LIL3, a Light-Harvesting Complex Protein, LinksTerpenoid and Tetrapyrrole Biosynthesisin Arabidopsis thaliana1[OPEN]

Daniel Hey2, Maxi Rothbart2, Josephine Herbst, Peng Wang, Jakob Müller, Daniel Wittmann,Kirsten Gruhl, and Bernhard Grimm*

Humboldt Universität zu Berlin, Lebenswissenschaftliche Fakultät, Institut für Biologie, AGPflanzenphysiologie, 10115 Berlin, Germany

ORCID IDs: 0000-0002-8749-8352 (D.H.); 0000-0001-8420-5326 (P.W.); 0000-0002-5711-6042 (J.M.); 0000-0002-9730-1074 (B.G.).

The LIL3 protein of Arabidopsis (Arabidopsis thaliana) belongs to the light-harvesting complex (LHC) protein family, which alsoincludes the light-harvesting chlorophyll-binding proteins of photosystems I and II, the early-light-inducible proteins, PsbS involvedin nonphotochemical quenching, and the one-helix proteins and their cyanobacterial homologs designated high-light-inducibleproteins. Each member of this family is characterized by one or two LHC transmembrane domains (referred to as the LHCmotif) to which potential functions such as chlorophyll binding, protein interaction, and integration of interacting partners intothe plastid membranes have been attributed. Initially, LIL3 was shown to interact with geranylgeranyl reductase (CHLP), an enzymeof terpene biosynthesis that supplies the hydrocarbon chain for chlorophyll and tocopherol. Here, we show another function of LIL3for the stability of protochlorophyllide oxidoreductase (POR). Multiple protein-protein interaction analyses suggest the directphysical interaction of LIL3 with POR but not with chlorophyll synthase. Consistently, LIL3-deficient plants exhibit substantialloss of POR as well as CHLP, which is not due to defective transcription of the POR and CHLP genes but to the posttranslationalmodification of their protein products. Interestingly, in vitro biochemical analyses provide novel evidence that LIL3 shows highbinding affinity to protochlorophyllide, the substrate of POR. Taken together, this study suggests a critical role for LIL3 in theorganization of later steps in chlorophyll biosynthesis. We suggest that LIL3 associates with POR and CHLP and thus contributes tothe supply of the two metabolites, chlorophyllide and phytyl pyrophosphate, required for the final step in chlorophyll a synthesis.

The proteins of the light-harvesting complex (LHC)superfamily are characterized by up to four hydrophobicand a-helical membrane-spanning domains, one or twoof which represent LHC motifs because they includepotential chlorophyll (Chl)-binding sites (Klimmek et al.,2006). The most abundant members of the LHC proteinfamily are the light-harvesting Chl-binding proteins(LHCPs) of PSI and PSII, each of which has three trans-membranemotifs (Jansson, 1994;Montané andKloppstech,2000; Rochaix, 2014). The S subunit of PSII (PsbS),which possesses four transmembrane domains (includ-ing two LHC motifs), is involved in nonphotochemicalquenching of PSII (Li et al., 2000; Niyogi et al., 2005). In

addition, so-called light-harvesting-like (LIL) proteinsshare one or two LHC motifs with LHCPs, but they aremore diverse in size and function. At least six differentLIL proteins, with between one and three hydrophobica-helical motifs, have been identified in Arabidopsis(Arabidopsis thaliana; Klimmek et al., 2006). The cyano-bacterial high-light-inducible proteins (HLIPs) and theireukaryotic orthologs, the one-helix proteins (OHPs),each possess one LHC motif-containing transmembranesequence (Dolganov et al., 1995; Jansson et al., 2000;Andersson et al., 2003). The so-called stress-enhancedproteins (SEP1, SEP2, and SEP3/LIL3) contain two trans-membrane domains, one of which includes an LHCmotif (Heddad and Adamska, 2000). As in the LHCPs,three membrane-spanning LHC domains are found inthe early light-inducible proteins (ELIPs; Grimm andKloppstech, 1987; Green et al., 1991). Interestingly, one ofthe two isoforms of ferrochelatase, ferrochelatase II, alsopossesses a C-terminal LHC motif (Sobotka et al., 2011).Removal of this motif does not affect its catalytic activitybut results in dissociation of the ferrochelatase dimer andimpairs Chl synthesis (Sobotka et al., 2011). Thus, theLHCmotifs in these variousmembers of the superfamilyappear to have multiple functions.

Each member of the LHC superfamily is potentiallycapable of binding photosynthetic pigments and, con-sequently, may be affected by light absorption. How-ever, Chl and/or carotenoid binding have not been

1 This work was supported by the Deutsche Forschungsgemein-schaft (grant no. 936-14 to B.G., subproject of the Deutsche For-schungsgemeinschaft Research Unit FOR2092).

2 These authors contributed equally to the article.* Address correspondence to bernhard.grimm@rz.hu-berlin.de.The author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Bernhard Grimm (bernhard.grimm@rz.hu-berlin.de).

D.H., M.R., and B.G. designed the research; D.H., M.R., J.H., P.W.,J.M., K.G., and D.W. performed the experiments; D.H., M.R., J.H.,and P.W. analyzed the data; B.G. wrote the article.

