coordination of plastid protein import and nuclear …...genes in the nucleus (mulo et al., 2003)....

Coordination of Plastid Protein Import and NuclearGene Expression by Plastid-to-NucleusRetrograde Signaling1[W][OA]

Tomohiro Kakizaki2, Hideo Matsumura, Katsuhiro Nakayama, Fang-Sik Che,Ryohei Terauchi, and Takehito Inaba*

The 21st Century Centers of Excellence Program, Cryobiofrontier Research Center, Iwate University, Morioka,Iwate 020–8550, Japan (T.K., K.N., T.I.); Iwate Biotechnology Research Center, Kitakami, Iwate 024–0003,Japan (H.M., R.T.); and Department of Environmental Biology, Faculty of Bioscience, Nagahama Institute ofBioscience and Technology, Nagahama, Shiga 526–0829, Japan (F.-S.C.)

Expression of nuclear-encoded plastid proteins and import of those proteins into plastids are indispensable for plastidbiogenesis. One possible cellular mechanism that coordinates these two essential processes is retrograde signaling fromplastids to the nucleus. However, the molecular details of how this signaling occurs remain elusive. Using the plastid proteinimport2 mutant of Arabidopsis (Arabidopsis thaliana), which lacks the atToc159 protein import receptor, we demonstrate that theexpression of photosynthesis-related nuclear genes is tightly coordinated with their import into plastids. Down-regulation ofphotosynthesis-related nuclear genes is also observed in mutants lacking other components of the plastid protein importapparatus. Genetic studies indicate that the coordination of plastid protein import and nuclear gene expression is independentof proposed plastid signaling pathways such as the accumulation of Mg-protoporphyrin IX and the activity of ABAINSENSITIVE4 (ABI4). Instead, it may involve GUN1 and the transcription factor AtGLK. The expression level of AtGLK1 istightly correlated with the expression of photosynthesis-related nuclear genes in mutants defective in plastid protein import.Furthermore, the activity of GUN1 appears to down-regulate the expression of AtGLK1 when plastids are dysfunctional. Basedon these data, we suggest that defects in plastid protein import generate a signal that represses photosynthesis-related nucleargenes through repression of AtGLK1 expression but not through activation of ABI4.

Plastids are a diverse group of organelles that per-form essential metabolic and signaling functionswithin all plant cells. It is generally believed thatplastids originated from a unicellular photosyntheticbacterium that was taken up by a eukaryotic host cell(Dyall et al., 2004). During evolution, most of the genesencoded by the bacterial ancestor have been trans-ferred to the host nuclear genome; for example, theplastid genome of Arabidopsis (Arabidopsis thaliana)encodes fewer than 100 open reading frames (Martinet al., 1998). Consequently, plastid biogenesis is de-pendent on the import of nuclear-encoded plastid

proteins (Keegstra and Cline, 1999; Soll and Schleiff,2004; Kessler and Schnell, 2006; Inaba and Schnell,2008; Jarvis, 2008), the genes for which must be ex-pressed at an appropriate level. For example, many ofthe photosynthesis-related nuclear genes that are re-quired for chloroplast biogenesis are induced viaphotoreceptors, such as phytochrome, in response tolight quality and quantity (Terzaghi and Cashmore,1995), so that the photosynthesis-related proteinswill be available for import into the developing chlo-roplasts. Other types of plastids, according to theirspecific metabolic functions, need other sets ofnuclear-encoded proteins. Therefore, the expressionof specific sets of nuclear genes and the import of theirtranslation products are indispensable for plastid dif-ferentiation.

After they have been imported into plastids,nuclear-encoded plastid proteins combine with plastid-encoded proteins to form functional multiproteincomplexes, such as photosystems and metabolic en-zymes. Because plastids need a distinct set of proteinsin specific amounts dictated by their functional andmetabolic states, it is also critical for plastids to givefeedback information to the nucleus, so that nucleargene expression can be adjusted appropriately. Thisfine-tuning mechanism is important for determin-ing which nuclear-encoded proteins are imported,and consequently it regulates the differentiation of

1 This work was supported by the 21st Century Centers ofExcellence Program, Grant-in-Aid for Young Scientists (no.18780247), the Agricultural Chemical Research Foundation, thePresident of Iwate University, and the Japan Society for the Promo-tion of Science (postdoctoral fellowship for young scientists to T.K.).

2 Present address: National Institute of Vegetable and Tea Science,360 Kusawa, Ano, Tsu, Mie 514–2392, Japan.

* Corresponding author; e-mail [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policydescribed in the Instructions for Authors (www.plantphysiol.org) is:Takehito Inaba ([email protected]).

[W] The online version of this article contains Web-only data.[OA] Open Access articles can be viewed online without a sub-

scription.www.plantphysiol.org/cgi/doi/10.1104/pp.109.145987

Plant Physiology�, November 2009, Vol. 151, pp. 1339–1353, www.plantphysiol.org � 2009 American Society of Plant Biologists 1339 www.plantphysiol.orgon July 3, 2020 - Published by Downloaded from

Copyright © 2009 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

plastids in a tissue- and developmental state-dependentmanner.

To accomplish such regulation, plastids send signals(called plastid signals) to the nucleus via multiplesignal transduction pathways. To date, three distinctplastid processes—tetrapyrrole biosynthesis, plastidgene expression, and organellar redox reactions—havebeen shown to give rise to plastid signals (Nott et al.,2006). These plastid signals have been studied usingspecific inhibitors and controlled light conditions. Forexample, tetrapyrroles may act as a plastid signal thatis induced by norflurazon, an inhibitor of carotenoidsynthesis. Genetic studies have identified several ge-nomes uncoupled mutants (gun1 to gun5) that exhibitnorflurazon-insensitive LHCB expression, and manyof these mutants have lesions in the tetrapyrrolebiosynthetic pathway (Susek et al., 1993; Mochizukiet al., 2001; Larkin et al., 2003). However, there areconflicting reports on the mechanism of norflurazonaction (La Rocca et al., 2001; Strand et al., 2003;Mochizuki et al., 2008; Moulin et al., 2008). Anotherplastid process that affects nuclear gene expression isplastid gene expression: inhibition of plastid proteinsynthesis by lincomycin generates a plastid signal thatsuppresses the expression of photosynthesis-relatedgenes in the nucleus (Mulo et al., 2003). This signalingpathway is only active during the first 3 d of seedlingdevelopment (Beck, 2005). In contrast, a redox retro-grade signaling pathway seems to play an importantrole in light acclimation of mature plants (Beck, 2005).Thus, alteration of the functional state of the plastid byinhibitors or by light conditions generates a widevariety of plastid signals, depending on the develop-mental stage of the plant.

Analysis of mutants lacking plastid protein importreceptors has suggested that plastid biogenesis re-quires the coordination of plastid protein import withnuclear gene expression and that this coordinationis likely to involve retrograde signaling from plas-tids to the nucleus. The protein import receptors ofplastids are composed of two distinct families of GTP-binding proteins (Kessler and Schnell, 2006). One ofthese, the atTOC159 family, consists of four genes inArabidopsis—atTOC159, atTOC132, atTOC120, andatTOC90 (Bauer et al., 2000; Hiltbrunner et al., 2004)—and appears to comprise a set of functionally distinctreceptors. The plastid protein import2 (ppi2) mutant,which lacks the most abundant receptor, atToc159,exhibits a severe albino phenotype. The mutant fails toaccumulate representative photosynthesis-related pro-teins, even though nonphotosynthetic proteins are ac-cumulated normally (Bauer et al., 2000). This differencein accumulation is in part attributable to the ability ofatToc159 to recognize photosynthesis-related, but notnonphotosynthetic, proteins as its substrates (Smithet al., 2004). An intriguing observation is that photo-synthesis-related nuclear gene expression also seems tobe compromised in ppi2. For example, two representa-tive photosynthesis-related nuclear genes, LHCB andsmall subunit of Rubisco (SSU), appear to be down-

regulated at the mRNA level in ppi2, whereas twononphotosynthetic genes, atTOC34 and chorismatemutase1, are unaffected (Bauer et al., 2000). When anartificial substrate of atToc159, preSSU-GFP, was ex-pressed in ppi2 under the control of a constitutivepromoter, the expression of preSSU-GFP was no longerdown-regulated and ppi2 accumulated a significantamount of the unprocessed protein in the cytosol (Smithet al., 2004). These observations suggest that plastidprotein import and nuclear gene expression are tightlycoordinated to regulate the flow of nuclear-encodedplastid proteins in response to the functional anddevelopmental states of the plastids. However, themolecular mechanism that coordinates protein importand nuclear gene expression remains elusive.

In this study, we examine the role of plastid proteinimport in the regulation of nuclear gene expression.Using mutants defective in plastid protein import,we show that such defects cause down-regulation ofphotosynthesis-related nuclear gene expression. Thisregulation appears to be mediated by a plastid signal-ing pathway that may involve both GUN1 andAtGLK1 and is distinct from the ABA INSENSITIVE4(ABI4) pathway. We propose a novel mechanism bywhich plastids regulate nuclear gene expression ac-cording to the rate of plastid protein import and thefunctional state of plastids.

