Knockdown of glycine decarboxylase complex alters photorespiratory carbon isotope fractionation in Oryza sativa leaves

The disruption of photorespiration in GDC knockdown rice plants alters leaf photorespiratory ¹³CO₂ fractionation and carbon isotope exchange.

Abstract

The influence of reduced glycine decarboxylase complex (GDC) activity on leaf atmosphere CO₂ and ¹³CO₂ exchange was tested in transgenic Oryza sativa with the GDC H-subunit knocked down in leaf mesophyll cells. Leaf measurements on transgenic gdch knockdown and wild-type plants were carried out in the light under photorespiratory and low photorespiratory conditions (i.e. 18.4 kPa and 1.84 kPa atmospheric O₂ partial pressure, respectively), and in the dark. Under approximately current ambient O₂ partial pressure (18.4 kPa pO₂), the gdch knockdown plants showed an expected photorespiratory-deficient phenotype, with lower leaf net CO₂ assimilation rates (A) than the wild-type. Additionally, under these conditions, the gdch knockdown plants had greater leaf net discrimination against ¹³CO₂ (Δ_o) than the wild-type. This difference in Δ_o was in part due to lower ¹³C photorespiratory fractionation (f) ascribed to alternative decarboxylation of photorespiratory intermediates. Furthermore, the leaf dark respiration rate (R_d) was enhanced and the ¹³CO₂ composition of respired CO₂ (δ¹³C_Rd) showed a tendency to be more depleted in the gdch knockdown plants. These changes in R_d and δ¹³C_Rd were due to the amount and carbon isotopic composition of substrates available for dark respiration. These results demonstrate that impairment of the photorespiratory pathway affects leaf ¹³CO₂ exchange, particularly the ¹³C decarboxylation fractionation associated with photorespiration.

Introduction

In C₃ plants, Rubisco operates in the leaf mesophyll cells, where CO₂ and O₂ compete to react with ribulose-1,5-bisphosphate (RuBP). The carboxylation of RuBP results in the formation of two molecules of 3-phosphoglycerate (3-PGA) that are integrated into the Calvin–Benson cycle. Alternatively, the oxygenation of RuBP produces one molecule of 3-PGA and one 2-phosphoglycolate (2-PG). The 2-PG is primarily recycled via photorespiration through a complex and energy-consuming set of reactions, which spans the chloroplasts, cytosol, peroxisomes, and mitochondria (Bauwe et al., 2010; Betti et al., 2016). By scavenging 2-PG, photorespiration removes a strong inhibitor of enzymes in photosynthetic carbon metabolism (Anderson, 1971; Kelly and Latzko, 1976; Peterhansel et al., 2013b; Walker et al., 2016) and recovers up to one molecule of 3-PGA for every two molecules of 2-PG. Nevertheless, a minimum of one out of four 2-PG carbon atoms is released as CO₂ by the glycine decarboxylase complex (GDC) and can be lost by the plant (Bauwe, 2018).

The GDC is an atypical mitochondrial four-protein system, comprised of three enzymes (P-, T-, and L-protein) and the H-protein, which is a small lipoylated protein (Somerville and Ogren, 1982; Douce et al., 2001; Bauwe, 2018). GDC plays a critical role in the photorespiratory cycle by catalyzing the conversion of two molecules of glycine into serine and one molecule of CO₂ and NH₃ (Somerville, 2001; Maurino and Peterhansel, 2010). However, in the absence of the H component, the GDC cannot oxidize glycine (Douce et al., 2001; Parys and Jastrzębski, 2008), which can accumulate. In C₃ plants, the impaired activity of the H-subunit leads to a knockdown (KD) of GDC activity and a photorespiratory phenotype (Ewald et al., 2007). Plants with reduced GDC activity typically have lower rates of leaf photosynthesis, a depletion of Calvin cycle metabolites, an impairment of photorespiratory nitrogen re-assimilation, and the accumulation of photorespiratory metabolites (e.g. glycine) under current ambient CO₂ and O₂ partial pressures (Wingler et al., 2000; Timm and Bauwe, 2013; Lin et al., 2016).

This buildup of leaf photorespiratory metabolites can have a negative feedback effect on Calvin cycle activity. For example, glyoxylate produced by glycolate oxidation negatively impacts on the activation state of Rubisco (Wingler et al., 1999; Peterhansel et al., 2010). Additionally, disruption of the photorespiratory pathway may lead to alternative decarboxylation reactions of accumulated pools of photorespiratory intermediates, such as glyoxylate and hydroxypyruvate in the peroxysomes (Wingler et al., 1999, 2000; Tcherkez, 2006; Peterhansel et al., 2010), and an increase in the ratio of moles of photorespiratory CO₂ released per mole of O₂ reacting with RuBP (α; Cousins et al., 2008, 2011; Walker and Cousins, 2013; Timm et al., 2018). Furthermore, the accumulation of photorespiratory intermediates could also affect the rates of leaf CO₂ evolved in the dark (R_d, μmol CO₂ m⁻² s⁻¹) and the ¹³C composition of R_d (δ¹³C_Rd, ‰) (Ghashghaie et al., 2003; Tcherkez et al., 2003).

The multiple leaf metabolic reactions simultaneously consuming and releasing CO₂ in the light make it difficult to determine how changes in photorespiration affect rates of leaf net CO₂ assimilation (A), mesophyll CO₂ conductance (g_m), refixation of (photo)respired CO₂, and mitochondrial non-photorespiratory respiration rates (R_L). However, photosynthesizing leaves discriminate against ¹³C during CO₂ diffusion from the atmosphere to the chloroplast stroma (through both the air and liquid phases), and during carboxylation, photorespiration, and mitochondrial non-photorespiratory respiration processes, with a specific ¹³C fractionation for each diffusional or biochemical step (Evans et al., 1986). The observed leaf net discrimination against ¹³C in the light (Δ_o, ‰) can be modeled with four ¹³C fractionation terms (‰): Δ_i, which accounts for the ¹³C discrimination during CO₂ diffusion from the atmosphere to the intercellular air space and for the Rubisco ¹³C fractionation (~29‰, Ubierna and Farquhar, 2014; von Caemmerer et al., 2014); Δ_gm, which accounts for the ¹³C discrimination during CO₂ diffusion in the liquid phase to chloroplast stroma and depends on the magnitude of g_m; and Δ_f and Δ_e which are associated with photorespiration and mitochondrial non-photorespiratory respiration activity, respectively (von Caemmerer and Evans, 1991; Flexas et al., 2008; Tazoe et al., 2011; Evans and von Caemmerer, 2013). Δ_f is primarily attributed to the glycine–serine reaction catalyzed by GDC, which releases CO₂ depleted in ¹³C compared with substrate and tends to decrease Δ_o (Farquhar et al., 1982; Ghashghaie et al., 2003; Lanigan et al., 2008). In contrast, Δ_e may increase or decrease Δ_o in relation to the difference between ¹³C composition (‰) of CO₂ entering the leaf chamber during measurements and in the plant growth chamber (Gillon and Griffiths, 1997; Ghashghaie et al., 2003).

The photorespiratory fractionation (f, ‰) estimated in vivo in multiple C₃ species ranges between 8‰ and 16.2‰ relative to photosynthetic products (Ghashghaie et al., 2003; Evans and von Caemmerer, 2013), with 11‰ predicted from the theory (Tcherkez, 2006). However, under photorespiratory conditions, when Rubisco oxygenation exceeds the capacity of the photorespiratory recycling of 2-PG or in the presence of disruption of the photorespiratory pathway, Δ_f and f may vary due to changes in α associated with alternative decarboxylation of photorespiratory intermediates (Cousins et al., 2008, 2011; Walker and Cousins, 2013). Alternative photorespiratory bypasses may occur in the chloroplasts (e.g. glyoxylate may be enzymatically reduced back to glycolate or further oxidized to CO₂, but with no RuBP regenerated; see Kebeish et al., 2007), peroxysomes (non-enzymatic decarboxylation of glyoxylate to formate using H₂O₂ as oxidizing agent may lead to formation of serine; catalase may be also involved as reported in Wingler et al., 1999), mitochondria (enzymatic oxidation of glycolate to glyoxylate with release of CO₂ and synthesis of glycine; see Niessen et al., 2007), and cytosol (enzymatic reduction of hydroxypiruvate to glycerate; see Timm et al., 2008).

