of open PSII reaction centres were not sensitive to changes in the blue-light fraction. Instead, these parameters were probably more influenced by the total light intensity (Han et al., 2021; Fu and Walker, 2022), which was kept constant throughout the experiments.

The blue-light-driven reductions in iWUE and WUE observed here are consistent with previous findings for sunflower and kale under similar experimental conditions (Zhen and Bugbee, 2020), as well as lettuce (Lactuca sativa) grown under varying blue-light fractions (Clavijo-Herrera et al., 2018). However, other studies have reported increased WUE at low blue-light levels (∼23–33%B) in lettuce and sweet basil (Ocimum basilicum) (Pennisi et al., 2019a; 2019b; Zhou et al., 2022), while no clear pattern was observed in tomato (Solanum lycopersicum) and eggplant (Solanum melongena) (Di et al., 2021; Zhou et al., 2022). These discrepancies might stem from differences in total intensity of the mixed blue and red lights across the studies (Pennisi et al., 2020; Modarelli et al., 2022), highlighting the need for further research, especially on tree species, to clarify whether the effects of blue light on WUE are maintained irrespective of the light intensity.

Components of mesophyll conductance respond differently to changes in the blue-light fraction

The blue-light responses of the total g_m (estimated as g_m13: Table 1) were well correlated with those of A_n across the %B levels in both species (Fig. 5A), suggesting at least an apparent coordination between them. This supports previous reports of blue-light-induced reductions in g_m observed using the fluorescence-based ‘constant J method’ (Loreto et al., 2009; Pallozzi et al., 2013; Momayyezi and Guy, 2017). However, our results further revealed that different individual components contributed to the changes in g_m13 (and consequently A_n) observed in alder and oak under varying blue light.

For alder, the blue-light-induced decrease in g_m13 was specifically linked to a decline in g_cm (Fig. 4), resulting in a significant correlation between A_n and g_cm, but not g_m18 (Fig. 5). This indicates that diffusional CO₂ resistance in the mesophyll primarily arose from the cytosol to the sites of carboxylation in the chloroplast, rather than the cell wall or plasma membrane. Conversely, for oak, it was decreasing g_m18 rather than g_cm that drove the reduction in g_m13 when %B exceeded 10% (Fig. 4), with g_m18 significantly correlated with A_n (Fig. 5). This indicates that in oak, the cell wall and plasma membrane play a crucial role in regulating CO₂ diffusion under changing blue-light conditions.

Several mechanisms could explain these species-specific responses. We propose that in alder, chloroplasts are more likely to exhibit avoidance movement as blue-light intensity increases in the mesophyll due to its thinner leaves (Brodersen and Vogelmann, 2010). This movement reduces the surface area of chloroplasts exposed to intercellular spaces, thereby lowering g_cm (Tholen et al., 2008; Ko et al., 2020). In contrast, the thicker leaves of oak might attenuate blue light in the mesophyll (Brodersen and Vogelmann, 2010), making chloroplast avoidance less likely. Instead, the blue-light-driven decline in g_m18 in oak suggests alternative mechanisms such as the down-regulation of CO₂-permeable aquaporins in the plasma membrane (Ben Baaziz et al., 2012; Balarynová and Fellner, 2019; Diao et al., 2025) or reduced carbonic anhydrase activity, which could increase ionic CO₂ transport resistance across the plasma membrane (Tholen and Zhu, 2011; Momayyezi and Guy, 2017). Additionally, hormonal regulation might play a role, as several phytohormones respond to changes in blue light and influence the CO₂ diffusion function of membranes (Pang et al., 2023).

Mesophyll conductance imposes a remarkable limitation on photosynthesis at high levels of blue light

We analysed photosynthetic limitations using the differential method, which has been suggested as the best approach compared with alternatives such as the elimination methods (Deans et al., 2019). For the first time, this analysis was conducted across the full spectrum of blue-to-red light ratios, revealing the varying contributions of biochemical, mesophyll, and stomatal limitations on photosynthesis (Fig. 8). In both species, V_cmax imposed the greatest overall limitation, and together with g_m13 it contributed considerably more to photosynthetic constraints than g_s, especially under high blue-light fractions. Notably, the integrated relative limitations imposed by g_m13 were 37% and 27% for alder and oak, respectively, highlighting its important role in constraining photosynthesis under blue-light illumination.