RESULTS

Expression of Photosynthesis-Related Nuclear Genes IsSpecifically Impaired in ppi2 Plants

It has been shown that atToc159 is a membrane-bound GTPase and appears to play a critical role in theimport of nuclear-encoded photosynthesis-relatedproteins (Kessler et al., 1994; Schnell et al., 1994; Baueret al., 2000; Smith et al., 2004). Therefore, ppi2-1 (orig-inally referred to as ppi2; Bauer et al., 2000), whichlacks the atToc159 protein import receptor, fails toaccumulate major photosynthesis-related proteins(e.g. LHCP and SSU) but not nonphotosynthetic plastidproteins (Bauer et al., 2000). Interestingly, mRNAaccumulation for a couple of nuclear-encoded photo-synthesis-related proteins was also shown to be com-promised in the ppi2-1mutant (Bauer et al., 2000). Thisobservation suggests that plastid protein import andnuclear gene expression are tightly coordinated toregulate the flow of nuclear-encoded plastid proteinsin response to the functional and developmental statesof the plastids.

To further pursue the link between plastid proteinimport and nuclear gene expression, we expanded ouranalysis of expression in the ppi2-1mutant to include adozen nuclear-encoded photosynthesis-related andnonphotosynthetic plastid proteins. Total proteinswere extracted from wild-type (Wassilewskija [Ws])and ppi2-1 leaves, resolved by SDS-PAGE, and probedwith antibodies against various proteins (Fig. 1A;

Kakizaki et al.

1340 Plant Physiol. Vol. 151, 2009 www.plantphysiol.orgon July 3, 2020 - Published by Downloaded from



Supplemental Fig. S1). As a control, the membraneswere also probed with antibody against actin. Asshown in Figure 1, A and B, the levels of all of thephotosynthesis-related proteins examined were de-creased in ppi2-1 to less than 20% of the level in thewild type, supporting the idea that atToc159 is respon-sible for the import of photosynthesis-related proteins.To further investigate the mechanisms that regulatethe accumulation of photosynthesis-related proteins inppi2-1, we measured the mRNA levels of the respec-tive genes using real-time PCR. This analysis revealedthat the expression of each gene was suppressed inppi2-1 compared with the wild type (Fig. 1B; theexpression level in the wild type was set to 1). Thesedata indicate that the accumulation of photosynthesis-

related proteins in ppi2-1 plastids is correlated withtheir mRNA expression level in the nucleus.

We next examined the accumulation of protein (Fig.1C; Supplemental Fig. S1) and mRNA (Fig. 1D) ofnonphotosynthetic plastid proteins to investigatewhether the down-regulation of nuclear-encoded plas-tid proteins in ppi2-1 is specific to photosynthesis-related proteins. Five proteins exhibiting variousactivities were chosen for the analysis. As shown inFigure 1, C and D, all of the proteins except thecarboxyl transferase a-subunit of acetyl-CoA carbox-ylase (ACC-CTa) accumulated to normal or higherlevels in the mutant compared with the wild type.Real-time PCR analysis revealed that the mRNA ac-cumulation for these proteins was not down-regulated

Figure 1. Expression profiles of nuclear-encodedplastid proteins in the wild type and ppi2. A andC, Accumulation of photosynthesis-related pro-teins (A) and nonphotosynthetic proteins (C) inwild-type (Ws) and ppi2-1 plants. Total proteinextracts were resolved by SDS-PAGE and probedwith antisera against the proteins indicated to theleft. B and D, Quantitative analysis of mRNA andprotein levels in ppi2-1 plants. The mRNA levelswere analyzed by real-time PCR and normalizedto the level of 18S ribosomal RNA. The proteinlevels were quantified using triplicate immuno-blots (Supplemental Fig. S1) and normalized toactin levels. In the case of POR protein, the actinsignal shown in C (Supplemental Fig. S1B) wasused for normalization. In the case of ACC-CTaand atTic110 proteins, the actin signal shown in A(Supplemental Fig. S1A) was used for normaliza-tion. The ratio of ppi2-1 to the wild type (ppi2/WT) is shown. Each bar, plotted here on a logscale, represents the mean of at least three inde-pendent samples. Error bars represent 1 SD. E,Classification of genes down- or up-regulated inthe ppi2-1 mutant as determined by SuperSAGE.The charts show the classification of the 1,000most down-regulated (left) and the 1,000 mostup-regulated (right) genes in the ppi2-1mutant (ascompared with the wild type). The tags wereclassified based on the Gene Ontology annota-tion, which assigns cellular components (P ,0.05, Fisher’s exact test). The abundance of eachclass of genes (as a percentage of the total down-or up-regulated genes) is shown in parentheses(down/up). Total RNA and proteins were purifiedfrom the third to sixth leaves of wild-type andppi2-1 plants grown on plates. LSU, Large subunitof Rubisco; OE23, 23-kD oxygen-evolving com-plex protein; POR, protochlorophyllide oxidore-ductase; Hsp93, 93-kD heat shock protein; IEP37,37-kD inner envelope protein; Tic110, 110-kDprotein of Tic protein translocation apparatus.

Communication between Plastid and Nucleus

Plant Physiol. Vol. 151, 2009 1341 www.plantphysiol.orgon July 3, 2020 - Published by Downloaded from



in the ppi2-1 mutant (Fig. 1D). The discrepancy be-tween the levels of ACC-CTa mRNA and ACC-CTaprotein might be an indirect effect of the slightlyreduced expression of the plastid-encoded b-subunit(ACC-CTb) in ppi2-1 (Fig. 1D).

These results were further confirmed by examiningthe phenotype of an additional atTOC159 mutant(referred to as ppi2-2), a T-DNA insertional mutant inthe Columbia (Col-0) background from the GABI-Katcollection (Supplemental Fig. S2). In this mutant,which has no detectable atTOC159, the expressionprofile of nuclear genes encoding plastid proteins wassimilar to that of ppi2-1, except for a slight decrease inIEP37 expression (Fig. 2A). Taken together, these dataindicate that the ppi2 mutation compromises the accu-mulation of nuclear-encoded photosynthesis-relatedproteins and that the decrease in protein accumulationis strongly correlated with the down-regulation ofnuclear-encoded gene expression.

Impact of the Protein Import Defect on the Nuclear

Transcriptome in the ppi2 Mutant

The lack of atToc159 in Arabidopsis results in asevere albino, seedling-lethal phenotype, as high-lighted by analysis of the ppi2-1 mutant (Bauer et al.,2000). This phenotype implies that the defect inatToc159 causes dramatic changes in nuclear geneexpression in the ppi2-1 mutant. Therefore, to investi-gate the global impact of the atToc159 defect onnuclear-encoded gene expression, we performedSuperSAGE (for Super Serial Analysis of Gene Ex-pression) analysis using wild-type and ppi2-1 plants.SuperSAGE is a powerful tool to examine global geneexpression profiles quantitatively (Matsumura et al.,2003). As summarized in Supplemental Figure S3,126,238 tags (29,238 unique tags) and 204,218 tags(43,580 unique tags) were identified from wild-typeand ppi2-1 cDNAs, respectively. To identify genescorresponding to each tag sequence, The ArabidopsisInformation Resource 7 gene annotation database wasqueried with 26-bp tags using the BLASTN program.Then, the 2,000 genes most significantly affected by theppi2-1mutation (1,000 each of up- and down-regulatedgenes) were classified based on the Gene Ontologyterms (http://www.arabidopsis.org/tools/bulk/go/index.jsp; Supplemental Tables S1 and S2). Amongthe down-regulated genes, approximately 30% en-coded proteins destined for chloroplasts or plastids(Fig. 1E). In contrast, only approximately 5% of theup-regulated genes encoded proteins targeted to chlo-roplasts or plastids (Fig. 1E). Given the fact that only10% to 15% (approximately 3,500) of nuclear genes arepredicted to encode plastid proteins, it appears thatthis class of genes is preferentially down-regulated inthe ppi2-1mutant. These data also suggest that there isa tight coordination between plastid protein importand the nuclear transcriptome and that this regulationis critical for proper accumulation of plastid proteinsand plastid biogenesis.

Figure 2. Expression profiles of nuclear-encoded plastid proteins in amutant allelic to ppi2-1 and in other mutants defective in plastidprotein import. A, Quantitative analysis of nuclear-encoded geneexpression in the ppi2-2 mutant by real-time PCR (Col-0 background;allelic to ppi2-1). B, Immunoblot analysis of plastid proteins in theppi2-2 mutant. C, Quantitative analysis of nuclear-encoded genes inthe attoc132 attoc120+/2 mutant by real-time PCR. D, Immunoblotanalysis of plastid proteins in the attoc132 attoc120+/2 mutant. E,Quantitative analysis of nuclear-encoded genes in the attic20-I mutantby real-time PCR. F, Immunoblot analysis of plastid proteins in theattic20-I mutant. In each case, transcript levels (A, C, and E) wereanalyzed by real-time PCR and normalized to the transcript levels ofACTIN2. The expression level in the wild type was set to 1. Each bar,plotted here on a log scale, represents the mean of at least threeindependent samples. Error bars represent 1 SD. Asterisks above the barsindicate significant differences (P , 0.05) between the wild type andthe respective mutant, as determined by Student’s t test. The ppi2-2,attic20-I, and Col-0 plants were grown on plates. attoc132 attoc120+/2 and Ws were grown on soil.