The aim of the present study was to test how changes in carbon flux through the photorespiratory pathway influenced leaf CO₂ and ¹³CO₂ isotope exchange, both in the light and in the dark, in transgenic plants of Oryza sativa with the GDC H-subunit KD in mesophyll cells. Both gdch-KD and wild-type (WT) plants were grown under low photorespiratory conditions (atmospheric CO₂ partial pressure of 184.2 Pa) to minimize any pleiotropic effects. In the light, measurements of leaf–atmosphere CO₂ and stable carbon isotope exchange were performed under low photorespiratory and photorespiratory conditions (atmospheric O₂ partial pressure of 1.84 kPa or 18.4 kPa, respectively, and CO₂ partial pressure of 27.6 Pa). The disruption of the photorespiratory pathway in the gdch-KD plants was characterized by leaf photosynthetic traits, Δ_o, Δ_f, f, α, R_d, and δ¹³C_Rd, compared with the WT.

Materials and methods

Plant material

Generation of GDC-H knockdown transgenic rice lines

The generation and the characterization of three Oryza sativa gdch-KD transgenic lines, including gdch-38, was previously described by Lin et al. (2016). Line gdch-38 was selected for analysis in the present study since in Lin et al. (2016) it had shown a more consistent photorespiratory-deficient phenotype under different O₂:CO₂ growing and measuring conditions compared with the other two gdch-KD lines. Untransformed O. sativa cv. IR64 line A009 (WT) was used as negative control for comparison with the gdch-KD line.

Plant growth conditions

Two batches of 10 transgenic gdch-38 line (T₄ generation) and 10 WT plants of O. sativa cv. IR64 were grown consecutively in a controlled-environment growth chamber (G_ch; Bigfoot series, BioChambers Inc., Winnipeg, MB, Canada) at the School of Biological Sciences at Washington State University, Pullman, WA (USA). All plants were individually grown in 4 liter free drainage pots; soil, irrigation, and fertilization were as in Giuliani et al. (2013).

The daily photoperiod was 14 h, from 8.00 h to 22.00 h standard time. Light was provided by F54T5/841HO Fluorescent 4100 K and 40 W halogen incandescent bulbs (Philips) and was supplied in a bell-shaped pattern; that is, with increasing photosynthetic photon flux density (PPFD) during the first 2 h, a maximum PPFD of 600 μmol photons m⁻² s⁻¹ incident on the plant canopy for 10 h, and decreasing PPFD in the last 2 h. Air temperature (t_air) was set at 22 °C in the dark period; after switching on the light, t_air tracked the PPFD pattern; that is, it ramped during the first 2 h from 22 °C to 26 °C, then 26 °C for 10 h, and decreased to 22 °C in the last 2 h photoperiod. Air relative humidity was maintained at ~70%, corresponding to a maximum air vapor pressure deficit (VPD) of ~1.6 kPa. During the light period, the CO₂ partial pressure (pCO₂) in the G_ch atmosphere was elevated to 184.2 Pa (2000 μmol mol⁻¹). The ¹³C composition of the atmospheric CO₂ during the light period (δ¹³C_Gch) was −41.6‰ and −30.6‰ for the first and second batch of grown plants, respectively. The δ¹³C_Gch was determined as described in Supplementary Methods S1 at JXB online, and was a proxy of the ¹³C composition of the CO₂ in the tank used (during the second plant growing cycle a new tank was needed and no tank with ¹³CO₂ composition comparable with the previous one was available).

Leaf biochemical analysis

Protein content

Protein immunoblot analysis was performed to determine the leaf abundance of GDC H-, P-, and T-subunits in fully expanded leaves of 4- to 5-week-old transgenic gdch-KD and WT plants. For each genotype, two separate protein extractions were performed, each one using the leaf tissue collected from two plants, according to Koteyeva et al. (2015). Protein concentration was determined for each extract with an RC DC protein quantification kit (Bio-Rad, Hercules, CA, USA) and 20 µg of protein per extract were separated by 10% (w/v) SDS–PAGE for the GDC P-subunit or 15% (w/v) for GDC H- and T-subunits. Proteins were then transferred to a nitrocellulose membrane and immunoblots (n=2 for both gdch-KD and WT) were performed according to Koteyeva et al. (2015) with primary antibodies for anti-Pisum sativum L. GDC H-, P-, and T-subunits (1:10 000) raised in rabbit (courtesy of Dr D. Oliver, Iowa State University). The L-subunit was not detected because antibodies were unavailable. The band intensities were quantified with ImageJ 1.37 software (NIH, USA).

Malate content

The leaf portions used for photosynthesis analysis in gdch-KD and WT plants (n=5) were sampled immediately after the leaf–atmosphere gas exchange measurements and frozen in liquid N₂. Malate content per unit leaf surface area (mmol malate m⁻²) was then determined with a spectrophotometry-based assay as described by Hatch (1979), with modifications by Edwards et al. (1982).

Leaf physiological analysis

Coupled measurements of leaf–atmosphere CO₂, H₂O, and ¹³CO₂ exchange

Measurements were performed in Pullman, WA, USA with a mean atmospheric pressure of 92.1 kPa. Two LI-6400XT portable gas analyzers (LI-COR Biosciences, Lincoln, NE, USA; detecting ¹²CO₂) operating as open systems were coupled to a tunable diode laser absorption spectroscope, which detects ¹²CO₂ and ¹³CO₂ isotopologs (TDLAS model TGA200A, Campbell Scientific, Inc., Logan, UT, USA; Bowling et al., 2003; Barbour et al., 2007; Ubierna et al., 2011; Stutz et al., 2014; Sun et al., 2014). Additional technical information on the system setup are available in Supplementary Methods S2.

For the leaf photosynthesis measurements, each LI-COR was equipped with a 2×3 cm leaf chamber (L_ch) assembled with an LED light source (6400-02B; LI-COR Biosciences). Alternatively, leaf dark respiration measurements were performed using an 8×10 cm custom-built L_ch having an adaxial glass window, and with a volume of ~100 cm³ (Barbour et al., 2007, based on Sharkey et al., 1985). The chamber had a hollowed stainless steel frame sealed with a closed-cell foam gasket and was connected to a circulating water bath for temperature control in the lumen. Before dark respiration measurements, the leaf portions included in the L_ch were exposed to the light, which was supplied by a LI-COR 6400-18 light source placed adjacent to the glass window.

Protocol for coupled measurements of leaf–atmosphere CO₂, H₂O, and ¹³CO₂ exchange

The mid to distal portions of two fully expanded leaves from the same stem on 4- to 5-week-old plants (n=4 for gdch-KD; n=5 for WT) grown under δ¹³C_Gch of −41.6‰ were used for leaf photosynthetic measurements. The leaves were positioned to cover the 6 cm² L_ch section area. Measurements were taken from 10.00 h until 16.00 h standard time under an O₂ partial pressure (pO₂) of 18.4 kPa (approximately the current atmospheric pO₂) and 1.84 kPa, pCO₂ (C_a) of 27.6 Pa, and ¹³C composition of CO₂ (from a pressurized tank) entering the L_ch (δ_in) of −48.0‰. PPFD was set at 1500 µmol photons m⁻² s⁻¹, leaf temperature (t_leaf) at 25 °C, and leaf to air VPD was kept between 1.0 kPa and 1.5 kPa. The airflow rate through the LI-COR system was 300 µmol s⁻¹ (~0.48 l min⁻¹). In particular, a C_a below current ambient pCO₂ (which was ~37 Pa) was chosen to amplify, under a pO₂ of 18.4 kPa, the signals of the photorespiratory-deficient phenotype in the gdch-KD plants compared with the WT.

Under each experimental O₂ condition, leaf portions were acclimated for ~30 min and data were recorded for ~30–40 min. The rate of net CO₂ assimilation per unit (one side) leaf surface area (A, µmol CO₂ m⁻² s⁻¹), stomatal conductance to CO₂ diffusion (g_sC, μmol CO₂ m⁻² s⁻¹ Pa⁻¹), intercellular pCO₂ (C_i, Pa), and the ratio C_i/C_a were determined.

