Photosynthesis of Root Chloroplasts Developed in Arabidopsis Lines Overexpressing GOLDEN2-LIKE Transcription Factors

Abstract

In plants, genes involved in photosynthesis are encoded separately in nuclei and plastids, and tight cooperation between these two genomes is therefore required for the development of functional chloroplasts. Golden2-like (GLK) transcription factors are involved in chloroplast development, directly targeting photosynthesis-associated nuclear genes for up-regulation. Although overexpression of GLKs leads to chloroplast development in non-photosynthetic organs, the mechanisms of coordination between the nuclear gene expression influenced by GLKs and the photosynthetic processes inside chloroplasts are largely unknown. To elucidate the impact of GLK-induced expression of photosynthesis-associated nuclear genes on the construction of photosynthetic systems, chloroplast morphology and photosynthetic characteristics in greenish roots of Arabidopsis thaliana lines overexpressing GLKs were compared with those in wild-type roots and leaves. Overexpression of GLKs caused up-regulation of not only their direct targets but also non-target nuclear and plastid genes, leading to global induction of chloroplast biogenesis in the root. Large antennae relative to reaction centers were observed in wild-type roots and were further enhanced by GLK overexpression due to the increased expression of target genes associated with peripheral light-harvesting antennae. Photochemical efficiency was lower in the root chloroplasts than in leaf chloroplasts, suggesting that the imbalance in the photosynthetic machinery decreases the efficiency of light utilization in root chloroplasts. Despite the low photochemical efficiency, root photosynthesis contributed to carbon assimilation in Arabidopsis. Moreover, GLK overexpression increased CO₂ fixation and promoted phototrophic performance of the root, showing the potential of root photosynthesis to improve effective carbon utilization in plants.

Introduction

Photosynthetic electron transfer reactions in plants take place in thylakoid membranes inside chloroplasts, in which light energy drives electron transport between a series of multisubunit protein complexes including PSII and PSI. Core reaction centers of PSII and PSI are surrounded by peripheral light-harvesting complexes, LHCII and LHCI, respectively, that capture light energy and transfer it to Chls in the reaction centers. Electrons excited in PSII are transferred stepwise to the plastoquinone pool, Cyt b₆f complex, plastocyanin and PSI, where another charge separation creates a strong reductant capable of reducing NADP⁺ (Rochaix 2011). Whereas photochemical reactions are essential to produce ATP and NADPH required for carbon fixation, excess light energy over the capacity for electron transport leads to the formation of harmful reactive oxygen species and eventually damages cells (Murchie and Niyogi 2011). Therefore, plants strictly regulate construction of the photosynthetic machinery and biogenesis of chloroplasts during photosynthetic organ development.

In plants, the genetic contribution to photosynthesis is shared between the nuclear and plastid genomes (Martin et al. 2002). Whereas genes for LHC proteins and pigment biosynthesis pathways reside in the nuclear genome, those encoding PS core subunits are mainly in the plastid genome. Therefore, chloroplast biogenesis depends on close cooperation between the nuclear and plastid genomes. We recently reported that nuclear-encoded photosynthesis genes form a tight co-expression network with key Chl biosynthetic genes, and suggested that there is a central transcriptional regulation system governing construction of Chl–protein complexes (Masuda and Fujita 2008, Kobayashi et al. 2012b). Golden2-like (GLK) transcription factors are proposed to be involved in the positive regulation of these photosynthesis-associated nuclear genes based on the evidence that Arabidopsis thaliana GLKs (GLK1 and GLK2) directly up-regulate many of those genes (Waters et al. 2009). Whereas loss of GLK activity leads to reduced Chl accumulation in photosynthetic tissues of Arabidopsis (Fitter et al. 2002), rice (Oryza sativa; Wang et al. 2013), tomato (Solanum lycopersicum; Powell et al. 2012) and Physcomitrella patens (Yasumura et al. 2005), increased expression of GLKs induces Chl accumulation and chloroplast biogenesis in non-foliar tissues such as Arabidopsis roots (Kobayashi et al. 2012a), rice calluses (Nakamura et al. 2009) and tomato fruits (Powell et al. 2012), suggesting that GLKs play a crucial role in chloroplast biogenesis during organ development.

Plant roots usually grow underground as heterotrophic organs and depend on aerial leaves for energy, although roots of some epiphytic plants turn green and perform active photosynthesis (Aschan and Pfanz 2003). In Arabidopsis roots, chloroplast development is essentially suppressed even under light conditions, and Chl accumulation is observed only in the upper part of the primary root near the hypocotyl junction (Kobayashi et al. 2012a). Recently we revealed that an auxin/cytokinin signaling pathway is involved in regulation of chloroplast development in the root through GLKs and another transcription factor, LONG HYPOCOTYL5 (HY5) (Kobayashi et al. 2012a). Consistent with the lack of Chl, the expression levels of GLK1 and GLK2 are very low in roots compared with those in leaves (Fitter et al. 2002). Overexpression of GLK genes results in remarkable accumulation of Chl in roots, reaching nearly 10% of the amounts of Chl observed in wild-type leaves (Kobayashi et al. 2012a). This suggests that overexpression of GLK genes derepresses chloroplast development in roots via transcriptional activation of photosynthesis-associated nuclear genes. However, given that GLKs are nuclear transcription factors and various essential proteins in the electron transport chain are encoded in the plastid genome, how the overall photosynthetic machinery is organized in the roots of the wild type and GLK overexpressors (GLK_OX) is largely unknown.

In this study, we investigated chloroplast development and photosynthetic functions in the roots of wild-type and GLK_OX Arabidopsis. Considering that photosynthesis in the roots contributes to the carbon economy in some species, our findings not only help to elucidate the global regulation of chloroplast biogenesis but also could lead to enhancement of crop photosynthesis.

Results

Chloroplast development in roots of GLK overexpressors

Consistent with our previous report that overexpression of GLK enhances Chl accumulation in roots (Kobayashi et al. 2012a), both GLK overexpressors, GLK1_OX and GLK2_OX, had visibly green roots compared with the yellowish roots of the wild type (Fig. 1A). A HY5-deficient mutant (hy5-215), which accumulates no Chl in roots (Usami et al. 2004, Kobayashi et al. 2012a), had albino roots. As reported previously (Kobayashi et al. 2012a), wild-type roots showed Chl accumulation only in the stele of the mature primary root (Fig. 1B). In both GLK1_OX and GLK2_OX roots, the stele of the primary root was the major site of Chl accumulation; however, Chl accumulation was also detected in the outer cell layers, endodermis, cortex and epidermis (Fig. 1B). Moreover, GLK_OX plants accumulated Chl in lateral roots, where Chl accumulation was hardly detected in the wild type (Supplementary Fig. S1). These data suggest that overexpression of GLKs triggers ectopic development of chloroplasts in the root.