V _cmax was calculated under 21% O₂, and under these conditions there is a possibility of overestimating it due to a lag between the oxygenase reaction by Rubisco and the CO₂ release from photorespiration (Sakoda et al., 2021). If such an overestimation occurred, the true contribution of V_cmax to photosynthetic limitation could be higher; however, given that R_photo showed no significant response to changes in %B (Fig. 2C, F), it probably played a minimal role in shaping the observed V_cmax response to varying blue-light fractions (Fig. 7A, D).

Interestingly, neither species exhibited lower c_c/c_a at higher %B compared with lower %B (Fig. 6C, F), despite the notable declines in both g_s and g_m13 in alder when %B exceeded 30% (Figs 2B, E, 4A, D). This was consistent with the photosynthetic limitation analysis, indicating that stomatal factors were not the primary cause of the observed reduction in A_n with increasing %B. Moreover, our findings imply that the parallel trends between A_n and g_m13 are only apparent and do not reflect reduced CO₂ availability for photosynthesis due to a reduction in total mesophyll conductance (Loreto et al., 2009; Márquez and Busch, 2024; Diao et al., 2024b). These findings highlight the species-specific roles of mesophyll conductance in regulating photosynthetic efficiency under blue light, reinforcing the need to consider g_m as a critical factor in plant responses to varying light spectra.

Conclusions

We examined the integrated roles of photosynthetic, stomatal, mesophyll, and biochemical responses to different blue-to-red ratios in two tree species, revealing previously unrecognized species-specific mechanisms. Oak exhibited greater tolerance of photosynthesisand stronger stimulation of stomatal conductance under increased blue light, probably reflecting its adaptation to high-radiation, water-limited environments where maintaining stomatal aperture under blue light might help to maximize carbon gain. In contrast, alder showed tighter coupling between photosynthesis and stomatal conductance, with photosynthetic performance being more sensitive to changes in blue-light fractions. Interestingly, leaf iWUE decreased with increasing blue light in both species but through different mechanisms: in alder, decreased A_n was linked to photoprotective processes (increased NPQ and decreased Φ_PSII), while in oak, the decline in iWUE was mainly attributable to increased g_s, prioritizing carbon gain over water conservation. More importantly, both species showed that V_cmax and g_m primarily limited A_n under blue light, with different components of g_m responsible for these limitations: the chloroplast membrane in alder and the cell wall and plasma membrane in oak, indicating different anatomical or biochemical controls over CO₂ diffusion in the mesophyll. By elucidating these species-specific mechanisms, our study not only challenges conventional assumptions based on results for herbaceous species or C₃ versus C₄ comparisons, but also provides novel insights into the different roles of mesophyll and stomatal conductance in shaping leaf water-use efficiency, thereby enhancing our understanding of the physiological strategies adopted by tree species under varying light environments.

Supplementary data

The following supplementary data are available at JXB  online.

Fig. S1. Density plot of the ratios of photon flux density for blue, green, and red light to photosynthetically active radiation.

Fig. S2. Illustration of autecological niches and leaf morphologies for grey alder and holm oak.

Fig. S3. Responses of photochemical fluorescence quenching to increasing blue-light fraction.

Fig. S4. Responses of Δ¹³C_obs to increasing blue-light fraction and its relationship with iWUE.

Fig. S5. Relationships between g_s and E, and between WUE and iWUE.

Author contributions

HD, MS, MML, and MH-P: conceptualization; HD and RTWS: methodology; HD: investigation, data curation, writing—original draft; MML, AG, RTWS, and MS: resources; HD, MML, MH-P, AG, RTWS, and MS: writing—review & editing; HD, MML, and MS: funding acquisition.

Funding

This work was supported by Velux Stiftung (project no. 1733). MML was supported by the Swiss National Science Foundation (SNSF, grant no. 213367). We also acknowledge funding from the SNSF (grant no. 189724, to AG and MS) and WSL (innovative research project PPF2020, to MML) for the instrumentation.

Data availability

The data supporting the findings of this study are available from the corresponding author, Haoyu Diao, upon request.