For leaf dark respiration measurements, gdch-KD and WT plants (n=4) grown at a δ¹³C_Gch of both −41.6‰ and −30.6‰ were used. Two plants per day (one gdch-KD and one WT) were taken out of the G_ch at 9.30 h standard time and the mid to distal portions of 8–9 fully expanded leaves, similar to those used for the photosynthetic analysis, were enclosed in the custom-built L_ch to cover the section area of ~76 cm². Leaf portions were first exposed to a PPFD of 750 µmol photons m⁻² s⁻¹ for 20 min, 500 µmol photons m⁻² s⁻¹ for 15 min (at t_leaf of 25 °C), and 100 µmol photons m⁻² s⁻¹ for 5 min (at t_leaf of 30 °C). Measurements were taken under a pO₂ of 1.84 kPa or 18.4 kPa for plants grown at a δ¹³C_Gch of −41.6‰ or −30.6‰, respectively. C_a was set at 35.0 Pa, and the airflow rate through the LI-COR was changed from 700 µmol s⁻¹ to 500 µmol⁻¹, and from 500 µmol s⁻¹ to 350 µmol s⁻¹ tracking the decreasing PPFD. A CO₂ cartridge from a set of cartridges with δ¹³C from −6.2‰ to −4.8‰ was used, one per day, as CO₂ source (the mean δ_in for all experimental conditions is shown in Supplementary Table S1). The different (higher) δ¹³CO₂ composition entering the L_ch with respect to the G_ch (−41.6‰) was chosen to have the leaf carbon assimilates produced in the L_ch with dissimilar (higher) δ¹³C signatures compared with those previously produced in the G_ch. After 40 min of leaf light exposure, darkness was imposed in the L_ch. Leaf CO₂ evolution was measured at a pO₂ of 18.4 kPa and t_leaf of 30 °C for 195 min to determine the dynamics of the dark respiration rate per unit (one side) of leaf surface area (R_d, µmol CO₂ m⁻² s⁻¹) and corresponding δ¹³C (δ¹³C_Rd, ‰). The t_leaf was set at 30 °C to enhance the precision of the dark measurements. Additionally, three plants (n=3) of the gdch-KD line and of the WT were taken out of the growth chamber at 12.00 h standard time 3 d after their use for measurements, and darkened at 25 °C for 24 h. Subsequently, leaf dark CO₂ evolution was measured at a t_leaf of 30 °C and a pO₂ of 18.4 kPa to determine R_d(24h) (µmol CO₂ m⁻² s⁻¹) and δ¹³C_Rd(24h) (‰). The blade portions used for dark measurements on WT and gdch-KD plants were sampled and dried in a ventilated oven at 55 °C for 48 h to determine leaf dry mass per (one side) unit of leaf surface area (LMA, g m⁻²).

For each gdch-KD and WT plant used for leaf photosynthesis measurements, the ¹³C signature of leaf dry matter (δ¹³C_dm, ‰) and leaf total N content as a fraction (%) of dry matter were determined as described in Supplementary Methods S3, and the leaf total N content per unit leaf surface area (g m⁻²) was calculated. The descriptions, values, and units of abbreviations and symbols are listed in Table 1.

Table 1.

Description of the abbreviations, symbol, value (as in Evans and von Caemmerer, 2013), and unit of the environmental parameters and leaf variables used in this study

Abbreviation	Description
Gch	Growth chamber
GDC	Glycine decarboxylase complex
gdch-KD	Transgenic GDC H-subunit knockdown
Lch	Leaf chamber
LEDR	Light-enhanced dark respiration
NH3	Ammonia
NH4+	Ammonium cation
PDH	Pyruvate dehydrogenase
RuBP	Ribulose 1,5-bisphosphate
TCA	Tricarboxylic acid
2-PG	2-phosphoglycolate
3-PGA	3-phosphoglycerate
Symbol	Environmental parameters/leaf variables	Value and unit
A	Net CO2 assimilation rate per unit (one side) leaf surface area	µmol CO2 m−2 s−1
a	13C fractionation during CO2 diffusion (in air) through stomata	4.4‰
b 3	Rubisco 13C fractionation	29.0‰
C a	CO2 mole fraction or CO2 partial pressure set in the leaf chamber	µmol mol−1; Pa
C c	CO2 mole fraction or CO2 partial pressure in the chloroplast	µmol mol−1; Pa
C i	CO2 mole fraction or CO2 partial pressure in the intercellular air space	µmol mol−1; Pa
C in	CO2 mole fraction entering the leaf chamber	µmol mol−1
C out	CO2 mole fraction leaving the leaf chamber	µmol mol−1
C s	CO2 mole fraction at the leaf surface	µmol mol−1
f	Photorespiratory 13CO2 fractionation	‰
g m	Mesophyll conductance to CO2 diffusion from the substomatal cavity to the chloroplast stroma	µmol CO2 m−2 s−1 Pa−1
g sC	Stomatal conductance to CO2 diffusion	µmol CO2 m−2 s−1 Pa−1
LMA	Leaf dry mass per (one side) unit surface area	g m−2
pCO2	Partial pressure of CO2	Pa
pO2	Partial pressure of O2	kPa
PPFD	Photosynthetic photon flux density	µmol photons m−2 s−1
R d	Dark respiration rate per unit (one side) leaf surface area	µmol CO2 m−2 s-1
R d(24h)	R d after 24 h dark	µmol CO2 m−2 s−1
R d(30min)	R d after 30 min dark	µmol CO2 m−2 s−1
R d(3h)	R d after 3 h dark	µmol CO2 m−2 s−1
R d(6min)	R d after 6 min dark	µmol CO2 m−2 s−1
R L	Light mitochondrial non-photorespiratoy respiration rate per unit (one side) leaf surface area	µmol CO2 m−2 s−1
t	Correction factor for ternary effects	‰
t air	Air temperature	°C
t leaf	Leaf temperature	°C
VPD	Vapor pressure deficit	kPa
α	Moles of CO2 released in the photorespiratory pathway per mole of O2 reacting with RuBP	mol CO2 mol−1 O2
Δe	13C discrimination associated wtih mitochondrial non-photorespiratory respiration	‰
Δf	13C discrimination associated with photorespiration	‰
Δgm	13C discrimination associated with mesophyll conductance to CO2 diffusion	‰
Δi	13C discrimination due to carboxylation, boundary layer and stomatal CO2 diffusion	‰
Δo	Observed (instantaneous) leaf net discrimination against 13CO2 in the light	‰
Γ	CO2 compensation point	µmol mol−1; Pa
Γ*	CO2 compensation point in absence of mitochondrial non-photorespiratory respiration	µmol mol−1; Pa
δin	δ13C of CO2 entering the leaf chamber	‰
δout	δ13C of CO2 leaving the leaf chamber	‰
δRdGch_substr	Fractional contribution of respiratory substrates from Gch carbon assimilates to δ13C of dark-evolved CO2	‰/‰
δRdLch_substr	Fractional contribution of respiratory substrates from Lch carbon assimilates to δ13C of dark-evolved CO2	‰/‰
δ13C	13C composition of CO2	‰
δ13Cdm	13C signature of leaf dry matter	‰
δ13CGch	13C composition of atmospheric CO2 in the growth chamber during the photoperiod	‰
δ13CLch_Ph	Representative δ13C of carbon assimilates produced in the Lch	‰
δ13CRd	δ13C of CO2 evolved by leaves in the dark	‰
δ13CRd(24h)	δ13C of CO2 evolved by leaves after 24 h dark	‰
δ13CRd(30min)	δ13C of CO2 evolved by leaves after 30 min dark	‰
δ13CRd(3h)	δ13C of CO2 evolved by leaves after 3 h dark	‰
δ13CRd(6min)	δ13C of CO2 evolved by leaves after 6 min dark	‰

Leaf net ¹³CO₂ discrimination in the light and mesophyll conductance to CO₂ diffusion