[OPEN] Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.17.00505

Plant Physiology�, June 2017, Vol. 174, pp. 1037–1050, www.plantphysiol.org � 2017 American Society of Plant Biologists. All Rights Reserved. 1037

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

demonstrated for all of the different groups. Althoughthey are thought to be localized to thylakoid mem-branes, they seem to function in diverse processes inchloroplasts. LHCPs bind Chl a and b as well as xan-thophylls and lutein, enhance the light-harvesting ca-pacity and energy transfer toward the reaction center ofthe two photosystems by increasing their antenna size,and provide photoprotection against excessive light(Horton, 2012; Ruban et al., 2012; Pribil et al., 2014;Allahverdiyeva et al., 2015). The ELIPs bind Chl andcarotenoids and are encoded by genes that are tran-siently induced during deetiolation and abiotic stress(Heddad and Adamska, 2000; Montané and Kloppstech,2000). The homologous HLIPs and OHPs have beenassigned multiple functions ranging from roles in re-sponses to temperature stress and nutrient deficiency,to Chl metabolism, in transient pigment storage duringthe assembly of Chl-binding proteins of the antenna orcore complexes, or during the repair of the photosyn-thetic complexes (Hernandez-Prieto et al., 2011; Yaoet al., 2012; Komenda and Sobotka, 2016).

In Arabidopsis, two SEP3 isoforms (here designatedLIL3) are encoded by nuclear genes (here termed LIL3.1and LIL3.2). A molecular function in protein-proteininteraction was assigned to LIL3 when LIL3 doublemutants (lil3.1/lil3.2) were found to accumulate smallamounts of geranylgeranylated Chl instead of wild-type amounts of phytylated Chl. Lack of LIL3 wasshown to destabilize the geranylgeranyl reductase(CHLP) in plastid membranes, which is encoded by thenuclear gene CHLP (At1g74470; Tanaka et al., 2010).These authors concluded that the physical interactionof LIL3 with CHLP tethers the enzyme to the plastidmembranes and prevents its rapid degradation. CHLPcatalyzes the reduction of geranylgeranyl pyrophos-phate before or after esterification with chlorophyllide(Chlide) and, thus, provides the hydrophobic alkylchain found in Chls and in the vitamin E derivativestocopherols and tocotrienols. LIL3 has been detectedin protein fractions of etiochloroplasts of barley(Hordeum vulgare), which also contain protochlorophyll(Reisinger et al., 2008). LIL3 proteins also are presentin fractions of Arabidopsis chloroplasts that containChlide and geranylgeranyl pyrophosphate, whichsupports the findings mentioned above regardingLIL3’s involvement in late steps of terpenoid and Chlbiosynthesis (Mork-Jansson et al., 2015a). Finally, theassociation of LIL3 with Chl has been demonstrateddirectly using a thermophoresis approach (Mork-Jansson et al., 2015b).

The mature LIL3 isoforms can be dissected intothree different structural domains. An extended water-soluble N-terminal globular domain (Fig. 1) is followedby the two transmembrane domains and a hydrophilicC-terminal segment that protrudes into the stroma ofthe plastid. In previous reports (Tanaka et al., 2010;Takahashi et al., 2014; Mork-Jansson et al., 2015a,2015b), LIL3 was proposed to serve as a membraneanchor for other proteins and as a binding site for Chl.Specifically, it was proposed that LIL3 contributes to

the formation of diverse protein complexes involved inChl synthesis or to the assembly of Chl into Chl-bindingproteins during their integration into the thylakoidmembrane and subsequent assembly into photosyn-thetic protein complexes.

In this context, it is interesting that the expression ofthe CHLP protein supplemented with the transmem-brane domain of either LIL3 or ascorbate peroxidase issufficient to anchor and stabilize CHLP in the absenceof LIL3 (Takahashi et al., 2014). The authors concludedthat CHLP function could be restored in lil3 mutants,although the oligomeric CHLP complexes observed inthe presence of LIL3 might not be rescued. Thus, boththe extent to which LIL3 is dispensable and the functionof its LHC membrane domain remain open.

Chl is synthesized via the tetrapyrrole biosynthesis(TBS) pathway, which also provides other macrocyclicend products, such as heme, phytochromobilin, andsiroheme, in plants. These tetrapyrroles associate withvarious proteins and act either as redox cofactors or aschromophores in multiple cellular processes (Tanakaet al., 2010; Tanaka and Tanaka, 2011; Brzezowski et al.,2015). In the Mg branch of TBS, Mg porphyrin is con-verted into a chlorine ring structure by the reductionof protochlorophyllide (Pchlide) to Chlide a. In angio-sperms, this reduction is catalyzed by the light-dependent protochlorophyllide oxidoreductase (POR).In Arabidopsis, three POR isoforms are encoded inPORA, PORB, and PORC genes. Chl a synthesis isterminated by the esterification of Chlide with phytylpyrophosphate or geranylgeranyl pyrophosphate bythe Chl synthase (CHLG; At3g51820). This reactionlinks TBS to the plastid-localized branch of the terpe-noid synthesis pathway and results in the formation ofa rather hydrophobic Chl, which can be attached to thepigment-binding proteins of PSI and PSII. The enzymesof the Mg branch of Chl synthesis and terpenoid syn-thesis also are found to be attached to plastid mem-branes. They are either indirectly associated with thesemembranes or have thylakoid membrane-spanningdomains for integration into their lipophilic surround-ings (Tanaka and Tanaka, 2007; Tanaka et al., 2011). Itremains unclear how the supply of substratemoieties tosuccessive enzymes during the synthesis of Chls andtocopherols is coordinated.

