Phosphorylation of GENOMES UNCOUPLED 4 Alters Stimulation of Mg Chelatase Activity in Angiosperms^,

GENOMES UNCOUPLED 4, a positive regulator of tetrapyrrole biosynthesis, is phosphorylated at serine 264 within a conserved C-terminal motif leading to altered stimulation of Mg chelatase in angiosperms.

Abstract

GENOMES UNCOUPLED 4 (GUN4) is a positive regulator of light-dependent chlorophyll biosynthesis. GUN4 activates Mg chelatase (MgCh) that catalyzes the insertion of an Mg²⁺ ion into protoporphyrin IX. We show that Arabidopsis (Arabidopsis thaliana) GUN4 is phosphorylated at Ser 264 (S264), the penultimate amino acid residue at the C terminus. While GUN4 is preferentially phosphorylated in darkness, phosphorylation is reduced upon accumulation of Mg porphyrins. Expression of a phosphomimicking GUN4(S264D) results in an incomplete complementation of the white gun4-2 null mutant and a chlorotic phenotype comparable to gun4 knockdown mutants. Phosphorylated GUN4 has a reduced stimulatory effect on MgCh in vitro and in vivo but retains its protein stability and tetrapyrrole binding capacity. Analysis of GUN4 found in oxygenic photosynthetic organisms reveals the evolution of a C-terminal extension, which harbors the phosphorylation site of GUN4 expressed in angiosperms. Homologs of GUN4 from Synechocystis and Chlamydomonas lack the conserved phosphorylation site found in a C-terminal extension of angiosperm GUN4. Biochemical studies proved the importance of the C-terminal extension for MgCh stimulation and inactivation of GUN4 by phosphorylation in angiosperms. An additional mechanism regulating MgCh activity is proposed. In conjunction with the dark repression of 5-aminolevulinic acid synthesis, GUN4 phosphorylation minimizes the flow of intermediates into the Mg branch of the tetrapyrrole metabolic pathway for chlorophyll biosynthesis.

Photosynthetic organisms produce the greatest diversity of tetrapyrroles for multiple biochemical and regulatory processes. The metabolic pathway of tetrapyrrole biosynthesis (TBS) supplies chlorophyll (Chl) for photosynthesis, heme for redox reaction, siroheme for nitrogen and sulfur assimilation, phytochromobilin for the control of photomorphogenesis, and in cyanobacteria and red algae phycobilins for light harvesting. This essential pathway starts with the synthesis of 5-aminolevulinic acid (ALA), the first precursor unique to the TBS pathway. Condensation of eight ALA molecules ultimately leads to protoporphyrin IX (Proto), into which Mg²⁺ and Fe²⁺ are inserted to produce Chl and heme, respectively. Chelation of Mg²⁺ is catalyzed by the ATP-consuming Mg chelatase (MgCh), which consists of the three subunits CHLH, CHLD, and CHLI, with the H-subunit binding the substrate (Fuesler et al., 1981; Walker and Weinstein, 1991; Gibson et al., 1999; Gräfe et al., 1999; Jensen et al., 1999; Karger et al., 2001; Lundqvist et al., 2010; Chen et al., 2015a). To prevent the accumulation of free and phototoxic metabolic intermediates, TBS is subject to tight transcriptional and posttranslational control in response to endogenous and environmental stimuli (Mochizuki et al., 2010; Tanaka et al., 2011; Brzezowski et al., 2015). Posttranslational control of TBS involves multiple mechanisms, including assembly of bi- and multi-molecular protein complex and redox control (Stenbaek and Jensen, 2010; Czarnecki and Grimm, 2013; Richter and Grimm, 2013). TBS is coordinated with the synthesis of apoproteins, which assemble with tetrapyrrole end products as cofactors and pigments (Plumley and Schmidt, 1995). These synchronized processes are under spatio-temporal control and are likely organized by means of auxiliary and assembly factors (Reisinger et al., 2008; Tanaka et al., 2010; Komenda and Sobotka, 2016). The first regulatory proteins of TBS have been shown to control metabolic flux through the pathway by modulating protein stability, enzyme activation, and inactivation (Meskauskiene et al., 2001; Tanaka et al., 2010; Czarnecki et al., 2011; Czarnecki and Grimm, 2012; Nishimura et al., 2013; Apitz et al., 2016).

One of the TBS regulators is GENOMES UNCOUPLED 4 (GUN4). It is essential for Chl accumulation during photoperiodic growth, but not for Chl synthesis per se (Larkin et al., 2003; Wilde et al., 2004; Sobotka et al., 2008; Peter and Grimm, 2009; Formighieri et al., 2012; Brzezowski et al., 2014). GUN4 has a stimulatory impact on MgCh via a proposed mechanism, which involves its interaction with CHLH and binding of both the enzyme’s substrate Proto and product Mg Proto (MgP; Larkin et al., 2003; Davison et al., 2005; Verdecia et al., 2005; Adhikari et al., 2009, 2011; Chen et al., 2015b; Kopečná et al., 2015). Low light-grown wild-type plants accumulate less GUN4 than plants under increasing light intensities, indicating that the GUN4 content likely correlates with the need of MgCh activity (Peter and Grimm, 2009). It has been suggested that GUN4’s action on MgCh significantly reduces the Mg²⁺ concentration needed for the chelation reaction (Davison et al., 2005), but the mechanistic basis for its effect on the MgCh reaction remains unclear even though the crystal structures of GUN4 and CHLH was resolved (Davison et al., 2005; Verdecia et al., 2005; Chen et al., 2015a, 2015b; Tarahi Tabrizi et al., 2015). Analysis of rice (Oryza sativa/ Os) GUN4 suggests that the C-terminal region of GUN4 is important for stimulation of in vitro MgCh activity (Zhou et al., 2012).

GUN4 was identified in a screen for gun mutants impaired in the retrograde signaling pathway between chloroplasts and the nucleus when seedlings were treated with norflurazon, an inhibitor of carotenoid biosynthesis (Susek et al., 1993; Mochizuki et al., 2001). Five of the six known gun mutants encode mutant proteins involved in TBS (Susek et al., 1993; Mochizuki et al., 2001; Larkin et al., 2003; Koussevitzky et al., 2007; Woodson et al., 2011) so that tetrapyrrole metabolites and/or enzymes of the pathway were proposed to play a role in communicating the plastid status toward the nucleus for modulation of the nuclear gene expression (Woodson et al., 2011; Jarvis and López-Juez, 2013; Chan et al., 2016). Particularly for GUN4, a role in singlet oxygen specific signaling is discussed (Brzezowski et al., 2014; Tarahi Tabrizi et al., 2016). However, apart from the gun-mediated retrograde signaling, various stimuli and signals, which originate from different plastid localized processes like redox control during photosynthesis, plastid gene expression, and tetrapyrrole or carotenoid biosynthesis, were identified (Chan et al., 2016; Singh et al., 2015; Leister, 2016).

Three allelic gun4 mutants have been reported. The gun4-1 mutation causes a substitution of Leu 88 with Phe, which impacts the stability of GUN4 (Larkin et al., 2003). The gun4-3 allele harbors a T-DNA within the coding region of the C terminus, leading to truncated and unstable GUN4. Both mutants show a strong reduction of GUN4 content, resulting in reduced MgCh stimulation and, ultimately, reduced Chl level. The allelic gun4-2 null mutant is not viable under photoperiodic low-light conditions and on soil, but accepts growth on sugar-supplied media at constant dim light. Under those conditions, the gun4-2 accumulates only traces of Chl (Larkin et al., 2003; Peter and Grimm, 2009). In contrast, overproduction of GUN4 leads to stimulation of ALA synthesis reaction and MgCh activity in tobacco (Peter and Grimm, 2009).

Posttranslational protein phosphorylation is one possibility to regulate activity, subcellular localization, and stability of proteins involved in cellular processes (Adams, 2001). Light-dependent phosphorylation modulates the activity of proteins involved in photosynthesis (Bellafiore et al., 2005; Bonardi et al., 2005; Pesaresi et al., 2009; Rochaix et al., 2012). Analysis of 27 phospho-proteomic data sets of Arabidopsis (Arabidopsis thaliana) protein extracts identified over 60,000 phosphorylation sites in 8,141 proteins (van Wijk et al., 2014) with over 300 identified to be plastid localized (Schönberg et al., 2014). Important chloroplast processes, including metabolism, photosynthesis, gene expression, and signaling, are often controlled by protein phosphorylation in acclimation to environmental changes (Bellafiore et al., 2005; Bonardi et al., 2005; Salinas et al., 2006; Baginsky and Gruissem, 2009). Among these targets, several TBS proteins were also identified in phosphoproteomics approaches (Baginsky and Gruissem, 2009; Brzezowski et al., 2015). Interestingly, GUN4 was also identified as a target for protein kinase activity in a phosphoproteomics study (Reiland et al., 2009).

Here, we provide confirmation for the phosphorylation site of GUN4 and elucidate the regulatory role of GUN4 phosphorylation for MgCh activation and the entire TBS pathway. We have explored the role of nonphosphorylatable and phosphomimicking variants of GUN4 by expressing them in gun4-2 knockout mutants. Continuous simulation of GUN4 phosphorylation results in reduced MgCh activity and consequently in lower Chl content in gun4-2 lines expressing phosphomimicking GUN4(S264D). Bioinformatics analysis uncovered that the extension of the C-terminal segment of GUN4 emerged during the evolution of oxygenic photosynthetic organisms. In parallel, a GUN4 phosphorylation site was developed that modulates MgCh activity in angiosperms but not in cyanobacteria or green algae. Thus, our studies reveal phosphorylation of GUN4 as a new control mechanism that reduces MgCh activity at the branch point of the TBS pathway when the rate of Chl synthesis is diminished.