The observed leaf net discrimination against ¹³CO₂ in the light (Δ_o, ‰) was calculated by mass balance from the TDLAS measurements according to Evans et al. (1986). Under photorespiratory conditions (18.4 kPa pO₂), the ¹³CO₂ fractionation for photorespiration (f, ‰) in the gdch-KD plants was calculated based on Evans and von Caemmerer (2013). Briefly, the value of f was determined by modeling the leaf net discrimination against ¹³CO₂ (Δ_o) as a function of the ¹³C discrimination fractions associated with CO₂ diffusion from the atmosphere to the intercellular air space and with carboxylation (Δ_i), with CO₂ diffusion in liquid phase to chloroplast stroma (Δ_gm), mitochondrial non-photorespiratory respiration (Δ_e), and photorespiration (Δ_f). The equation Δ_o=Δ_i−Δ_gm−Δ_f−Δ_e can be rearranged so that Δ_f=Δ_i−Δ_o−Δ_gm−Δ_e and f can be estimated by substituting Δ_f with ${\mathrm{\Delta }}_{\mathrm{f}}=\frac{1+t}{1-t}\left(f\frac{{\mathrm{\Gamma }}^{\ast }}{C_{\mathrm{a}}}\right)$to get $\frac{1+t}{1-t}\left(f\frac{{\mathrm{\Gamma }}^{\ast }}{C_a}\right)={\mathrm{\Delta }}_{\mathrm{i}}-{\mathrm{\Delta }}_{\mathrm{o}}-{\mathrm{\Delta }}_{{\mathrm{g}}_{\mathrm{m}}}-{\mathrm{\Delta }}_{\mathrm{e}}$. An f value of 16.2‰ was taken from Evans and von Caemmerer (2013) and assumed for WT plants. The input parameters needed to calculate f include the leaf mitochondrial respiration rate in the light (R_L, µmol CO₂ m⁻² s⁻¹), the CO₂ compensation point in the absence of mitochondrial non-photorespiratory respiration (Γ*, μmol mol⁻¹), and mesophyll CO₂ conductance (g_m, mol CO₂ m⁻² s⁻¹). Values of R_L at a t_leaf of 25 °C were modeled for both genotypes from the corresponding R_d at 30 °C after 3 h in the dark [R_d(3h), μmol CO₂ m⁻² s⁻¹] following leaf photosynthesis under atmospheric pO₂ of 18.4 kPa using the temperature response function in Bernacchi et al. (2001). The Γ* was modeled based on von Caemmerer (2000), as described in Supplementary Methods S4, and was significantly different (P<0.05) between WT and gdch-KD plants, 45.0±1.7 SE (n=4) μmol mol⁻¹ and 53.3±0.6 SE (n=3) μmol mol⁻¹, respectively. Finally, g_m was estimated based on leaf–atmosphere CO₂ and ¹³CO₂ exchange data, according to Evans and von Caemmerer (2013). Specifically, ¹³C-based g_m was calculated in the gdch-KD and WT plants at 1.84 kPa pO₂, but only in WT plants under 18.4 kPa pO₂, using an f value of 16.2‰. The ¹³C-based g_m cannot be calculated in gdch-KD plants at 18.4 kPa pO₂ because g_m and f are not independent variables in the applied procedure. Therefore, at 18.4 kPa, the g_m values of gdch-KD plants were set the same as for the WT. This assumes that the ¹³C-based g_m integrates the within-leaf resistances affecting CO₂ movement across the cell wall, plasma membrane, and the chloroplast membranes, and that this cumulative resistance does not differ between gdch-KD and WT plants. This assumption is supported by the fact that the ¹⁸O-based g_m, which was determined by analysis of leaf–atmosphere ¹⁸O exchange according to Ubierna et al. (2017), Kolbe and Cousins (2018), and Sonawane and Cousins (2019), was not significantly different between the gdch-KD and WT plants at 18.4 kPa pO₂ (Supplementary Table S2). The ¹⁸O-based g_m is not strictly associated with the biochemistry of photosynthesis as is the ¹³C-based g_m and therefore cannot be used to estimate f. The values of ¹³C-based g_m for gdch-KD and WT plants at each pO₂ were used to calculate the corresponding pCO₂ in the chloroplasts (C_c, Pa) by applying Fick’s first law.

The Γ* was defined in terms of Rubisco kinetic properties according to Jordan and Ogren (1984), and the estimate of CO₂ released per O₂ reacting with RuBP (α) was determined for the gdch-KD plants versus α set equal to 0.5 in the WT as described in Supplementary Methods S5. A sensitivity analysis for the dependency of f on Γ* and α is also described in Supplementary Methods S5.

¹³C composition of leaf dark-evolved CO₂ and contributions of leaf chamber and growth chamber assimilates to substrates feeding leaf dark respiration

The ¹³C composition of the dark-evolved CO₂ determining R_d (δ¹³C_Rd, ‰) was calculated according to Barbour et al. (2007) as described by Evans et al. (1986).

The substrates feeding leaf dark respiration were from carbon assimilates produced in the L_ch and in the G_ch. Given δ¹³C_Rd(i) as the mean values of δ¹³C for dark-evolved CO₂ at time i from light–dark transition, the fractional contribution of L_ch assimilates to δ¹³C_Rd(i) (^δRdL_{ch_substr(i)}, ‰/‰) was calculated for gdch-KD and WT plant types after leaf photosynthesis under both O₂ levels as

$${\displaystyle \begin{array}{l}{\displaystyle {}^{{}^{\delta \mathrm{R}\mathrm{d}}}{\mathrm{L}}_{\mathrm{c}\mathrm{h}\mathrm{\_}\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}(i)}=\frac{({\mathrm{\delta }}^{13}{C}_{\mathrm{R}\mathrm{d}\left(i\right)}-{\mathrm{\delta }}^{13}{C}_{\mathrm{R}\mathrm{d}\left(24\mathrm{h}\right)})}{({\mathrm{\delta }}^{13}{C}_{\mathrm{L}\mathrm{c}\mathrm{h}\mathrm{\_}\mathrm{P}\mathrm{h}}-{\mathrm{\delta }}^{13}{C}_{\mathrm{R}\mathrm{d}\left(24\mathrm{h}\right)})}}\end{array}}$$ (1)

where δ¹³C_Rd(i) was determined by steps of 3 min over 195 min in the dark; δ¹³C_Rd(24h) is the mean δ¹³C_Rd after 24 h in the dark as shown in Supplementary Table S1; and δ¹³C_{Lch_Ph} (‰) is the representative δ¹³C of gdch-KD or WT carbon assimilates produced in the L_ch at a pO₂ of 1.84 kPa or 18.4 kPa before the light–dark transition (values are shown in Supplementary Table S1). The assumptions underlying Equation 1 and the calculation of δ¹³C_{Lch_Ph} are reported in Supplementary Methods S6.

Based on the total fractional contributions of L_ch and G_ch carbon assimilates to δ¹³C_Rd equal to 1.0, the complementing fractional contribution of G_ch assimilates to δ¹³C_Rd(i) [^δRdG_{ch_substr(i)}, ‰/‰] was calculated for both plant types after leaf photosynthesis under both O₂ levels as

$$\begin{array}{l}{}^{\delta \text{Rd}}{\text{G}}_{\text{ch\_substr}(i)}=1{-}^{\text{δRd}}{\text{L}}_{\text{ch\_substr}(i)}\hfill \end{array}$$ (2)

In addition, to make a combined analysis of the data collected in the two O₂ experimental conditions possible, the δ¹³C_Rd generated from plants grown at the more depleted ^δ13C_Gch were edited to cancel out the bias in the δ¹³C_Gch effect on δ¹³C_Rd with respect to the other batch of plants. In particular, the δ¹³C_Rd following leaf photosynthesis at the lower O₂ experimental level were edited through the procedure described in Supplementary Methods S7.

Leaf CO₂ compensation points in the presence of R_L

Leaf–atmosphere gas exchange measurements were taken with an LI-6400XT portable gas analyzer equipped with the 2×3 cm L_ch on gdch-KD and WT plants (n=4) at a PPFD of 1500 µmol photons m⁻² s⁻¹, t_leaf of 25 °C, leaf to air VPD between 1.0 kPa and 1.5 kPa, C_a decreasing from 35.0 Pa to 3.7 Pa, and at a pO₂ of 1.84 kPa or 18.4 kPa. For each leaf, a least square regression analysis of the response of A (µmol CO₂ m⁻² s⁻¹) to C_i (Pa) was applied to the initial slope (for C_i≤9.2 Pa) to determine the CO₂ compensation point in the presence of R_L (Γ, Pa).

Statistical analysis

Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA). A linear mixed effects model (PROC MIXED) was used with plant type (gdch-KD and WT) and O₂ level (pO₂ of 1.84 kPa and 18.4 kPa) as fixed factors and leaves as random factor nested within plant type. The effects of plant type, O₂ level, and plant type×O₂ level interaction on A, g_sC, C_i, C_i/C_a, Δ_o, C_c, Γ, R_d(6m), δ¹³C_Rd(6m), R_d(30m), δ¹³C_Rd(30m), R_d(3h), and δ¹³C_Rd(3h) were assessed. A PROC MIXED procedure was applied as a one-way ANOVA to determine the plant type (fixed factor) effect on the following traits: total N content, malate content, Γ*, Δ_o, Δ_i, Δ_i−Δ_o, Δ_gm, Δ_e, Δ_f, R_d(24h), δ¹³C_Rd(24h), LMA, and δ¹³C_dm. A one-sample t-test (P<0.05) was applied to test the difference of f or α modeled for the gdch-KD plants compared with a constant f or α value assumed in the WT. A two-sample t-test (P<0.05) was applied to test the difference between gdch-KD and WT g_m at a pO₂ of 1.84 kPa, and WT g_m at the two O₂ levels. For each plant type, a three-parameter non-linear model was fit to the R_d and δ¹³C_Rd responses determined over the 195 min in the dark after leaf photosynthesis at each O₂ experimental level. In particular, the δ¹³C_Rd values associated with the lower O₂ level had been first edited as described in Supplementary Methods S7, and then used for the analysis. The fitting model y=θ₁e^–θ2x+θ₃ was employed where x are minutes from 0 to 195 by steps of three, and y are R_d or δ¹³C_Rd values; θ₁, θ₂, and θ₃ are the range, slope, and lower asymptote (or floor) parameters, respectively, which were determined using non-linear least squares with the iterative Gauss–Newton algorithm. Specifically, for R_d or δ¹³C_Rd responses, the range parameter corresponds to the difference between initial and lower asymptote values, and the slope is the exponential rate of change. An extra sum of squares F-test was applied to define the significance (P<0.05) of the effects of plant type and O₂ level (main effects), and plant type×O₂ level interactions on the three parameters of R_d or δ¹³C_Rd fitting models.