Fig. 1.

Chl accumulation in GLK_OX roots. (A) Color phenotype of roots of 21-day-old plants. GLK_OX plants have greenish roots compared with the yellowish wild-type and albino hy5-215 roots. (B) Confocal microscopy of Chl fluorescence in the primary root. Chl autofluorescence micrographs were merged with differential interference contrast images. Arrows indicate boundaries between the stele and endodermis. Bar = 50 µm.

Next, we compared plastid morphology between wild-type and GLK_OX plants by observing plastids of primary roots at approximately 4.0 cm from the root–hypocotyl junction with transmission electron microscopy (TEM) (Fig. 2). In the wild type, although the thylakoid membrane networks were poorly formed in the root chloroplasts as compared with those in the leaf chloroplasts, grana stacks were relatively well developed (Fig. 2A, D). The development of thylakoid membrane networks with a highly stacked grana structure was remarkably enhanced in root chloroplasts of both GLK_OX lines (Fig. 2B, C). Particularly in the GLK1_OX root chloroplasts, thylakoid membranes were extensively stacked and formed very thick and wide grana structures. In addition, chloroplasts of GLK_OX roots accumulated starch grains, which were hardly detected in wild-type roots, indicating elevated photosynthetic activity of root chloroplasts in these overexpressors. We also found that chloroplasts in GLK1_OX roots were larger than those in wild-type roots (Fig. 2E), presumably due to the pronounced development of thylakoid membrane structures. Furthermore, in GLK1_OX and GLK2_OX roots, the number of chloroplasts per cell was increased and cells containing multiple chloroplasts were observed more frequently than in wild-type roots (Fig. 2F). These results suggest that GLK overexpression not only enhances chloroplast development via activating biogenesis of thylakoid membranes but also induces chloroplast division in the root. It should be noted that the visible phenotype of root gravitropism which is regulated through amyloplast sedimentation was not affected in GLK1_OX roots, suggesting that certain plastids retained their roles in these root tissues.

Fig. 2.

Plastid development in GLK_OX roots. (A–C) Plastid ultrastructure in the primary root cells of wild-type (WT) (A), GLK1_OX (B) and GLK2_OX seedlings (C). (D) A typical chloroplast in a mature Arabidopsis leaf cell. Bars = 0.5 µm. (E) Histogram of plastid size in the outer cells of the stele determined from primary root cross-sections from WT and GLK_OX seedlings (n = 42, 87 and 64 for the WT, GLK1_OX, and GLK2_OX, respectively). (F) Histogram of plastid number in the outer cells of the stele determined from primary root cross-sections (n = 60, 45, and 54 for the WT, GLK1_OX, and GLK2_OX, respectively).

Gene expression analysis in roots of GLK overexpressors

In aerial organs, GLK acts as a transcriptional activator by directly binding to the promoters of nuclear LHC and Chl biosynthetic genes (Waters et al. 2009). Consistent with this, quantitative reverse transcription–PCR (qRT–PCR) analysis revealed that these target genes were highly up-regulated in roots of both GLK_OX lines (Fig. 3A). In general, transcripts responded more strongly to GLK1_OX than to GLK2_OX in roots, as expected from the relatively greener pigmentation in GLK1_OX roots (Fig. 1). In GLK1_OX, transcript levels of LHC genes (LHCA4 and LHCB6) were increased >10-fold, while those of Chl biosynthetic genes (HEMA1, CHLH and CHL27) were increased >5-fold (Fig. 3A).

Fig. 3.

Quantitative RT–PCR analysis of chloroplast biogenesis-associated gene expression in roots of wild-type and GLK_OX seedlings. Expression levels of nuclear-encoded genes for Chl biosynthesis and light harvesting (A), plastid-encoded genes (B), nuclear-encoded genes for sigma factors (C) and GATA-type nuclear transcription factor genes (D). Data are presented as the fold difference from wild-type root samples after normalization to the reference gene ACTIN8. Values are the means ± SE from three independent experiments.

To examine whether GLK overexpression also affects plastid gene expression, we investigated the expression levels of PsaA, PsbA and RpoB in roots of both GLK_OX lines (Fig. 3B). The PsaA and PsbA genes, which encode the core proteins of PSI and PSII, respectively, are transcribed by plastid-encoded RNA polymerase (PEP), whereas RpoB, which itself encodes the β-chain of PEP, is transcribed by nuclear-encoded plastid RNA polymerase (NEP) (De Santis-MacIossek et al. 1999). In both GLK1_OX and GLK2_OX roots, transcript levels of PsaA and PsbA increased 4-fold compared with those in the wild-type roots. The transcript abundance of RpoB also increased in both GLK_OX lines, but was lower than those of PsaA and PsbA.

We then examined the expression of genes for sigma factors (SIGs), which are nuclear-encoded transcriptional initiation factors required for the binding of PEP to specific promoters of plastid genes (Fig. 3C). The Arabidopsis genome encodes six SIGs that are localized in plastids and activate subsets of plastid gene promoters in a partly redundant manner (Schweer et al. 2010). As observed for nuclear- and plastid-encoded photosynthetic genes, the expression of all SIG genes, except for SIG5, was up-regulated in both GLK_OX lines. Among them, SIG4 in GLK1_OX showed the most prominent up-regulation. Because two GATA transcription factors, GNC and CGA1, were recently reported to induce chloroplast development in non-photosynthetic tissues (Chiang et al. 2012), we also analyzed the expression of GNC and CGA1 in GLK_OX roots (Fig. 3D). Although GNC and CGA1 have not been identified as direct targets of GLK factors, expression of these genes was increased in GLK_OX roots. Our results suggest that GLK overexpression up-regulates not only their primary target genes involved in Chl biosynthesis and light harvesting but also GLK-non-targeted genes associated with chloroplast development in the root.

Pronounced accumulation of peripheral PS proteins in roots of GLK overexpressors

To clarify further the effects of GLK overexpression on chloroplast biogenesis in roots, we examined the levels of photosynthetic proteins in GLK_OX roots by immunoblot analysis. A 10 µg aliquot of total membrane protein from roots of wild-type and GLK_OX plants was analyzed together with a dilution series (0.2, 1 and 5 µg) of membrane proteins from wild-type leaves (Fig. 4). Since GLK1_OX roots contain roughly 10% of the Chl present in wild-type leaves on a fresh weight basis (Kobayashi et al. 2012a), comparison between 10 µg of membrane protein from GLK_OX roots and 1 µg from wild-type leaves is appropriate. As shown in Fig. 4, the levels of all membrane photosynthetic proteins were higher in GLK_OX roots than in wild-type roots. On the other hand, differences in the protein levels were small between GLK1_OX and GLK2_OX roots, except in CP43 and LHCB6. Compared with the wild-type leaf samples containing 1 µg of protein, the levels of PsbO and LHC proteins (LHCA1, LHCA2, LHCB1, LHCB6 and total LHCII) in GLK_OX root samples containing 10 µg of protein were higher, while those of D1, D2 and CP43 were comparable or lower. Moreover, the amount of PsaA/PsaB, the core reaction center proteins of PSI, was very low even in GLK_OX roots. These results show that GLK overexpression induces the accumulation of nuclear-encoded peripheral photosynthetic proteins more strongly than that of plastid-encoded reaction center proteins in PSI and PSII.