TBS is subject to tight transcriptional and posttrans-lational control, which, among other things, pre-vents the accumulation of photoreactive tetrapyrroleintermediates and directs the appropriate amounts ofmetabolites into the branched TBS pathway. The un-impaired flow of metabolites is thought to be based onprotein-protein interactions during the course of tetra-pyrrole metabolism (Wang and Grimm, 2015). Ulti-mately, the dynamic formation of protein complexes isachieved by successive protein-protein interactions,which lead to assembly, disassembly, and reformation inalternative combinations (Zhang et al., 2015). This con-cept of multienzyme complexes is based on the need forsmooth and balanced transfer of metabolites from oneenzyme to the next without the release of potentially

1038 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

toxic intermediates. The stability of macromolecularprotein assemblies is facilitated through direct physicalinteractions among the proteins. This interplay is oftensupported by nonenzymatic chaperones and auxil-iary factors (Tanaka et al., 2010; Albus et al., 2012;Mork-Jansson et al., 2015a).In this work, we have extended our previous studies

on the posttranslational organization of TBS with theaim of elucidating the function of LIL3 as an auxiliaryfactor in other steps in TBS, and we examined LIL3’sinteractions with other proteins involved in Chl bio-synthesis. We observed that lil3.1/lil3.2 double mutantsnot only contain reduced amounts of CHLP but also ofPOR and other proteins of TBS, implying that LIL3has a broader role in the organization of the Chl

biosynthesis pathway. Based on an analysis of the in-teractions of LIL3, an updatedmodel for the function ofLIL3 in the coordination of terpenoid and Chl synthesisis presented.

RESULTS

LIL3 Deficiency Destabilizes Chl-Synthesizing Enzymes

In 12-d-old lil3.1 and lil3.2 seedlings grown undershort-day conditions (10-h-light/14-h-dark photope-riod, 100 mmol photons m22 s21), the Chl content wasnormal, but the Chl a/b ratio was increased to morethan 5 in lil3.1 (Fig. 2A). Immunoblot analysis revealedthat steady-state levels of several enzymes involved in

Figure 1. Alignment and domain structure of the Arabidopsis LIL3 proteins. Protein sequences of the two LIL3 proteins werealigned using Jalview 2.0 (Waterhouse et al., 2009). Numbers indicate amino acid positions, and protein domains are highlightedby different background shadings. The first transmembrane helix (TMH) carries a conserved LHC motif, which is indicated by ablack line below the respective sequence.

Figure 2. Characterization of lil3.1, lil3.2, andlil3.1/lil3.2 mutants in comparison with thewild type (WT; ecotype Nossen [Nos]). A, TotalChl contents and Chl a/b ratios in lil3 singlemutants. Chl was extracted with alkaline ace-tone and quantified byHPLC. FW, Fresh weight.B, Steady-state levels of proteins/enzymes in-volved in Chl biosynthesis. Equal amounts oftotal leaf protein were fractionated by SDS-PAGE, transferred onto a membrane, and probedwith specific antibodies. The asterisk next toPORC marks the top band as an unspecific sig-nal. ALA-BS, 5-Aminolevulinic acid biosynthe-sis. C, Quantitative reverse transcription-PCRanalysis of the expression in lil3 single anddouble mutants of tetrapyrrole biosynthesisgenes related to LIL3 and POR function. TheACTgene was used as a reference, and data werenormalized to expression levels in wild-typeplants.

Plant Physiol. Vol. 174, 2017 1039

LIL3-POR Interaction

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

Chl synthesis are lower in the lil3.1/lil3.2 double mutantthan in wild-type seedlings, while little or no changewas detectable in either of the single mutants (Fig. 2B).Apart from CHLP, the content of PORA and PORBwasstrikingly reduced, while levels of PORC, CHLG,GUN4, CHL27, and LHCB1 (light-harvesting Chl-bindingproteins of PSII) were depressed to a lesser extent. GUN4is a positive regulator of Chl synthesis, which stimu-lates Mg chelatase activity (Larkin et al., 2003; Peter andGrimm, 2009; Adhikari et al., 2011). CHL27 is an es-sential subunit of the Mg protoporphyrin mono-methylester (MgPME) oxidative cyclase (the enzymethat precedes POR in the pathway) and has been pro-posed to contain the catalytic, substrate-binding do-main (Tottey et al., 2003).

The reduced levels of enzymes of the TBS pathwayin the lil3.1/lil3.2 double mutant (Fig. 2B) could be dueto changes in nuclear gene expression in response toLIL3 deficiency or altered plastid-derived retrogradesignaling. Analysis of transcript levels indicated thatlil3.1 contains up to 20% of the wild-type amount ofthe LIL3.1 transcript. Using the available antibodies forthe two LIL3 isoforms, a weak immunoreactive bandwas indeed detectable with the anti-LIL3.1 antibody inboth lil3.1 and the lil3.1/lil3.2 double mutant (Fig. 2B).The top band in the PORC gel is caused by unspecificbinding of the antibody. GUN4 is known to appear atleast in two different sizes, as already described byLarkin et al. (2003). These observations are in agree-ment with previous suggestions that lil3.1 is a knock-down mutant (Tanaka et al., 2010). In lil3.2, the LIL3.2transcripts were not detectable (Fig. 2C). Thus, lil3.2 isa null mutant. The anti-LIL3.2 antibody labels anotherweak band with a slightly different mobility fromwild-type LIL3.2. This must represent a cross-reactingprotein of unknown nature (Fig. 2B). Inactivation ofeither of the two LIL3 genes did not modify the ex-pression of the other paralog. No significant changesin transcription levels of the other genes analyzedwere found in either of the single lil3 mutants, whilesome noteworthy alterations were noted in the doublemutant (Fig. 2C; Supplemental Fig. S1, A–C). Specifi-cally, the CHLH, LHCA1, and FC2 transcripts accu-mulated to levels at least 2-fold higher than in thecontrols. The CHLH gene encodes the Mg chelatasesubunit H, and FC2 codes for the dominant ferroche-latase isoform 2 in chloroplasts. In contrast, both lil3.1and the doublemutant contained only 50% of thewild-type level of the FC1 transcript. But, insofar as anti-bodies were available for these studies, immunoblotanalysis of the corresponding products of these genesindicated no changes in protein content. We concludefrom these results that the reduced contents of PORand a few other proteins of Chl synthesis do notcorrelate with the levels of transcription of their cor-responding genes and are due to posttranslationalmodifications. Moreover, these data point to partialinhibition of Pchlide reduction, which prompted usto take a closer look at plants deficient in both LIL3isoforms.

