Impairment of Lhca4, a subunit of LHCI, causes high accumulation of chlorophyll and the stay-green phenotype in rice

Impairment of Lhca4, a subunit of LHCI, causes high chlorophyll content and LHCII accumulation in rice, suggesting a novel functional interaction between LHCI and LHCII.


Introduction
Chlorophyll synthesis and breakdown are strictly regulated in plants not only because it is an essential photosynthetic molecule, but also because in its free form it photo-oxidatively damages cells (Tanaka et al., 2011). Mutants that retain the greenness of leaves under senescence-inducing conditions are called stay-green mutants, and they are classified into two types: functional and non-functional. Functional stay-green mutants retain the green of leaves via delayed senescence, whereas the greenness of leaves is not necessarily correlated with leaf functionality in the non-functional mutants. The non-functional stay-green mutants can be further classified (Thomas and Howarth, 2000). The Type C mutants show impairment in chlorophyll degradation, with the majority of the mutations being in the genes that encode chlorophylldegrading enzymes, such as the Chla-degrading enzyme SGR/ NYE1, Chlb-degrading enzymes NYC1 and NOL, and pheophytinase PPH/NYC3 (Ren et al., 2007;Kusaba et al., 2007;Park, 2007;Morita et al., 2009;Sato et al., 2009;Schelbert et al., 2009;Shimoda et al., 2016). In addition, mutants of a chloroplast protein THF1/NYC4 and a small subunit of the PSII core complex, PsbM/CytG, show the stay-green phenotype via inhibition of the degradation of chlorophyllprotein complexes (Huang et al., 2013;Yamatani et al., 2013;Kohzuma et al., 2017). Although the genes associated with these mutations do not encode chlorophyll-degrading enzymes, the mutants are classified as Type C because chlorophyll degradation is impaired during senescence. Another class of non-functional stay-green mutants, Type E, accumulates high levels of chlorophyll in the pre-senescent leaves, resulting in longer retention of greenness than in those of the wild-type. In other words, Type E can be designated as a mutant with impairments in the proper regulation of chlorophyll accumulation.
Chlorophyll is contained in the photosystems in the thylakoid membranes. PSII uses light energy for the extraction of electrons from water, resulting in the production of oxygen. The photo-excited electrons are transferred to PSI through the cytochrome b 6 f complex and are used to produce NADPH. The proton gradient formed during the electron transfer process is used for the production of ATP. The light-harvesting complex I (LHCI) is an antenna complex for PSI, and consists of four subunits, Lhca1-Lhca4, in algae and land plants. Lhca1/Lhca4 and Lhca2/Lhca3, respectively, form dimers. The light-harvesting complex II (LHCII) is an antenna complex for PSII, encoded by Lhcb1-Lhcb6. Lhcb1, 2, and 3 are the major LHCII subunits and form trimers. Lhcb4, 5, and 6 are minor LHCII subunits that exist as monomers. Higher plants contain Chla and Chlb. Chla is a major chlorophyll found in all chlorophyll-protein complexes, whereas Chlb is found only in the light-harvesting complexes, LHCI and LHCII.
For optimal photosynthesis, the performances of PSI and PSII should be balanced (Minagawa, 2011). Photosynthetic organisms have several mechanisms to tune the performances of PSI and PSII. State transition is one such mechanism, wherein trimeric LHCII moves to PSI via phosphorylation by STN7 kinase, and functions as an antenna of PSI under 'PSII-excess' light conditions (Depège et al., 2003;Bellafiore et al., 2005). This condition is called State 2, while State 1 is induced by a 'PSI-excess' situation. State transition is a short-term acclimation under changing light conditions. In addition, plants perform long-term acclimation as well. For example, under lengthy PSII-excess conditions, Arabidopsis increases PSI by increasing the psaA and psaB mRNA levels (Pesaresi et al., 2009). LHCII is typically considered an antenna of PSII, except under specific conditions, such as State 2. However, accumulating evidence suggests that LHCII acts as an antenna of PSI even under moderate light conditions (Wientjes, et al., 2013;Grieco et al., 2015). LHCII could be an efficient antenna for PSI, and trimeric LHCII interacts with PSI even in State 1, where LHCII phosphorylation is absent (Benson et al., 2015;Grieco et al., 2015).
In this study, we isolated a novel rice mutant that retains green of leaves in the field in autumn. This stay-green mutant, delayed yellowing1 (dye1), accumulates a higher amount of chlorophyll than the wild-type in the pre-senescent leaves, suggesting that it is a Type E mutant. To our knowledge, this is the first empirical report on Type E stay-green mutants. Positional cloning revealed that DYE1 encodes Lhca4, a subunit of LHCI. Recent reports using Lhca mutants in Arabidopsis have revealed various functions of LHCI in addition to its role as an antenna of PSI, such as its involvement in state transition (Benson et al., 2015;Bressan et al., 2016). Interestingly, Lhcb1 was highly accumulated among the chlorophyll-protein complexes in the pre-senescent dye1-1 leaves, suggesting that there is a novel functional interaction between LHCI and LHCII.