Results

Leaf GDC protein and malate content

Leaves of gdch-KD plants had 21% (±2 SE) H-protein content compared with the WT, while the P- and T-protein content was 77% (±6 SE) and 83% (±2 SE) of that of the WT, respectively (Fig. 1; n=2). Malate content in leaf samples taken immediately after measurements of leaf photosynthesis were 0.49±0.08 mmol m⁻² and 0.38±0.02 mmol m⁻² (mean ±SE; n=5) in gdch-KD and WT plants, respectively (P=0.29).

Fig. 1.

Immunoblot analysis for GDC P-, T-, and H-subunits in mature leaves of gdch-KD compared with WT plants. The protein molecular weight of each subunit (kDa) is shown. Subunit protein abundances for gdch-KD plants are mean percentage values of the WT (n=2).

Leaf physiological analysis

Leaf photosynthetic responses

There was no observable difference in growth phenotypes between the gdch-KD and WT plants when they were grown under 184.2 Pa pCO₂ (2000 μmol mol⁻¹). However, at approximately current ambient CO₂ and O₂ partial pressures, the net rate of CO₂ assimilation (A) was lower in the gdch-KD compared with the WT but there was no significant difference in A between plant types under low photorespiratory conditions, when pO₂ was reduced to 1.84 kPa (Table 2). There was, however, a significant effect of O₂ level on g_sC (negative effect) and C_i/C_a (positive effect) but not a plant type effect (Table 2). There was a significant plant type×O₂ level interaction on Δ_o, which showed higher values for the gdch-KD compared with the WT at a pO₂ of 18.4 kPa, but no difference at a pO₂ of 1.84 kPa (Table 2). There was no significant plant type effect on g_m at a pO₂ of 1.84 kPa (P=0.586), and no O₂ effect on g_m in the WT (P=0.701; Table 2). There was, however, a significant effect of plant type on C_c, which showed comparable values in the gdch-KD and WT plants at 1.84 kPa pO₂ and higher values in the gdch-KD plants compared with the WT at 18.4 kPa pO₂ (modeled based on equal g_m in both transgenic and WT plants). In addition, O₂ level had a positive effect on C_c (Table 2). The Γ in the gdch-KD compared with WT plants was significantly lower under 1.84 kPa pO₂ but higher under 18.4 kPa pO₂ (Table 3). There was no significant difference in leaf N content between plant types, with means of 2.30±0.08 SE g m⁻² and 2.31±0.14 SE g m⁻² in gdch-KD and WT leaves, respectively (n=4).

Table 2.

Leaf photosynthetic traits estimated on gdch-KD and WT plants under approximately current ambient and below current ambient O₂ levels (pO₂ of 18.4 kPa and 1.84 kPa, respectively) at C_a of 27.6 Pa.

Plant-type	pO2	A	g sC	C i	C i /C a	g m a	C c	Δo
	(kPa)	(µmol CO2 m−2 s−1)	(µmol CO2 m−2 s−1 Pa−1)	(Pa)		(µmol CO2 m−2 s−1 Pa−1)	(Pa)	(‰)
gdch-KD	1.84	24.7±1.4	3.47±0.53	18.1±0.8	0.66±0.03	4.56±0.43	12.6±1.4	14.9±0.7
	18.4	6.3±0.3	1.37±0.11	22.1±0.2	0.80±0.01		20.6±0.2	23.7±0.5
WT	1.84	21.6±1.2	2.78±0.43	18.5±0.9	0.67±0.03	3.80±0.76	12.6±0.5	14.3±0.8
	18.4	14.3±0.8	2.45±0.31	20.5±0.5	0.74±0.02	4.07±0.14	17.0±0.6	17.8±0.3
Significance	Plant type	P=0.050	P=0.629	P=0.411	P=0.411	–	P=0.031	P=0.003
	pO2	P<0.001	P=0.018	P=0.003	P=0.003	–	P=0.000	P<0.001
	Plant type×pO2	P=0.002	P=0.057	P=0.171	P=0.171	–	P=0.033	P=0.003

Values are the mean±SE (n=4). Significance (P<0.05) of the effects of plant type, pO₂, and plant type×pO₂ interaction were evaluated by SAS PROC MIXED.

^a No significant differences were evaluated by a two sample t-test (significance for P<0.05) between the gdch-KD and WT g_m values at a pO₂ of 1.84 kPa (P=0.586), and the WT g_m values at the two pO₂ values (P=0.701).

Table 3.

CO₂ compensation points (Γ) determined under low photorespiratory (1.84 kPa pO₂) and photorespiratory (18.4 kPa pO₂) conditions on gdch-KD and WT plants

Plant-type	pO2	Γ
	(kPa)	(Pa)
gdch-KD	1.84	0.25±0.06
	18.4	5.48±0.04
WT	1.84	0.62±0.12
	18.4	4.54±0.18
Significance	Plant type	P=0.055
	pO2	P<0.001
	Plant type×pO2	P=0.001

Values are the mean ±SE (n=3 for gdch-KD at a pO₂ of 18.4 kPa; n=4 otherwise). Significance (P<0.05) for the effects of plant type, pO₂, and plant type×pO₂ interaction was evaluated by SAS PROC MIXED.

Δ_o plotted versus C_i/C_a showed a similar response in the gdch-KD and WT plants at 1.84 kPa pO₂ but was significantly greater in the gdch-KD compared with WT plants at 18.4 kPa pO₂ (Fig. 2; see Table 2 for statistical analysis). The gdch-KD plants had significantly lower Δ_gm, Δ_f, and Δ_e compared with the WT under a pO₂ of 18.4 kPa (Fig. 3A). Additionally, the gdch-KD plants had a significantly lower f with mean values of 3.4±0.5‰ SE (n=4) compared with 16.2‰ in the WT under approximately current ambient pO₂ (Fig. 3B; P<0.001). A significantly higher α was determined at 18.4 kPa pO₂ in gdch-KD plants, with a mean value of 0.59±0.01 SE (n=3), versus 0.5 assumed for the WT (P<0.01). There was a negative linear dependency of f on Γ* and on α (Supplementary Methods S5; Fig. S1A and B, respectively); however, there was a positive linear dependency of f on g_m and R_L, with a greater sensitivity to g_m (Supplementary Fig. S2A ands B, respectively).

Fig. 2.

Leaf ¹³CO₂ net discrimination in the light (Δ_o) versus C_i/C_a under a pO₂ of 1.84 kPa and 18.4 kPa for individual gdch-KD and WT plants. The line represents the leaf ¹³CO₂ net discrimination modeled in relation to C_i/C_a as Δ¹³C_mod=a+(b₃−a)×C_i/C_a (Farquhar et al., 1982) where a=4.4‰ and b₃=29.0‰. Δ¹³C_mod is a proxy of Δ_i as described by Evans and von Caemmerer (2013). Open symbols are for gdch-KD and filled symbols for WT plants. Circles are for a pO₂ of 1.84 kPa and squares for a pO₂ of 18.4 kPa.

Fig. 3.