Fig. 4.

Differential accumulation of membrane photosynthetic proteins in roots of GLK1_OX and GLK2_OX. Immunoblot analysis of photosynthetic proteins in 10 µg of total membrane protein from root samples of wild type (WT), GLK1_OX (1_OX) and GLK2_OX (2_OX) compared with those in a dilution series (5, 1 and 0.2 µg) of total membrane protein from WT leaves.

Pigment composition in roots of GLK overexpressors

To evaluate the balance among PSI, PSII and their antenna complexes in root chloroplasts, we compared pigment compositions in root samples with those in wild-type leaves (Table 1). As reported previously (Kobayashi et al. 2012a), Chl contents were increased in roots of both GLK_OX lines, particularly GLK1_OX, compared with those in wild-type roots. In parallel with the Chl accumulation, the amount of carotenoids was also substantially increased in both GLK_OX lines. In all root samples, the proportions of Chl b and carotenoids, which are pigments in LHC antennae, relative to Chl a were higher than those in wild-type leaves (Table 1), confirming that there is a higher LHC antenna/reaction center ratio in root chloroplasts.

Table 1.

Pigment composition in roots of the wild type, GLK1_OX and GLK2_OX, and leaves of the wild type

	Chl a (nmol g−1 FW)	Chl b (nmol g−1 FW)	Chl a/b	Carotenoids
				µg g−1 FW	µg µmol−1 Chl a
Wild-type leaf	2,252.7 ± 173.0	715.7 ± 59.5	3.16 ± 0.04	367.7 ± 29.6	163.4 ± 1.2
Wild-type root	19.5 ± 1.6	7.5 ± 0.6	2.67 ± 0.15	8.2 ± 0.6	423.6 ± 13.5
GLK1OX root	134.4 ± 7.9	53.9 ± 2.6	2.49 ± 0.03	58.3 ± 2.7	435.6 ± 6.3
GLK2OX root	62.4 ± 4.5	24.0 ± 1.3	2.60 ± 0.09	24.5 ± 1.1	396.1 ± 11.2

Values are means ± SE (n = 3 for wild-type leaves, 10 for wild-type roots and 6 for GLK1_OX and GLK2_OX roots).

To characterize further the PS–LHC complexes in GLK1_OX roots, precise photosynthetic pigment analysis was performed by HPLC. In GLK1_OX roots, the amount of xanthophylls relative to Chl a was higher than that in wild-type leaves, whereas the proportion of β-carotene was almost equivalent between these two samples (Table 2), consistent with the data showing that LHC antennae are actively formed in root chloroplasts (Fig. 4, Table 1). To estimate the PSI/PSII ratio, we examined the amount of Chl a′ and pheophytin (Phe) a, which are specific to PSI (Kobayashi et al. 1988, Jordan et al. 2001) and PSII (Klimov et al. 1977a, Klimov et al. 1977b, Zouni et al. 2001), respectively. The PSI/PSII ratio estimated from proportions of Chl a′ and Phe a to Chl a was lower in GLK1_OX roots than in wild-type leaves, supporting the result obtained by immunoblot analyses that PSI is accumulated less in GLK1_OX roots (Fig. 4).

Table 2.

Molar ratios of photosynthetic pigments in wild-type leaves and GLK1_OX roots

	Chl a/Chl a′	Chl a/Phe a	PSI/PSII	β-Carotene/Chl a	Total xanthophyll/Chl a
Wild-type leaf	546 ± 88	99 ± 19	0.37 ± 0.08	0.14 ± 0.04	0.68 ± 0.06
GLK1OX root	565 ± 110	70 ± 9	0.26 ± 0.06	0.12 ± 0.01	1.29 ± 0.10

Values are means ± SD from five independent samples.

PSs formed in root chloroplasts are different from those in leaf chloroplasts

To examine the state of Chl–protein complexes in the roots of wild-type and GLK_OX plants, Chl fluorescence spectra were measured at 77K (Fig. 5A). In the emission spectra of wild-type leaves, there were shoulder bands at 684 and 691 nm, which primarily originate from CP43 and CP47 in PSII, respectively (Govindjee 1995). In addition, there was an emission peak at 732 nm, which can be attributed to the PSI–LHCI complex. Since the peaks originating from the PSI–LHCI complex were also detected in all root samples, the spectra were normalized to the fluorescence emission maximum of PSI–LHCI (Fig. 5A). In all root samples, the two shoulder bands originating from PSII were present, but they were slightly shifted to 683 and 689 nm. This shift indicates the presence of LHCII complexes that are excitonically uncoupled from PSII (Hölzl et al. 2009). For PSI, the emission maximum of root samples was also shifted to lower wavelengths (729–731 nm), indicating the presence of LHCI proteins that were either weakly or not coupled to PSI reaction centers, an effect that is usually observed in plants suffering from a decrease in PSI reaction center (Stöckel et al. 2006).

Fig. 5.

Comparison of PS complexes in roots of the wild type and GLK_OX lines with those in wild-type leaves. Characteristics of PS complexes were examined in samples from roots of wild type (WT R), GLK1_OX (GLK1_OX R) and GLK2_OX (GLK2_OX R) and from wild-type leaves (WT L). (A) Chl fluorescence emission spectra at 77K. Vertical dashed lines represent emission peaks at 684, 691 and 732 nm observed in WT L samples. Slight blue shifts in these peaks were observed in all root samples. Representative data from multiple independent experiments are shown (n > 3). (B and C) Transient fluorescence induction kinetics of Chl in the absence (B) or presence (C) of 40 µM 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU). Values are means from three independent experiments. Two inflections, labeled J and I, are observed between the levels O (origin) and P (peak) only in the absence of DCMU (B).