RESULTS

GUN4 Is Rapidly Phosphorylated at S264 in the Dark

To monitor the dynamic phosphorylation of GUN4, we developed an in vitro kinase assay using Arabidopsis (At) leaf extracts and recombinant GUN4 proteins. P³² from radioactive labeled ATP was incorporated into recombinant GUN4 by plant extracts when incubated in the light, but the signal was strongly enhanced when the kinase assay was performed in the dark (Fig. 1A). Furthermore, immunological analysis of both light and dark assays with an antibody specific for the phosphorylated C-terminal segment of GUN4 (pGUN4) resulted in signals that reflected the overall levels of radioactivity detected (Fig. 1A). Previous mass spectrometric analysis had shown that GUN4 is phosphorylated at Ser 264 (S264), the penultimate residue in the protein (Reiland et al., 2009). No phosphorylation of the substitution mutants AtGUN4(S264I) and AtGUN4(S264D) by plant extracts was detectable, neither with the radioactivity assay nor the pGUN4 antibody. Hence, S264 is the only phosphorylation site in AtGUN4 (Fig. 1B). The GUN4 phosphorylation was induced within minutes of dark incubation (Fig. 1C), indicating a rapid posttranslational GUN4 modification in darkness.

Figure 1.

In vitro analysis of GUN4 phosphorylation. A, Kinase assay with recombinant GUN4(WT) and Col-0 plant extracts incubated in light (L) or dark (D). B, Kinase assay with recombinant GUN4(WT), GUN4(S264I), or GUN4(S264D) incubated in dark. For A and B, the soluble fraction, obtained by centrifugation subsequent to the incubation of the kinase assay, was analyzed. C, Kinetic analysis of GUN4 phosphorylation in the dark (minutes of dark incubation). Total fractions of the kinase assay were analyzed. D, Kinase assay with recombinant GUN4 and fractions of purified Arabidopsis chloroplasts. Kinase assays were carried out with intact (total), stroma, or thylakoid preparations of isolated Arabidopsis chloroplasts (top panel). Purity of fractions was determined by western blotting and incubation with antibodies raised against compartment-specific proteins (bottom, GSAT: Glu-1-semialdehyde 2,1-aminotransferase (soluble); LHCB1: light-harvesting Chl binding protein b1 (thylakoid bound); fullGUN4: amount of endogenous GUN4). The presence of the recombinant substrate was proven by Coomassie stain of a protein gel or a western blot using an antibody raised against the full-length Arabidopsis (At)GUN4 (fullGUN4). Phosphorylation was visualized using an autoradiograph (P³²) or an antibody raised against the phosphorylated C terminus of AtGUN4 (pGUN4).

The GUN4 kinase (GUK) was predominantly localized in thylakoid membranes of Col-0 chloroplasts, which correlates with the GUN4 localization (Fig. 1D). The CHLI subunit possesses the ATPase activity required for Mg chelation, and CHLD harbors conserved amino acids (aa) essential for ATP hydrolysis (Hansson and Kannangara, 1997; Jensen et al., 1998, 1999; Lundqvist et al., 2010). But a phosphate group transfer to GUN4 during an in vitro MgCh reaction could be excluded (Supplemental Fig. S1).

Analysis of the Redox Dependency of GUN4 Phosphorylation and Subplastidic Distribution of Phosphorylated GUN4

A difference in the redox poise between light- and dark-exposed chloroplast extracts was expected (Buchanan and Balmer, 2005; Rochaix et al., 2012; Dietz and Hell, 2015). Redox control of the activities of the dark-induced GUK was therefore conceivable after the transition from light to dark. Hence, GUN4 phosphorylation was assayed in the presence of dithiothreitol (DTT) and/or hydrogen peroxide (H₂O₂), both in light and darkness (Fig. 2, A and B). Although the addition of DTT seems to stimulate phosphorylation of GUN4 in both light- and dark-incubated extracts, the ratio between light- and dark-dependent GUK activity remained similar to that in control reactions without additives. In the presence of DTT, GUK activity was, like in the control reactions, reduced in light compared to dark treatment (Fig. 2, A and B). DTT treatment resulted in phosphorylation of only one immunoreactive GUN4 protein band within the soluble and membrane fractions of the kinase assay (band B). In contrast to the control reactions, only band B was detectable with both the fullGUN4 and pGUN4 antibody (Fig. 2, A and B; most pronounced difference for pGUN4 visible under dark conditions). This indicates that reducing conditions do seem to prevent further GUN4 modification by plant extracts (see next chapter).

Figure 2.

Occurrence of different GUN4 species after incubation of recombinant GUN4 with plant extracts and effect of additives on GUN4 kinase activity. A to C, In vitro kinase assay with recombinant GUN4(WT) and Col-0 plant extracts incubated in light or dark. Fractions of the assay mixtures were obtained by centrifugation of the assay mixtures subsequent to incubation. Pellet (A) and soluble (B) fractions of the kinase assay incubated with or without 1 mm H₂O₂ and 1 mm DTT. L, light; D, dark incubation. C, Analysis of the soluble (S) and pellet fraction (P) of a kinase assay performed in light (L) and dark (D). For the pellet fraction, a volume corresponding to 10 times that of the soluble fraction was analyzed. The presence of the recombinant substrate was proven by western blotting using an antibody raised against the full-length Arabidopsis (At)GUN4 (fullGUN4). Phosphorylation was visualized using an antibody raised against the phosphorylated C terminus of AtGUN4 (pGUN4). The GUN4 species with different mobility during SDS-PAGE are labeled (A–C).

Addition of H₂O₂ or inhibition of photosynthetic electron transport with 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU; Supplemental Fig. S2A) had no effect on the GUK activity in the assay mixtures in either light or dark (Fig. 2, A and B). In addition, the subplastidic distribution of recombinant GUN4 after the kinase assay was not influenced by the phosphorylation status. Independent from high and low phosphorylation rates of GUN4 in dark and light, respectively, the same amount of recombinant GUN4 was found to be bound to the membranes after the kinase assay (Fig. 2A).

The Three Immunoreactive GUN4 Variants Are Phosphorylated in the Kinase Assay

Analysis of total plant extracts with antibodies against the full-length GUN4 identifies three immunoreactive GUN4 variants of different sizes, which are distinct from the GUN4 precursor (Larkin et al., 2003; Peter and Grimm, 2009; Fig. 6). When we analyzed the soluble and pellet fraction of a GUN4 kinase assay, we observed that the recombinant GUN4 also migrates in three different variants after incubation with plant extracts (Fig. 2, A–C). We found a dominant immunoreactive GUN4 band after incubation of recombinant GUN4 with plant extracts in both the soluble and pellet fraction of a kinase assay (Fig. 2C, band B). An additional GUN4 variant of slightly higher mobility was found in the soluble fraction (band C, difference from band B approximately 1 kD), while the membrane fraction contained an additional, but slower migrating GUN4 variant (band A; Fig. 2, A–C). At least the occurrence of the faster migrating GUN4 band C in the soluble fraction was a result of a thus-far-unknown enzymatic activity and was found to be independent of the GUN4 phosphorylation (Supplemental Fig. S2B). In all conducted experiments, we always observed a strong phosphorylation of GUN4 represented by the variant B independent of its subplastidial origin. In resemblance to the GUN4 variants A and B, the alternative GUN4 form represented by protein band C was phosphorylated, although to a minor extent. We observed a varying degree of GUN4 modification leading to the C variant and, thus, the intensity of the band C detected with the pGUN4 also varies among different experiments (Figs. 1 and 2; Supplemental Fig. S2B). Because the pGUN4 antibody is highly specific and did not detect nonphosphorylated GUN4 species, that is GUN4(S264I) or GUN4(WT) without incubation with plant extracts (Fig. 1; Supplemental Fig. S2B), the observed phosphorylation signal for GUN4 band C was not the result of unspecific binding of the peptide antibody to the nonphosphorylated GUN4 C terminus. Interestingly, supplement of DTT to the plastid extracts prevented the separation of GUN4 detected in the membrane and soluble fraction of the kinase assay, and thus only the B variant can be detected after DTT treatment (Fig. 2, A and B). The obvious posttranslational modifications leading to a changed mobility and the occurrence of three immunoreactive GUN4 bands (Larkin et al., 2003; Peter and Grimm, 2009) reflects an unknown physiological role for each of the GUN4 variants. Because the established kinase assay with plastid extracts and recombinant GUN4 induces the additional GUN4 modification(s), we provide an additional tool to study the GUN4 modifications in the future.

Figure 6.

Gene expression profiling and western-blot analysis of gun4-2 mutants expressing the GUN4(S264D). A qPCR analysis of genes encoding enzymes involved in TBS. Values are represented as 2^−ΔΔC(t) relative to Col-0. qPCR results were visualized as a heat map using a tool provided by http://bar.utoronto.ca. n = 4 biological replicates for each genotype. B, Western-blot analysis of TBS enzymes. Analyses were performed with 14- to 21-d-old Col-0 plants and complementation lines expressing GUN4-WT, GUN4-I, or GUN4-D in the gun4-2 background. GluTR/HEMA1/HEMA2, GLUTAMYL-tRNA REDUCTASE 1 and 2; GSAT/GSA1/GSA2: GLU-1-SEMIALDEHYDE 2,1-AMINOTRANSFERASE 1 and 2; CHLH: MgCh subunit H; CHLD: MgCh subunit D; CHLI/CHLI1/CHLI2: MgCh subunit I-1 and I-2; CHLM: Mg-PROTOPORPHYRIN IX METHYLESTER TRANSFERASE; CHL27/CHL27 aerobic MgPMME cyclase subunit; CHLG: CHLOROPHYLL SYNTHASE; FC1/FC2: FERROCHELATASE 1 and 2; LHCB1/ LHCB1.2: LIGHT HARVESTING CHLOROPHYLL BINDING PROTEIN b1. The tree immunoreactive GUN4 bands, designated as A, B, and C, are labeled.