Kakizaki et al.




Defects in Plastid Protein Import Result in

Down-Regulation of Photosynthesis-Related Genesin the Nucleus

The correlation between protein import defectsand the down-regulation of photosynthesis-relatednuclear genes was further examined using additionalmutants: attoc132 attoc120+/2 and tic20-I. The attoc132attoc120+/2 mutant has fewer atToc132/120 proteinimport receptors, which, unlike atToc159, are specificfor nonphotosynthetic proteins (Ivanova et al., 2004;Kubis et al., 2004). The tic20-I mutant lacks a putativechannel component of the protein translocationmachinery at the inner envelope membrane of chlo-roplast. In this mutant, the import of both nonpho-tosynthetic and photosynthesis-related proteins iscompromised (Teng et al., 2006). As shown in Figure2, C and E, the expression of two photosynthesis-related nuclear genes,OE23 and LHCB1, was impairedin both mutants. However, the effect on the expressionof the nonphotosynthetic genes IEP37 and pyruvatedehydrogenase E1a subunit (E1a) was relatively moder-ate (Fig. 2, C and E). The accumulation of each proteinwas proportional to themRNA expression level (Fig. 2,D and F). These results suggest that plants withreduced import of nonphotosynthetic proteins alsospecifically down-regulate the expression of photo-synthesis-related nuclear genes.Since all of these mutants exhibit pale or albino

phenotypes in the early stages of cotyledon develop-ment (Bauer et al., 2000; Ivanova et al., 2004; Teng et al.,2006), it was possible that the arrest of plastid devel-opment at the cotyledon stage led to the severe growthdefects in the adult stage. That is, the down-regulationof nuclear gene expression in these mutants could be adevelopmental effect rather than a direct effect ofdefects in protein import. To obtain more evidence thatthe inhibition of plastid protein import results directlyin the down-regulation of nuclear gene expression, wetook advantage of atTOC159-silenced plants. Theseplants have a transgene expressing sense atTOC159from a constitutive promoter (Bauer et al., 2002), andmost of them look normal during the early stages ofgrowth. However, as the plants grow older, some oftheir leaves become variegated due to cosuppressionof the endogenous atTOC159 gene (Fig. 3, B–D). Giventhe late onset of the variegated phenotype (Fig. 3C), itwas assumed that normal proplastid-to-chloroplastdifferentiation had been completed during the cotyle-don stage. We analyzed mRNA and protein accumu-lation in green versus albino sectors of these plants(Fig. 3, C and D, arrows and arrowheads). In the albinosectors, the accumulation of nuclear-encoded photo-synthesis-related proteins was specifically impaired,and this observation was in part attributable to thedown-regulation of photosynthesis-related nucleargene expression (Fig. 3, E and F). In contrast, neithermRNA nor protein accumulation for the nonphoto-synthetic proteins, E1a and IEP37, was significantlyaffected.

Altogether, these data indicate that defects in pro-tein import into chloroplasts result in the down-regulation of photosynthesis-related nuclear geneexpression regardless of the import substrates of thedefective import apparatus. They also suggest thatdefects in protein import into plastids, such as thosecaused by the ppi2 mutation, induce the generationof plastid signal that suppresses the expression ofphotosynthesis-related genes in the nucleus.

Figure 3. Expression profiles of nuclear-encoded plastid proteins in atransgenic line exhibiting atTOC159 silencing. A to D, Representativephenotypes of transgenic plants carrying a 35S::atTOC159 transgene.A, A 14-d-old wild-type plant. B, A 14-d-old transgenic plant showingnormal greening. C and D, A 25-d-old (C) and a 33-d-old (D) transgenicplant showing a variegated phenotype. Green and white sectors typicalof those used for further analysis are indicated by arrowheads andarrows, respectively. Bars = 1 cm. E, Immunoblot analysis of totalprotein extracts from green sectors (lane 1) and albino sectors (lane 2) oftransgenic plants. The immunoblots were probed with antisera againstthe proteins indicated at the left. The membrane was also probed withanti-actin antibody as a loading control. F, Gene expression profile ingreen and albino sectors from 25-d-old 35S::atTOC159 transgenicplants. Transcript levels were analyzed by real-time PCR and normal-ized to the transcript levels of ACTIN2. The expression level in thegreen sectors was set to 1. Each bar, plotted here on a log scale,represents the mean of at least three independent samples. Error barsrepresent 1 SD. Asterisks above the bars indicate significant differences(P , 0.05) between green and albino sectors, as determined byStudent’s t test.





Plastid Mg-Protoporphyrin IX Levels May

Not Be Correlated with the Repression ofPhotosynthesis-Related Nuclear Genes in ppi2

The down-regulation of nuclear gene expression inresponse to defects in plastid protein import suggeststhat there are retrograde signaling pathways thatcoordinate plastid protein import and nuclear geneexpression. It has been proposed that intermediates inthe tetrapyrrole synthesis pathway act as retrogradesignals, and several GUN genes have been identifiedthat are involved in the synthesis of tetrapyrrole and inthe signaling activities of pathway intermediates(Vinti et al., 2000; Mochizuki et al., 2001; Larkinet al., 2003; Strand et al., 2003). According to onescenario, the accumulation of Mg-protoporphyrin IX(Mg-ProtoIX) results in the repression of photosynthesis-related nuclear genes when chloroplasts are damagedby norflurazon treatment (La Rocca et al., 2001; Strandet al., 2003). Although other reports show that the levelof Mg-ProtoIX is not correlated to the level of LHCBexpression (Mochizuki et al., 2008; Moulin et al., 2008),it is still of interest to test whether GUN genes andtetrapyrroles are responsible for the down-regulationof photosynthesis-related nuclear genes in plants de-fective in plastid protein import.

We first examined the expression of GUN4, a keyregulator of Mg-ProtoIX synthesis (Larkin et al., 2003),in the ppi2-2mutant. We switched from ppi2-1 to ppi2-2for these experiments, as the ppi2-2 mutant is in theCol-0 background and can be used for crossing withother mutants in the same background. Real-time PCRanalysis revealed that the expression of GUN4 in4-week-old ppi2-2 was approximately 30% that in thewild type (Fig. 4A). Furthermore, GUN4 protein wasdramatically decreased in 4-week-old ppi2-2 mutant(Fig. 4B). We observed similar results in other plastidprotein import mutants (Fig. 2, D and F). BecauseGUN4 stimulates Mg-chelatase activity (Larkin et al.,2003), these results suggest that the synthesis of Mg-ProtoIX would also be significantly impaired in thesemutants. Consistent with this idea, the level of Mg-ProtoIX in the ppi2-2mutant was much lower than thatin wild-type plants (Fig. 4C). Therefore, the down-regulation of photosynthesis-related nuclear genes inppi2 mutants is unlikely to be caused by hyperaccu-mulation of Mg-ProtoIX in the mutant.

Coordination of Plastid Protein Import and Nuclear GeneExpression Requires Novel Components Other

Than ABI4

Given our results indicating that GUN4 and thelevel of Mg-ProtoIX may not be involved in the down-regulation of photosynthesis-related nuclear genes inppi2, we next focused on the GUN1 protein, a penta-tricopeptide repeat protein in plastids. GUN1 has beenproposed to act as an intermediate between GUN4 andABI4 and is positioned downstreamofwell-characterizedplastid signals such as redox state and plastid gene

expression signals (Koussevitzky et al., 2007). Thisraises the possibility that in the import mutants, someplastid signal other than Mg-ProtoIX could act torepress nuclear gene expression via GUN1. Indeed,the ppi2-2 mutant has normal levels of GUN1 expres-sion (Fig. 4A); therefore, the mutant would not beexpected to be impaired in this step of plastid signal-ing. We next examined the effect of a gun1 mutationon nuclear gene expression in ppi2. We attemptedto generate a double mutant of gun1-101 and ppi2-2(gun1-101 ppi2-2). Interestingly, when the progeny ofa gun1-101 ppi2-2+/2 (gun1/gun1;atTOC159/attoc159)line were examined, we were unable to recover albino(gun1-101 ppi2-2) plants. Instead, we found that ap-

Figure 4. Expression of GUN genes and levels of Mg-ProtoIX in theppi2-2 mutant. A, Expression of GUN1 and GUN4 in ppi2-2. TotalRNAwas isolated from 4-week-old seedlings and analyzed by real-timePCR. The expression level in the wild type was set to 1. Each bar,plotted on a log scale, represents the mean of at least three independentsamples. Error bars represent 1 SD. The asterisk above the bar indicates asignificant difference (P , 0.05) between the wild type and therespective mutant, as determined by Student’s t test. B, GUN4 proteinaccumulation in 4-week-old Col-0 and ppi2-2 plants. C, Steady-statelevels of Mg-ProtoIX in 4-week-old Col-0 and ppi2-2 plants. FW, Freshweight.