Materials and methods
Plant material dye1-1 was isolated from a rice M 2 population (Oryza sativa L. 'Nipponbare') irradiated with carbon ion beams (1.6 GeV). dye1-2 was isolated from the N-methyl-N-nitrosourea-mutagenized Nipponbare pool by using TILLING-based screening (Suzuki et al., 2008). The plants were cultivated in pots under field conditions.

Photosynthetic parameters
Foliar chlorophyll content was measured non-destructively using a SPAD-502 Plus instrument (KONICA MINOLTA; http://www. konicaminolta.jp, last accessed 24 December 2017). For pigment extraction, leaves were ground in a mortar in liquid nitrogen and extracted using 80% acetone. Chla and Chlb levels were determined as described by Porra et al. (1989). Maximum quantum yield (F v /F m ) and the carbon fixation rate were measured using a LI-6400XT portable photosynthesis system (LI-COR; http://www.licor.com/, last accessed 24 December 2017) at 30 °C, with a CO 2 concentration of 400 ppm, humidity of 70-80%, and photon flux density of 2000 μmol m -2 s -1 . Oxidation of P700 was measured using a Dual-PAM-100 chlorophyll fluorescence and P700 photosynthesis analyser (Walz; http://www.walz.com/, last accessed 24 December 2017). The antenna function of LHCI was estimated as the time required to reach two-thirds of the maximum P700 oxidation with far-red light (intensity 2). State transition (F r ) was determined as described by Lunde et al. (2000) using a JUNIORPAM fluorometer (Walz). F r is calculated using the following formula: . F i and F ii are the fluorescence in the presence of PSI light in States 1 and 2, respectively, whereas F i´ and F ii´ are the fluorescence in the absence of PSI light in State 1 and 2, respectively.
Quantitative RT-PCR Total RNA was extracted from the leaves of the wild-type and dye1-1 plants using a total RNA extraction kit (RBC Bioscience; http://www.rbcbioscience.com/, last accessed 24 December 2017). First-strand cDNA was synthesized from 500 ng total RNA using ReverTra ACE qPCR RT Master Mix with gRNA Remover (TOYOBO; http://www.toyobo.co.jp/, last accessed 24 December 2017). The transcript level was determined by quantitative RT-PCR using a KAPA SYBR FAST qPCR kit (KAPA Biosystems; http:// www.kapabiosystems.com/, last accessed 24 December 2017) and a Rotor-Gene Q real-time PCR cycler (Qiagen; http://www.qiagen. com/, last accessed 24 December 2017). The primers used for amplification are listed in Supplementary Table S1 at JXB online.