Analysis of GUN4 Phosphorylation Capacity in Known Protein Kinase and gun Mutants

To prove whether GUN4 is phosphorylated by known plastid-localized protein kinases, we made use of the available knockout mutants of well-characterized plastid protein kinases. Extracts of stn7/8 (Wunder et al., 2013a, 2013b), CHLOROPLAST SENSOR KINASE (csk-1; Puthiyaveetil et al., 2008), and THYLAKOID ASSOCIATED KINASE 1 (tak1-1; Snyders and Kohorn, 1999; Snyders and Kohorn, 2001) mutants display wild-type-like phosphorylation of GUN4 (Fig. 3, A and B), indicating that GUN4 is unlikely to be a substrate for these kinases.

Figure 3.

In vitro analysis of GUN4 phosphorylation in different genotypes. A, Kinase assay with recombinant GUN4(WT) and extracts of Arabidopsis mutants for tak1-1 and csk-1. B, GUN4 kinase assay with extracts from Arabidopsis knockout mutants and overexpressor line of stn7/stn8 (single and double mutants; oeSTN7/8, overexpressing lines). C, GUN4 kinase activity in Col-0 and Ler-0 wild type, gun4-1, gun4-3, and gun5-1. The kinase assay was performed using a one-leaf-kinase-assay protocol (see Methods). Relative signal intensities for the pGUN4 antibody were determined by densitometry. L, light; D, dark incubation. A to C, Only the soluble fraction, obtained by centrifugation after incubation of the assay mixture, was analyzed. The presence of the recombinant protein substrate was proven by Coomassie stain of SDS gels or Ponceau stain of western-blot membranes or a western blot using an antibody raised against the full-length Arabidopsis (At)GUN4 (fullGUN4). Phosphorylation was visualized using an antibody raised against the phosphorylated C terminus of AtGUN4 (pGUN4).

Surprisingly, GUK activity was increased by up to 2-fold in gun5 (chlh mutant) and in two independent gun4 knockdown mutants relative to wild-type controls (Fig. 3C). All three mutant lines are characterized by reduced steady-state levels of Mg porphyrins and accumulation of Chl (Supplemental Fig. S3A). To examine a putative correlation between Chl accumulation and GUK activity, we also analyzed pale-green mutants of the TBS pathway, such as the hy2 (allelic to gun3) and chl27 mutants, which encode the phytochromobiline synthase and the catalytic protein of the Mg protoporphyrin IX monomethylester (MgPMME) oxidative cyclase, respectively. Additionally, we tested GUK activity in extracts of mutants involved in the chloroplast signal recognition particle-dependent insertion pathway of light-harvesting Chl-binding (LHC) proteins and for reductants involved in TBS, here an ntrc mutant encoding the NADPH-dependent thioredoxin reductase (Supplemental Fig. S3B). All mutant extracts showed a wild-type-like phosphorylation of recombinant GUN4 in light and dark. Hence, a reduced Chl content was not per se responsible for up-regulated GUK activity in gun4 and gun5 mutants.

Impact of Porphyrin Intermediates on GUN4 Phosphorylation

Because the previously reported GUN4 functions are closely related to its binding capacity of Proto and MgP, we next tested the impact of tetrapyrrole intermediates on GUK activity in a kinase assay with wild-type plastid extracts supplemented with Proto or MgP. GUN4 phosphorylation was reduced in assay mixtures supplemented with MgP but not with Proto (Fig. 4A). Attenuation of GUK activity in the presence of elevated levels of tetrapyrrole intermediates was also observed in kinase assays using extracts of Arabidopsis wild-type plants, which were fed with ALA. These extracts contained increased steady-state levels of porphyrins and showed reduced GUK activity compared to the reactions with extracts from control plants (Fig. 4B; Supplemental Fig. S3C). While strongly repressed in light (reduced by approximately 70% compared to the control reaction) and strongly reduced in dark in comparison to the control reaction (2-fold), it is remarkable that the GUK activity was still inducible in darkness when extracts of ALA-fed plants were used (plus approximately 30% compared to the light sample of non-ALA-fed plant protein extracts; Fig. 4B). This is indicative of an additional potential factor regulating GUK activity in the dark, independent from porphyrin intermediates. Nevertheless, these findings might reveal that posttranslational control of GUK activity probably implements at least sensing of steady-state levels of Mg porphyrins.

Figure 4.

Porphyrin contents correlate inversely with GUN4 kinase activity. A, GUN4 kinase assay with Col-0 plant extracts in the presence of Proto and MgP (10 µM) performed in light. w/o, control reaction. B, GUN4 kinase activity in Col-0 plants with elevated endogenous porphyrin and Mg porphyrin levels. Plants were incubated with (+) or without (–) 1 mm ALA for 14 h in dark prior to the preparation of extracts. The kinase assay was performed using the one-leaf-kinase assay protocol in light (L) or dark (D). For A and B, soluble fractions of the kinase assays were analyzed. Relative signal intensities for the pGUN4 antibody were determined by densitometry. The presence of the recombinant substrate was proven by Ponceau stain of western-blot membranes or using an antibody raised against the full-length Arabidopsis (At)GUN4 (fullGUN4). Phosphorylation was visualized using an antibody raised against the phosphorylated C terminus of AtGUN4 (pGUN4).

Loss of Its Phosphorylation Site Does Not Alter GUN4 Function

We assumed that an exchange of the phosphorylated aa of GUN4 (S264I/; Fig. 1B) would significantly alter the function of AtGUN4, as observed for other phosphorylated proteins (Baginsky and Gruissem, 2009; Willig et al., 2011; Wang et al., 2015). A successful complementation of the white gun4-2 knockout mutant (Supplemental Fig. S3A) was achieved with pGUN4::GUN4(S264I) (GUN4-I lines; Supplemental Fig. S4). The expression of GUN4-I in gun4-2 led to lines phenotypically indistinguishable from gun4-2/GUN4-WT lines complemented with pGUN4::GUN4(WT) (Supplemental Fig. S4). The transformants expressed GUN4 at least to wild-type levels with a clear correlation between GUN4 mRNA and protein in all transformants analyzed (Supplemental Fig. S5, A and B), indicating that phosphorylation does not affect GUN4 stability and its transcript levels in planta. Expression of either GUN4 variant (GUN4-WT and GUN4-I) in gun4-2 did not modify the levels of ALA synthesis (Supplemental Fig. S4D), Mg porphyrin intermediates (Supplemental Fig. S4, E–G), Chl (Supplemental Fig. S4, H and J), and heme (Supplemental Fig. S4J) in comparison to wild type. All parameters analyzed were restored to wild-type level when compared to the white gun4-2 null mutants, which were extensively analyzed previously (Supplemental Fig. S3A; Larkin et al., 2003; Peter and Grimm, 2009). Under normal growth conditions, RNA and protein content of key enzymatic steps of TBS were unchanged in GUN4-WT and GUN4-I lines (Supplemental Fig. S5). It can be concluded that the metabolite flow was not substantially modified and, therefore, the GUN4 function in TBS remains unaffected when it cannot be phosphorylated in GUN4-I lines.

Expression of Phosphomimic GUN4(S264D) in gun4-2 Only Partially Complements the Knockout Phenotype

Homozygous gun4-2 lines expressing GUN4-D (pGUN4::GUN4(S264D)), which mimics the constitutively phosphorylated species (Léger et al., 1997; Willig et al., 2011), showed retarded growth and a pale-green leaf phenotype (Fig. 5, A–C, H). Although GUN4-D transcript levels are comparable to those of GUN4 in Col-0, reduced ALA synthesis (Fig. 5D), reduced flux of intermediates through the Mg branch of TBS (Fig. 5, E–G), and reduced Chl contents (60%–70% of the wild-type and control lines) were observed (Fig. 5H). The white phenotype of the gun4-2 null mutant (Supplemental Fig. S3A) was only partially restored, although the mutant expressed similar GUN4-D transcript and protein content. Like gun4 and gun5 mutants (Supplemental Fig. S3A), gun4-2/GUN4-D lines exhibit a chlorotic phenotype in consequence of reduced flow of intermediates and a reduction of total Chl content (Fig. 5H). Notably, the pool of noncovalently bound heme was unaffected in the lines expressing GUN4-WT, GUN4-I, or GUN4-D (Fig. 5J).

Figure 5.

Characterization of gun4-2 mutants expressing the phosphomimetic GUN4(S264D). A, Phenotypes of 6- to 7-week-old Col-0 plants and three independent gun4-2/GUN4(S264D) lines (referred to as GUN4-D) grown under short-day conditions. B, PCR-based verification of the mutant background of lines shown in A using primers for the wild-type allele and the SALK_026911 T-DNA insertion (T-DNA). C, Relative expression levels of GUN4 quantified by qPCR. Values are represented as 2^−ΔΔC(t) relative to Col-0 from four independent biological replicates. Data are given as mean ± sem. D, Rates of ALA synthesis of 14-d-old seedlings grown under SD conditions. E to G, Steady-state levels of Mg porphyrins of the pathway. PChlide, protochlorophyllide. H to J, Chl level, Chl a/b ratio, and heme level. For C to J, 14- to 21-d-old Col-0 plants and different complementation lines expressing GUN4-WT, GUN4-I, or GUN4-D in the gun4-2 background were analyzed. Data for intermediates and end products of the TBS pathway are given as mean ± sd, n = 4 biological replicates for each genotype.

