GUN4, a Regulator of Chlorophyll Synthesis and Intracellular Signaling

Plastids are semiautonomous organelles such as the chloroplast. Plastid genomes encode 60 to 80 proteins in higher plants, and more than 3500 nuclear genes are predicted to encode chloroplast proteins in Arabidopsis(1, 2). The developmental and metabolic status of plastids affects the expression of nuclear genes that encode plastid proteins, and distinct plastid-to-nucleus signaling pathways have been hypothesized on the basis of physiological, genetic, and molecular studies. Plastid signals are also important for efficient metabolism and proper leaf development (3,4).

We carried out a genetic screen in Arabidopsis to identify components of the plastid-to-nucleus signaling pathways. Our screen employed the herbicide Norflurazon, which blocks chloroplast development, and the use of Lhcb reporter genes, which are severely repressed in tissues lacking mature chloroplasts. We identified five mutants (gun1 to gun5) that derepress Lhcb in the absence of normal chloroplast development. gun2 through gun5 mutations act in one pathway, whereas gun1 defines a separate pathway (5). gun2, gun3, and gun5mutations affect plastid enzymes that synthesize tetrapyrroles (such as heme, chlorophyll, and phytochromobilin) and reduce levels of the chlorophyll precursor magnesium-protoporphyrin IX (Mg-Proto) (5, 6). Buildup of Mg-Proto in the plastid is sufficient to regulate the expression of a large number of nuclear-encoded chloroplastic proteins whose functions are associated with photosynthesis (4, 6).gun4 is a chlorophyll-deficient mutant that also affects the Mg-Proto pathway (5).

We mapped the gun4-1 allele to a 99-kb interval on the bottom of chromosome 3. A bacterial artificial chromosome (BAC) fragment was used to rescue the pigmentation and gene expression phenotypes of gun4-1 (Fig. 1A) (7). An open reading frame (ORF) was identified on the complementing BAC fragment (GenBank accession no. NM_115802) that encoded a previously uncharacterized protein with a putative chloroplast transit peptide (Fig. 1B). A missense mutation that caused a Leu⁸⁸ → Phe⁸⁸ (L88F) substitution in the derived amino acid sequence was identified ingun4-1 (Fig. 1B) (7). GUN4-related proteins were found only in species that carry out oxygenic photosynthesis (7). Arabidopsis and rice GUN4 proteins have small divergent COOH-terminal extensions that are absent in bacterial and red algae chloroplast relatives (Fig. 1B). Most species have one GUN4-related gene, but Synechocystis andNostoc have three and four, respectively. One of theSynechocystis homologs is more similar to GUN4 than the other two (fig. S1). Three of the four Nostoc homologs are fused to other ORFs (GenBank accession nos. BAB73525, BAB78327, BAB76648, and BAB77506).

• Download high-res image
• Open in new tab
• Download Powerpoint

The identity of GUN4 was confirmed by expressing the candidate gene in gun4-1 under the control of the strong and constitutive 35S promoter. gun4-1 transformants containing the 35S::GUN4 transgene had either a wild-type appearance or a noticeable lack of chlorophyll (Fig. 1C), and there was a correlation between the presence of chlorophyll and the presence of the encoded protein (Fig. 1D). These data indicated that this candidate is GUN4 and that cosuppression ofGUN4 in the transformants blocked chlorophyll accumulation. The gun4-1 mutant contained less GUN4 protein than the wild type (Fig. 1D), but similar amounts of GUN4 mRNA are present ingun4-1 and wild-type plants (7), which indicates that the L88F substitution affects protein stability. A minor fraction of GUN4 protein routinely migrated more slowly than the bulk of GUN4 (Fig. 1D) (7). The predominant electrophoretic form of GUN4 migrated similarly to recombinant GUN4 Δ1-69, which was expressed in Escherichia coli(7). The ratio between the more slowly migrating species and the predominant form of GUN4 appeared to be approximately maintained when GUN4 was overexpressed or underexpressed (Fig. 1D).