Leaf ¹³CO₂ net discrimination and discrimination fractions in the light, and ¹³CO₂ photorespiratory fractionation for gdch-KD versus WT plants determined based on Evans and von Caemmerer (2013). (A) Observed leaf net ¹³CO₂ discrimination in the light (Δ_o), and modeled ¹³C discrimination fractions for gdch-KD (n=3) and the WT (n=4) at an atmospheric pO₂ of 18.4 kPa. Δ_i is the additive ¹³CO₂ discrimination during CO₂ diffusion from atmosphere to intercellular air space and due to carboxylation; Δ_i−Δ_o is comprised of three terms: Δ_gm, which is the ¹³CO₂ fractionation fraction during CO₂ diffusion in the liquid phase to chloroplast stroma, and Δ_e and Δ_f, which are the ¹³C fractionation fractions associated with light mitochondrial non-photorespiratory respiration and photorespiration, respectively. Δ_f was calculated as Δ_f=Δ_i−Δ_o−Δ_gm−Δ_e. Values are mean ±SE. (B) ¹³CO₂ fractionation for photorespiration (f) in gdch-KD plants calculated at an atmospheric pO₂ of 18.4 kPa from ${\mathrm{\Delta }}_{\mathrm{f}}=\frac{1+t}{1-t}\left(f\frac{{\mathrm{\Gamma }}^{\ast }}{C_{\mathrm{a}}}\right)$ versus f of 16.2‰ in the WT. Values for gdch-KD plants are the mean ±SE (n=3). Significance (P<0.05) of the effect of plant type on the variables in (A) was evaluated by SAS PROC MIXED as a one-way ANOVA; * for 0.01<P<0.05; ** for P<0.01. Significance in (B) was evaluated by one-sample t-test (P<0.05). ** for P<0.01.

Leaf dark respiration responses

In the gdch-KD and WT plants, the R_d showed a hyperbolic decrease over the 3 h in the dark after leaf light exposure under different O₂ levels, with a noticeable rapid decline in the first hour; however, R_d was higher following leaf photosynthesis at a pO₂ of 18.4 kPa compared with 1.84 kPa. (Fig. 4). A significant positive O₂ effect on R_d responses of gdch-KD and WT plants was inferred based on significantly higher floor (μmol CO₂ m⁻² s⁻¹; P<0.0001) and range (μmol CO₂ m⁻² s⁻¹; P<0.0001) parameters after leaf photosynthesis at pO₂ of 18.4 kPa compared with 1.84 kPa in the non-linear model fit to the R_d responses (Supplementary Tables S3, S4). In particular, based on the spot measurements, a significant positive O₂ effect was determined on R_d(6min) (together with a significant plant type effect), R_d(30min), and R_d(3h) (Table 4). A significant plant type effect on R_d responses was inferred based on a statistically larger range (P=0.023) and less steep rate of exponential change (slope, μmol CO₂ m⁻² s⁻¹ min⁻¹; P<0.0001) for the gdch-KD versus WT plants (Supplementary Tables S3, S4). After leaf photosynthesis at a pO₂ of 18.4 kPa, a significant plant type effect on R_d (with higher R_d determined in the gdch-KD plants versus the WT; see Fig. 4) was driven by the significantly less steep R_d slope (P<0.001) in gdch-KD compared with WT plants. In addition, following leaf photosynthesis at a pO₂ of 18.4 kPa, a change in R_d for ~75% of the R_d range occurred in WT plants within the first 30 min after light–dark transition; in contrast, this fractional variation took ~90 min in the gdch-KD plants (Fig. 4). The mean values of R_L inferred from R_d(3h) were 0.59±0.03 SE μmol CO₂ m⁻² s⁻¹ for gdch-KD and 0.56±0.03 SE μmol CO₂ m⁻² s⁻¹ for WT plants after leaf photosynthesis under a pO₂ of 1.84 kPa (n=4). In contrast, R_L was 0.98±0.12 SE μmol CO₂ m⁻² s⁻¹ for gdch-KD and 0.82±0.09 SE μmol CO₂ m⁻² s⁻¹ for WT plants after leaf exposure to a pO₂ of 18.4 kPa (n=4). For the R_L values, a non-significant plant type effect and a significant effect of the O₂ level can be inferred from the significance of R_d(3h) (see Table 4). In addition, no significant difference in leaf dry mass per unit surface area (LMA) was determined between gdch-KD and WT plants, with values of 43.6±2.7 SE g m⁻² and 44.5±1.5 SE g m⁻² (n=4), respectively.

Fig. 4.

Dynamics of leaf dark respiration rate (R_d) determined during ~3 h in the dark on gdch-KD (open symbols) and WT (filled symbols) plants after leaf photosynthesis under a pO₂ of 1.84 kPa (circles) or 18.4 kPa (squares). Symbols correspond to the mean ±SE (n=4) determined every 3 min.

Table 4.

Leaf dark respiration rates (R_d) at 30 °C and ¹³CO₂ composition of dark-evolved CO₂ (δ¹³C_Rd) determined on gdch-KD versus the WT after 6 min [R_d(6min) and δ¹³C_Rd(6min); n=4], 30 min [R_d(30min) and δ¹³C_Rd(30min); n=4], 3 h [R_d(3h) and δ¹³C_Rd(3h); n=4], and 24 h [R_d(24h) and δ¹³C_Rd(24h); n=3] in the dark following leaf exposure to light under approximately current ambient and below current ambient O₂ levels (pO₂ of 18.4 kPa and 1.84 kPa, respectively)

Plant-type	pO2	R d(6min)	δ13CRd(6min)	R d(30min)	δ13CRd(30min)	R d(3h)	δ13CRd(3h)	R d(24h)	δ13CRd(24h)
	(kPa)	(µmol CO2 m-2 s-1)	(‰)	(µmol CO2 m−2 s−1)	(‰)	(µmol CO2 m−2 s−1)	(‰)	(µmol CO2 m−2 s−1)	(‰)
gdch-KD	1.84	1.48±0.05	−39.2±0.9*	1.04±0.06	−47.3±1.6*	0.81±0.04	−56.0±0.7*	0.79±0.02	−58.0±0.5*
WT	1.84	1.23±0.04	−40.6±1.2*	1.04±0.01	−47.2±1.7*	0.77±0.04	−54.1±0.7*	0.69±0.02	−58.6±1.0*
Significance								P=0.045	P=0.705
gdch-KD	18.4	2.59±0.29	−45.6±1.7	1.98±0.20	−54.1±0.9	1.34±0.17	−55.6±1.2	0.69±0.06	−58.1±0.1
WT	18.4	2.13±0.07	−43.2±0.9	1.47±0.08	−50.4±2.8	1.12±0.13	−52.5±1.8	0.74±0.08	−58.6±0.6
Significance	Plant type	P=0.042	P=0.723	P=0.110	P=0.349	P=0.215	P=0.132	P=0.596	P=0.410
	pO2	P=0.0001	P=0.032	P=0.0009	P=0.035	P=0.004	P=0.429	–	–
	Plant type×pO2	P=0.407	P=0.238	P=0.069	P=0.375	P=0.326	P=0.753	–

Values are the mean ±SE; the asterisks indicate means from δ¹³C_Rd values edited according to Supplementary Method S7. Significance (P<0.05) of the effects of plant type, pO₂, and plant type×pO₂ interaction was evaluated by SAS PROC MIXED. The effect of plant type on R_d(24h) and δ¹³C_Rd(24h) was evaluated at a pO₂ of 1.84 kPa or 18.4 kPa by one-way ANOVA (significance for P<0.05).

In gdch-KD and WT plants, the δ¹³C_Rd estimated over the 3 h after light–dark transition showed a negative hyperbolic pattern, with most of the δ¹³C_Rd variation occurring in the first 30 min (Fig. 5A, B). A tight positive correlation between R_d and δ¹³C_Rd over the 3 h dark period was determined after leaf photosynthesis at a pO₂ of 1.84 kPa for both plant types (r>0.90). After leaf photosynthesis at a pO₂ of 18.4 kPa, a positive correlation between R_d and δ¹³C_Rd with r=0.75 and r=0.78 was determined for gdch-KD and the WT, respectively. Statistical analysis of a non-linear model fit to the δ¹³C_Rd responses showed a significantly lower δ¹³C_Rd range after leaf photosynthesis at a pO₂ of 18.4 kPa (‰; P<0.0001) compared with 1.84 kPa pO₂. In contrast, the floor parameter was non-significantly different between the O₂ levels (Supplementary Tables S5, S6). These statistical results indicate a significant effect of the O₂ level during previous leaf light exposure on the δ¹³C_Rd values of both plant types over 3 h in the dark (with lower δ¹³C_Rd at a pO₂ of 18.4 kPa, and higher δ¹³C_Rd at a pO₂ of 1.84 kPa). The spot δ¹³C_Rd measurements also showed significantly lower δ¹³C_Rd(6min) and δ¹³C_Rd(30min) after leaf photosynthesis at a pO₂ of 18.4 kPa compared with 1.84 kPa (Table 4). There was no significant plant type effect on the δ¹³C_Rd over the 3 h in the dark (Table 4; Supplementary Table S6) and there was no difference for δ¹³C_dm between gdch-KD and the WT (Supplementary Table S1).

Fig. 5.