LIL3 Inactivation by Virus-Induced Gene Silencing

Based on the normal Chl content of each of the singlelil3 mutants, we inferred that each LIL3 isoform cancompensate partially for loss of the other. The severelyreduced growth rate of the lil3.1/lil3.2 double mutantsprecluded detailed biochemical analysis of chloroplastprotein complexes and tetrapyrrole intermediates.Therefore, wemade use of virus-induced gene silencing(VIGS) to inactivate LIL3 gene expression in 12-d-oldArabidopsis seedlings. One week after infection, de-veloping leaves showed a uniform pale-green pheno-type in comparison with control plants expressingVIGS constructs for GFP inactivation (Fig. 3A). LIL3transcript levels were reduced to 40% (LIL3.1) and 10%(LIL3.2) of those in controls (Fig. 3B). As a result of geneinactivation, the total amount of LIL3 protein wasstrongly diminished (Fig. 3C). As in the case of the LIL3

Figure 3. Characterization of VIGS-LIL3.1/LIL3.2 plants. A, Phenotypeof Arabidopsis plants with rosette leaves 3 weeks after infiltration with apTRV2-Lil3.1/Lil3.2 construct. Leaves of VIGS-LIL3 transformants arepale green compared with control leaves. B, Quantitative PCR (qPCR)analysis of VIGS-LIL3.1/LIL3.2 plants, confirming the down-regulationof both LIL3 genes. The ACT gene was used as a reference, and datawere normalized to expression in VIGS-GFP control plants. C, Steady-state levels of selected proteins/enzymes involved in Chl biosynthesis.Similar amounts of total protein were fractionated by SDS-PAGE,transferred onto a membrane, and probed with specific antibodies.

1040 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

T-DNA insertion mutants, CHLP, PORA, and PORBcontents also were reduced drastically (below our de-tection limits), while CHLG content was similar to thecontrol value.Short-day-acclimated lil3 mutants are already char-

acterized by reduced rates of ALA synthesis (Tanakaet al., 2010). The drop in ALA synthesis and the lowlevels of CHLP and POR observed in our VIGS seed-lings prompted us to determine their Chl content andthe steady-state levels of Chl precursors (Fig. 4). TheChl content was reduced by approximately 50% asa result of VIGS-induced LIL3 silencing (Fig. 4A).Amounts of tocopherols, the phytyl chain-containingtocochromanols, also were depressed by 50%, and notocotrienols were detectable (Fig. 4B), which is com-patible with the loss of CHLP. Levels of Mg porphyrinsand chlorins, as well as rates of ALA synthesis, alsowere strongly diminished (Fig. 4, C and D). Less than20% of the Chl was phytylated in the LIL3-VIGS plants(Fig. 4, E and F). Conversely, these plants containedenhanced amounts of geranylgeranylated Chl a and band of all intermediates formed during the three-stepreduction of geranylgeranylated Chl to phytylated Chl(Fig. 4E). During the reduction step catalyzed by CHLP,three double bonds present in the geranylgeranylmoiety

are reduced to single bonds, which lead to the formationof the phytol chain.

LIL3-Mediated Protein-Protein Interactions

The reduced content of POR in the lil3.1/lil3.2mutantand VIGS-LIL3 plants can be explained if LIL3 acts tostabilize not only CHLP (Tanaka et al., 2010) but alsoPOR. Bimolecular interaction between the two LIL3isoforms and PORB, therefore, was explored using thebimolecularfluorescence complementation (BiFC) assay.Gene constructs encoding protein fusions of the targetproteins to either of the two halves of YFP were tran-siently expressed in Nicotiana benthamiana and revealedplastid-localized interaction of the two transgenic pro-teins (Fig. 5A). BiFC assays with LIL3 and the two otherisoforms PORA/B and PORC detected similar positiveinteractions within chloroplasts (Supplemental Fig.S2A). Moreover, a POR-GluTR interaction was found bythe same approach. This interaction is consistentwith thefindings reported by Kauss et al. (2012), who proposedthe formation of a protein complex consisting of CHL27,POR, FLU, and GluTR in chloroplasts of dark-grownplants. The POR reaction provides Chlide, one of the

Figure 4. Levels of selected tetrapyrroles in VIGS-LIL3.1/LIL3.2 plants. After extraction of freeze-dried leaf samples with basicacetone, levels of Chls and other intermediates were quantified by HPLC using pure standards. A, Total Chl contents of VIGS-LIL3.1/LIL3.2 plants and Chl a/b ratio. B, Steady-state levels of tocopherol a. C, Steady-state levels of Chl precursors. MME, Mg proto-porphyrin IX-monomethylester; MgP, Mg protoporphyrin IX. D, Rates of ALA synthesis, the rate-limiting step of Chl biosynthesis, inVIGS-LIL3.1/LIL3.2 plants. Measurement of ALA synthesis is accomplished by blocking the subsequent enzymatic step of ALAdehydratase and in vitro conversion of ALA into a pyrrole. E, Characterization of Chl a and b species by HPLC. Similar amounts oftotal Chl (1.5 mg) were separated on a Prontosil 200-3-C30 column. Geranylated and phytylated Chl species could be separated dueto their different retention times. Identification of Chl a and bwas performed based on their fluorescence properties (Chl a, excitationat 440 nm, emission at 660 nm; Chl b, excitation at 460 nm, emission at 660 nm). Numbers 1 to 4 indicate different Chl species,where 1 represents phytylated Chl and 2 to 4 represent geranylated Chls. Labels (a) and (b) indicate cross-detected Chl species (i.e.Chl b, whichwas detectedwith the Chl a detection parameters, andvice versa) due to the similar fluorescence emission properties. F,Amounts of phytylated Chl are given in percentages (representing elution peak 1). DW, Dry weight; FW, fresh weight.