Kakizaki et al.




proximately 25% of the seeds in gun1-101 ppi2-2+/2siliques were aberrantly shaped (Supplemental Fig.S4). These data suggest that the gun1-101 ppi2-2 doublemutant has defects in embryogenesis and is not viable.When developing embryos were examined by lightmicroscopy, every embryo in gun1-101 ppi2-2+/2 si-liques appeared to be normal until the late heart stage(Fig. 5, A and B). Subsequently, about 25% of theembryos started to twist or bend backwards at thewalking stick embryo stage (Fig. 5, C and D). Thus,these phenotypes are likely to be associated with thelethality of the gun1-101 ppi2-2 double mutant.To further examine the role of GUN1 in the down-

regulation of nuclear gene expression in ppi2, werecovered gun1-101 ppi2-2 embryos from ovules con-taining late heart stage embryos, cultured them on abasal medium, and obtained seedlings of gun1-101ppi2-2 (Fig. 5F). The embryo-rescued gun1-101 ppi2-2double mutant seedlings appeared to be more sicklythan embryo-rescued ppi2-2 single mutant seedlings(Fig. 5, compare E and F). To characterize the geneexpression profile of the double mutant, we measuredthe mRNA and protein accumulation of nuclear-encoded plastid proteins. Although the accumulation

of photosynthesis-related proteins in the double mu-tant was not increased compared with the ppi2-2mutant (Fig. 5H), derepression of two photosynthesis-related genes (LHCB1 and GUN4) in the gun1-101 ppi2-2mutant was observed (Fig. 5G). These data suggest thatGUN1 may play a role in coordinating nuclear geneexpression and plastid protein import.

To further clarify whether the down-regulation ofnuclear gene expression in ppi2mutants is mediated byGUN1-ABI4 or some other, novel pathway, we exam-ined the role of ABI4 in the regulation of gene expres-sion in ppi2. It has been proposed that ABI4 acts as anegative regulator of photosynthesis-related nucleargene expression and that the competitive binding ofABI4 to theG-box in response to plastid signals inhibitsthe induction of LHCB expression (Koussevitzky et al.,2007). We examined the accumulation of ABI4 mRNAin3-d-old seedlings and rosette leaves by real-timePCRanalysis (Fig. 6A). Interestingly, the accumulation ofABI4 transcripts was much higher in ppi2-2 seedlingscomparedwith thewild type. Thismay be explained bythe retarded growth of ppi2-2 plants, in that 3-d-oldmutant seedlings are less mature than wild-type onesand ABI4 expression normally declines with age. Con-

Figure 5. Analysis of the gun1-101 ppi2-2 double mutant obtained by embryo rescue. A to D, Embryogenesis of the gun1-101ppi2-2 double mutant viewed with a light microscope with a differential interference contrast unit. A, Early heart stage. B, Lateheart stage. C, Walking stick stage. D, Mature embryo stage. E and F, Visible phenotypes of ppi2-2 (E) and gun1-101 ppi2-2 (F)plants obtained by embryo rescue. Seedlings were cultured for 2 weeks. Bars = 1 mm. G, Quantitative analysis of geneexpression in embryo-rescued (ER) ppi2-2 and gun1-101 ppi2-2 seedlings (2 weeks old). The mRNA levels were analyzed byreal-time PCR and normalized to the levels of ACTIN2. The expression level in wild-type seedlings (1 week old) was set to 1. Eachbar, plotted here on a log scale, represents the mean of at least three independent samples. Error bars represent 1 SD. Asterisksabove the bars indicate significant differences (P, 0.05) between ppi2-2 and gun1-101 ppi2-2, as determined by Student’s t test.H, Accumulation of plastid proteins in embryo-rescued seedlings (2 weeks old). Total protein extracts were resolved by SDS-PAGE and probed with antisera against the proteins indicated to the left.





sistent with the previous observation that the expres-sion of ABI4 is restricted to developing embryos andseedlings (Finkelstein et al., 1998; Soderman et al.,2000), we could not detect ABI4 transcripts in eitherwild-type or ppi2-2 rosette leaves (Fig. 6A).

Because ABI4 is postulated to act as a negativeregulator in plastid signaling pathways, the hyper-accumulation of ABI4 in ppi2 led us to speculate thatthis may account for the mutant’s decreased expres-sion of photosynthesis-related nuclear genes. To assessthe role of ABI4 in the down-regulation of nucleargene expression in ppi2, we generated abi4 ppi2 doublemutants. We employed two different abi4 alleles (Fig.6B). The abi4-1mutant has a 1-bp deletion, leading to aframeshift, and has been confirmed to be a loss-of-function mutant (Finkelstein et al., 1998). We alsoidentified another mutant, a T-DNA insertional mu-tant referred to as abi4-2 (Supplemental Fig. S5), fromthe Salk Institute Genomic Analysis Laboratory col-lection (Alonso et al., 2003). ABI4 transcripts werebarely detectable in both mutants (Fig. 6C), confirming

that they both appear to be loss-of-function mutants.When ppi2-2 was crossed with abi4-1, the resultingabi4-1 ppi2-2 double mutant seedlings showed expres-sion of photosynthesis-related nuclear genes (such asLHCB1, SSU1A, and GUN4) that was at least as low asthat of ppi2-2 single mutant seedlings (Fig. 6D). In thecase of LHCB1 and GUN4, the repression was moreextreme in abi4-1 ppi2-2 than in ppi2-2. Similar resultswere obtained from another double mutant, abi4-2ppi2-2 (Fig. 6E).

Taken together, these data suggest that the down-regulation of nuclear gene expression in ppi2 is mainlymediated by a novel pathway that is independent ofABI4. Instead, this pathway is likely to involve GUN1and transcription factors other than ABI4.

AtGLK1 May Coordinate Plastid Protein Import andNuclear Gene Expression

Recently, it has been suggested that the binding ofABI4 to G-boxes prevents the binding of G-box bind-

Figure 6. The role of ABI4 in the regulation of nuclear gene expression in ppi2mutants. A, Expression analysis of ABI4 in Col-0and ppi2-2. Total RNAwas extracted from 3-d-old seedlings or rosette leaves. Transcript levels were analyzed by real-time PCRand normalized to the transcript levels of ACTIN2. The expression level in 3-d-old Col-0 seedlings was set to 1. Each bar, plottedon a log scale, represents the mean of at least three independent samples. Error bars represent 1 SD. B, Diagram of the substitution(CS8104) and T-DNA insertion (SALK_080095) mutations in the ABI4 gene. LB, Left border. C, Expression analysis of ABI4transcripts in two abi4mutants by RT-PCR. Total RNAwas purified from 3-d-old seedlings. RTwas carried out in the presence (+)or absence (2) of reverse transcriptase. ACTIN2 was amplified as a positive control. D and E, Expression profiles of nuclear-encoded plastid proteins in abi4 ppi2 double mutants. Total RNA was isolated from 3-d-old seedlings. Transcript levels wereanalyzed by real-time PCR and normalized to those of ACTIN2. The expression level in Col-0 was set to 1. Each bar is plotted ona log scale and represents the mean of at least three independent samples. Error bars represent 1 SD.

Kakizaki et al.




ing factors (GBFs), leading to the down-regulationof nuclear gene expression (Koussevitzky et al.,2007). In this model, the activities of positive (GBFs)and negative (ABI4) factors are key regulators ofphotosynthesis-related nuclear gene expression. Ourresults (presented above) suggested that there may beother positive and negative regulators in addition toABI4 and GBFs. To identify such potential regulators,we surveyed a data set obtained by SuperSAGEfor transcription factors that are significantly down-regulated (less than 25% of wild-type level) in ppi2-1.We reasoned that such transcription factors might actas positive regulators of photosynthesis-related nu-clear gene expression in response to plastid signals inthe wild type. As shown in Table I, we identified 16transcription factor genes that are suppressed in ppi2-1(P , 0.05, Fisher’s exact test). Intriguingly, AtGLK1, atranscriptional activator that interacts with GBF1 andGBF3 (Tamai et al., 2002), was identified among thesetranscription factors. The GLK gene family is alsoknown to affect chloroplast biogenesis (Fitter et al.,2002). To examine whether the expression of AtGLK1is associated with the down-regulation of nuclear geneexpression in mutants defective in plastid proteinimport, we examined the expression level of AtGLK1in various mutants by real-time PCR analysis (Fig. 7, Aand B). The ppi2, attoc132 attoc120+/2, and attic20-Imutants all showed decreased AtGLK1 expression.Because Arabidopsis has a functionally redundantgene, AtGLK2 (Fitter et al., 2002), we examined theexpression of AtGLK2 in ppi2 and confirmed that theexpression of AtGLK2 was also impaired (Fig. 7C). Incontrast, the expression of G-box-binding factors, suchas GBF1 and HY5, was unaffected or up-regulated inthe mutant (Fig. 7C).We next investigated the link between AtGLK1 and

known plastid signaling pathways. A recent paper

reported that GLK genes are responsive to norflurazonand lincomycin treatments (Waters et al., 2009). Whenwild-type plants were treated with norflurazon,AtGLK1 expression was significantly down-regulated(Fig. 7B). In contrast, the level of AtGLK1 in the gun1-101 mutant treated with norflurazon was not de-creased to the same level as in norflurazon-treatedwild-type plants (Fig. 7D). A similar result was ob-tained for the gun1-101 ppi2-2 double mutant. Thegun1-101 ppi2-2 double mutant exhibited a muchhigher AtGLK1 expression level compared with theppi2-2 single mutant (Fig. 5G). Thus, GUN1 may actas a negative regulator of AtGLK1 expression whenplastids are dysfunctional, leading to subsequentdown-regulation of photosynthesis-related genes inthe nucleus.