Protein analysis
Total protein was extracted from a 100-mg (FW) sample of leaves from both the wild-type and dye1-1, using 400 µl of 2×SDS buffer [0.125 M Tris, pH 6.8, 4% SDS, 4% mercaptoethanol, 1% bromophenol blue (BPB), 20% glycerol]. The extracted proteins were diluted to one-fifth concentration using 1×SDS buffer (62.5 mM Tris, pH 6.8, 2% SDS, 2% mercaptoethanol, 0.5% BPB, 10% glycerol) and subjected to SDS-PAGE with or without boiling. The antibodies against Lhca1-Lhca4, and Lhcb1 for western blot analysis were purchased from Agrisera (http://www.agrisera.com/en/info/home. html, last accessed 24 December 2017), while the anti-PsaF antibody was provided by Y. Takahashi (Graduate School of Natural Science and Technology, Okayama University, Japan). The antibody against D1 was described previously by Kato et al. (2012). Detection of each protein was performed using an ECL Prime western blotting detection system (GE Healthcare; http://www3.gehealthcare.com/, last accessed 24 December 2017) and an ODYSSEY Fc imaging system (LI-COR). Quantification of band intensity in the western blot analysis was performed using Image Studio Ver 5.2 (LI-COR). An SDS-PAGE gel was stained by Coomassie Brilliant Blue R-250 to detect the Rubisco large subunit. Blue native PAGE analysis was performed using thylakoids solubilized in 1% β-dodecyl-maltoside, as described by Yamatani et al. (2013).

Positional cloning
Twenty-two stay-green F 2 segregants from a cross between dye1-1 and the japonica rice Gimbozu EG4 were used for coarse-mapping. For fine-mapping, about 3000 F 2 plants from a cross between dye1-1 and a line from the chromosome segment substitution line (CSSL) between Koshihikari and Kasalath (SL224) were used (Ebitani et al., 2005). The DNA markers used in positional cloning are described in Supplementary Table S2.

Whole-genome sequencing
Whole-genome sequencing of dye1-1 was performed with HiSeq2000 (Illumina; https://www.illumina.com/, last accessed 24 December 2017). Three candidate mutations against the Nipponbare genome sequence with quality over 50 were detected within the 43.1-kb DYE1 candidate region. Among the three candidate mutations, the G-to-A substitution in the second exon of Lhca4 was the only 'homozygous' mutation.

Isolation of a rice stay-green mutant, dye1
In a screening of rice mutants in the field, we isolated one that showed delayed yellowing during natural senescence. This recessive mutant, named delayed yellowing1-1 (dye1-1), was greener than the wild-type cultivar, Nipponbare, 5 weeks after heading, when most leaves are senescent (Fig. 1A). Measurement of the chlorophyll content of flag leaves during the ripening period showed that dye1-1 had higher chlorophyll contents not only 6 weeks after heading (SPAD units: 9.28 ± 0.83 in Nipponbare, 20.96 ± 1.48 in dye1-1) but also 1 week before heading, i.e. before leaf senescence set in (23.98 ± 0.96 in Nipponbare, 30.54 ± 1.04 in dye1-1) (Fig. 1B). The shoots of Nipponbare and dye1-1 plants were harvested 1 week before heading and there was no significant difference between their weights (Fig. 1C).