To rule out that the chlorotic phenotype of the gun4-2/GUN4-D lines was not a consequence of altered expression or stability of enzymes involved in TBS, we analyzed the expression of several genes encoding proteins of TBS. In particular, enzymes that are rate-limiting for ALA-formation—such as GLUTAMYL-tRNA REDUCTASE (GluTR) or GLU-1-SEMIALDEHYDE 2,1-AMINOTRANSFERASE (GSAT)—and enzymes essential for the synthesis of Mg porphyrins (like the MgCh subunits CHLH, CHLD, and CHLI; the MgP methyltransferase CHLM; and the aerobic cyclase subunit CHL27) accumulate to similar levels in lines expressing GUN4-D (Fig. 6, A and B). The sole significant difference between control lines and lines expressing the phosphomimetic GUN4-D concerns the LHCB1 proteins (Fig. 6B), whose levels in the photosynthetic antenna correlate with the lower amounts of Chl and the higher Chl a/b ratio in these lines (Fig. 5I). As revealed by unchanged LHCB1.2 transcript level in gun4-2/GUN4-D (Fig. 6A), the reduction of LHCB1 is not a result of reduced expression of the mRNA. Similarly, differences in the expression of genes for TBS proteins were not observed. In conclusion, the alterations in the metabolism of TBS found in gun4-2/GUN4-D are explained neither by modified transcriptional control (Fig. 3A) nor by reduced posttranslational stability of the enzymes in the pathway (Fig. 3B).

Comparison of gun4-2/GUN4-D Lines with Known gun4 Mutants Reveals an Impact on MgCh Activity

A lower MgCh activity in two allelic knockdown mutants, gun4-1 and gun4-3, has been explained by the diminished expression of GUN4 and posttranslational destabilization of GUN4 (Larkin et al., 2003; Davison et al., 2005; Peter and Grimm, 2009). GUN4-D lines also exhibited a chlorotic phenotype, although the amount of GUN4 present was not reduced relative to wild-type plants (Fig. 5C and 6B; Supplemental Fig. S6C). A comparative analysis of wild-type, gun4-1, and gun4-3 as well as gun4-2 lines expressing GUN4-WT, GUN4-I, and GUN4-D under different light intensities (Supplemental Figs. S6 and S7) uncovered a modulated function of the GUN4-D variant when compared with the wild-type controls. The reduction in green pigmentation of gun4-1, whose GUN4 content was drastically reduced, was more pronounced than that seen in gun4-3 (Supplemental Fig. S6, A–C). Interestingly, the pale green gun4-2/GUN4-D lines phenotypically resembled gun4-3 more than gun4-1 under the different light intensities (Supplemental Fig. S6, A and B) even though GUN4-D levels were at least wild-type-like (Supplemental Fig. S6, C and D). Attenuated GUN4 function also correlated with the altered levels of TBS intermediates (Supplemental Fig. S7, A–C) and end-products in gun4-1, gun4-3, and gun4-2/GUN4-D under various growth conditions (Supplemental Fig. S7, D–H). Expression of GUN4-D in gun4-2 led, on average, to 85%, 70%, and 50% wild-type Chl content in low, normal, and high light, respectively (Supplemental Fig. S7, D–G). In contrast to the reduced Chl content, the heme content was not altered in all variants, which is again consistent with GUN4’s primary function in Chl rather than heme synthesis (Supplemental Fig. S7H). The observed differences in the chlorotic gun4 lines analyzed here did not result from differential expression of the MgCh subunits or any of the analyzed TBS enzymes under varying light intensities (Supplemental Fig. S6D).

Phosphorylation Alters GUN4-Dependent Stimulation of MgCh in Vitro and in Vivo

Given the phenotypic resemblance between the GUN4-D lines and gun4-1 and gun4-3, we quantified the effect of phosphorylated GUN4 on MgCh activity with purified recombinant subunits from rice (Supplemental Fig. S8A) using a continuous fluorometric assay. In the absence of recombinant GUN4, in vitro Mg chelation was low compared to assays supplemented with GUN4. Addition of either OsGUN4 or AtGUN4 led to a 3- to 4-fold stimulation of rice MgCh activity (Supplemental Fig. S8B). While recombinant AtGUN4(WT) and GUN4(S264I) stimulated MgCh in vitro to the same extent, GUN4(S264D)-dependent stimulation reached only 50% of control MgCh activity (Fig. 7A). The ability of GUN4(S264D) to bind Proto and MgP in vitro was not altered by the phosphomimetic substitution at the phosphorylation site when compared to the GUN4(WT) protein (Supplemental Fig. S9, A and B). Hence, the reduced stimulatory effect of phosphorylated GUN4 on MgCh activity was not caused by altered affinity for the substrate and product of the MgCh reaction. MgCh activities analyzed in plant extracts from wild-type and gun4-2 lines expressing GUN4-WT, GUN4-I, and GUN4-D further confirmed an altered stimulation of MgCh by GUN4(S264D). Three independent gun4-2/GUN4-D lines showed a 50% to 60% reduction in in vivo MgCh activity compared to the control lines analyzed (Fig. 7B).

Figure 7.

Influence of phosphorylated GUN4 on in vitro and in vivo MgCh activity. A, In vitro MgCh assay using recombinant MgCh subunits from rice together with GUN4(WT), GUN4(S264I), and GUN4(S264D) from Arabidopsis. Data are given as mean ± sd. B, In vivo MgCh activity assayed in extracts of Col-0 and complementation lines expressing GUN4-WT, GUN4-I, or GUN4-D in the gun4-2 background.

In summary, the introduction of a negatively charged aa next to the C terminus of GUN4, which mimics the phosphorylation of the wild-type Ser residue, reduces the stimulating effect of GUN4 on Mg chelation without modifying its ability to bind porphyrins.

The C-Terminal Segment of GUN4 Was Extended during Evolution and the p-Site of GUN4 Is Conserved in Angiosperms

The data presented so far revealed a physiological role of the phosphorylated GUN4 C-terminus for the stimulation of the MgCh reaction. A multiple sequence alignment of GUN4 homologs indicated that the C-terminal segment was extended during the evolution of oxygen-dependent photosynthetic organisms (Supplemental Fig. S10). Only angiosperms are characterized by a conserved C-terminal region with a phosphorylation site as identified in AtGUN4. Interestingly, apart from the carnivorous plant Genlisea aurea whose UniProt-KB entry is incomplete, only angiosperm GUN4 homologs have retained a phosphorylatable aa residue (Ser or Thr) among the last five aa (Fig. 8B; Supplemental Fig. S10). Additionally, the phosphorylation site was found in the highly conserved motif K-P/T-D/N-Y-S-F (Supplemental Fig. S11).

Figure 8.

Significance of the extended C-terminal segment of GUN4 and conservation of the phosphorylation site in angiosperms. A, Superimposition of the experimentally obtained crystal structures of cyanobacterial GUN4 (Synechocystis sp. PCC 6803 / 1Y6I/ green and T. elongatus / 1Z3X/ blue) on a model of Arabidopsis GUN4 (Phyre2/ orange). The highly conserved LRG motif and the phosphorylation site (p-site S264) found in angiosperm GUN4 sequences are indicated (see also Supplemental Fig. S10). B and C, Extension of the CTL of GUN4 during evolution of oxygenic photosynthetic organisms. The CTL was determined using an algorithm that counts the number of aa within the C terminus after the highly conserved GUN4 core domain (see “Materials and Methods”). B, Phosphorylatable aa (Ser or Thr) located within the five C-terminal residues are indicated by purple circles. C, Statistically significant differences in the CTL between the different groups of organisms are indicated by an asterisk (Student’s t test, P < 0.02). The numbers of sequences available for each group are indicated at the bottom of each bar.

Superimposition of two x-ray structures of GUN4 available from cyanobacteria (Synechocystis sp. PCC 6803 and Thermosynechococcus elongates; Davison et al., 2005; Verdecia et al., 2005) on a predicted model of AtGUN4 indicated that the C-terminal segment of AtGUN4 forms a distinct protrusion, approximately 42 aa long, that projects from a common GUN4 core fold (Fig. 8A). The compact structure of the core domain terminates at helix 8 (Supplemental Fig. S10; Fig. 8A), which begins next to a Leu-Arg-Gly (LRG) motif that is essential for porphyrin binding (Davison et al., 2005; Chen et al., 2015b). Using a bioinformatics approach, we determined the length of the C-terminal segment by counting the aa residues from the highly conserved LRG and a His-Pro motif at the end of helix 8 (Supplemental Fig. S10; Fig. 8, B and C). The most ancient aerobic photosynthetic organisms (cyanobacteria) express a GUN4 homolog with a mean C-terminal length (CTL) of 5 ± 1.5 aa. The CTL significantly increases to 9 ± 5 aa within the group of diatoms and brown algae and was further extended to approximately 32 ± 15 aa within the clade of green algae. In angiosperms, the C-terminal segment of GUN4 is 44 ± 7 aa long (Fig. 8C). To examine whether the CTL of GUN4 has a direct influence on its stimulatory function for MgCh, we assayed recombinant AtGUN4 variants C terminally truncated by 3, 6, 9, and 12 aa and found a progressive loss of their stimulatory impact on MgCh activity (Fig. 9, D and E).

Figure 9.