Plant Physiol. Vol. 174, 2017 1041

LIL3-POR Interaction

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

two substrates of CHLG. We made several attempts touncover LIL3-CHLG interactions using BIFC and yeasttwo-hybrid assays but were unable to demonstrate anystable interaction (Supplemental Fig. S2).

The LIL3-POR interaction was confirmed using addi-tional biochemical strategies. The split-ubiquitin-mediatedtwo-hybrid approach in yeast allows one to analyze in-teractions among membrane-bound proteins. Interactionbetween all members of either protein family, LIL3 andPOR, was demonstrated, independent of the combinationof the respective fusionproteins considered (SupplementalFig. S2B). Heterodimer formation of POR isoforms wasverified by the BIFCmethod (Supplemental Fig. S2) and isconsistent with the reported tendency of POR to oligo-merize (Szenzenstein et al., 2010; Yuan et al., 2012).

Additional confirmationof thePOR-LIL3 interactionwasachieved in pull-down experiments (Fig. 5B). RecombinantHis-tagged LIL3 was incubated with DDM-solubilizedchloroplast extracts. After intensive washing, the boundproteinswere eluted from the affinity column. In contrast toan eluate from the control assay with recombinant RIBA1as the bait, a plastid-localized bifunctional protein in ribo-flavin biosynthesis (Hiltunen et al., 2012), PORA/B andPORC were immunologically identified among the chlo-roplast proteins that bound specifically to LIL3 (Fig. 5B).

Dissection of LIL3 Domains and Interaction of POR withSpecific Domains

To identify the LIL3domain that interactswith the PORisoforms (Supplemental Fig. S2B), the two LIL3 variantswere dissected into the N-terminal water-soluble half(LIL3 glob) and the membrane-spanning domain (LIL3H1H2) with and without the C-terminal segment (LIL3H1H2+C and LIL3 H1H2, respectively). Gene constructswere designed for the expression of these peptides and forparallel expression of recombinant POR paralogs in yeast.Experiments revealed an interaction of the N-terminalglobular LIL3 peptide with all three POR isoforms, whilethemembrane domains and theC terminus togetherwereinsufficient to mediate POR binding. In contrast, all threeLIL3 peptide variants failed to interact with CHLG inyeast two-hybrid assays (Supplemental Fig. S2B).

Stability of LIL3 in PORB-Deficient Lines

As LIL3.1 and LIL3.2 deficiency correlates with re-duced contents of POR (Fig. 2B) and CHLP (Tanakaet al., 2010), we asked whether POR deficiency alsomight affect LIL3 expression or accumulation. AnArabidopsis CHLP knockout line is not available, butwe selected four allelic PORB knockout mutants (Fig.6A), which carry their T-DNA insertions at differentsites in the PORB gene (Supplemental Fig. S3A). InArabidopsis, three isoforms of POR (PORA–PORC)exist, of which PORB is the dominant isoform in greentissue (Armstrong et al., 1995). The allelic porBmutantsdisplayed slightly retarded growth rates and developedfewer leaves than wild-type seedlings over the samegrowth period. The mutants accumulated either noPORB transcripts or far fewer than the control line(Supplemental Fig. S3B). The presence of the T-DNAinsertions in the coding regions precludes the detection

Figure 5. In vitro and in planta analyses of the LIL3-POR interaction. A,BiFC assay for interaction between LIl3.1, LIl3.2, and PORB. The indi-cated proteins fused to the N- and C-terminal halves of splitYFP weretransiently expressed in N. benthamiana leaves. G1 indicates fusion ofthe target proteins to theN-terminal splitYFP half, whereasG3 stands forfusion to the C-terminal splitYFP half. Yellow fluorescence indicatinginteraction was detected by confocal microscopy. In addition, to con-firm that the interactions occurred inside the chloroplast, Chl auto-fluorescence (chl) was recorded. Bars = 20 mm. B, Pull-down assayswith recombinant His-tagged LIL3.1 as bait and solubilized wild-typechloroplasts (cp) as prey. A recombinant RibA1 protein was used as acontrol. The addition or omission of the given components to the dif-ferent assays is indicated by + or 2, respectively. Bait proteins wereimmobilized on Ni-NTA agarose and incubated with solubilized wild-type chloroplasts. Eluted proteins were separated by SDS-PAGE, trans-ferred onto a membrane, and probed with specific antibodies againstPOR isoforms and CHLP. The immune-reacting protein band at slightlyhigher molecular weight, which was detected by anti-CHLPantibody inthe RibA1 assay, might be due to a cross-reacting protein bound toRibA1.