Positional cloning of DYE1
To isolate the DYE1 gene, we obtained an F 2 population between dye1-1 and Gimbozu EG4, a japonica strain that harbors over a 1000 copies of mPing, a mobile MITE in rice (Nakazaki et al., 2003;Naito et al., 2006). Transposed mPings can be used as sequence characterized amplified region (SCAR) makers for gene mapping (Monden et al., 2009). Coarse-mapping of 22 plants showing the stay-green phenotype selected from this segregating population detected a linkage between DYE1 and the mPing-SCAR marker MK8-6 (77.6 cM) on Chromosome 8 (Fig. 3A). For fine-mapping, we generated an F 2 population from a cross between dye1-1 and a CSSL that had its Chromosome 8 replaced with that of the indica cultivar Kasalath. Genotyping around 3000 F 2 plants and their progeny revealed several recombinants near the DYE1 candidate region (see Supplementary Fig. S1). Analysis of the genotype and phenotype of the recombinants revealed  that DYE1 is located between the derived cleaved-amplified polymorphic sequence (dCAPS) markers SNP-3 _RC and dCAPS5 (Fig. 3A). This 43.1-kb candidate region contains eight functional genes. Whole-genome sequencing of dye1-1 using Illumina HiSeq2000 revealed a G-to-A substitution in the second exon of Lhca4 (Os08g0435900), which was the only mutation in the candidate region. This single-base change could cause amino acid substitution from Glu to Lys at position 146 from the first Met in Lhca4, which is a subunit of LHCI. This amino acid residue is highly conserved among the Lhca and Lhcb subunits, and is involved in pigment binding (Melkozernov and Blankenship, 2003). Taken together with the fact that rice has only one copy of the Lhca4 gene in its genome, it suggests that this amino acid substitution causes a severe impairment of Lhca4 function (see Supplementary  Fig. S2) (Melkozernov and Blankenship, 2003;Klimmek et al., 2006). Western blot analysis revealed that the content of the Lhca4 apoprotein was severely reduced in dye1-1, suggesting that E146K substitution drastically reduces the stability of the Lhca4 protein ( Supplementary Fig. S3).
dye1-2, another allele of dye1, isolated by TILLING-based screening (Suzuki et al., 2008) of the Nipponbare mutant population, was found to have a single-base change, causing substitution from Val to Met at position 107 from the first Met (Fig. 3A). dye1-2 showed higher chlorophyll content than Nipponbare 1 week before heading, and the stay-green phenotype during natural senescence, albeit to a weaker degree than dye1-1 (Fig. 1B).
A complementation experiment was designed to confirm that DYE1 encodes Lhca4, and was performed via the Agrobacterium-mediated transformation method, wherein dye1-1 was transformed with a 6-kb genomic fragment carrying the entire coding region of the wild-type Lhca4 gene. This showed that Lhca4 content was significantly reduced in dye1-1, but the complementation lines accumulated normal amounts of Lhca4, as expected (see Supplementary Fig. S3). In these complementation lines, the chlorophyll contents of the flag leaves 1 week before heading were similar to those in Nipponbare (Fig. 3B). These results confirmed that DYE1 encodes Lhca4.

Photosynthetic properties of dye1
We examined the photosynthetic properties of dye1-1 using flag leaves at heading, since DYE1 encodes a subunit of LHCI. Both Chla and Chlb contents were higher in dye1-1 than in Nipponbare (1.85 ± 0.09 and 0.57 ± 0.02 nmol mg -1 FW for Chla and Chlb, respectively, in Nipponbare; 2.66 ± 0.08 and 0.79 ± 0.04 nmol mg -1 FW for Chla and Chlb, respectively, in dye1-1), but the Chla/b ratio was similar (see Supplementary Table S3). F v /F m was slightly higher in dye1-1 than in Nipponbare (Supplementary Table S3). To examine the structure of PSI-LHCI super-complexes, blue native PAGE analysis was performed using mature leaves (Fig. 4). In dye1-1, two bands emerged (indicated by red arrows in Fig. 4), which were not observed in Nipponbare. The upper band is thought to correspond with the PSI-LHCI super-complex lacking Lhca4, and the lower band corresponds with the PSI core complex lacking all Lhca subunits. The PSI core complex band was prominent in dye1-1, suggesting a severely defective organization of the PSI-LHCI super-complex. Consistent with this observation, the kinetics of P700 oxidation induced by far-red light, which reflects the antenna function of LHCI (Gobets and van Grondelle, 2001;Bonente et al., 2012), was found to be much slower in dye1-1 (Supplementary Table S3). Interestingly, the F r value, an indicator of state transition, was very low in dye1-1 and moderately low in dye1-2, which was consistent with the severity of their phenotype (Fig. 5) (Lunde et al., 2000). A similar observation has been reported in Lhca4 and other Lhca subunit mutants in Arabidopsis (Benson et al., 2015). It is likely that the reduced state transition in dye1-1 was due to the reduced function of LHCI, and was not specific to Lhca4.