¹³CO₂ composition associated with R_d (δ¹³C_Rd) determined during ~3 h in the dark in gdch-KD (open symbols) and WT (filled symbols) plants. (A) δ¹³C_Rd after leaf photosynthesis under a pO₂ of 1.84 kPa edited (see Supplementary Methods S7) to remove the effect of a lower atmospheric δ¹³CO₂ compared with (B) while growing the plants. (B) Distributions of δ¹³C_Rd after leaf photosynthesis under a pO₂ of 18.4 kPa. Symbols correspond to the mean ±SE (n=4) determined every 3 min.

Over the 3 h after the light–dark transition, the fractional contribution of L_ch assimilates to δ¹³C_Rd (^δRdL_{ch_substr}, ‰/‰) showed a decreasing hyperbolic pattern for both gdch-KD and WT plants (Fig. 6), with no significant differences between plant types and O₂ levels.

Fig. 6.

Distributions over ~3 h in the dark of the fractional contributions (total contribution equal to 1.0) to δ¹³C_Rd of recent L_ch carbon assimilates (^δRdL_ch__substr, ‰/‰) and G_ch assimilates (^δRdG_ch__substr, ‰/‰) for gdch-KD (open symbols) and WT (filled symbols) after leaf light exposure at a pO₂ of (A) 1.84 kPa and (B) 18.4 kPa. The first values are at 3 min after light–dark transition. Symbols correspond to the mean determined every 3 min (n=4). Continuous lines represent logarithmic trend lines (R² >0.90) for gdch-KD (lower) and the WT (higher), respectively.

Discussion

Altered photorespiratory metabolism and leaf photosynthetic traits

Based on leaf protein analysis, gdch-KD plants had ~21, 77, and 83% of GDC H-, P-, and T-protein abundance, respectively, compared with the WT. Previous studies reported how GDC activity is linearly correlated with H-protein accumulation (Wingler et al., 1997; Lin et al., 2016). Additionally, in agreement with Lin et al. (2016), the gdch-KD plants in the current study showed an expected photorespiratory-deficient phenotype. Under photorespiratory conditions, a disruption of the photorespiratory pathway negatively affects the rate of net CO₂ assimilation (A) due to accumulation of metabolites that inhibit the Calvin–Benson cycle and restrict RuBP regeneration (Wingler et al., 2000). Specifically, leaf glycine level is a sensitive indicator of altered photorespiratory carbon flow (Blackwell et al., 1988; Timm et al., 2012). For example, gdch-KD mutants of Arabidopsis, barley, and rice had substantial increases in leaf contents of glycine under ambient pO₂ (Bauwe and Kolukisaoglu, 2003; Lin et al., 2016). This accumulation of glycine and its precursors (P-glycolate, glycolate, and glyoxylate) in the gdch-KD plants has been suggested to alter photorespiratory carbon metabolism (Peterhansel et al., 2010, 2013a). These changes have important implications for understanding and modeling leaf carbon metabolism, because they may influence the stoichiometry of CO₂ released per oxygenation reaction (α) and the CO₂ compensation point (Γ) (see Cousins et al., 2008, 2011; Walker and Cousins, 2013).

In the present study, the gdch-KD plants had greater Γ compared with the WT under 18.4 kPa pO₂, as previously reported by Lin et al. (2016). This may be partially due to enhanced R_L, which Igamberdiev et al. (2004) and Bykova et al. (2005) reported was needed to compensate for the lack of photorespiratory regulation of redox and energy balance (Igamberdiev et al., 2001). The increase in Γ in the gdch-KD plants could also be associated with a higher α leading to an increasing Γ* compared with the WT. It has been previously suggested that α increased in Arabidopsis mutants lacking peroxysomal hydroxypuruvate reductase (Cousins et al., 2008) and the peroxysomal malate dehydrogenase (Cousins et al., 2011). However, these previous publications did not determine whether these disruptions to photorespiration influenced leaf CO₂ isotope exchange.

Leaf net ¹³C discrimination in the light and photorespiratory ¹³C fractionation

Under leaf photorespiratory conditions, the change in leaf net discrimination against ¹³CO₂ (Δ_o) in the gdch-KD plants compared with WT plants was caused by a higher C_i/C_a, lower Δ_gm, greater Δ_e, and lower Δ_f (Fig. 3A). However, the lower Δ_gm in the gdch-KD plants with respect to the WT was due to a reduction in A, since WT g_m was applied to both plant types (see Δ_gm equation in Evans and von Caemmerer, 2013). In addition, a difference in Δ_e between gdch-KD and WT plants was related to proportional changes in R_L/(A+R_L) and (C_i−Γ*) (see Δ_e equation in Evans and von Caemmerer, 2013).

The term Δ_f is dependent on $\frac{{\mathrm{\Gamma }}^{\ast }}{C_{\mathrm{a}}}$, but, despite the higher Γ* in the gdch-KD relative to WT plants, it was lower in the transgenic plants caused by the lower f. In WT plants, f is primarily attributed to the ¹³C discrimination associated with the decarboxylation of glycine catalyzed by GDC (Tcherkez et al., 2004; Tcherkez, 2006). While Rooney (1988) determined an in vitro f of 7–8‰ for Glycine max (soybean), Igamberdiev et al. (2001, 2004) reported f for several species between 9.8 and 13.7‰, and Ghashghaie et al. (2003) reported an f of >9–11‰ for Senecio species. Additionally, Evans and von Caemmerer (2013) determined an in vivo f of 16.2‰ in Nicotiana tabacum. Based on Farquhar et al. (1982), and according to O’Leary (1988) and Tcherkez (2006),

$${\displaystyle \begin{array}{ll}{\displaystyle f=}& ({\mathrm{\delta }}^{13}{\mathrm{C}}_{\mathrm{g}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{i}\mathrm{n}\mathrm{e}}-{\mathrm{\delta }}^{13}{\mathrm{C}}_{\mathrm{p}\mathrm{h}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{p}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{d}}{\mathrm{\_}}_{\mathrm{C}\mathrm{O}2})\\ & /(1+{\mathrm{\delta }}^{13}{\mathrm{C}}_{\mathrm{p}\mathrm{h}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{p}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{\_}\mathrm{C}\mathrm{O}2}/1000)\end{array}}$$ (3)

where glycine is assumed to have the same ¹³C signature of recently fixed carbon. Therefore, an increase in δ¹³C of the released CO₂ during photorespiration corresponds to a linear decrease in f. There is also a negative linear dependency of f on Γ* and α, as shown in Supplementary Methods S5; Fig. S1A, B.

In C₃ plants, most of the CO₂ released by photorespiration tends to be through GDC (Badger, 1985; Bauwe et al., 2010). However, previous reports have suggested that alternative reactions can release CO₂ when the flux of glycolate into the photorespiratory cycle exceeds its metabolic capacity, or when the traditional photorespiratory pathway has been genetically disrupted (Cousins et al., 2008, 2011; Timm et al., 2008; Peterhansel et al., 2013a). The GDC multienzyme system requires all subunits to function (Douce et al., 2001); in the present study, since a low level of the H-subunit was determined in gdch-KD plants, some residual activity for GDC is expected in the transgenic plants. A change in ¹³C fractionation associated with the knockdown of GDC activity is therefore unlikely because the products of the glycine decarboxylation reaction (NH₄⁺, NADH, and methylene-tetrahydrofolate) can be readily processed by downstream reactions in the glycolate pathway (Bauwe et al., 2010; Maurino and Peterhansel, 2010). Thus, the higher α and lower f in the gdch-KD compared with the WT suggest an increased flow of photorespiratory carbon through alternative decarboxylation reactions, independent of the GDC, and a buildup of photorespiratory metabolites.