1042 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

of PORB on immunoblots (Fig. 6B). Thus, the immu-noreactive bands that are still detectable in the porBextracts either represent PORA or result fromunspecificbinding of unknown proteins to anti-PORA/B anti-bodies. Moreover, in the porB mutants, chlorophyllcontents and chlorophyll a/b ratios were similar to thoseof the wild type (Fig. 6, A and C), while the steady-statelevels of Mg Proto and MgPME were lower. Interest-ingly, the levels of Pchlide were 2-fold higher in at leasttwo of the four porBmutants, and in three mutants, theChlide contents were reduced by as much as 75% rel-ative to the wild-type level (Fig. 6E). These changespoint to a defect in the enzymatic reduction of Pchlide,as a consequence of the lack of PORB and the inabilityof other POR isoforms to compensate for its loss. No-tably, the rate of ALA synthesis is hardly affected at all,indicating that regulatory feedback at the level of PORis not detectably disturbed (Fig. 6D).Comparative immunoblot analysis of proteins required

for chlorophyll synthesis revealed several interesting dif-ferences between porB mutants and wild-type seedlings.Interestingly, the content of the two LIL3 isoforms, CHLP

and CHLG, was higher in porB than in the wild type.These proteins belong to the late steps of chlorophyllsynthesis. Quantification of the immunoblot signals ofLIL3.1 revealed an up to 40% increase in the steady-statelevel in the porB lines compared with the wild type (Fig.6B). We also examined whether the lack of PORB ex-pression modulates the expression of LIL3 and genes in-volved in tetrapyrrole biosynthesis, but the analyzedgenes showed unaltered expression levels (SupplementalFig. S3B). This indicates that themodifiedprotein contentsof TBS enzymes are not explained by a different nucleargene expression in the porB mutants but rather by post-translational effects.

Loss of LIL3 Impairs Photosynthetic Performance and theAccumulation of Photosynthetic Complexes in theThylakoid Membrane

As mentioned above, LIL3-VIGS plants showed 50%reduced chlorophyll contents compared with controlplants (Fig. 4A). Interestingly, the amounts of the major

Figure 6. Characterization of porB T-DNA in-sertion lines. A, Phenotypes of four indepen-dent T-DNA insertion lines for the PORB gene.Plants were grown for 4 weeks under short-dayconditions. B, Steady-state levels of proteins/enzymes involved in Chl biosynthesis. Similaramounts of total leaf protein were fractionatedby SDS-PAGE, transferred onto a membrane,and probed with specific antibodies. Chem-iluminescence signals of LIL3.1 blots werequantified, and numbers (in percentage of thewild-type [WT] signal) are given above eachlane. C, Total Chl contents and Chl a/b ratios ofporB mutants. D, Rates of ALA synthesis in theporB mutant lines. E, Steady-state levels of Chlprecursors. Col-0, Columbia-0; FW, fresh weight;MME, Mg protoporphyrin IX-monomethyl ester;MgP, Mg protoporphyrin IX.

Plant Physiol. Vol. 174, 2017 1043

LIL3-POR Interaction

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

LHC proteins Lhca1 and Lhcb1 appeared to be unal-tered and slightly reduced, respectively (Fig. 7A). Thus,we analyzed the protein complexes of the thylakoidmembrane using blue native (BN)-PAGE to obtain amore detailed view of the effects of LIL3 deficiency.Extracts with similar amounts of chlorophyll were sol-ubilized with 1% DDM and separated on native gels(Fig. 7B). We found that silencing of both LIL3 isoformsdrastically changes the composition of the proteincomplexes found in the thylakoid membrane (Fig. 7B).PSII-LHCII supercomplexes were almost absent,whereas PSI monomers and the PSII dimer were clearlyreduced compared with the control. In contrast, mono-meric LHCs accumulated. Separation of the proteincomplexes in the second dimension revealed a sub-stantial loss of PSII core subunits (Fig. 7C), whereas theamount of subunits of the other complexes remainedunchanged.

We further analyzed the photosynthetic performanceof these plants using room temperature Chl fluores-cence spectroscopy (Fig. 7D). LIL3-VIGS plants clearlyshowed a high-Chl fluorescence phenotype (Meureret al., 1996), and, as a result of the elevated basal fluo-rescence, the maximum photochemical efficiency ofPSII in the dark-adapted state (Fv/Fm) decreased from0.856 0.005 to 0.6246 0.049. In addition, the operatingefficiency of PSII decreased from 0.791 6 0.006 to0.508 6 0.044, whereas nonphotochemical quenchingmore than doubled from 0.0546 0.011 to 0.1336 0.024in LIL3-silencing plants compared with wild-typeplants. To assess the relative photosynthetic activityof PSII and PSI, we performed 77K spectroscopy (Fig.7E). These experiments revealed that PSI also was se-verely impaired, as is clearly indicated by the strongdecrease in the PSI emission peak around 730 nm.

LIL3.1 Shows High Affinity for Pchlide andOther Tetrapyrroles

Its membership of the LHC protein family charac-terizes LIL3 as a potential Chl-binding protein. LIL3 hasa typical LHC-binding motif (Fig. 1), which, in princi-ple, should enable it to bind tetrapyrroles. To comple-ment the results relating to the binding of LIL3 toenzymes as well as regulatory and auxiliary factors oflate Chl synthesis, we explored its binding capacity forthe substrates and products of POR (Fig. 8A). Purifiedrecombinant LIL3.1 (Fig. 8B) was incubated with sev-eral tetrapyrrole intermediates and Chls. The bindingconstants were determined by monitoring the quench-ing of intrinsic Trp fluorescence (Fig. 8C). Upon addi-tion of the ligands, fluorescence quenching increasedwith increasing concentrations of the tetrapyrrole me-tabolites. LIL3.1 bound free Chl and intermediates inalmost equimolar concentrations. Data analysis wasperformed assuming one specific binding site, as fittingof the data with equations representing two bindingsites resulted in anomalously high affinities for thesecond site. Moreover, in particular for Pchlide and

Chls, high levels of nonspecific binding were observed,which were taken into account during curve fitting ofall ligands (Table I). The lowest KD was determined forLIL3.1-Pchlide binding (251 6 7 nM). Affinities for Chland the other analyzed Chl precursors were somewhatlower. A KD of about 350 nM for Chl fits well with a