Analysis of chlorophyll-protein complexes in dye1
The results of western blot analyses, performed for a number of photosynthetic proteins (Fig. 6), showed that the Lhca4 content was severely reduced in pre-senescent (at heading) and senescent (4 weeks after heading) leaves of dye1-1. E146K substitution is thought to influence the stability of the Lhca4 protein, as mentioned above. Other Lhca subunits, Lhca1-Lhca3, were also reduced in content in the pre-senescent and senescent leaves of dye1-1. It is very likely that the impairment of Lhca4 destabilizes other Lhca subunits because of defects in the proper formation of the PSI-LHCI supercomplex. Slight increases in the subunits of the PSII core, D2 (1.4-fold) and CP47 (1.3-fold), were observed in the dye1-1 flag leaves at heading, suggesting that the amount of PSII core complex increased slightly. In addition, a slight increase in the PSI core subunits PsaF (1.5-fold) and PsaH (1.5-fold) were observed in the dye1-1 flag leaves at heading, suggesting a slight increase in the PSI core complex. Interestingly, a more prominent increase was observed in a trimeric LHCII subunit in dye1-1. The content of Lhcb1 in the flag leaves at heading in dye1-1 was 2.6-fold that of the equivalent leaves in Nipponbare. In contrast, there was no significant increase in the content of Lhcb5, a monomeric LHCII subunit, in dye1-1. Levels of several chlorophyll-protein complexes were elevated in dye1-1, suggesting that the higher chlorophyll content in the pre-senescent dye1-1 leaves was due to an increase in such complexes, particularly trimeric LHCII.
In rice, there are two copies of the Lhcb1 gene, Lhcb1a and Lhcb1b. qPCR analysis revealed no significant up-regulation of Lhcb1a and Lhcb1b expression in dye1-1, suggesting that the increase in Lhcb1 protein in dye1-1 was not regulated at   the mRNA level (Fig.7). Similarly, there was no increase in the mRNA of the PSI core subunit PsaA in dye1-1. As it is known that LHCII accumulation is regulated by chlorophyll content, particularly Chlb, the expression of the genes involved in chlorophyll synthesis was investigated. The mRNA level of HemA1, which encodes the rate-limiting enzyme of tetrapyrrole synthesis, Glu-tRNA reductase, was not significantly different between Nipponbare and dye1-1 (Tanaka et al., 2011) (see Supplementary Fig. S4). Similarly, the mRNA level of CAO, which encodes the Chlb-synthesizing enzyme chlorophyllide a oxygenase, was comparable between Nipponbare and dye1-1 (Supplementary Fig. S4).