To further confirm that AtGLK1 is indeed a positiveregulator in the plastid signaling pathway that coor-dinates plastid protein import and photosynthesis-related nuclear gene expression,we generated transgenicplants overexpressing AtGLK1 in a ppi2-2 background.Interestingly, these plants had pale green cotyledonsand mature leaves (Fig. 8A), unlike the albino pheno-type typical of ppi2 mutants. Likewise, the chlorophyllcontent was partially restored in these transgenicplants (Fig. 8B). To compare the effects of AtGLK1overexpression with those of the abi4 mutation (Fig. 6,D and E), we isolated mRNA from 7-d-old seedlings ofthe overexpression lines. As expected, the level ofAtGLK1 in these seedlings was more than 10-foldhigher than in the wild type (Fig. 8C). In addition,the expression of some photosynthesis-related genesin the AtGLK1 overexpression lines was partially re-stored from the ppi2-2 phenotype (Fig. 8C). Notably, allof the LHCB genes examined were significantly up-regulated in the overexpression line. Although thefunction of AtGLK1 is still subject to debate (Savitch

Table I. Transcription factors that were suppressed in ppi2-1

Tag Sequence (5# to 3#)aTag Count Arabidopsis Genome

Initiative CodeProtein

Wild Type ppi2-1

AAAGGAAGGTTCGTCAAGAGAT 35 2 AT1G68520.1 COL6 (CONSTANS-LIKE6)AACAGTATGACATCAACGGTGA 26 6 AT5G44260.1 Zinc finger (CCCH-type) family proteinGAGATATTGGTTTCTTCTATTA 15 6 AT2G36990.1 SIG6 (RNA POLYMERASE SIGMA-70 FACTOR)GTGGTATCTTTAGTGGGTTCGG 13 3 AT5G24930.1 COL4 (CONSTANS-LIKE4)TTATATCTCACTTTACAAATTT 9 1 AT1G25440.1 COL16 (CONSTANS-LIKE16)CGATGAGAAGACTATGATAAGT 8 0 AT1G02340.1 HFR1 (LONG HYPOCOTYL IN FAR-RED)GTTGACATCAATCAATCATCAA 8 3 AT2G18300.2 Basic helix-loop-helix (bHLH) family proteinTGAATCGGCAGCAGAACTCGAG 8 1 AT4G27410.3 RD26 (RESPONSIVE TO DESICCATION26)TGACTAATAAGGAGACCCTTTT 7 0 AT2G20570.1 GLK1 (GOLDEN2-LIKE1)AATGATGATGAGGTTACTTTGA 7 0 AT2G47890.1 COL13 (CONSTANS-LIKE13)ATTAAGTGGCTTGAAATGTAAA 6 2 AT1G19350.1 BES1 (BRI1-EMS-SUPPRESSOR1)TTTATGCTTCAGCAGCACACGC 6 0 AT5G57660.1 COL5 (CONSTANS-LIKE5)TAACTGTGAATCTGAAGATTCA 5 0 AT1G68190.1 Zinc finger (B-box-type) family proteinGTCTGATACAGTTTGATCCAGT 5 0 AT2G02080.1 ATIDD4 (ARABIDOPSIS THALIANA INDETERMINATE

DOMAIN4)CTAAGTATCATTGAAGAGGAGC 5 1 AT5G46690.1 BHLH071 (BETA HLH PROTEIN71)CATACTGGAGTCAAAGTCAATG 5 1 AT5G60850.1 OBP4 (OBF-BINDING PROTEIN4)

aTag represented as a 22-bp sequence excluding the NlaIII site (CATG).





et al., 2007; Waters et al., 2009), our data seem tosupport the proposal that AtGLK1 coordinates theexpression of photosynthetic genes. In contrast to thegun1-101 ppi2-2 double mutant, the increased photo-synthetic gene expression was correlated with therestoration of protein accumulation in the transgenicline (Fig. 8D). The accumulation of translocon com-ponents (Toc and Tic proteins) was not affected, indi-cating that the recovery of photosynthesis-relatedproteins was not due to an unexpected up-regulationof protein import capacity.

Taken together, these data suggest that AtGLK1mayact as a positive regulator in a plastid-to-nucleussignaling pathway that coordinates plastid proteinimport and nuclear gene expression in response tothe functional state of plastids. Furthermore, GUN1down-regulates the expression of AtGLK1 when plas-tids are dysfunctional, leading to the suppression ofnuclear-encoded photosynthesis-related genes.

DISCUSSION

Plastid protein import and the expression of nuclear-encoded plastid proteins are both integral parts ofplastid biogenesis. However, the mechanism by whichthese two processes are coordinated remains to becharacterized. In this study, we examined the linkbetween plastid protein import and nuclear geneexpression. As summarized by the model in Figure

9, we found two important links between the twoprocesses. First, we provided empirical evidence thatplastid dysfunction caused by defects in plastid pro-tein import generates a plastid signal. Analysis ofmutants lacking plastid protein import receptors re-vealed that inhibition of plastid protein import resultsin down-regulation of nuclear gene expression (Figs.1 and 2). We also observed up-regulation of atTIC110expression in the ppi2-1mutant (Fig. 1D), and this maybe an attempt by the plant to compensate for proteinimport defects. Second, we showed that the plastidsignal generated by import defects acts through sup-pression of GLK1 expression but not activation of ABI4(Fig. 9). The abi4 ppi2 double mutant did not exhibitrestoration of LHCB expression compared with theppi2 single mutant. In contrast, the gun1 ppi2 mutantshowed recovery of both AtGLK1 and LHCB expres-sion. Thus, suppression of GLK1 expression is medi-ated by the activity of GUN1 (Fig. 9). Overall, thesedata suggest that plastid dysfunction caused specifi-cally by defects in plastid protein import may act as asignal to coordinate plastid protein import and theexpression of nuclear-encoded photosynthetic pro-teins.

The protein import receptors of plastids are com-posed of two distinct families of GTP-binding pro-teins. The function of these proteins has beenelucidated mostly by genetic studies, in which it hasbeen assumed that the substrate proteins of eachreceptor are likely to be specifically decreased in

Figure 7. Gene expression of AtGLK1 and othertranscription factors in the ppi2 mutants. A,AtGLK1 mRNA accumulation in the two ppi2mutants. B, AtGLK1 mRNA accumulation in im-port machinery mutants and plants treated withnorflurazon (NF). C, Accumulation of AtGLK2and other transcription factors in the ppi2-2 mu-tant. D, Accumulation of LHCB1 and AtGLK1mRNA in the wild type (Col-0) and gun1-101treated with 1 mM norflurazon. After growth undernormal conditions for 3 d, plants were treatedwith norflurazon or dimethyl sulfoxide (DMSO)under continuous light (approximately 125 mmolm22 s21) for 5 d, and the mRNA levels wereexamined using real-time PCR. The expressionlevel in the corresponding wild type was set to 1.Each bar is plotted on a log scale and representsthe mean of at least three independent samples.Error bars represent 1 SD. Asterisks above the barsindicate significant differences (P , 0.05) be-tween the wild type and the respective mutant, asdetermined by Student’s t test.

Kakizaki et al.




mutants for that receptor. This is apparent in the caseof the ppi2 mutants, which lack atToc159: atToc159appears to be a receptor for photosynthesis-relatedproteins, and ppi2-1 fails to accumulate them (Baueret al., 2000; Smith et al., 2004). Likewise, proteomeanalysis of an attoc132 single mutant coincides wellwith the conclusion from biochemical studies thatatToc132 is a receptor for nonphotosynthetic proteins(Ivanova et al., 2004; Kubis et al., 2004). However,our results underscore the complexity of the regula-tion of plastid biogenesis by plastid protein import.