To evaluate the functionality of the photosynthetic apparatus in root chloroplasts of GLK_OX plants, we analyzed the transient kinetics of Chl fluorescence using a logarithmic timing series (Fig. 5B). Wild-type leaves displayed a typical polyphasic fluorescence rise exhibiting the origin-inflection-intermediary peak–peak (O-J-I-P) transient (Govindjee 1995). Similar kinetics was observed in leaves of both GLK_OX lines (Supplementary Fig. S2), demonstrating that GLK overexpression does not affect electron transport of PSII in leaves. The O-J-I-P transient was also observed in all root samples, indicative of functional electron transport in the root PSII. However, as reported previously (Kobayashi et al. 2012a), the O-J transition, which constitutes the photochemical phase of the Chl a fluorescence rise, occurred with a higher fluorescence yield in the wild-type roots than in the leaves, suggesting that electron transport from Q_A to Q_B is insufficient in the root chloroplasts. Moreover, such high Chl fluorescence at the O-J transition was also observed in roots of GLK_OX plants, indicating that GLK overexpression does not improve the efficiency of photosynthetic electron transport in the root chloroplasts.

We next measured Chl fluorescence kinetics in the presence of 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) (Fig. 5C). DCMU inhibits electron transfer from Q_A to Q_B, which results in a rapid reduction of total Q_A (Govindjee 1995). In wild-type roots, the fluorescence rise was much faster than that in leaves, indicating that Q_A reduction occurs more rapidly in root chloroplasts than in leaf chloroplasts. This phenomenon was more pronounced in the GLK1_OX roots, consistent with the finding that GLK1_OX enhances the formation of antennae relative to reaction centers in root chloroplasts.

Photosynthetic electron flow in root chloroplasts

To evaluate photosynthetic activity in root chloroplasts, we analyzed Chl fluorescence using pulse amplitude modulation (PAM) techniques (Maxwell and Johnson 2000). First we determined the maximum quantum efficiency of PSII (F_v/F_m) in the dark (Fig. 6), which represents the intrinsic photosynthetic efficiency of PSII. As observed previously (Kobayashi et al. 2012a), F_v/F_m in dark-adapted roots of the wild type was slightly lower than that in leaves, suggesting a slight decrease in maximal efficiency of light utilization in the root PSII. In GLK_OX plants, particularly GLK1_OX, F_v/F_m levels were further decreased in roots, although those in leaves were unchanged. Next we analyzed the light–response curves of Chl fluorescence from PSII. In all root samples, the photochemical quantum yield of PSII (Φ_II) was lower than that in wild-type leaves under middle to high photosynthetically active radiation (PAR) (Fig. 7A). In addition, GLK1_OX roots had lower Φ_II even under low PAR (Supplementary Fig. S3A). The Φ_II can be viewed as a product of two components, the coefficient of photochemical quenching (qP) and the maximum quantum efficiency of open PSII (F_v′/F_m′). Based on the ‘puddle’ model (Kramer et al. 2004), qP represents the Q_A redox status and thus the openness of PSII. In all root samples, qP was not very different from that in the wild-type leaves (Fig. 7B), except in the GLK_OX roots under low PAR (Supplementary Fig. S3B). When the fraction of open PSII was estimated using another coefficient (qL) on the basis of the ‘lake’ model (Kramer et al. 2004), PSII in the roots was found to be in an even more oxidized state than that in the leaves (Supplementary Fig. S3F). On the other hand, F_v′/F_m′ was substantially decreased in all root samples compared with that in wild-type leaves (Fig. 7C; Supplemental Fig. S3C). These data suggest that the low Φ_II in roots resulted mainly from low PSII quantum efficiency and not PSII acceptor-side limitation.

Fig. 6.

Maximum quantum yield of PSII (F_v/F_m). F_v/F_m levels were compared between leaves and roots from the wild type, GLK1_OX and GLK2_OX. Values are means ± SE (n = 5 for leaves and 7 for roots in each line).

Fig. 7.

Light–response curves of Chl fluorescence parameters. Wild-type leaves (WT L) and roots of the wild type (WT R), GLK1_OX (GLK1_OX R) and GLK2_OX (GLK2_OX R) were dark adapted for 5 min prior to the measurements and exposed for 3 min to each light intensity. (A) Effective quantum yield of PSII (Φ_II). (B) Coefficient of photochemical quenching (qP), a measure of the redox state of the PSII acceptor side. (C) Maximum PSII quantum yield under light conditions (F_v′/F_m′). (D) Quantum yield of non-regulated energy dissipation (Φ_NO). (E) Quantum yield of regulated energy dissipation (Φ_NPQ). (F) Coefficient of non-photochemical quenching (qN). Data are means of multiple experiments (n = 5 for wild-type leaves and 7 for each root sample).

Absorbed light energy by LHCII–PSII can be divided into Φ_II, quantum yield of light-induced non-photochemical quenching (Φ_NPQ) and quantum yield of non-light-induced non-photochemical quenching (Φ_NO) (Kramer et al. 2004). Whereas Φ_NO levels were similar in all samples (Fig. 7D), except for GLK1_OX roots under low PAR (Supplementary Fig. S3D), Φ_NPQ levels in all root samples were higher than those in wild-type leaves (Fig. 7E; Supplementary Fig.S3E). The Φ_NPQ component represents the fraction of regulated heat dissipation. In fact, in all root samples, the coefficient of non-photochemical quenching (qN) was also higher than that in wild-type leaves (Fig. 7F).

To explore this further, we evaluated the photoinhibitory component (qI) of qN, which is characterized by its very slow relaxation kinetics in the range of hours (Murchie and Niyogi 2011). Total qN was determined in plant samples after a 10 min exposure to actinic light (420 µmol photons m⁻² s⁻¹). After relaxation for 15 min in the dark, the remaining qN components, corresponding to qI, were determined. The contribution of qI to the total qN in root samples was similar to that in wild-type leaves (Fig. 8), suggesting that the rapidly reversible component related to the xanthophyll cycle is the main contributor to the increase in non-photochemical quenching in roots. In the leaves, there were no large differences in the light–response curves of Chl fluorescence between wild-type and GLK_OX plants (Supplementary Fig. S4).

Fig. 8.

Contribution of photoinhibition to total non-photochemical quenching (qN). Wild-type leaves (WT L) and roots of the wild type (WT R), GLK1_OX (GLK1_OX R) and GLK2_OX (GLK2_OX R) were dark adapted for 15 min and then exposed to light stress (420 µmol photons m⁻² s⁻¹) for 10 min. First, total qN was measured at the end of the light stress. After relaxation of the rapidly reversible qN component with additional dark treatment for 15 min, the remaining qN component was determined as the photoinhibition-related qN component (qI). Values are the means ± SE from three independent experiments.