Figure 7. Photosynthetic characterization of VIGS-LIL3.1/LIL3.2 plants.A, Steady-state levels of the major light-harvesting proteins Lhca1/Lhcb1. Similar amounts of total leaf protein were fractionated by SDS-PAGE, transferred onto a membrane, and probed with specific anti-bodies. Two independent biological samples for VIGS-LIL3.1/LIL3.2 areshown. B, BN-PAGE of thylakoids isolated from VIGS plants. Equalamounts of thylakoids (equivalent to 10mg of Chl) were solubilizedwith1% DDM and fractionated on native gels. C, Separation of thylakoidmembrane proteins into the second dimension. Lanes from BN-PAGEwere denatured and placed on top of SDS-PAGE gels. After the run, gelswere stained with Coomassie Brilliant Blue. D, Room temperature Chlfluorescence analysis of LIL3.1/3.2 VIGS plants. E, The 77K Chl spec-troscopy of LIL3.1/3.2 VIGS plants. Data were normalized to the fluo-rescence at the PSII emission peak at 688 nm.

1044 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

previously reported value of 230 nM obtained usingmicroscale thermophoresis (Mork-Jansson et al., 2015b).The specificity of LIL3.1’s affinity for tetrapyrroles wasconfirmed when binding assays with FMN yieldeda KD value of 7,600 nM and complete quenching of Trpfluorescence (data not shown). Interestingly, the spe-cific quenching of the intrinsic Trp fluorescence did notexceed 40% when tetrapyrroles were used as ligands.We suspected that quenching might occur only atspecific Trp residues and performed a structure pre-diction for LIL3.1 using the Phyre2 server (Kelley et al.,2015). LIL3.1 possesses nine Trp residues, of whichthree are clustered near the beginning of the trans-membrane helix 1 (Trp-161, Trp-162, and Trp-164),which harbors the LHC motif (Supplemental Fig. S4).With one possible exception (Trp-112), the other Trpresidues are localized in the N-terminal globular do-main or in the C-terminal region, respectively. Wepropose that fluorescence quenching occurs only atthese four Trp residues, which are localized nearest tothe proposed binding site (i.e. LHCmotif). This wouldexplain the incomplete fluorescence quenching andconfirms the specificity of tetrapyrrole binding to theLHC motif.

DISCUSSION

LIL3 Stabilizes POR

Previous observations of the metabolic phenotype oflil3.1/lil3.2 double mutants revealed that LIL3 is re-quired specifically for the reduction of geranylgeranylpyrophosphate to phytyl pyrophosphate. LIL3 servesto stabilize CHLP, an enzyme that is involved inmodifying the hydrophobic alkyl chain of the ger-anylgeranyl moiety by reducing the three double bondsto single bonds in preparation for the subsequent syn-thesis of Chls and tocopherols. As two LIL3 isoformsare present in Arabidopsis, we asked whether otherenzymes of Chl biosynthesis also might depend onLIL3. As shown here, the abundance of several en-zymes that mediate late steps in Chl biosynthesis isindeed decreased in lil3 double mutants and VIGS-LIL3plants (Figs. 2A and 3C). In particular, the POR contentwas reduced markedly. Comparison of the expressionlevels of transcripts coding for enzymes of Chl bio-synthesis with the steady-state amounts of their pro-ducts confirms that these effects are due to decreases inprotein stability (Fig. 2, B and C; Supplemental Fig. S1).These findings suggest that both LIL3 isoforms mightinteract with the enzymes themselves.

Three alternative methods established that LIL3 in-teracts directly with POR isoforms. (1) BiFC assaysconfirmed the plastid-localized interaction (Fig. 5A;Supplemental Fig. S2A). (2) Pull-down experimentsrevealed selective purification of POR proteins fromthe plastid extracts when immobilized LIL3 was usedas the bait (Fig. 5B). (3) Bimolecular interaction alsowas demonstrated in the yeast two-hybrid system

Figure 8. Analysis of the pigment-binding properties of recombi-nant LIL3.1. A, Schematic description of the flow of intermediatesthrough the magnesium branch of the tetrapyrrole biosyntheticpathway (starting from protoporphyrin IX), indicating the role of thePOR protein. B, Assessment of the purity of the LIL3.1 protein usedfor Trp fluorescence quenching. A 2-mg aliquot of the recombinantprotein was subjected to SDS-PAGE and stained with CoomassieBrilliant Blue. MW, Molecular weight. C, Trp fluorescence quench-ing analysis of recombinant LIL3.1 protein incubated with increas-ing amounts of the indicated ligands. The rates of relative quenching(indicated as 1 2 Frel) measured at the corresponding ligand con-centrations are shown. The concentration of the LIL3.1 protein wasset to 360 nM. Data analysis was performed assuming one specificbinding site plus unspecific binding. Mean values from three inde-pendent experiments are shown.

Plant Physiol. Vol. 174, 2017 1045

LIL3-POR Interaction

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

(Supplemental Fig. S2B). Based on these data, we con-clude that the stability of CHLP and POR depends onLIL3 expression and that its interaction with theseproteins mediates their selective localization at theplastid membranes. These results also are consistentwith the stabilization of transgenic CHLP variants,which were supplemented with the transmembranedomain of LIL3 (Takahashi et al., 2014). These trans-formants showed complementation of the lil3.1/lil3.2double mutant.

Functional Role of LIL3 in Chl Synthesis

Interestingly, CHLP and POR share a commonfunction in that they catalyze the formation of the twosubstrates required for the subsequent enzymatic stepimplemented by Chl synthase. The fact that LIL3 as-sociates with both enzymes strongly suggests that itfacilitates the coordination of their action in the mem-brane, so that their products are simultaneously madeavailable to Chl synthase. Therefore, we also testedwhether LIL3 interacts with CHLG (Supplemental Fig.S2B). However, BiFC assays, yeast two-hybrid experi-ments, and pull-down experiments all failed to provideevidence for stable physical contact between CHLG andLIL3. Thus, a direct involvement of LIL3 in the final stepin the synthesis of Chl a could not be demonstrated.