As shown in Figure 2C, significant reduction ofatToc132/atToc120 receptor abundance in the attoc132attoc120+/2 mutant results in an unexpected decreasein photosynthesis-related proteins, apparently due torepression of photosynthesis-related nuclear genes.Although the atToc132/atToc120 receptor preferen-tially binds nonphotosynthetic proteins, it also bindsphotosynthesis-related proteins, albeit more weakly(Ivanova et al., 2004). Feedback regulation of photo-synthesis-related nuclear genes in response to the lackof atToc132/atToc120 may reduce the competition of

Figure 8. Characterization of the ppi2-2 mutantoverexpressing AtGLK1. A, Representative phe-notype of the ppi2-2mutant overexpressing a full-length AtGLK1 cDNA (35S::AtGLK1). Left, Emptyvector control; right, 35S::AtGLK1. Bars = 1 mm.B, Total chlorophyll content in 1-week-old seed-lings. Error bars represent SD (n = 3). FW, Freshweight. C, Expression profiles of genes encodingplastid proteins in the ppi2-2mutant overexpress-ing AtGLK1. Transgenic plants were grown onplates containing hygromycin for 1 week. Tran-script levels were analyzed by real-time PCR andnormalized to the levels of ACTIN2. Each bar isplotted on a log scale and represents the mean ofat least three independent samples. Error barsrepresent 1 SD. Asterisks above the bars indicatesignificant differences (P , 0.05) between ppi2-2carrying empty vector (black bars) and ppi2-2carrying 35S::AtGLK1 lines (dark and light graybars), as determined by Student’s t test. D, Accu-mulation of plastid proteins in the ppi2-2 mutantoverexpressing AtGLK1 (1 week old). Total pro-tein extracts were resolved by SDS-PAGE andprobed with antisera against the proteins indi-cated to the left.





photosynthesis-related proteins for these bindingsites, allowing plastids to preferentially incorporateessential nonphotosynthetic proteins. Thus, the regu-lation of nuclear gene expression based on the status ofplastid protein import is likely to be a strategy forplastids to maintain essential functions under stress.

Down-regulation of nuclear gene expression iscaused by specific signals from plastids. However,the primary signaling molecules that are generated byprotein import defects remain to be characterized. Ithas been proposed that the accumulation of Mg-ProtoIX may act as a signaling molecule derivedfrom plastids when plants are treated with norflur-azon. However, we could not detect any accumulationof Mg-ProtoIX in the ppi2-2mutant compared with thewild type (Fig. 4C). Thus, it appears that a signalingmolecule other than Mg-ProtoIX is involved in thecoordination of plastid protein import and nucleargene expression. It should be noted that recent papersshowed, consistent with our results, that the level ofMg-ProtoIX was not correlated with the level of LHCBexpression in norflurazon-treated plants (Mochizukiet al., 2008; Moulin et al., 2008).

Because ppi2 exhibits severe defects in plastid bio-genesis, signaling molecules generated as a result ofplastid dysfunction might act as primary signals in themutant. We also cannot exclude the possibility thatprotein import defects are communicated by otherknown plastid signals, such as redox signals (reactiveoxygen species). A unique signal that may be spe-cifically generated by protein import defects is theaccumulation of precursor proteins in the cytosol.Although we have not measured the levels of unpro-cessed precursors in ppi2 and other mutants, othershave observed the accumulation of such precursors inother protein import mutants, such as tic21 (Teng et al.,2006). If unprocessed precursors act as a signals gen-erated by protein import defects, we speculate thatthey may act through additional signaling pathwaysthat do not rely on the functional state of plastids.

According to existing literature and our results here,plastid signaling pathways activated by redox, nor-flurazon, lincomycin, and defects in protein importconverge at GUN1 in chloroplasts (Fig. 9). Down-stream of GUN1, however, the signaling pathway forprotein import defects utilizes GLK1 but not ABI4 (Fig.9). It is not clear whether the redox, norflurazon, andlincomycin pathways are also mediated by GLK1. Arecent paper reported that the expression of AtGLKgenes is sensitive to norflurazon and lincomycin(Waters et al., 2009). Consistent with this observation,we confirmed that the expression of AtGLK1 is re-pressed by norflurazon (Fig. 7B). Thus, we favor themodel that the norflurazon signaling pathway utilizesboth ABI4 and GLK1. The next question is how GUN1discriminates between different upstream plastid sig-nals and is able to send the signal generated by defectsin protein import specifically to the GLK1 pathway.Wespeculate that additional components act togetherwith GUN1 to specify downstream pathways. Forexample, distinct factors may associate with GUN1 inresponse to different signals. It is also possible thatGUN1 is modified differently depending on the typeof signal, allowing GUN1 to associate with differentdownstream components.

Although AtGLK1 appears to be important for theinduction of photosynthesis-related nuclear genes, theexpression of those genes in ppi2-2 lines overexpress-ing AtGLK1 was not fully restored to the wild-typelevel (Fig. 8C). This suggests that additional factors areinvolved in the plastid-to-nucleus signaling pathway.One possibility is that a negative regulator, like ABI4,which is more abundant in ppi2 (Fig. 6A), keeps thegenes partially repressed even in the presence of anexcess of a positive factor (AtGLK1). Recently, HY5 hasbeen identified as a negative regulator of photosynthesis-related nuclear genes when plastid biogenesis is ar-rested by lincomycin treatment (Ruckle et al., 2007). Inthis study, we observed a remarkable induction of HY5in the ppi2-1 background (Fig. 7C). It is notable thatHY5acts as a negative regulator in the plastid signalingpathway, whereas it functions as a positive regulator inphotomorphogenetic pathways (Oyama et al., 1997).Thus, it is not surprising if the induction of HY5 isassociated with the decrease of photosynthesis-relatednuclear gene expression in the ppi2 mutant. Accordingto our SuperSAGE analysis, there are many othertranscription factors up- or down-regulated in ppi2.The combinatorial interplay between these transcrip-tion factors may play critical roles in fine-tuningphotosynthesis-related nuclear gene expression inresponse to the functional state of the plastids.

In summary, we demonstrated that plastid proteinimport and nuclear gene expression are tightly coor-dinated and that this coordination may be in partmediated by GUN1 and the transcription factor AtGLK.In contrast to other transcription factors that havebeen identified so far in retrograde signaling, AtGLKmay act as a positive regulator of photosynthesis-related nuclear gene expression. These results provide

Figure 9. Model for the regulation of LHCB expression by plastidsignals. Multiple plastid signals, including that generated by defects inplastid protein import, converge at GUN1. Subsequently, the plastidsignal generated by import defects acts through the suppression ofGLK1 expression but not the activation of ABI4. Although plastidsignals generated by norflurazon and lincomycin have been shown toact through the GUN1-ABI4 pathway, they may also act through theGUN1-GLK1 pathway.

Kakizaki et al.




evidence that plastid signaling pathways constitutea complex signaling network composed of positiveand negative regulators. Further identification of sig-naling components should provide insight into thesignaling network that coordinates nuclear and plastidfunctions.

MATERIALS AND METHODS

Plant Material and Growth Conditions

All experiments were performed on Arabidopsis (Arabidopsis thaliana)

accessions Col-0 and Ws. The T-DNA insertion lines (Supplemental Table S3)

were obtained from GABI-Kat (Rosso et al., 2003), the collections of the Salk

Institute Genome Analysis Laboratory (Alonso et al., 2003), and the Syngenta

Arabidopsis Insertion Library (Sessions et al., 2002). Plants were grown on soil

(expanded vermiculite) or on 0.5% agar medium containing 1% Suc and 0.53Murashige and Skoog salts at pH 5.8. To synchronize germination, all seeds

were kept at 4�C for 2 d after sowing. Plants were grown under continuous

white light (80 mmol m22 s21, unless specified) at 22�C and 50% relative

humidity in a growth chamber (LHP-350S; NK system). The plants grown on

soil were watered with Hyponex (Hyponex Japan) diluted 1,000-fold in water.

For norflurazon treatment in Figure 7, plants were grown under normal

conditions for 3 d and then treated with 1 mM norflurazon or dimethyl

sulfoxide under continuous light (approximately 125 mmol m22 s21) for 5 d.

RNA Isolation and Real-Time PCR Analysis

Total RNAwas extracted from leaf tissues of wild-type and mutant plants

using an RNeasy Plant Mini Kit (Qiagen) according to the manufacturer’s

instructions. Briefly, cDNA was synthesized using the PrimeScript reverse

transcription (RT) reagent kit (TaKaRa) with random hexamer and oligo(dT)

primers. Real-time PCR was performed on a Thermal Cycler Dice Real-Time

System (TaKaRa) using SYBR Premix Ex Taq (TaKaRa) as described previously

(Nakayama et al., 2007; Okawa et al., 2008). Primers used for real-time PCR are

listed in Supplemental Table S4. The transcript level of each gene was

normalized to that of 18S ribosomal RNA or ACTIN2. Negative controls for

PCR were performed with the RNA templates prior to cDNA synthesis to

check for genomic DNA contamination.

RT-PCR for ABI4

Homozygous abi4-1 and abi4-2 plants were identified by genomic PCR

using specific primers (Supplemental Tables S3 and S5). Total RNA was

extracted from 3-d-old seedlings of Col-0, abi4-1, and abi4-2. From 500 ng

of RNA for each sample, single-stranded cDNA was synthesized using an

oligo(dT) primer and PrimeScript reverse transcriptase (TaKaRa). RT-PCR anal-

ysis was performed with the same gene-specific primers for ABI4 and ACTIN2

that were used for real-time PCR (Supplemental Table S4).