In vitro analysis of truncated GUN4. A, Expression and purification of recombinant wild-type and truncated Arabidopsis GUN4. The numbers of aa removed from the C terminus are indicated. Purified proteins were analyzed with a nonphosphospecific GUN4 antibody raised against the nonphosphorylated C-terminal segment of GUN4 (np-pGUN4) and with the fullGUN4 antibody. B, In vitro MgCh assay using recombinant MgCh subunits from rice and GUN4(WT) and GUN4 variants from Arabidopsis truncated by 3, 6, 9, and 12 aa. Data are given as ± sd.

Because an extended CTL and the aa composition of the GUN4 C-terminal region are conserved only in angiosperms, we hypothesized that the C-terminal region of GUN4 has acquired a specific function during evolution of organisms with oxygenic photosynthesis. A comparative analysis of AtGUN4 and GUN4 homologs from cyanobacteria (Synechocystis PCC 6803 [Syn]) and green algae (Chlamydomonas reinhardtii [Cr]) showed that neither recombinant SynGUN4 nor CrGUN4 stimulated rice MgCh (Fig. 10A). The porphyrin-binding properties of SynGUN4 and CrGUN4 were similar to those of AtGUN4 (Supplemental Fig. S11). Kinase assays with cell extracts from Syn and Cr and crude extracts from At seedlings, in combination with their respective endogenous GUN4 homolog, confirmed that, at least under our assay conditions, only the recombinant AtGUN4 was strongly phosphorylated (Fig. 10B).

Figure 10.

In vitro analysis of Synechocystis, Chlamydomonas, and Arabidopsis GUN4 homologs. A, In vitro MgCh assay using recombinant MgCh subunits from rice in combination with Syn, Cr, and At GUN4 homologs. Inset: Purified recombinant GUN4 homologs (1 µg) used in the MgCh assay analyzed by SDS-PAGE. Data are given as ± sd. B, Kinase assay using recombinant GUN4 homologs from Syn, Cr, and At and native extracts of Syn and Cr cultures and At plants (GUN4+extract). Top: Radioactive signal (P³²) after the kinase assay. Bottom: Coomassie-stained gels used to control for equal amounts of substrate. The positions of the three different recombinant proteins on the SDS-PA gels are indicated on the right. Background phosphorylation within the extracts (first three lanes) and potential autophosphorylation of the recombinant substrates (three lanes in the middle) are shown. The three lanes on the right side show the phosphorylation signal of the protein extracts from the organisms combined with the recombinant GUN4 substrate.

DISCUSSION

GUN4 stimulates the activity of the MgCh complex, affects the enzymes of ALA synthesis by an unknown feedback mechanism and is involved in plastid-derived retrograde signaling (Mochizuki et al., 2001; Larkin et al., 2003; Peter and Grimm, 2009). Because of these diverse functions, it is comprehensible that GUN4 action is posttranslationally controlled during development and growth of seedlings. Previously published studies convincingly demonstrated the need of GUN4 for proper MgCh function in various photosynthetic organisms (Mochizuki et al., 2001; Larkin et al., 2003; Wilde et al., 2004; Sobotka et al., 2008; Peter and Grimm, 2009; Formighieri et al., 2012; Brzezowski et al., 2014). Nevertheless, the current knowledge cannot explain how the stimulatory GUN4 is posttranslationally controlled to enable rapid adjustment of MgCh activity to the daily changing needs of Chl biosynthesis or an attenuation of its activity during dark periods. Our recent experiments demonstrated that phosphorylation modulates GUN4 function on MgCh activity with a successive impact on the activity of ALA synthesis.

GUN4 Is Phosphorylated by a Plastid Membrane Localized Kinase

The newly established in vitro kinase assay revealed that recombinant GUN4(WT) is rapidly phosphorylated by Arabidopsis Col-0 extracts. Two GUN4 substitution mutants, GUN4(S264I) and GUN4(S264D), were not phosphorylated by the same extracts (Fig. 1B), which confirms the previously identified (by mass spectrometry) phosphorylation site of AtGUN4 (Reiland et al., 2009). Additionally, we found that the GUK is predominantly localized at the membrane fraction of chloroplasts, where GUN4 (Fig. 1D), the MgCh (Fuesler et al., 1984; Gibson et al., 1996; Larkin et al., 2003; Mochizuki et al., 2010), and other enzymes of the Chl branch are mainly associated or inserted (Tanaka and Tanaka, 2007; Mochizuki et al., 2010). This correlation argues for the requirement of proximity of the GUK and its protein substrate enabling rapid control of GUN4 by phosphorylation. Although ATP hydrolysis takes place during MgCh catalysis (Jensen et al., 1996, 1998, 1999; Adams and Reid, 2013), a phosphate transfer to GUN4 by the MgCh subunits H, D, and I was not observed in vitro (Supplemental Fig. S1). The observed deregulation of GUK activity in gun4 and gun5 mutants (Fig. 3C) points to a functional connection of GUK, GUN4, and MgCh under physiological conditions. As other chlorotic mutants of deficient TBS and impaired LHC integration into the thylakoid membrane (chloroplast signal recognition particle mutants) did not show altered GUN4 phosphorylation (Supplemental Fig. S3B), the enhanced GUK activity in gun4 and gun5 mutants was not a consequence of reduced Chl content. It is rather likely that either MgCh activity itself determines GUK activity (see below) or that expression of GUK is altered in the MgCh mutants.

An initial attempt to search for the GUK revealed that none of the analyzed known kinases, like STN7/8, CSK, or TAK1, are responsible for GUN4 phosphorylation. Further biochemical studies on unknown plastid-localized protein kinases are needed to identify the GUK. As soon as the identity of the GUK is unraveled, the transcriptional and posttranscriptional regulation of GUK and its physiological significance on the whole TBS pathway can be analyzed.

The search for the GUK will be facilitated by the use of the antibody against the phosphorylated C-terminal peptide, which was not yet sensitive enough to detect phosphorylated GUN4 in plastid extracts. Future studies will be performed to uncover and quantify phosphorylated GUN4 in planta under light and dark conditions and with extracts from putative protein kinase mutants.

Factors Influencing GUN4 Phosphorylation

Kinase assays performed in light and dark suggested that GUN4 phosphorylation is controlled by a factor that represses GUK activity in light but induces GUN4 phosphorylation in dark (Figs. 1–4). One of the most obvious parameters differing in plastid extracts incubated in light and dark is the redox status. Because the well-known STN7 kinase was shown to be regulated by a redox shift within the photosynthetic machinery (Bellafiore et al., 2005; Bonardi et al., 2005; Rochaix et al., 2012) and enzymatic activities of TBS are regulated in a redox-dependent manner (Jensen et al., 2000; Ikegami et al., 2007; Luo et al., 2012; Richter and Grimm, 2013; Pérez-Ruiz et al., 2014), it was assumed that the GUK and thus GUN4 phosphorylation c­ould be regulated in a similar way. However, application of reductants like DTT or oxidants like hydrogen peroxide or DCMU, which inhibits photosynthetic electron transfer leading to oxidized photosynthesis-associated redox-regulators, does not seem to influence GUK activity in light or dark.

A second explanation for the induced or repressed GUK activity in light- and dark-exposed plant extracts is the different activity of the TBS pathway. In light, the pathway has its optimal activation state with high MgP and MgPMME steady-state levels (Supplemental Fig. S3C). In dark, the flow of intermediates is rapidly attenuated by the FLU-protochlorophyllide-dependent repression of GluTR activity (Meskauskiene et al., 2001; Meskauskiene and Apel, 2002; Lee et al., 2003; Richter et al., 2010; Kauss et al., 2012). By inhibition of ALA synthesis, the rate-limiting step of TBS, the steady-state levels of MgP and MgPMME, were below the detection limit and protochlorophyllide is the only porphyrin, which accumulates in dark due to the light dependency of protochlorophyllide oxidoreductase (POR).

Interestingly, addition of MgP, but not Proto, to a kinase assay repressed the GUK activity (Fig. 4A). Additionally, GUK activity was also repressed when a kinase assay with extracts of ALA-fed Col-0 plants was performed (Fig. 4B). ALA feeding leads to excessive steady-state levels of porphyrins (up to 10- to 20-fold increase) compared to nonfed plants grown in light (Supplemental Fig. S3C). The existence of an as-yet-unknown factor regulating GUK activity is revealed by the observation that GUK activity was still induced in darkness compared to the light incubation when extracts of ALA-fed plants were used (Fig. 4B). Thus, it is not excluded that an additional factor independently from Mg porphyrins contributes to the regulation of GUK activity. Nevertheless, the correlation of elevated Mg porphyrin levels and lower GUK activity prompted us to suggest that Mg porphyrins affect the GUN4 phosphorylation more in the presence than in the absence of light. However, an effect of MgP, the product of the MgCh reaction, on GUK activity is further supported by the finding that gun4 and gun5 mutants with reduced MgP level (Supplemental Fig. S3A) showed a 2- to 3-fold induction of GUK activity (Fig. 3C). Thus, it could be speculated that elevated MgP level repressed GUK activity and reduced MgP level-stimulated GUK activity in addition to other factors.

In conclusion, the results suggest that GUK activity is also modulated either directly or indirectly by porphyrins. However, regulation of the GUK by TBS intermediates needs further studies to explore the interdependence between the flow of TBS intermediates, unknown factors, and GUK activity. Future studies will also show to what extent unknown factor(s) (e.g. a phosphatase) implement additional control of the GUK activity.