Discussion
dye1 was isolated as a delayed-yellowing mutant of rice in a field experiment. The pre-senescent mature leaves of dye1 contained a higher level of chlorophyll compared with those of the wild-type, suggesting that this elevated chlorophyll content causes the 'stay-green' phenotype of the senescent leaves.
However, dye1 showed comparable levels of CO 2 assimilation rate and expression of senescence-inducible genes in the senescent leaves, suggesting that it is a non-functional stay-green mutant. Taken together, the results suggest that dye1 is a Type E stay-green mutant (Thomas and Howarth, 2000). To our knowledge, this is the first empirical report on such a mutant.
Positional cloning revealed that DYE1 encodes Lhca4, a subunit of the PSI antenna complex LHCI. While dye1 is the only Lhca mutant reported in rice so far, two Lhca mutants have been reported in Arabidopsis, including a quadruplemutant of Lhca1-Lhca4 (ΔLhca) (Ganeteg et al., 2004;Benson et al., 2015;Bressan et al., 2016). These mutants are reported to show reduced state transition, and slightly higher F v /F m values and Chla/b ratios (Benson et al., 2015;Bressan et al., 2016), which are common to dye1-1. In addition to these characteristics, dye1-1 showed higher chlorophyll content, which has not been described in the Lhca mutants in Arabidopsis. The fact that dye1-1, a mutant of a chlorophyll-protein complex, has a higher chlorophyll content than the wild-type is thought to be due to a higher accumulation of other chlorophyll-protein complexes, particularly the trimeric LHCII.
dye1-1 has an E146K amino acid substitution in Lhca4. This residue corresponds to E154 in Arabidopsis Lhca4, which is a pigment-binding site conserved not only among the Lhca subunits, but also among the Lhcb subunits of different species (Melkozernov and Blankenship, 2003). Furthermore, western blot analysis revealed that the Lhca4 apoprotein content was severely reduced in dye1-1. Taken together, these results suggest a severe impairment of Lhca4 function in dye1-1. Interestingly, the phenotype of the Lhca4 mutant is the severest among the single Lhca-subunit mutants in Arabidopsis (Benson et al., 2015). This is partly because the impairment of Lhca4 drastically influences the stability of other LHCI subunits and the organization of the PSI-LHCI super-complex. Indeed, the kinetics of P700 oxidation by far-red light was much slower in dye1-1. Nevertheless, the biomass and CO 2 assimilation efficiency in dye1-1 were comparable with those of the Nipponbare pre-senescent leaves.
Plants adapt to an imbalance in PSI/PSII through shortand long-term acclimations. In Arabidopsis, long-time exposure of PSII-activating light dramatically increases the transcription of psaA/pasB and the accumulation of PSI core proteins (Pesaresi et al., 2009). Because the LHCI antenna was reduced in dye1, the activity of PSI is thought to be continuously low. However, only a slight increase in the PSI core and no increase in psaA transcript accumulation were observed in dye1-1, suggesting that the mechanism of long term-acclimation in rice is different from that in Arabidopsis. In contrast, higher LHCII accumulation was observed in dye1-1. It is possible that the increase in trimeric LHCII accumulation is a kind of compensation for a reduced PSI antenna. LHCII has been classically thought to be the antenna of PSII, but increasing evidence suggests that it could be a major antenna of PSI, even under normal conditions. Indeed, Bressan et al. (2016) suggested that LHCII can compensate for the deficiency of LHCI, although not perfectly. The mRNA levels of Lhcb1a and Lhcb1b did not increase in dye1-1, suggesting that the increase in Lhcb protein was regulated not at the mRNA level, but at the protein level. For instance, it is suggested that the gunrelated retrograde signaling, which regulates Lhcb transcription, is not involved in this phenomenon (Kleine and Leister, 2016). Lhcb protein accumulation is known to be regulated by chlorophyll content, particularly by Chlb (Bellemare et al., 1982;Horn and Paulsen, 2004;Kusaba et al., 2007); however, expression of the enzyme genes HemA1 and CAO involved in chlorophyll synthesis did not differ significantly between Nippponbare and dye1-1, even though these enzymes are also under post-translational control (Nakagawara et al., 2007;Tanaka et al., 2011). Thus, the detailed mechanism behind the increase in LHCII content in dye1-1 remains unclear. An increase in the LHCII content in dye1-1 might be a compensation for the low activity of PSI, considering that the biomass and carbon assimilation rate of dye1-1 were comparable to those of the wild-type, despite very low antenna function of LHCI. Consistent with this idea, Lhca4-antisense plants in Arabidopsis show considerable growth defects under field and controlled conditions (Ganeteg et al., 2004). However, the possible involvement of reduced state transition cannot be excluded. In any case, our observation that the impairment of Lhca4 caused increased LHCII accumulation suggests that LHCI and LHCII functionally interact in a previously unknown manner.

Supplementary data
Supplementary data are available at JXB online. Fig. S1. Genotype and phonotype of recombinant individuals in the DYE1 candidate region. Fig. S2. Alignment of Lhca4 proteins from different species. Fig. S3. Western blot analysis of Lhca4 in the complementation line. Fig. S4. Expression of HemA1 and CAO in dye1-1. Table 1. Primers used in the quantitative RT-PCR Table 2. Information on dCAPS markers used in the positional cloning of DYE1. Table 3. Analysis of the photosynthetic properties of the pre-senescent flag leaves in dye1-1.