Antibodies and Immunoblotting

Rabbit polyclonal antibody against GUN4 was produced using a NusA-

GUN4 fusion protein as the antigen (Supplemental Fig. S6). Antibodies

against atTic110 (Inaba et al., 2003, 2005), SSU (Inaba et al., 2005), LHCP (Payan

and Cline, 1991), ACC-CTa (Kozaki et al., 2000), and LSU, OE23, POR, E1a,

Hsp93, and IEP37 were described elsewhere. Proteins were detected using

chemiluminescence reagents with a luminoimage analyzer (AE-6972C;

ATTO). For some experiments, the signal was quantified using image acqui-

sition software (CS Analyzer; ATTO).

SuperSAGE Library Construction and Tag Sequencing

To obtain the SuperSAGE libraries, we followed the original SuperSAGE

protocol described by Matsumura and colleagues (Matsumura et al., 2003;

Terauchi et al., 2008) with somemodification. Total RNAwas isolated from the

third to sixth leaves of Ws and the ppi2-1 mutant using RNAiso reagent

(TaKaRa). Biotinylated cDNAs from each sample were bound to streptavidin-

coated magnetic beads and digested with NlaIII. The streptavidin-bound

cDNA was washed and divided into two portions in separate tubes. Linkers

containing a binding site for EcoP15I (a type III restriction enzyme) but

different sites for PCR primers were ligated to the NlaIII cleavage site at the 5#ends of the bead-bound cDNA fragments in each pool. Both pools of cDNAs

were digested with EcoP15I to release SuperSAGE tags from the beads, then

the released tag pools were combined and linker-flanked ditags were formed

by blunt-end ligation. Following amplification of the ditags by linker-specific

PCR, the SuperSAGE library was directly sequenced with a large-scale

pyrosequencing method (454 Life Sciences). The sequences of the linkers

and PCR primers are listed in Supplemental Table S6.

SuperSAGE Data Analysis and Tag-to-Gene Assignment

To identify the genes from which tags were obtained, each 26-bp tag was

annotated against The Arabidopsis Information Resource 7 database using the

BioEdit program (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). The

data used in this paper have been deposited in the National Center for

Biotechnology Information Gene Expression Omnibus (http://www.ncbi.

nlm.nih.gov/geo) with accession numbers GSM308557 and GSM308558.

Quantitation of Mg-ProtoIX

Mg-ProtoIX separation and quantitation were done as described previ-

ously (Strand et al., 2003) with slight modifications. Leaf material was

homogenized in 0.67 mL of 0.1 M potassium phosphate buffer (pH 7.8) with

a Bead Smash 12 (Wakenyaku) homogenizer and centrifuged at 10,000g for 3

min at 4�C. The residue was resuspended in 0.67 mL of acetone:0.1 M NH4OH

(8:2, v/v) and then homogenized and centrifuged. The residue was resus-

pended again in methanol:0.1 M NH4OH (9:1, v/v), and the same procedures

described above were repeated. The collected supernatants were mixed and

centrifuged prior to HPLC analysis on a CapcelPak-C18 column (Shiseido).

The extracts were eluted with a linear gradient of solvent A (30% [w/v]

acetonitrile and 2.5 mM tetrabutyl ammonium phosphate, pH 7.0) to solvent B

(90% [w/v] acetonitrile). Column eluent was monitored by fluorescence

detection (420-nm excitation, 595-nm emission), and Mg-ProtoIX was identi-

fied and quantified using chemically synthesized standards (Frontier Scien-

tific). Fifty microliters each of standard solutions of Mg-ProtoIX from 1028 to

10212M were injected into the HPLC system; the 10212

M solution was the limit

for preparing the calibration line in the system used.

Embryo Rescue

For embryo-rescue experiments, immature ovules (containing late heart

stage embryos) were excised from developing siliques and cultivated on a

basal medium (13 Murashige and Skoog salts, 3% Glc, 0.55 mM myoinositol,

5 mM thiamine hydrochloride, and 0.8% agar) under continuous light as

described (Baus et al., 1986).

Microscopy

To observe embryo development, ovules were dissected from siliques of

appropriate ages and cleared in chloral hydrate:water:glycerol (8:2:1) over-

night. Differential interference contrast images were obtained using an Axio

Imager A1 microscope (Zeiss).

Construction of Vector for OverExpression of AtGLK1and Plant Transformation

A full-length cDNA of AtGLK1 was amplified by PCR using primers that

specifically anneal with AtGLK1 (GLK1_F_SpeI, 5#-ACTAGTACAAGTGAA-

GATTGATCGATG-3#; GLK1_R_NheI, 5#-GCTAGCGCCTACGGATCTTGTG-

CGTTTCAG-3#) and subcloned into the SpeI andNheI sites of the binary vector

pCAMBIA1303. The GLK1/pCAMBIA1303 plasmid was transformed into

Agrobacterium tumefaciens (GV3101) by electroporation and was introduced

into ppi2-2+/2 plants by the floral-dip protocol (Clough and Bent, 1998).

Chlorophyll Extraction and Quantification

Seven-day-old seedlings (20–40 mg fresh weight) were ground in liquid

nitrogen with a mortar and pestle. The resulting powder was suspended in





80% (v/v) acetone and centrifuged at 3,000 rpm at room temperature.

Absorbance of the supernatants was measured at 646.6 and 663.6 nm, and

total chlorophyll content was calculated using the following formula: mg total

chlorophyll mL21 = (17.76 3 A646.6) + (7.34 3 A663.6).

Sequence data from this article can be found in the GenBank/EMBL data

libraries under accession numbers GSM308557 and GSM308558.

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Figure S1. Immunoblot analysis of plastid proteins in ppi2

and wild-type plants.

Supplemental Figure S2.Analysis of the T-DNA insertion allele (ppi2-2) of

atTOC159.

Supplemental Figure S3. Overview of SuperSAGE analysis.

Supplemental Figure S4. Characterization of the gun1-101 ppi2-2 double

mutant.

Supplemental Figure S5. Schematic diagram of the T-DNA insertion site

in ABI4.

Supplemental Figure S6. Production of antibody against recombinant

GUN4 protein.

Supplemental Table S1. Annotation of down-regulated tags in ppi2-1.

Supplemental Table S2. Annotation of up-regulated tags in ppi2-1.

Supplemental Table S3. Primers used for genotyping of T-DNA insertion

lines.

Supplemental Table S4. Gene-specific primers used in real-time PCR.

Supplemental Table S5. CAPS marker for genotyping of abi4-1.

Supplemental Table S6. Sequence of the primers and linkers used for

SuperSAGE library construction.

ACKNOWLEDGMENTS

Part of this work was initiated in Prof. Danny Schnell’s laboratory

(University of Massachusetts, Amherst), and we thank Prof. Schnell for his

kind support and discussion. Antibodies were kindly provided by Drs. Felix

Kessler (atToc75), Alice Barkan (OE23), Ken Cline (LHCP), Doug Randall

(E1a), and Ken Keegstra and John Froehlich (Hsp93). We also thank Ms.

Kumiko Okawa, Hiromi Morii, and Yuko Suzuki for their valuable discus-

sions and technical assistance.

Received August 10, 2009; accepted August 30, 2009; published September 2,

2009.

LITERATURE CITED

Alonso JM, Stepanova AN, Leisse TJ, Kim CJ, Chen H, Shinn P, Stevenson

DK, Zimmerman J, Barajas P, Cheuk R, et al (2003) Genome-wide inser-

tional mutagenesis of Arabidopsis thaliana. Science 301: 653–657

Bauer J, Chen K, Hiltbrunner A, Wehrli E, Eugster M, Schnell D, Kessler

F (2000) The major protein import receptor of plastids is essential for

chloroplast biogenesis. Nature 403: 203–207

Bauer J, Hiltbrunner A, Weibel P, Vidi PA, Alvarez-Huerta M, Smith MD,

Schnell DJ, Kessler F (2002) Essential role of the G-domain in targeting

of the protein import receptor atToc159 to the chloroplast outer mem-

brane. J Cell Biol 159: 845–854

Baus AD, Franzmann L, Meinke DW (1986) Growth in vitro of arrested

embryos from lethal mutants of Arabidopsis thaliana. Theor Appl Genet

72: 577–586

Beck CF (2005) Signaling pathways from the chloroplast to the nucleus.

Planta 222: 743–756

Clough SJ, Bent AF (1998) Floral dip: a simplified method for Agro-

bacterium-mediated transformation of Arabidopsis thaliana. Plant J 16:

735–743

Dyall SD, Brown MT, Johnson PJ (2004) Ancient invasions: from endo-

symbionts to organelles. Science 304: 253–257

Finkelstein RR, Wang ML, Lynch TJ, Rao S, Goodman HM (1998) The

Arabidopsis abscisic acid response locus ABI4 encodes an APETALA 2

domain protein. Plant Cell 10: 1043–1054

Fitter DW, Martin DJ, Copley MJ, Scotland RW, Langdale JA (2002) GLK

gene pairs regulate chloroplast development in diverse plant species.