Known Properties of GUN4 Are Not Influenced by Phosphorylation

Previous studies suggested that GUN4 and MgCh subunits are more efficiently recruited to the plastid membrane system when the flow of intermediates within the pathway is accelerated (Adhikari et al., 2009). This finding is in agreement with a spatio-temporal organization of the pathway (Czarnecki and Grimm, 2012), suggesting that association of GUN4 at the thylakoid membranes controls the activity of the MgCh complex. GUN4 found in the membrane and soluble fraction of a kinase assay was always more phosphorylated in the dark. The relative distribution of phosphorylated and nonphosphorylated GUN4 in membrane and soluble fractions of a kinase assay was similar (Fig. 2, A and B), and therefore we assume that the subplastidic distribution or the binding of GUN4 to plastid membranes is not determined by the phosphorylation status.

Protein phosphorylation is one mechanism to control stability, turnover, and activity of proteins in all cellular compartments (Adams, 2001; Willig et al., 2011; Trotta et al., 2016; Zhao et al., 2016). Our studies revealed that the stability of GUN4 is not influenced by its phosphorylation status. Under different growth conditions, the GUN4 content in gun4-2 complementation lines expressing either the nonphosphorylatable GUN4(S264I) or phosphomimicking GUN4(S264D) followed the amount of mRNA expressed (Figs. 5 and 6; Supplemental Figs. S4 and S5). Hence, it is unlikely that the chlorotic phenotype of the gun4-2/GUN4-D lines is explained by a temporarily and/or locally enzymatic degradation of GUN4, which might be induced by phosphorylation.

Furthermore, modified structural integrity of the GUN4 by the Ser to Asp exchange can be excluded. Firstly, the phosphorylation site lies within a C-terminal extension of AtGUN4, which has no structural connection to the common GUN4 core domain found in Syn and AtGUN4 (Fig. 8A). Secondly, and more important, the ability of GUN4 to bind Proto and MgP in vitro was not altered by the phosphomimetic substitution at the phosphorylation site (Supplemental Fig. S9, A and B), and apparent dissociation constants (K_d) for Proto and MgP of all GUN4 variants tested were similar to published K_d values of cyanobacterial and AtGUN4 homologs (Larkin et al., 2003; Davison et al., 2005; Verdecia et al., 2005; Adhikari et al., 2009). Furthermore, the binding of porphyrins takes place within a core domain, which is common to all GUN4 proteins and structurally independent from the C-terminal extension harboring the phosphorylation site (Verdecia et al., 2005; Chen et al., 2015b; Kopečná et al., 2015). Thus, it is unlikely that the substitution of Ser to Asp interferes with the structural integrity of GUN4 in vivo.

In summary, two of the main properties of GUN4 were not altered by the introduction of a negatively charged aa next to the C terminus: the Proto and MgP binding, which are discussed to have an important role in stimulating MgCh (Verdecia et al., 2005; Kopečná et al., 2015) and the subplastidic distribution of GUN4 in the stroma and at the thylakoid membrane (Adhikari et al., 2009).

Phosphorylation of GUN4 Alters the Stimulatory Impact on MgCh

As some key features of GUN4 are likely not influenced by phosphorylation, it was surprising that the gun4-2/GUN4-D lines showed a chlorotic phenotype that was comparable to the known gun4 knockdown mutants under various growth conditions (Supplemental Figs. S6 and S7; Larkin et al. 2003; Peter and Grimm, 2009). Although expressed to at least wild-type-like levels, GUN4(S264D) only partially complements the white seedling-lethal gun4-2 knockout mutants under standard photoperiodic growth conditions (Supplemental Fig. S3A; Larkin et al., 2003; Peter and Grimm, 2009). The wild-type-like expression and accumulation of TBS enzymes involved in the early and late steps of Chl formation indicated that a posttranslational rather than a transcriptional defect is responsible for the diminished Chl accumulation of gun4-2/GUN4-D. Enzymatic analysis unraveled strong reduction of MgCh activity in gun4-2/GUN4-D lines in comparison to control lines (Fig. 7B). Additionally, the recombinant GUN4(S264D) stimulates rice MgCh subunits only by 50% compared to control reactions (Fig. 7A). The observed reduction of MgCh activity in planta was accompanied by a strong reduction of ALA synthesis capacity in gun4-2/GUN4-D (Fig. 5D). Results of previous studies on gun4 and MgCh mutants in Arabidopsis and tobacco interpreted this down-regulation as protective mechanisms to prevent accumulation of phototoxic Proto when MgCh is blocked (Papenbrock et al., 2000a, 2000b; Peter and Grimm, 2009). Because the same regulatory connection was observed, a perturbation of MgCh reaction in gun4-2/GUN4-D lines is supported. It is worth mentioning that the gun4-2/GUN4-I lines did not show the opposite effect of MgCh activity stimulation. This might be explained by the fact that the optimal activation state of GUN4, and thus of MgCh, is achieved when GUN4 is not phosphorylated in light-grown (wild-type) plants (see below). A previous study on STN7 also revealed that the knockout of the phosphorylation sites does not alter STN7 properties and that nonphosphorylatable STN7 behaves like the wild-type protein (Willig et al., 2011). Hence, knockout of phosphorylation sites does not ultimately lead to adverse effects, especially not when the phosphorylation has a negative impact on a protein function.

GUN4 binds to the MgCh H subunit (Larkin et al., 2003) and stimulates MgCh activity (Larkin et al., 2003; Wilde et al., 2004; Davison et al., 2005; Verdecia et al., 2005; Sobotka et al., 2008; Luo et al., 2012) by a yet-unknown mechanism. Although crystal structures of GUN4 (Davison et al., 2005; Verdecia et al., 2005; Chen et al., 2015b; Tarahi Tabrizi et al., 2015) and CHLH (Chen et al., 2015a) were resolved, a model neither for simple protein-protein interaction nor for a mode of action of GUN4 on the activity of MgCh was presented. Based on our current knowledge, we suggest that the interaction of GUN4 with the H subunit is negatively affected by the phosphorylation of the GUN4 C-terminus. An additional negative charged phosphate group at the free and flexible C terminus of GUN4 (Fig. 8A) might influence binding of the H subunit as it was observed for other protein-protein interactions in photosynthetic organisms (Kim et al., 2004; Bellafiore et al., 2005). This view is even supported by the identification of a protein-protein interaction interface surrounding the C-terminal phosphorylation site (https://www.predictprotein.org/ Yachdav et al., 2014). When this interface is missing (Fig. 9; Zhou et al., 2012) then GUN4 less efficiently stimulates the MgCh reaction. However, other explanations take into account that either the recruiting of the other MgCh subunits D and I, the MgCh complex stability itself or the binding of Proto to the H subunit is altered when phosphorylated GUN4 is present.

In summary, the chlorotic phenotype of gun4-2/GUN4-D is explained by a perturbed MgCh reaction that leads to down-regulation of ALA synthesis and finally results in reduced flow of intermediates within the TBS pathway.

Evolution of the Phosphorylation Site of GUN4 Is a Prerequisite to Regulate MgCh Activity in Angiosperms

The bioinformatics analysis of GUN4 homologs from different photosynthetic active organisms unraveled the evolution of a C-terminal extension within the group of angiosperms. In addition, only species within the clade of angiosperms express a GUN4 with the highly conserved C-terminal motif harboring the phosphorylation site (Fig. 8; Supplemental Fig. S10). Despite the fact that the CTL of GUN4 from angiosperms determines its stimulatory impact on GUN4 (Fig. 9; Zhou et al., 2012), we found that GUN4 is phosphorylated only in angiosperms. Furthermore, the Syn and Cr GUN4 homologs lacking the C-terminal extension failed to stimulate rice MgCh in vitro (Fig. 10). At first view this was rather unexpected, because a stimulation of cyanobacterial or green algae MgCh by the individual GUN4s occurs in vitro and in vivo (Larkin et al., 2003; Wilde et al., 2004; Sobotka et al., 2008; Peter and Grimm, 2009; Formighieri et al., 2012; Brzezowski et al., 2014). On the other hand, this finding indicates that due to the independent evolution of, for example, cyanobacteria and angiosperms, species-specific determinants for the interaction of MgCh and GUN4 have evolved. Among others, one of those factors might be the significantly shorter C-terminus of GUN4 in organisms that are evolutionarily older than angiosperms (Fig. 8). To our knowledge, this cross-species complementation of MgCh subunits and GUN4 was not performed before and certainly needs further attention to understand the evolutionary differences in MgCh stimulation by GUN4. However, the evidence supports the idea that the C-terminal extension of GUN4 evolved as a prerequisite to regulate MgCh activity by phosphorylation of GUN4 in the light-dependent TBS pathway of angiosperms.

Conclusion

Why is regulation of GUN4’s action on MgCh by phosphorylation required in angiosperms? One of the main differences between angiosperms and evolutionarily older organisms is the existence of a dark operating POR (Reinbothe et al., 2010). Dark operating POR allows Chl synthesis in the dark, and its activity prevents the accumulation of photoreactive porphyrin intermediates in darkness (e.g. Kada et al., 2003). During dark incubation, angiosperms would accumulate harmful amounts of phototoxic porphyrins when suppression of ALA synthesis is not functioning (Meskauskiene et al., 2001; Vavilin and Vermaas, 2002; Kim et al., 2008; Reinbothe et al., 2010). Due to the existence of the FLU-dependent repression of ALA synthesis (Meskauskiene et al., 2001), angiosperms avoid the accumulation of excessive amounts of protochlorophyllide (Fig. 11). Nevertheless, a significant rate of ALA formation is observed, which is suggested to provide precursors for the synthesis of heme in darkness (Papenbrock et al., 1999; Richter et al., 2010; Czarnecki et al., 2011). Heme is an essential cofactor of several enzymes, including those that act in the mitochondrial electron transport chain even when light is limiting. Then it is questionable how the differential allocation of the tetrapyrrole precursor Proto into the heme and Chl branch of TBS is regulated in darkness and how the dominating MgCh activity is decelerated (Papenbrock et al., 1999).