We also evaluated the redox state of PSI by analyzing changes in P700⁺ levels during moderate light (126 µmol photons m⁻² s⁻¹) exposure (Klughammer and Schreiber 2008). Although the P700 signal from wild-type roots was too low to evaluate, clear P700⁺ signal was detected in GLK1_OX roots (Fig. 9) and used for comparison with those from leaves of the wild type and GLK1_OX. In both wild-type and GLK1_OX leaves, the photochemical quantum yield of PSI (Φ_I), which represents the fraction of open P700, increased rapidly after light exposure and maintained high steady-state levels. In contrast, Φ_I in GLK1_OX roots decreased after illumination and remained at lower levels. We then analyzed the quantum yield of non-photochemical energy dissipation in PSI due to donor-side limitations (Φ_ND) or acceptor-side limitations (Φ_NA). Φ_ND represents the fraction of P700⁺, whereas Φ_NA represents the fraction of P700 that cannot be oxidized in a given state (Pfundel et al. 2008). In GLK1_OX roots, Φ_ND levels stayed higher than those in wild-type and GLK1_OX leaves. Meanwhile, steady-state levels of Φ_NA were not markedly different among all samples, although inverse profiles of Φ_NA were detected between GLK1_OX roots and leaf samples during the first few seconds, as observed for Φ_I. These data indicate that the efficiency of light utilization in PSI decreased in GLK1_OX roots mainly due to donor-side limitation as compared with that in leaves.

Fig. 9.

Quantum yield of PSI. Slow induction kinetics of the PSI quantum yields in wild-type leaves (WT L), and GLK1_OX leaves (GLK1_OX L) and roots (GLK1_OX R) were measured for 10 min under actinic light (126 µmol photons m⁻² s⁻¹). (A) Photochemical quantum yield of PSI (Φ_I). (B and C) Non-photochemical quantum yield of energy dissipation due to PSI donor-side limitation (Φ_ND) (B) and PSI acceptor-side limitation (Φ_NA) (C). Data are means of two independent measurements.

Photosynthetic activity and photoautotrophic growth of roots in GLK overexpressors

Finally, we evaluated the effects of GLK overexpression on overall photosynthetic performance in the root. For this experiment, roots were excised from 21-day-old seedlings and their photosynthetic activities were measured. Fig. 10A shows the light–response curve of net O₂ evolution under a high CO₂ (5%) condition. In both GLK_OX lines, O₂ evolution activity overcame the light compensation point at <100 µmol photons m⁻² s⁻¹, whereas wild-type roots did not achieve net O₂ production even at the highest light condition (400 µmol photons m⁻² s⁻¹). Furthermore, roots of GLK_OX plants exhibited light-dependent CO₂ fixation activity under an atmospheric CO₂ (0.039%) condition (Fig. 10B). Both GLK_OX lines had much higher CO₂ fixation activity than the wild type at 200 µmol photons m⁻² s⁻¹. We also tested the hy5-215 mutant, which cannot accumulate Chl in the root (Fig. 1A) (Usami et al. 2004, Kobayashi et al. 2012a), as a negative control. As expected, neither O₂ evolution nor CO₂ fixation activities were detected in the hy5-215 roots.

Fig. 10.

Photosynthetic activity of roots. (A) Light–response curve of net O₂ evolution activity in roots. The zero level represents the light compensation point. (B) Net CO₂ fixation activity in roots under dark or light (200 µmol photons m⁻² s⁻¹) conditions. Numbers in the graph indicate the gross CO₂ fixation rate (µmol CO2 min⁻¹ g FW⁻¹) in each root sample calculated by subtracting the CO₂ fixation rate in the dark from that in the light. (C) Change in dry weight of detached roots. Roots detached from 21-day-old seedlings were further incubated for 7 d without a carbon source under dark or light (60 µmol photons m⁻² s⁻¹) conditions. The dry weight of the 28-day-old detached roots was normalized to the initial weight of the 21-day-old root. Values are means ± SE from three independent experiments for A and B, and from seven independent experiments for C.

To investigate photoautotrophic growth activity, excised roots from 21-day-old seedlings were further incubated on Murashige and Skoog (MS) medium in the absence of any available carbohydrates for 7 d in the dark or light (60 µmol photons m⁻² s⁻¹). After a 7 d incubation, dry weights of the excised roots were compared with the initial root weights of 21-day-old seedlings (Fig. 10C). The change in dry weight after incubation without carbon sources represents the balance between photosynthesis and respiration activities. In the hy5-215 mutant, dry weight decreased to the same extent in dark- and light-treated roots, consistent with the absence of photosynthetic activity in hy5-215 roots (Fig. 10A, B). A decrease in dry weight compared with the initial weight was also observed in wild-type roots under both light and dark conditions, although light treatment slightly reduced the decrease. In both GLK_OX lines, in contrast, dry weights of light-treated roots were much larger than those of dark-treated roots and comparable with the initial levels, suggesting that carbon fixation by photosynthesis entirely compensates for the loss due to the respiration in these roots. These results show that the roots of GLK_OX plants have the ability to perform photosynthesis that can maintain their weights without sugar. It should be noted, however, that the growth of whole GLK_OX plants was the same or less than that of the wild type, probably due to excessive Chl accumulation in leaves in these plants (Waters et al. 2008) which was not advantageous for regulated photosynthetic growth.

Discussion

Coordinated up-regulation of genes for chloroplast biogenesis in the nucleus and plastids

Consistent with the previous report that nuclear genes for enzymes involved in Chl biosynthesis and LHC proteins are direct targets of GLK factors in Arabidopsis (Waters et al. 2009), the expression of these genes was strongly up-regulated in GLK_OX roots (Fig. 3A). In addition, genes that were not identified as direct targets of GLKs such as GNC, CGA1 and most genes for SIGs were also up-regulated in GLK_OX roots (Fig. 3C, D). Although we cannot exclude the possibility that these genes are directly targeted by GLKs in the root, it is likely that the induction of the primary GLK targets secondarily influences the expression of GLK-non-targeted genes related to chloroplast biogenesis. Similar results were obtained in rice GLK1_OX callus, in which counterparts of Arabidopsis GLK-non-targeted genes, such as most nuclear-encoded PS genes, were globally up-regulated (Nakamura et al. 2009). Such global up-regulation of chloroplast-related genes would contribute to chloroplast development in the GLK_OX roots. As an example, GNC and CGA1, which were up-regulated in GLK_OX roots, induce development and division of chloroplasts in the hypocotyl and the root (Chiang et al. 2012), consistent with our data in Figs. 1 and 2. Because GLK_OX roots substantially accumulated the gene products of not only GLK direct targets (LHCA1, LHCB1 and LHCB6) but also non-direct targets (LHCA2 and PSBO) (Fig. 4), an intact peripheral LHC antenna system nearly equivalent to that in wild-type leaves could be constructed in GLK_OX roots. Furthermore, carotenoids substantially accumulated in GLK_OX roots together with Chl (Tables 1, 2), although the genes involved in carotenoid biosynthesis have not been identified as GLK targets (Waters et al. 2009). These results suggest a global influence of GLK-induced Chl–LHC antenna formation in roots.