To determine whether GUN4 is essential for chlorophyll accumulation, a null allele, gun4-2, was isolated that had a transferred DNA (T-DNA) insertion near the center of theGUN4 ORF (GenBank accession no. BH251364) (Fig. 1B) (7); gun4-2 did not accumulate GUN4 protein (Fig. 1D). gun4-2 mutants were very small and developed yellow or white tissue when grown under standard conditions (Fig. 1E) (8). gun4-2 produced a number of leaves and flowerlike structures after approximately 2 months on sucrose-containing media (Fig. 1F) and was pale green in dim light (Fig. 1E). These results indicated that, although GUN4 is required for chlorophyll accumulation under normal growth conditions,GUN4 is not essential for chlorophyll synthesis.

The effect on chlorophyll accumulation (Fig. 1, C, E, and F), the putative transit peptide in the derived amino acid sequence, and the similarity to hypothetical red algae chloroplast and cyanobacterial proteins (Fig. 1B) suggested that GUN4 was a chloroplast protein. Analysis of purified and fractionated Arabidopsischloroplasts by immunoblotting with affinity-purified antibodies to GUN4 (9) confirmed chloroplast localization (Fig. 2A). We found GUN4 in the stromal fraction in addition to the envelope and thylakoid membranes (Fig. 2A). The difference in the amount of GUN4 in envelope and thylakoid membranes may be overestimated (Fig. 2A), because envelopes contain less protein than thylakoids (10). Because GUN4 was predicted to be a soluble protein, our analysis of fractionated chloroplasts suggested that GUN4 might be tethered to chloroplast membranes by specific protein-protein interactions.

• Download high-res image
• Open in new tab
• Download Powerpoint

If GUN4 stably binds other proteins, identification of associated proteins might suggest its function. Each chloroplast fraction was analyzed by gel filtration chromatography (9) to determine whether GUN4 was a component of a protein complex. The elution profile of stromal GUN4 resembled the elution profile of recombinant GUN4 Δ1-69, which was expressed in E. coli and purified (Fig. 2B) (9). Mature GUN4 is 22 kD, and the elution profile of recombinant and stromal GUN4 is consistent with an elongated monomer or an extremely compact dimer. The solubilized thylakoid peak was broad, and the predominant species was 500 kD (Fig. 2B). The solubilized envelope form of GUN4 eluted near the void fraction (Fig. 2B), and the exclusion limit of the column was 1.3 MD.

We purified a GUN4 complex from solubilized thylakoids, and copurifying proteins were identified by matrix-assisted laser desorption/ionization–time of flight (MALDI-TOF) mass spectrometry (9). The 500-kD thylakoid complex disintegrated into a 230-kD complex and free GUN4 during this procedure (7), resulting in two copurifying proteins and an excess of GUN4 in the final fraction. The copurifying proteins were identified as the 144-kD Proto-binding subunit of Mg-chelatase, commonly called ChlH (11), and a fragment of ChlH (Fig. 2C).GUN5 was previously found to encode ChlH, andgun4gun5 double mutants have a severe chlorophyll-deficient phenotype (5). However, GUN4 is not required for Mg-chelatase activity in vitro or in vivo (Fig. 1E) (11), suggesting that either GUN4 regulates Mg- chelatase or GUN4 has a different function that involves ChlH.

Mg-chelatase is composed of three subunits (ChlH, ChlD, and ChlI) and catalyzes the first step in the chlorophyll branch of the plastid tetrapyrrole biosynthetic pathway (Fig. 3A) (11).Synechocystis was chosen as a system to test the hypothesis that GUN4 regulates Mg-chelatase, because SynechocystisMg-chelatase is well characterized (12) andSynechocystis has three GUN4 relatives. SynGUN4a, theSynechocystis ORF that is most similar to GUN4 (fig. S1), and Synechocystis Mg-chelatase subunits were expressed inE. coli and purified (Fig. 3B) (9). SynGUN4 markedly stimulated Mg-chelatase in a standard assay that employed deuteroporphyrin IX (Deutero), a more water-soluble derivative of Proto (Fig. 3C). The stimulatory effect of SynGUN4a was saturable and increased the reaction rate threefold (fig. S2). These data suggest that SynGUN4a activates Mg-chelatase by binding to ChlH.