Plant J 31: 713–727

Hiltbrunner A, Grunig K, Alvarez-Huerta M, Infanger S, Bauer J, Kessler

F (2004) AtToc90, a new GTP-binding component of the Arabidopsis

chloroplast protein import machinery. Plant Mol Biol 54: 427–440

Inaba T, Alvarez-Huerta M, Li M, Bauer J, Ewers C, Kessler F, Schnell DJ

(2005) Arabidopsis tic110 is essential for the assembly and function of the

protein import machinery of plastids. Plant Cell 17: 1482–1496

Inaba T, Li M, Alvarez-Huerta M, Kessler F, Schnell DJ (2003) atTic110

functions as a scaffold for coordinating the stromal events of protein

import into chloroplasts. J Biol Chem 278: 38617–38627

Inaba T, Schnell DJ (2008) Protein trafficking to plastids: one theme, many

variations. Biochem J 413: 15–28

Ivanova Y, Smith MD, Chen K, Schnell DJ (2004) Members of the Toc159

import receptor family represent distinct pathways for protein targeting

to plastids. Mol Biol Cell 15: 3379–3392

Jarvis P (2008) Targeting of nucleus-encoded proteins to chloroplasts in

plants. New Phytol 179: 257–285

Keegstra K, Cline K (1999) Protein import and routing systems of chloro-

plasts. Plant Cell 11: 557–570

Kessler F, Blobel G, Patel HA, Schnell DJ (1994) Identification of two GTP-

binding proteins in the chloroplast protein import machinery. Science

266: 1035–1039

Kessler F, Schnell DJ (2006) The function and diversity of plastid protein

import pathways: a multilane GTPase highway into plastids. Traffic 7:

248–257

Koussevitzky S, Nott A, Mockler TC, Hong F, Sachetto-Martins G, Surpin

M, Lim J, Mittler R, Chory J (2007) Signals from chloroplasts converge

to regulate nuclear gene expression. Science 316: 715–719

Kozaki A, Kamada K, Nagano Y, Iguchi H, Sasaki Y (2000) Recombinant

carboxyltransferase responsive to redox of pea plastidic acetyl-CoA

carboxylase. J Biol Chem 275: 10702–10708

Kubis S, Patel R, Combe J, Bedard J, Kovacheva S, Lilley K, Biehl A,

Leister D, Rios G, Koncz C, et al (2004) Functional specialization

amongst the Arabidopsis Toc159 family of chloroplast protein import

receptors. Plant Cell 16: 2059–2077

Larkin RM, Alonso JM, Ecker JR, Chory J (2003) GUN4, a regulator of

chlorophyll synthesis and intracellular signaling. Science 299: 902–906

La Rocca N, Rascio N, Oster U, Rudiger W (2001) Amitrole treatment of

etiolated barley seedlings leads to deregulation of tetrapyrrole synthesis

and to reduced expression of Lhc and RbcS genes. Planta 213: 101–108

Martin W, Stoebe B, Goremykin V, Hapsmann S, Hasegawa M, Kowallik

KV (1998) Gene transfer to the nucleus and the evolution of chloro-

plasts. Nature 393: 162–165

Matsumura H, Reich S, Ito A, Saitoh H, Kamoun S, Winter P, Kahl G,

Reuter M, Kruger DH, Terauchi R (2003) Gene expression analysis of

plant host-pathogen interactions by SuperSAGE. Proc Natl Acad Sci

USA 100: 15718–15723

Mochizuki N, Brusslan JA, Larkin R, Nagatani A, Chory J (2001)

Arabidopsis genomes uncoupled 5 (GUN5) mutant reveals the involve-

ment of Mg-chelatase H subunit in plastid-to-nucleus signal transduc-

tion. Proc Natl Acad Sci USA 98: 2053–2058

Mochizuki N, Tanaka R, Tanaka A, Masuda T, Nagatani A (2008) The

steady-state level of Mg-protoporphyrin IX is not a determinant of

plastid-to-nucleus signaling in Arabidopsis. Proc Natl Acad Sci USA

105: 15184–15189

Moulin M, McCormac AC, Terry MJ, Smith AG (2008) Tetrapyrrole

profiling in Arabidopsis seedlings reveals that retrograde plastid nu-

clear signaling is not due to Mg-protoporphyrin IX accumulation. Proc

Natl Acad Sci USA 105: 15178–15183

Mulo P, Pursiheimo S, Hou CX, Tyystjarvi T, Aro EM (2003) Multiple

effects of antibiotics on chloroplast and nuclear gene expression. Funct

Plant Biol 30: 1097–1103

Nakayama K, Okawa K, Kakizaki T, Honma T, Itoh H, Inaba T (2007)

Kakizaki et al.




Arabidopsis Cor15am is a chloroplast stromal protein that has cryopro-

tective activity and forms oligomers. Plant Physiol 144: 513–523

Nott A, Jung HS, Koussevitzky S, Chory J (2006) Plastid-to-nucleus

retrograde signaling. Annu Rev Plant Biol 57: 739–759

Okawa K, Nakayama K, Kakizaki T, Yamashita T, Inaba T (2008)

Identification and characterization of Cor413im proteins as novel com-

ponents of the chloroplast inner envelope. Plant Cell Environ 31:

1470–1483

Oyama T, Shimura Y, Okada K (1997) The Arabidopsis HY5 gene encodes a

bZIP protein that regulates stimulus-induced development of root and

hypocotyl. Genes Dev 11: 2983–2995

Payan LA, Cline K (1991) A stromal protein factor maintains the solubility

and insertion competence of an imported thylakoid membrane protein.

J Cell Biol 112: 603–613

Rosso MG, Li Y, Strizhov N, Reiss B, Dekker K, Weisshaar B (2003) An

Arabidopsis thaliana T-DNA mutagenized population (GABI-Kat) for

flanking sequence tag-based reverse genetics. Plant Mol Biol 53: 247–259

Ruckle ME, Demarco SM, Larkin RM (2007) Plastid signals remodel light

signaling networks and are essential for efficient chloroplast biogenesis

in Arabidopsis. Plant Cell 19: 3944–3960

Savitch LV, Subramaniam R, Allard GC, Singh J (2007) The GLK1

‘regulon’ encodes disease defense related proteins and confers resis-

tance to Fusarium graminearum in Arabidopsis. Biochem Biophys Res

Commun 359: 234–238

Schnell DJ, Kessler F, Blobel G (1994) Isolation of components of the

chloroplast protein import machinery. Science 266: 1007–1012

Sessions A, Burke E, Presting G, Aux G, McElver J, Patton D, Dietrich B,

Ho P, Bacwaden J, Ko C, et al (2002) A high-throughput Arabidopsis

reverse genetics system. Plant Cell 14: 2985–2994

Smith MD, Rounds CM, Wang F, Chen K, Afitlhile M, Schnell DJ (2004)

atToc159 is a selective transit peptide receptor for the import of nucleus-

encoded chloroplast proteins. J Cell Biol 165: 323–334

Soderman EM, Brocard IM, Lynch TJ, Finkelstein RR (2000) Regulation

and function of the Arabidopsis ABA-insensitive4 gene in seed and

abscisic acid response signaling networks. Plant Physiol 124: 1752–1765

Soll J, Schleiff E (2004) Protein import into chloroplasts. Nat Rev Mol Cell

Biol 5: 198–208

Strand A, Asami T, Alonso J, Ecker JR, Chory J (2003) Chloroplast to

nucleus communication triggered by accumulation of Mg-protopor-

phyrinIX. Nature 421: 79–83

Susek RE, Ausubel FM, Chory J (1993) Signal transduction mutants of

Arabidopsis uncouple nuclear CAB and RBCS gene expression from

chloroplast development. Cell 74: 787–799

Tamai H, Iwabuchi M, Meshi T (2002) Arabidopsis GARP transcriptional

activators interact with the Pro-rich activation domain shared by G-box-

binding bZIP factors. Plant Cell Physiol 43: 99–107

Teng YS, Su YS, Chen LJ, Lee YJ, Hwang I, Li HM (2006) Tic21 is an

essential translocon component for protein translocation across the

chloroplast inner envelope membrane. Plant Cell 18: 2247–2257

Terauchi R, Matsumura H, Kruger DH, Kahl G (2008) SuperSAGE: the

most advanced transcriptome technology for functional genomics. In G

Kahl, K Meksem, eds, The Handbook of Plant Functional Genomics.

Wiley-VCH, Weinheim, Germany, pp 37–54

Terzaghi WB, Cashmore AR (1995) Light-regulated transcription. Annu

Rev Plant Physiol Plant Mol Biol 46: 445–474

Vinti G, Hills A, Campbell S, Bowyer JR, Mochizuki N, Chory J, Lopez-

Juez E (2000) Interactions between hy1 and gun mutants of Arabidopsis,

and their implications for plastid/nuclear signalling. Plant J 24: 883–894

Waters MT, Wang P, Korkaric M, Capper RG, Saunders NJ, Langdale JA

(2009) GLK transcription factors coordinate expression of the photo-

synthetic apparatus in Arabidopsis. Plant Cell 21: 1109–1128





coordination of plastid protein import and nuclear …...genes in the nucleus (mulo et al., 2003)....

Documents