Figure 11.

Model for the impact of phosphorylated GUN4 on the TBS pathway and regulation of the GUN4 kinase in organisms harboring light-dependent POR. The MgCh reaction is regulated by GUN4 phosphorylation in organisms that lack light-independent POR. In angiosperms, the stimulatory impact of GUN4 on MgCh is modulated by a kinase that responds to varying steady-state levels of Mg-porphyrins in light or dark. For further explanation, see the text.

We suggest a function for GUN4 phosphorylation at this metabolic and regulatory step: in the dark, stronger phosphorylation of GUN4 leads to a slowdown of MgCh activity. Thus, less Proto is consumed within the Chl branch in physiological situations where heme synthesis has to be favored and Chl synthesis is hampered (Fig. 11). Then, reduced MgP steady-state levels correlate with increased activity of GUK (see above), resulting in more phosphorylated GUN4 (Fig. 11). This mechanism helps to distribute porphyrins between the heme and the Chl branch, prevents the accumulation of porphyrins, and balances the flow of intermediates in response to changing activities of the TBS pathway.

MATERIALS AND METHODS

Growth Conditions and Genotypes Used

If not otherwise indicated, physiological experiments were performed with Arabidopsis (Arabidopsis thaliana; At) plants grown on soil for 14 to 21 d in short-day (SD) conditions under 120 to 130 µmol photons m⁻² s⁻¹. The different genotypes used in this study are summarized in Supplemental Table S1. Synechocystis sp. PCC 6803 was grown at 600 µmol photons m⁻² s⁻¹ (Lars Bähr, Humboldt-Universität zu Berlin, Institute of Biology). The cell wall-deficient Chlamydomonas wild-type strain CC-3395 (cwd, arg7-8, mt-; Chlamydomonas Center, University of Minnesota, St. Paul, MN) used in kinase assays was grown on Tris-acetate phosphate medium supplemented with 50 µg/mL Arg under 50 µmol photons m⁻² s⁻¹ at 23°C.

Porphyrin, Pigment, and Heme Analysis

Intermediates and end products of the TBS pathway were extracted from ground and lyophilized plant material by incubating with acetone 0.2 m NH₄OH (9:1, v/v) at −20°C for at least 1 h. The supernatant (13,000 rpm, 10 min, 4°C) was removed, heme was extracted from the pellet using acetone:HCl:dimethyl sulfoxide (10:0.5:2, v/v/v), and the HPLC analysis was performed on Agilent HPLC systems, essentially as described previously (Schlicke et al., 2014; Scharfenberg et al., 2015).

RNA, cDNA, qPCR

RNA was extracted from frozen plant material using TRIsure (Bioline), and 2-µg aliquots of DNase (Thermo Fisher)-treated RNA were transcribed into cDNA using Moloney murine leukemia virus reverse transcriptase (Thermo Fisher Scientific) and an oligo(dT)₁₈ primer following the instructions of the enzyme supplier. Prior to quantitative PCR (qPCR), cDNA was diluted 5-fold. qPCR was carried out in 10-µL reactions in a 96-well-plate qPCR machine (Biorad) using 2× qPCR mastermix (BioTool), 1 µL of cDNA, and a set of qPCR primers published elsewhere (Richter et al., 2013; Schlicke et al., 2014). PEX4 (AT5G25760), SAND (AT2G28390), and ACTIN-2 (AT3G18780) were routinely used as reference genes. The qPCR results reported were obtained from at least four independent biological replicates, each containing a pool of 14- to 21-d-old seedlings. Calculation of relative gene expression was done with the Biorad CFX-manager software (1.6) and is given as 2^−ΔΔC(t). Portions of the qPCR results are presented as heat maps, created using a tool provided by http://bar.utoronto.ca.

Cloning

Cloning of DNA fragments for heterologous expression in, and purification of recombinant protein from, Escherichia coli expression strains was performed using standard protocols (Sambrook, 2001). All cloning PCRs were performed with Phusion polymerase (NEB). For PCRs using Chlamydomonas 4A+ (137c / Krishna K. Niyogi, University of California, Berkeley, CA; Cr) cDNA as template, the PCR buffer was supplemented with 1% dimethyl sulfoxide (v/v). Vectors and cloning primers used in this study are listed in Supplemental Table S2. AtGUN4 (AT3G59400) and CrGUN4 (Cre05.g246800) fragments were amplified from cDNAs prepared by reverse transcription of wild-type RNA without the predicted transit peptide (chloroP v1.1) and were cloned into the pQE80L expression vector (Qiagen). Base exchanges leading to aa substitutions in pQE_AtGUN4(S264I) and pQE_AtGUN4(S264D) were introduced using mutagenized reverse primers and pQE_AtGUN4(WT) vector as template (Supplemental Table S2). The Synechocystis (Syn) GUN4 (/YCF53) fragment was provided by Prof. Annegret Wilde (Freiburg University) and was cloned into pQE80L. Expression constructs of rice (Oryza sativa) MgCh subunits (OsCHLH/D/I/GUN4) in the pet28a vector were provided from Zhou et al. (2012).

Genomic DNA (gDNA) fragments used for complementation of the gun4-2 knockout line were amplified from Col-0 gDNA preparations with primers annealing 800 bp upstream (5′) and 600 bp downstream (3′) of the GUN4 exon (AT3G59400; Supplemental Table S2). The GUN4(WT) fragment was cloned into pCambia3301 using SmaI and PmlI restriction sites. Base exchanges in pCambia_AtGUN4(S264I)_gDNA and pCambia_AtGUN4(S264D)_gDNA were introduced via a site-directed mutagenesis PCR (Laible and Boonrod, 2009) using complementary primers and pCambia_GUN4(WT)_gDNA vector as template (Supplemental Table S2). The products of all cloning steps were verified by sequencing.

Complementation of gun4-2

Heterozygous gun4-2 (SALK026911; Larkin et al., 2003; Peter and Grimm, 2009) mutants were transformed with constructs encoding GUN4(WT) (GUN4-WT), GUN4(S264I) (GUN4-I), GUN4(S264D) (GUN4-D) in pCambia3301 using Agrobacterium tumefaciens GV2260. Primary transformants were selected using the phosphinothricin resistance cassette (bar) encoded in pCambia3301. Complementation lines in the homozygous gun4-2 background were selected by PCR-based screening using the following primer combinations: for SALK026911 T-DNA, AtGUN4_gDNA_Seq_fwd (5′ GATCTTCAAGCTATCGACAATCT) and the left border T-DNA primer (http://signal.salk.edu/); for the wild-type allele, AtGUN4_gDNA_Seq_fwd and AtGUN4 3’ UTR 650 rev (5′TCTGCAGTATCCAGGCGCGT).

Protein Extraction and Western-Blot Analysis

Total leaf protein was extracted from frozen plant material in protein extraction buffer (2% [w/v] SDS, 56 mm NaCO₃, 12% [w/v] Suc, 56 mm DTT, and 2 mm EDTA, pH 8.0) by heating the sample for 20 min at 70°C. After centrifugation, protein concentration was determined using the BCA kit (Thermo Fisher Scientific). Aliquots (15 µg) of protein were fractionated on 12% SDS-PAGE gels, transferred to nitrocellulose membranes, and treated with antibodies according to established protocols (Sambrook, 2001).

Phospho-Specific GUN4 Antibody

A phospho-specific antibody was raised against the phosphorylated C-terminal peptide (C-GADKRVFKTNY(pS)F) of GUN4 (pGUN4). Rabbit IgGs were purified by absorption against the phosphorylated and nonphosphorylated antigen, yielding a pGUN4 and nonphosphorylated pGUN4 antibody, respectively (AMS Biotechnology Europe). The pGUN4 antibody is at least 10 times less sensitive than the fullGUN4 antibody (Peter and Grimm, 2009). Therefore, no phosphorylated protein can be detected in total leaf protein extracts.

Expression and Purification of Recombinant Proteins

Heterologous expression of 6xHIS-tagged proteins was performed in the E. coli Rosetta expression strain, which was grown in Luria-Bertani medium. Starter cultures were grown at 37°C. Expression of transgenes was induced with 1 mm isopropyl-β-d-thiogalactopyranosid at an OD₆₀₀ of 0.4 to 0.6. Os/AtGUN4, and OsCHLI were expressed for 3 h at 30°C. SynGUN4, CrGUN4, OsCHLH, and OsCHLD were expressed overnight at 18°C. All recombinant proteins were purified from the soluble supernatant of E. coli lysates under native conditions using Ni-NTA matrix. Purification was performed as described in “The QIAexpressionist” (www.qiagen.com). All GUN4 proteins were concentrated and transferred into lysis buffer (50 mm NaH₂PO₄, 300 mm NaCl, 10 mm imidazole, pH 8.0) by centrifugation in Amicon Ultra-15 centrifugal filter units (Merck Millipore, 10 kDa MWCO). Glycerol was added to a final concentration of 5% (v/v) and aliquots were stored at −80°C. Purified MgCh subunits (OsCHLH, D, I) were dialyzed overnight into PBS (137 mm NaCl, 2.7 mm KCl, 10 mm Na₂HPO₄, 2 mm KH₂PO₄, pH 7.4) using SnakeSkin dialysis tubing (Thermo Fisher, 10 kD MWCO), concentrated and stored at −80°C. Purity and concentration of recombinant proteins was determined by SDS-PAGE and different amounts of bovine serum albumin as reference standard.