In addition to the induction of photosynthesis-associated nuclear genes, the expression of PEP-dependent photosynthetic genes (PsaA and PsbA) was up-regulated in roots of both GLK_OX lines (Fig. 3B). As proposed previously (Nakamura et al. 2009), it is likely that the increased gene expression of SIGs, which trigger PEP-dependent transcription (Schweer et al. 2010), leads to the up-regulation of the PEP-dependent genes in GLK_OX roots. Moreover, expression of the NEP-dependent RpoB gene, which encodes the β-chain of PEP, was increased in GLK_OX roots, highlighting the possibility that overall PEP activities are elevated in these roots. Our recent study suggests that thylakoid membrane biogenesis changes nucleoid morphology and globally activates plastid gene expression independently of photosynthesis (Kobayashi et al. 2013). Furthermore, one of the nuclear-encoded plastid RNA polymerases, RPOTmp, is reported to be tightly associated with thylakoid membranes (Azevedo et al. 2008). Thus, the thylakoid membrane biogenesis in GLK_OX root plastids may also positively influence global plastid gene expression. In addition, the increased plastid number per cell in GLK_OX roots (Fig. 2F) would contribute to the increased transcripts of plastid-encoded genes.

Reduced light utilization in the imbalanced PSs in root plastids

Protein analyses clearly revealed high abundance of LHC antenna complexes relative to the reaction centers in GLK_OX roots (Fig. 4). A similar tendency was observed in wild-type roots although the amount of each photosynthetic protein was much lower than that in GLK_OX roots. Moreover, low Chl a/b and high xanthophyll/Chl a ratios were detected not only in GLK_OX roots but also in the wild-type roots (Table 1), suggesting that chloroplasts developed in the root have relatively large antenna complexes per reaction center. Consistent with these data, enhanced grana formation was observed in root chloroplasts (Fig. 2). In particular, GLK1_OX root chloroplasts had substantially enlarged grana thylakoids (Fig. 2B). Based on the model that attractive forces between trimeric LHCs on closely appressed thylakoids cause membrane adhesion during grana formation (Standfuss et al. 2005), the excessive accumulation of LHC antenna systems would cause the hyperstacking of grana thylakoids in GLK_OX root chloroplasts. In support of this idea, hyperstacked grana structures were also observed in the rice non-yellow coloring 1 mutant, in which LHC–Chl b complexes are selectively accumulated during senescence due to the loss of Chl b-degrading activity (Kusaba et al. 2007). Because Q_A reduction in the presence of DCMU occurred more rapidly in all root samples than in wild-type leaves (Fig. 5C), the imbalance between the antennae and the reaction centers in root chloroplasts may lead to excessive input of light energy from antenna complexes to the PSII reaction centers, which may result in the inefficient electron transport to the quinone pool in root chloroplasts (Fig. 5B).

The PAM analyses revealed that Φ_II levels were considerably decreased in all root samples particularly at middle to high PAR (Fig. 7A). These changes accompanied remarkable increases in thermal dissipation of light energy (Fig. 7E), which occurs in the LHCII antenna (Murchie and Niyogi 2011). Thus, it is likely that the high antenna/reaction center ratio causes excess excitation of root PSII, thereby increasing thermal dissipation of light energy within the antenna and decreasing the PSII photochemical efficiency. This idea is supported by the substantial accumulation of carotenoids relative to Chl a in root chloroplasts (Table 1). Xanthophylls, in particular, showed a marked increase in GLK1_OX roots (Table 2), consistent with the high degree of non-photochemical quenching observed in the roots. It should be noted that the intrinsic PSII efficiency represented by F_v/F_m was also lower in all root samples than in wild-type leaves (Fig. 6), which may be caused by the partial uncoupling of LHCII from the PSII reaction center (Fig. 5A), and/or the inefficient electron transport from Q_A to Q_B (Fig. 5B). The decrease in intrinsic PSII efficiency could be partly responsible for the decrease in the actual PSII efficiency under the light. In addition, GLK1_OX roots showed a decrease in Φ_II even under low PAR (Fig. 7A; Supplementary Fig. S3A). As represented by the lower qP (Fig. 7B; Supplementary Fig. S3B) and qL (Supplementary Fig. S3F), PSII in GLK1_OX roots was in a more reduced state than in wild-type leaves and other root samples under low PAR, suggesting that electron transport downstream of PSII is retarded in GLK1_OX root chloroplasts. In the GLK1_OX root, Φ_NO levels, which represent the proportion of non-regulated energy dissipation by heat and fluorescence (Kramer et al. 2004), increased under low PAR (Fig. 7D; Supplementary Fig. S3D). Because regulated thermal energy dissipation is less functional under low PAR (Fig. 7E; Supplementary Fig. S3E), the excess input of light energy from enlarged antenna systems into the reaction centers may cause over-reduction of PSII and increase the energy lost in a non-regulated manner in GLK1_OX root chloroplasts.

Despite the similar up-regulation of PsaA and PsbA in GLK_OX roots (Fig. 3B), accumulation of the PSI reaction center (PsaA/PsaB) was much lower than that of PSII reaction center proteins D1 (PsbA) and D2 (PsbD) and core antenna CP43 (PsbC) (Fig. 4). This result is confirmed by the pigment analyses in GLK1_OX roots (Table 2). Considering the similar accumulation of LHCA and LHCB proteins in GLK_OX roots, the reason that the PSI reaction center failed to accumulate is not clear. The assembly of PSI involves assembly-dependent regulation of biosynthesis of the major chloroplast-encoded subunits (Wostrikoff et al. 2004). In fact, several factors are involved in assembly, stability and regulation of PSI complexes (Ozawa et al. 2009). Thus, it is possible that such essential factors for PSI accumulation may be deficient in roots even under GLK overexpression. P700 measurements showed that PSI in GLK1_OX roots is in a more oxidized state than in leaf chloroplasts due to donor-side limitations (Fig. 9), suggesting that LHCI–PSI in GLK1_OX root chloroplasts functions effectively and electron transport to PSI is rather retarded. In GLK1_OX roots, the low photochemical efficiency of PSII (Fig. 7A, C) may lead to the donor-side limitation in PSI even though PSII is more abundant than PSI (Fig. 4, Table 2). Alternatively, it is possible that the electron transport chain downstream of PSII, such as Cyt b₆f complexes or plastocyanin, is impaired in the GLK1_OX root as observed in the lower qP level (Supplementary Fig. S3B).