• Download high-res image
• Open in new tab
• Download Powerpoint

Formation of the ChlH-Proto complex is a rate-limiting step in Mg-chelation (12), and because GUN4 can bind ChlH, we tested the possibility that GUN4 affects ChlH-Deutero binding. The rate-limiting effect of ChlH-Proto complex formation has been demonstrated with preincubation experiments. For example, if ChlH is preincubated with Proto and then diluted into a Mg-chelatase reaction that contains low Proto concentrations, a much higher Mg-Proto synthesis rate is obtained as compared to the same reaction set up without a preincubation (12). This preincubation effect was observed with Deutero (Fig. 3D, compare traces 1 and 3). When ChlH was preincubated with SynGUN4a, ChlH used low Deutero concentrations much more effectively (Fig. 3D, compare traces 1 and 2). The reaction that used a ChlH-Deutero preincubation reached a plateau (Fig. 3D, trace 3), but the reaction that used a ChlH-SynGUN4a preincubation did not reach a plateau and eventually surpassed the reaction that used the ChlH-Deutero preincubation (Fig. 3D, compare traces 2 and 3). These results suggest that SynGUN4a increased the rate of ChlH-Deutero complex formation.

A reaction that used a ChlH-SynGUN4a-Deutero preincubation produced Mg-Deutero at a relatively high rate as expected (Fig. 3D, compare trace 5 to traces 1 to 3). However, a reaction that used a SynGUN4a-Deutero preincubation also produced Mg-Deutero at a relatively high rate (Fig. 3D, compare traces 4 and 5). If SynGUN4a were an allosteric regulator and SynGUN4a-ChlH interactions simply increased the affinity of ChlH for Deutero, a reaction that used a SynGUN4a-Deutero preincubation would produce Mg-Deutero at a rate that was intermediate between that shown in traces 1 and 2 (Fig. 3D). The high rate indicated by trace 4 suggests that SynGUN4a promotes Mg-chelatase–substrate interactions by binding Deutero. From these data it also seemed possible that SynGUN4a might bind Mg-Deutero.

Mg-Deutero cofractionated with glutathioneS-transferase (GST)–GUN4 Δ1-69 and GST-SynGUN4a but not GST on glutathione agarose (Fig. 4, A and B) (7), which indicates that Mg-Deutero stably associated with GUN4 Δ1-69 and SynGUN4a. Tetrapyrrole-binding proteins often alter the absorbance spectra of bound tetrapyrroles (13). Purified GUN4 Δ1-69 (Fig. 4C) red-shifted the B (Soret) B bands and the red-most Q bands of Deutero and Mg-Deutero by at least several nm (Fig. 4, C and D). The most dramatic effects were on the Mg-Deutero B band, which was shifted from 402 to 412 nm (Fig. 4D), and on the Deutero extinction coefficient, which was approximately doubled (Fig. 4C). The Deutero and Mg-Deutero binding activities of SynGUN4a were saturable (Fig. 4, E and F). A dissociation constant (K _d) value of 0.26 ± 0.029 μM was determined for Mg-Deutero (Fig. 4E), and aK _d value of 2.2 ± 0.3 μM was determined for Deutero (Fig. 4F). The data fit best with a model that predicts a single type of binding site (9).