MgCh Assay

The basic in vitro MgCh assay has been described previously (Zhou et al., 2012). A typical assay (100 µL) contained 2.5 µM OsCHLH, 1 µM OsCHLD, 1 µM OsCHLI, and 2.5 µM (Syn, Cr, At, Os) GUN4 proteins in assay buffer (50 mm Tricine-NaOH, 15 mm MgCl₂, 2 mm DTT, 1 mm ATP, and 5 µM Proto). OsCHLH, GUN4, and Proto, and OsCHLD and OsCHLI, were separately preincubated for 20 min on ice in assay buffer, and aliquots of the two mixtures were warmed for 5 min at 30°C and combined in the 96-well plate to start the assay. MgP formation was continuously monitored by spectrofluorometry using a 96-well plate reader mounted on a Hitachi F7000 fluorescence photometer. Excitation and emission wavelengths were set to 416 and 595 nm, respectively, with slit widths of 5 nm (ex) and 10 nm (em). Absolute amounts of MgP formed were calculated from a standard curve.

In vivo MgCh activity was assayed with crude extracts of 21-d-old Arabidopsis plants grown on soil in SD. Briefly, plant material was homogenized in 0.4 m sorbitol, 20 mm Tricine-KOH, 20 mm NaHCO₃ (pH 8.0), filtered through one layer of miracloth, and centrifuged at 500g for 8 min. The pellet was resuspended in assay buffer (0.3 m sorbitol, 20 mm Tricine-KOH, 5 mm MgCl₂, pH 8.0, Roche cOmplete protease inhibitor), and the assay was initiated by mixing the extract with an equal volume of assay buffer supplemented with 10 µM Proto, 10 mm ATP, 2 mm DTT, and 25 mm MgCl₂. MgP was extracted by mixing 100 µL of the assay mixture with 500 µL acetone:0.2M NH₄OH (9:1, v/v), subsequently centrifuged, and quantified by HPLC.

Kinase Assay

The protein kinase assay was performed with native extracts of Synechocystis, Chlamydomonas, and Arabidopsis plants using purified recombinant proteins as substrates. Then 5-to 6-week-old rosette plants were homogenized in 50 mL of homogenization buffer (50 mm HEPES, 330 mm sorbitol, 5 mm MgCl₂, 10 mm KCl, pH 7.6; Bellafiore et al., 2005), filtered through one layer of miracloth, and centrifuged for 5 min at 2,000g (4°C). The supernatant was discarded and the pellet was resuspended in 0.5 to 1 mL of kinase assay buffer (KAB; 50 mm HEPES, 100 mm sorbitol, 5 mm MgCl₂, 5 mm NaCl, pH 7.6, Roche cOmplete protease inhibitor). Chloroplast isolation and fractionation was carried out as described before (Richter et al., 2013). The chloroplast pellet was resuspended in KAB. Cell cultures of Synechocystis sp. PCC 6803 and Chlamydomonas CC-3395 were centrifuged and the pellet was resuspended directly in KAB. Total cell extracts of Synechocystis were obtained by sonification (5× 1 min, 30%, on ice) of resuspended cells. Extracts of Chlamydomonas CC-3395 were prepared by passing cell suspensions several times through a syringe. The Chl concentration of the extracts was determined in 80% acetone (Porra, 2002). A typical kinase assay (50 µL) contained 5 µg of recombinant protein and an extract volume corresponding to 10 µg Chl. To compare kinase activities derived from the different organisms and genotypes, amounts of total protein equivalent to that in the control (At) extract containing 10 µg of Chl were used. Kinase assays were performed either with 5 µCi of γ-P³²ATP and 0.4 mm (cold) ATP or with 1 mm ATP alone. Assay mixtures were set up in reaction tubes and incubated either in darkness (wrapped in aluminum foil) or under an additional light source (100 µmol photons m⁻² s⁻¹) for 30 min in a water bath at 30°C. Then the mixtures were centrifuged for 1 min (at maximum speed and RT) and 40 µL of the supernatant was mixed with 10 µL of 5× SDS-PAGE loading buffer and prepared for SDS-PAGE and western-blot analysis. The pellet fraction was stored at 20°C until further use. Protein phosphorylation was visualized either by exposing Coomassie-stained and dried SDS gels to a phosphor imager screen (radioactive assay) or using the western-blot procedure with the phosphospecific GUN4 antibody.

To assay numerous genotypes or treatments in parallel, we developed a one-leaf kinase assay. This assay uses 100 to 200 mg of fresh or frozen Arabidopsis leaf material. Fresh and frozen leaf samples were homogenized in 1 to 1.5 mL of homogenization buffer either with a reaction tube mortar or with a ball mill (Retsch, Germany), filtered, centrifuged, and resuspended in 50 to 100 µL of KAB. The amount of extract prepared by this technique was sufficient for three to four kinase assays.

GUK activity was assayed by incubating recombinant GUN4 in the presence of 1 mm DTT, 10 mm H₂O₂, 10 µM Proto, 10 µM MgP, 10 and 50 µM DCMU. Steady-state levels of porphyrins in planta were raised by feeding 21-d-old seedlings with 1 mm ALA in 20 mm Tris/HCl (pH 7.4) for 14 h in dark.

Binding Affinities

The binding affinities of Syn, Cr, and AtGUN4 for porphyrins were determined as described before (Karger et al., 2001; Larkin et al., 2003). An aliquot of each protein (0.5 µM) in quenching buffer (50 mm Tricine, 300 mm glycerol, 1 mm DTT, pH 7.9) was mixed with increasing amounts of porphyrins (0–25 µM Proto or MgP), and incubated for 5 min at room temperature. Trp fluorescence was measured with excitation and emission wavelengths set to 280 and 340 nm, respectively. Calculation of the apparent dissociation constants (K_d) was carried out as described (Karger et al., 2001).

In Silico Analysis of GUN4 Structure

The three-dimensional structures experimentally obtained for GUN4 homologs from Synechocystis sp. PCC 6803 (1Y6I; Verdecia et al., 2005) and Thermosynechococcus elongatus (1Z3X; Davison et al., 2005) were compared and aligned with a predicted structure forAtGUN4. The structure prediction was performed using the Phyre2 web portal in the intensive modeling mode (Kelley et al., 2015). Four separate predictions resulted in structures showing identical folding of the GUN4 core domain and variable positions of the C-terminal arm. Alignment and graphic output was produced using PyMol (The PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC).

Determination of the length of the CTL of GUN4 was based on a data set of 84 GUN4 FASTA sequences obtained from a protein blast search of the UniProt database (May 2015) using TPLGHLPLTNALRGTQLLKCVLSHPA peptide as the query. This sequence includes part of helix 8 of the common GUN4 core fold plus part of the C-terminal segment. Downstream processing was carried out with the help of the UniProt web portal and the NCBI taxonomy tool. The CTL was determined using a Phyton script, starting from the last conserved His-Pro (underlined in the original query sequence above) of the common GUN4 core fold and extending to the C terminus. In addition, the sequences where checked for the presence of a phosphorylatable aa (Ser or Thr) among the last five aa.

Miscellaneous

Physiological experiments were performed on at least four independent biological replicates, each containing a pool of 14- to 21-d-old seedlings. If not otherwise indicated, data are given as mean ± sd. In vitro assays and western blots were repeated at least two times and one representative result is shown.

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. GUN4 was not phosphorylated during an MgCh reaction.

• Supplemental Figure S2. Influence of DCMU on GUN4 kinase activity and posttranslational modification of GUN4.

• Supplemental Figure S3. Factors influencing GUN4 kinase activity.

• Supplemental Figure S4. Characterization of gun4-2 mutants expressing the nonphosphorylatable GUN4-I.

• Supplemental Figure S5. Gene expression profiling and western-blot analysis of gun4‐2 mutants expressing GUN4-WT and the nonphosphorylatable GUN4-I.

• Supplemental Figure S6. Comparative analysis of Col-0, gun4-1, gun4-3, and gun4-2 lines expressing GUN4-WT, GUN4-I, and GUN4-D grown under different light intensities.

• Supplemental Figure S7. Analysis of intermediates and end-products from Col-0, gun4-1, gun4-3, and gun4-2 lines expressing GUN4-WT, GUN4-I, or GUN4-D grown under different levels of light.

• Supplemental Figure S8. Purification of recombinant rice MgCh subunits and AtGUN4 and in vitro MgCh assay.

• Supplemental Figure S9. Analysis of the porphyrin-binding properties of recombinant Arabidopsis GUN4 variants by Trp fluorescence quenching.

• Supplemental Figure S10. Alignment of the aa sequences of GUN4 (homologs) from various photosynthetic organisms.

• Supplemental Figure S11. Analysis of the porphyrin-binding properties of recombinant Synechocystis (Syn) and Chlamydomonas (Cr) GUN4 homologs by Trp fluorescence quenching.

• Supplemental Table S1. Overview of genotypes used in this study.

• Supplemental Table S2. Cloning information.

Glossary

aa

amino acid

ALA

5-aminolevulinic acid

Chl

chlorophyll

CTL

C-terminal length

DCMU

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

DTT

dithiothreitol

gDNA

genomic DNA

GUK

GUN4 kinase

KAB

kinase assay buffer

LRG

Leu-Arg-Gly

MgCh

Mg chelatase

MgP

Mg protoporphyrin IX

MgPMME

Mg protoporphyrin IX monomethylester

POR

protochlorophyllide oxidoreductase

Proto

protoporphyrin IX

qPCR

quantitative PCR

S264

Ser 264

TBS

tetrapyrrole biosynthesis