GLK factors improve phototrophic growth in roots

Although the overexpression of GLKs could not improve photochemical efficiency in root chloroplasts, total photosynthetic activities measured in terms of O₂ evolution and CO₂ fixation greatly increased in those roots (Fig. 10A, B). Furthermore, GLK_OX roots showed the ability to perform photosynthesis that can maintain their weights without sugar (Fig. 10C), demonstrating that the mass production of photosynthetic machinery increases total CO₂ fixation in those roots. It is interesting to note that wild-type roots also showed gross O₂ evolution and CO₂ fixation activities although the level was much lower than in GLK_OX roots, demonstrating that root photosynthesis can contribute to carbon assimilation in Arabidopsis. Not only leaves but also other so-called non-photosynthetic organs such as stem, root, fruit and flower have the potential to perform photosynthetic CO₂ fixation (Aschan and Pfanz 2003). With the exception of aerial roots of some orchid species, which can serve as primary photosynthetic organs, most green photosynthesizing roots are considered to contribute to internal CO₂ recycling using respiratory-released CO₂ (Aschan and Pfanz 2003). Because Arabidopsis roots, which lack stomata, accumulate Chls inside the stele (Kobayashi et al. 2012a), the recycling of internally respired CO₂ may be the main role of photosynthesis there. As we reported previously, Chl accumulation in the wild-type Arabidopsis root occurs predominantly in the basal area near the hypocotyl junction and decreases toward the tip even when the light is equally irradiated to the whole root (Kobayashi et al. 2012a). Considering that only the upper parts of the root in the soil can receive light, the root base may have an inherent ability to develop chloroplasts and perform photosynthesis. As proposed in the case of photosynthesis in seeds and fruits (Kalachanis and Manetas 2010, Tschiersch et al. 2011), it is possible that the high antenna/reaction center ratio in root chloroplasts is of advantage in the low-light environments of roots growing in the soil.

In this study, we demonstrated that induction of Chl biosynthesis and chloroplast development by GLK overexpression increases photosynthesis in roots. Further increases in photosynthetic activity of roots achieved by improving the balance between antennae and reaction centers and the photochemical efficiency could promote overall biomass production in plants.

Materials and Methods

Plant materials and growth conditions

Arabidopsis thaliana wild type (Columbia), hy5-215 (Oyama et al. 1997), GLK1_OX and GLK2_OX (Waters et al. 2008) were grown vertically on solid medium containing 1× MS plant salt mixture (Wako), 1% (w/v) sucrose and 0.8% (w/v) gelrite (Wako), pH 5.7 at 23°C under continuous white light (60 µmol photons m⁻² s⁻¹) for 21 d unless stated otherwise.

Microscopic analyses

For Fig. 1B, primary roots at approximately 4.0 cm from the root–hypocotyl junction were examined using a confocal laser scanning microscope (LSM700; Carl Zeiss). Chl autofluorescence between 660 and 700 nm was detected under 488 nm laser excitation and merged with differential interference contrast images.

For ultrastructure analysis of root plastids, primary roots at approximately 4.0 cm from the root–hypocotyl junction were analyzed by TEM according to Toyooka et al. (2000) with modification (Kobayashi et al. 2012a). For quantification of the number and size of plastids, high-resolution TEM images prepared as described above were merged in a whole cross-section image of the stele of the primary root. All plastids observed in the outer three cell layers of the cross-section were counted and used for area quantification with ImageJ software (National Institutes of Health).

qRT–PCR analysis

Total RNA was extracted from roots of 21-day-old seedlings using the RNeasy Plant Mini kit (Qiagen). Genomic DNA digestion and reverse transcription were performed using the PrimeScript RT reagent Kit with gDNA Eraser (TAKARA BIO INC.) according to the manufacturer’s instructions. cDNA amplification was performed using the Thunderbird PreMix kit (Toyobo) and 200 nM gene-specific primers, listed in Supplementary Table S1. Thermal cycling consisted of an initial denaturation step at 95°C for 10 s, followed by 40 cycles of 5 s at 95°C and 30 s at 60°C. Signal detection and quantification were performed in duplicate using MiniOpticon (Bio-Rad). The relative abundance of all transcripts amplified was normalized to the constitutive expression level of ACTIN8 (Pfaffl 2001). Three independent biological experiments were performed for each root sample.

Immunoblot analysis

The membrane protein fraction was prepared from roots or leaves of 21-day-old seedlings as described previously (Kobayashi et al. 2013). A 10 µg aliquot of total membrane protein from each root sample was electrophoresed together with a dilution series (0.2–5 µg) of total membrane protein from wild-type leaves and electrotransferred onto PROTRAN nitrocellulose membranes (Schleicher & Schuell) as described (Kobayashi et al. 2007). Protein bands that reacted with primary antibodies were secondarily labeled with goat anti-rabbit IgG secondary antibody conjugated with horseradish peroxidase (Thermo Scientific) and detected using a chemiluminescence reagent (ImmunoStar LD, Wako) and an imager (ImageQuant LAS 4000 mini, GE Healthcare). Antibodies against PsaA/PsaB and total LHCII were kindly provided by R. Tanaka, Hokkaido University, Sapporo, Japan, and those against D1 and D2 were kindly provided by M. Ikeuchi, The University of Tokyo, Tokyo, Japan. Antibodies against PsbO, CP43, LHCA1, LHCA2, LHCB1 and LHCB6 were from AgriSera.

Spectroscopic pigment determination

Plant tissues crushed in liquid nitrogen were homogenized in 80% acetone, and debris was removed by centrifugation at 10,000×g for 5 min. The absorbance of the supernatant at 720, 663, 647, 645 and 470 nm was measured with an Ultrospec 2100 pro spectrophotometer (GE Healthcare Biosciences). The Chl (a and b) and carotenoid contents of the samples were determined according to Melis et al. (1987) and Lichtenthaler (1987), respectively.

HPLC pigment analysis

The leaf or root tissue was ground in a glass mortar for 1 min with approximately 30 ml of anhydrous Na₂HPO₄ (–4°C) as a desiccant. The ground material was transferred to a glass beaker, to which approximately 20 ml of chloroform (–20°C) was added, and the mixture was sonicated for 2 min at 4°C. The extract was then filtered and dried under a vacuum. The green solid material obtained by the above procedure was immediately redissolved in approximately 20 µl of chloroform. For carotene, Phe a, Chl a′ and Chl a analyses, a 3–5 µl aliquot of the extract was injected into a silica HPLC column (YMC-pak SIL, 250×4.6 mm i.d.) cooled to approximately 4°C in an ice–water bath. The sample was eluted isocratically with degassed hexane/2-propanol/methanol (100/0.7/0.2, v/v/v) at a flow rate of 0.9 ml min⁻¹, and was monitored with a JASCO UV-970 detector (λ = 425 nm) and a JASCO Multiwavelength MD-915 detector (λ = 300–800 nm) in series. For xanthophyll and Chl a analyses, an aliquot of 3–5 µl of extract was injected onto a reversed-phase HPLC column (Kaseisorb LC ODS 2000-3, 250×4.6 mm i.d.) cooled to 4°C in an ice–water bath. The pigments were eluted isocratically with degassed water/ethanol/methanol/2-propanol (3/86/13/1, v/v/v/v) at a flow rate of 0.3 ml min⁻¹, and monitored with a JASCO UV-2070 detector (λ = 425 nm) and a SHIMADZU Multiwavelength SPD-M10A VP photodiode array detector (λ = 300–800 nm) in series.