• Download high-res image
• Open in new tab
• Download Powerpoint

GUN4 is the first known example of a plastid Proto/Mg-Proto binding protein that stimulates Mg-chelatase. However, an unrelated protein has been reported to regulate upstream tetrapyrrole biosynthetic enzymes in a species that carries out anoxygenic photosynthesis (14). Although GUN4 is not an essential Mg-chelatase subunit in vitro or in vivo (Figs. 1E and 3C), reduced chlorophyll accumulation throughout the life cycle of all gun4 mutants indicates an essential function for GUN4 in plants. SynGUN4a is required for effective use of low Deutero concentrations and high rates of Mg-Deutero synthesis by Synechocystis Mg-chelatase (Fig. 3, C and D). The marked stimulatory effect of SynGUN4a-Deutero preincubations on Mg-chelatase activity (Fig. 3D) and the similar Deutero affinities of SynGUN4a and ChlH (Fig. 4F) (15) argue that SynGUN4a promotes Mg-chelatase–substrate interactions. In the presence of SynGUN4a, the activity of SynechocystisMg-chelatase appears qualitatively more similar to that of Mg-chelatase from crude chloroplast fractions, which possess a 30- to 160-fold lower Proto Michaelis constant (K _m) thanSynechocystis Mg-chelatase (12, 16,17). GUN4 probably contributes substantially to the relatively low Proto K _m values reported in crude chloroplast fractions, and without GUN4, higher plants probably do not effectively use low concentrations of Proto or synthesize Mg-Proto at high rates. SynGUN4a bound Mg-Deutero more tightly than Deutero (Fig. 4, E and F), and the SynGUN4a K _d value for Mg-Deutero binding was substantially lower than the 2.43 ± 0.46 μM K _d value reported for ChlH (15). Because SynGUN4a has a relatively high affinity for Mg-Deutero, we suggest that SynGUN4a may participate in product release and/or a separate function that involves binding the product of Mg-chelatase.

Mg-chelatase has been suggested to associate with the chloroplast inner envelope, but Mg-chelatase has not been definitively localized within the chloroplast (18). Because we purified a ChlH-GUN4 complex from thylakoid membranes (Fig. 2C), we suggest that a fraction of Mg-chelatase may associate with thylakoids. Although more work is required to unambiguously define the composition of the envelope and thylakoid GUN4 complexes, it seems likely that both thylakoid and envelope GUN4 complexes contain Mg- chelatase subunits. The heterogeneity of the GUN4 complexes extracted from thylakoid membranes (Fig. 2B) and the large size difference between the thylakoid and envelope complexes (Fig. 2B) may reflect the unstable interactions among Mg-chelatase subunits (11) or the association of a ChlH-GUN4 complex with other envelope proteins that might have catalytic or signaling functions.

The large pool of free GUN4 in the stroma (Fig. 2B) suggests additional functions for GUN4. For example, the tetrapyrrole-binding activity of GUN4 might protect Proto and Mg-Proto from catabolic enzymes, help target Mg-Proto to downstream chlorophyll biosynthetic pathway enzymes, or help protect plants from photooxidative damage. Proto and Mg-Proto interactions with O₂ produce reactive oxygen species in bright light, and Mg-Proto accumulates transiently at dawn (19). The photosensitivity of gun4-2 (Fig. 1E) and the absence ofGUN4 homologs in species that carry out anoxygenic photosynthesis support a role for GUN4 in photoprotection. However, we cannot rule out the possibility that the pool of free GUN4 is due to unstable ChlH-GUN4 interactions.

The precise role of GUN4 in plastid-to-nucleus signaling is not clear, because GUN4 both stimulates signal synthesis and binds the signal. Mg-Proto is thought to exit the plastid and interact with cytosolic signaling pathways (4–6). This model seems reasonable because related molecules (such as mitochondrial heme precursors, heme, phytochromobilin, and chlorophyll degradation products) are exported from plastids (20–23) and because tetrapyrrole trafficking is likely monitored by the cell. GUN4 may promote Mg-Proto export by stimulating Mg-Proto synthesis and/or by recruiting Mg-Proto to the plastid envelope.