Leaf dark respiration and ¹³C isotopic composition of dark-evolved CO₂

Leaves of C₃ plants in the first 30 min after light–dark transition largely respire metabolites (carbohydrates and organic acids) recently produced in the light (Cornic, 1973; Rademacher et al., 2002; Barbour et al., 2007; Werner et al., 2009; Werner and Gessler, 2011; Lehmann et al., 2015, 2016) and show high rates of CO₂ evolution (named as LEDR, light-enhanced dark respiration; Atkin et al., 1998). While the activity of pyruvate dehydrogenase (PDH) and metabolism in the tricarboxylic acid (TCA) cycle are the major mitochondrial decarboxylations in the dark, they are partially inhibited in the light (Ghashghaie et al., 2003; Tcherkez et al., 2005, 2008; Barbour et al., 2007). It has been suggested that LEDR mostly depends on a buildup of malate and fumarate in the light, which are then rapidly decarboxylated after the light–dark transition (Atkin et al., 1998; Barbour et al., 2007; Tcherkez et al., 2012). However, there is evidence for species-specific differences (Lehmann et al. 2016; Gessler et al., 2017). Overall, R_d in the gdch-KD and WT plants showed an expected negative hyperbolic pattern during 3 h in the dark following leaf photosynthesis. The results of the present study indicate that leaf photosynthesis under photorespiring conditions, before the light–dark transition, led to an additional buildup of TCA cycle substrates in the gdch-KD plants compared with the WT. In fact, the gdch-KD plants had a significantly higher R_d over 3 h dark following leaf photosynthesis under the approximately current ambient O₂ level (with leaf blades having no significantly different LMA), compared with the WT; this suggests a greater accumulation of metabolites in the light, in particular photorespiratory intermediates, as respiratory substrates to feed R_d. More precisely, a restricted photorespiratory pathway in the light may lead to an accumulation of 2-carbon metabolites in the gdch-KD plants.

The increase in LEDR has been reported to come from the decarboxylation of ¹³C heavier metabolites, primarily malate, and the decline in LEDR rates and δ¹³C_Rd over time due to a decrease in malate availability (Barbour et al. 2007; Gessler et al., 2009). In the gdch-KD plants, the cumulative leaf respired CO₂ over 30 min after leaf photosynthesis at a pO₂ of 18.4 kPa was 4.1 mmol CO₂ m⁻² higher with respect to the WT; theoretically, if this enhancement of R_d in the gdch-KD plants was due to malate alone this would require ~1 mmol malate m⁻². The non-significant differences in the leaf malate content determined during the light period between gdch-KD and the WT suggest that metabolites other than malate, such as photorespiratory intermediates, may have contributed to the greater LEDR rates in the gdch-KD plants compared with the WT. A substantial part of the malate in leaves is also stored in vacuoles, as observed in C₄ plants (Hatch, 1979; Arrivault et al. 2017), and not readily available for LEDR.

In the gdch-KD and WT plants presented here, the δ¹³C_Rd decreased during the 3 h dark period, tracking the decline in R_d (Fig. 5A, B). Over the 3 h of darkness, there was an increase in the contribution to δ¹³C_Rd from respiratory substrates generated during plant growth (^δRdG_{ch_substr}, ‰/‰) for both plant types and O₂ experimental conditions. Regardless of plant type or O₂ treatment, the ^δRdL_{ch_substr} went from ~50% after 6 min from the light–dark transition to ~30% after 30 min in the dark, while after 3 h in the dark it represented only ~10% [see Fig. 6; data of δ¹³C_Rd(3h) approaching δ¹³C_Rd(24h) are shown in Table 4]. Tcherkez et al. (2010) estimated on sunflower (Helianthus annuus) that recent assimilates provide 40–60% of the substrates for R_d (via a pool with a half-life of several hours) both in the light and in the dark. A similar contribution of recent assimilates to R_d was determined by Nogués et al. (2004) on French bean (Phaseolus vulgaris) leaves during ~2 h in the dark following illumination, which indicates that leaf respiration was fed by a mixture of recent and older substrates.

The tendency for a lower leaf δ¹³C_Rd in gdch-KD plants compared with the WT following leaf light exposure under photorespiratory conditions may partially depend on the higher Δ_o in the gdch-KD plants during leaf photosynthesis in the L_ch at approximately current ambient pO₂. A greater Δ_o would cause (recent) carbon assimilates synthetized in the L_ch (Supplementary data Table S1) to produce more depleted respiratory substrates and a lower δ¹³C_Rd in the gdch-KD compared with the WT. It is also possible that higher Δ_o in the gdch-KD compared with WT plants during growth under enriched atmospheric pCO₂ and current ambient pO₂ may have produced G_ch assimilates feeding R_d over 3 h after the light–dark transition with slightly lower δ¹³C compared with the WT.

The ¹³C fractionation during leaf dark respiration can change depending on species and environmental conditions (Ghashghaie et al., 2003; Priault et al., 2009; Werner et al., 2009; Lehmann et al., 2016). In addition, δ¹³C_Rd is influenced by the isotopic signatures of respiratory substrates, from diverse non-homogeneous isotope distributions in the substrates (positional effects) and the different relative activities of decarboxylation pathways. However, decreasing δ¹³C_Rd over time is mainly dependent on the origin of respiratory substrates, where CO₂ released from pyruvate decarboxylation is ¹³C enriched (compared with total organic matter) but relatively ¹³C depleted from acetyl-CoA metabolism through the TCA cycle (Tcherkez et al., 2003). Under continuous darkness and constant t_air, it has been shown that δ¹³C_Rd decreases due to a switch in respiratory substrates from carbohydrates to more ¹³C-depleted substrates such as lipids or proteins (Ghashghaie et al., 2003; Tcherkez et al., 2003).

The similar δ¹³C_Rd(24h) in gdch-KD versus WT plants implies that the long-term substrates for the TCA cycle produced in the G_ch were ¹³C isotopically similar. This is further supported by similar leaf δ¹³C_dm between the gdch-KD and WT plants. Interestingly, δ¹³C_Rd(24h) was more depleted than δ¹³C_dm, in agreement with Tcherkez et al. (2003) who had found that CO₂ evolved in the dark by French bean leaves in a condition of carbohydrate starvation had a lower δ¹³C than total leaf organic matter. This denotes potential changes in dark respiration substrates, such as carbohydrate oxidation producing ¹³C-enriched CO₂ and β-oxidation of fatty acids producing ¹³C-depleted CO₂ when compared with total organic matter.

Conclusions

Under photorespiratory conditions, the gdch-KD plants had altered ¹³C discrimination fractions in the light, with a lower Δ_f caused by a reduced f. This change in Δ_f and the lower Δ_gm lead to a higher Δ_o in the gdch-KD plants in comparison with the WT. The lower f in the gdch-KD plants was attributed to a greater α compared with the WT, suggesting the occurrence of alternative photorespiratory reactions in the GDC-impaired plants. In addition, the enhanced R_d in the gdch-KD compared with WT plants after photorespiratory leaf photosynthesis indicated that the photorespiratory disruption led to an additional buildup of metabolites in the light that were decarboxylated by the TCA cycle in the dark. The tendency for a more depleted δ¹³C_Rd in the gdch-KD plants compared with the WT after photorespiratory leaf photosynthesis was mainly ascribed to a higher Δ_o before the light–dark transition and differences in the δ¹³C of the substrates feeding R_d. These results indicate that an alteration in photorespiratory carbon metabolism can have a significant effect on leaf CO₂ exchange and ¹³CO₂ discrimination, both in the light and in the dark.

Supplementary data

Supplementary data are available at JXB online.

Methods S1. Estimate of the ¹³CO₂ composition of the growth chamber atmosphere during the light period.

Methods S2. Additional technical information on the system setup to measure online leaf atmosphere CO₂, H₂O, and ¹³CO₂ exchange.

Methods S3. Estimate of the ¹³C signature and total N content in the leaf biomass.

Methods S4. Estimate of the CO₂ compensation point in the absence of mitochondrial non-photorespiratory respiration.

Methods S5. Estimate of α, and evaluation of the sensitivity of f to Γ* and α (shown in Fig. S1).

Methods S6. Estimate of the fractional contribution of respiratory substrates from leaf chamber and growth chamber carbon assimilates to the ¹³C composition of dark-evolved CO₂.

Methods S7. Editing of the ¹³C composition of dark-evolved CO₂ for plants grown at an atmospheric ¹³C composition of −41.6‰.

Fig. S1 Sensitivity of the f parameter to Γ* and α.

Fig. S2. Sensitivity of the f parameter to g_m, R_L, and e'.

Table S1. Data used to calculate the fractional contributions of leaf chamber and growth chamber carbon assimilates to ¹³C composition of leaf dark-evolved CO₂.

Table S2. Values of ¹⁸O-based g_m.

Table S3. Statistics for the model used to fit leaf dark respiration rates.

Table S4. Significance for the model used to fit leaf dark respiration rates.

Table S5. Statistics for the model used to fit the ¹³C composition of leaf dark respiration rates.

Table S6. Significance for the model used to fit ¹³C composition of leaf dark respiration rates.

Author contributions

SK, SC, H-CL, RAC, WPQ, and JMH generated the transgenic plant material; SvC, RG, RTF, GEE, and ABC planned and designed the experiments; RG performed leaf gas and isotope exchange measurements and analysis; NK performed the biochemical analysis; RG, SvC, RTF, GEE, and ABC interpreted the data; and RG, ABC, and GEE developed the manuscript.