Chl fluorescence measurement

Fluorescence emission spectra of Chl proteins at 77K were obtained directly from plant tissues in liquid nitrogen using a spectrofluorometer under 435 nm excitation (RF-5300PC, Shimadzu).

For Chl fluorescence induction experiments (Fig. 5B, C), leaves or 1 cm segments of primary roots from the hypocotyl junction were excised and dark-incubated for 5 min before initiation of the experiments. Five segments of primary roots were used for an experiment in a batch. When required, the tissues were infiltrated with 40 µM DCMU and 150 mM sorbitol by depression before dark incubation. Chl fluorescence transients were measured in a logarithmic time series between 30 µs and 10 s after the onset of strong actinic light (1,650 µmol photons m⁻² s⁻¹) with an LED pump-probe spectrometer (JTS-10, BioLogic).

Photochemical efficiency analysis was performed using a PAM fluorometer (Junior-PAM, Walz) at room temperature. Plants were pre-incubated under dim light (∼5 µmol photons m⁻² s⁻¹) for >30 min before experiments. Then, single leaves or batches of 1 cm segments of the primary roots from the hypotocyl junction were dark-incubated for 5 min on the leaf clip with the MS medium. After measuring minimum Chl fluorescence (F_o) in the dark, maximal Chl fluorescence (F_m) was determined with a saturating pulse. After actinic light treatment for 3 min, stationary fluorescence (F) and maximum fluorescence under the actinic light (F_m′) were determined followed by the measurement of minimal fluorescence of illuminated samples (F_o′) after far-red light treatment. From these fluorescence yields, photosynthetic parameters were calculated according to the following equations (Van Kooten and Snel 1990, Maxwell and Johnson 2000): F_v/F_m = (F_m – F_o)/F_m, F_v′/F_m′ = (F_m′ – F_o′)/F_m′, Φ_II = (F_m′ – F)/F_m′, qP = (F_m′ – F)/(F_m′ – F_o′), qN = 1 – (F_m′– F_o′)/(F_m – F_o). The actual PSII efficiency Φ_II can be transformed into a product of the PSII openness (qP) and the quantum efficiency of open PSII (F_v′/F_m′): Φ_II = (F_m′ – F)/F_m′ = qP×F_v′/F_m′. The qL, Φ_NPQ and Φ_NO were determined according to the method of Kramer et al. (2004).

P700 absorbance measurement

The redox state of P700 was determined in a batch of leaves or roots from 21-day-old seedlings by measuring the absorbance change at 830–875 nm reference beams using Dual-PAM-100 (Waltz) at room temperature (Klughammer and Schreiber 2008). The measurements were performed using the automated induction and recovery program provided by the Dual-PAM software (Pfundel et al. 2008). The maximal P700 signal (P_m) was determined by application of a saturating pulse (15,000 µmol photons m⁻² s⁻¹) after far-red pre-illumination. Then actinic light (126 µmol photons m⁻² s⁻¹) was supplied for 10 min and saturating pulses were given every 20 s to determine the maximum P700 signal under the actinic light (P_m′). Each saturating pulse was followed by a 1 s dark interval to determine the minimum level of the P700⁺ signal (P_o). Quantum yields in PSI were calculated according to Pfundel et al. (2008) although the correct interpretation of these parameters remains debatable.

Oxygen evolution and CO₂ fixation analysis

For O₂ evolution and CO₂ fixation analysis, roots (approximately 0.2 g FW) excised from 21-day-old seedlings were used. The O₂ evolution rate was determined with a Clark-type oxygen electrode (LD2, Hansatech Instruments). Roots wetted with liquid MS medium were incubated in an air chamber with 5% CO₂ at 23°C. Changes in O₂ concentration in the chamber were monitored with the electrode under several light conditions (0–400 µmol photons m⁻² s⁻¹).

The CO₂ fixation rate was measured with a portable infrared gas analyzer (LI-6400, LI-COR) and a 6400-17 whole plant Arabidopsis chamber. During the measurement, temperature and CO₂ concentration in the chamber were kept at 23°C and 390 µl CO₂ l⁻¹, respectively.

Dry weight measurement

Roots detached from 21-day-old seedlings at the root–hypocotyl junction were collected to determine the initial weight or further incubated for 7 d on vertical MS medium without sucrose under continuous light (60 µmol photons m⁻² s⁻¹) or in the dark. Collected samples were dehydrated completely at 70°C for 48 h with dry silica gel and immediately weighed to avoid water absorption from the air.

Supplementary data

Supplementary data are available at PCP online.

Funding

This work was supported by the Ministry of Education, Sports, Science and Culture in Japan [Grants-in-Aid for Scientific Research on Priority Areas (Nos. 24570042, 22370016 and 24770055) and the Global Center of Excellence Program (K03)]; the Japan Society for the Promotion of Science [research fellowships of for young scientists] and a RIKEN post-doctoral fellowship (to K.K.); the Howard Hughes Medical Institute and the Gordon and Betty Moore Foundation [grant GBMF3070 to K.K.N.].

Glossary

Abbreviations

DCMU

3-(3,4-dichlorophenyl)-1,1-dimethylurea

F_v/F_m

maximum quantum efficiency of PSII

F_v′/F_m′

maximum quantum efficiency of open PSII

GLK

golden2-like

GLKox

GLK overexpressor

HY5

long hypocotyl 5

LHC

light-harvesting complex

MS

Murashige and Skoog

NEP

nuclear-encoded plastid RNA polymerase

O-J-I-P

origin-inflection-intermediary peak–peak

PAM

pulse amplitude modulation

PAR

photosynthetically active radiation

PEP

plastid-encoded RNA polymerase

Phe

pheophytin

qI

photoinhibitory component of qN

qL

coefficient of open PSII on the basis of the ‘lake’ model

qN

coefficient of non-photochemical quenching

qP

coefficient of photochemical quenching

qRT–PCR

quantitative reverse transcription–PCR

SIGs

sigma factors

TEM

transmission electron microscopy

Φ_I

photochemical quantum yield of PSI

Φ_II

quantum yield of PSII

Φ_NA

quantum yield of non-photochemical energy dissipation in PSI due to acceptor-side limitations

Φ_ND

quantum yield of non-photochemical energy dissipation in PSI due to donor-side limitations

Φ_NO

quantum yield of non-light induced non-photochemical quenching

Φ_NPQ

quantum yield of light-induced non-photochemical quenching

Disclosures

The authors have no conflicts of interest to declare.