A previous study of CHLG assembly into complexeswith HLIPs in cyanobacteria (Chidgey et al., 2014) alsoargues against a direct LIL3-CHLG interaction. Theauthors of that report presented evidence for the for-mation of a complex containing HLIPs and CHLG, to-gether with multiple chaperones and auxiliary factors,such as YCF39, in close proximity to the site of the finalsteps of cyanobacterial Chl synthesis and assembly ofphotosynthetic Chl-binding proteins of PSII (Chidgeyet al., 2014). Based on these results in cyanobacteria, itcannot be excluded that the channeling of pigments andthe supply and integration of photosynthetic proteinsin the thylakoid membrane require the additional as-sistance of proteins of the HLIP/OHP group.

It should be noted here that (like all other represen-tatives of the two-helix proteins of the LHC family)

LIL3 is absent in cyanobacteria. The current modelimplies that LIL3 facilitates the supply of substrates tothe plant CHLG by interacting with CHLP and POR(Fig. 9). We have demonstrated here that LIL3 makes asubstantial contribution to the stability of POR, as wellas CHLP, and thus helps to ensure a balanced flow ofmetabolites for Chl synthesis. However, our currentknowledge of the organization of late Chl biosynthesisremains fragmentary, and further studies on plant LIL3and OHPs and their involvement in Chl synthesis andChl supply for photosynthetic protein complexes areneeded. It is expected that the control of Chl synthesisand the biogenesis of Chl-binding proteins are morecomplex in higher plants. Chl synthesis serves for moreChl-binding proteins in plants than in cyanobacteria, asthe light-harvesting proteins also are assembled withChl. The integration of nucleus- and plastid-encoded

Figure 9. Working model for LIL3 function in the regulation of Chlbiosynthesis. In the current working model, LIL3 acts as a scaffold/membrane anchor protein not only for CHLP but also for POR.WhereasCHLP provides phytyl pyrophosphate, POR produces Chlide. Together,they provide the substrates for the reaction catalyzed by CHLG. Bystabilizing CHLPas well as POR, LIL3 also may help to direct the flow ofintermediates through the final steps of tetrapyrrole and terpenoid bio-synthesis toward Chl biosynthesis, as it would allow the synchronousaccumulation of precursors for the final Chl assembly. In addition, itsrelatively high affinity for Pchlide also would permit LIL3 to supply thePOR protein with its substrate.

Table I. Binding constants for LIL3.1 pigment binding obtained by Trp fluorescence quenching analysis

Quenching analysis was performed assuming one specific binding site plus unspecific binding(F ¼ Bmaxx

KDþx þNS$x, where x represents the ligand concentration). The dissociation constant (KD) representsthe ligand concentration at which half of the binding sites are occupied, and the term Ns is a pro-portionality constant representing the slope of the linear nonspecific binding.

Ligand KD [nM] Nonspecific Binding (Ns)

Protoporphyrin IX 397 6 15 0.0120 6 0.0009Mg protoporphyrin IX 624 6 30 0.0042 6 0.0013Mg protoporphyrin IX-monomethylester 615 6 19 0.0088 6 0.0009Protochlorophyllide 251 6 7 0.0345 6 0.0006Chl a 359 6 26 0.0284 6 0.0018Chl b 342 6 25 0.0393 6 0.0020

1046 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

Chl-binding proteins into thylakoids during the si-multaneous association with pigments and cofactorsprobably requires the combined action of LIL3 andOHP isoforms, apart from other auxiliary factors. Towhat extent ELIPs and other representatives of the LILfamily contribute to these processes is an open question(Hutin et al., 2003; Tzvetkova-Chevolleau et al., 2007;Hayami et al., 2015).

LIL3 and Its Tetrapyrrole-Binding Property

It remains a challenging task to assign further func-tions to LIL3, apart from its contribution to proteinstability in Chl biosynthesis (Tanaka et al., 2010) and theassembly of proteins into the photosynthetic complexes(Lohscheider et al., 2015; Mork-Jansson et al., 2015a).Functions of LIL3 in Chl and precursor transfer requireat least the physical interaction of LIL3 with otherproteins. Here, we show that Chl precursors bind toLIL3.1 (Fig. 8). It is particularly noteworthy that Pchlide,a substrate for its binding partner POR, binds to LIL3with a higher affinity than any other tetrapyrrole tested.The KD values of LIL3 for metabolites of Chl synthesisare in agreement with previous values reported byMork-Jansson et al. (2015b), while Takahashi et al. (2014)placed more emphasis on the membrane-anchoringfunction of LIL3 for its interacting protein partner. Fur-ther specific binding properties of LIL3 for tetrapyrroleswill be explored in the future, including the identifica-tion of the specific amino acid residues involved.Moreover, it will beworthwhile tofind outwhether LIL3also binds carotenoids, which would enable thequenching of transferred energy from light-absorbingtetrapyrroles. Taken together, our data support a rolefor LIL3 in the organization of Chl synthesis, in agree-ment with the prediction that LIL3 serves as an auxiliaryfactor for interim binding and channeling of Chl pre-cursors. Based on the different affinities, it can be as-sumed that LIL3 functions as a local sink for Pchlide inthe thylakoid membrane, thereby supplying POR withits substrate.

Physiological Consequences of LIL3 Deficiency on theAssembly of Photosynthetic Complexes

To assess the effects of a relative lack of LIL3 on theassembly and repair of photosynthetic complexes(Lohscheider et al., 2015; Mork-Jansson et al., 2015a),we performed BN-PAGE to examine photosystem-LHCcomplex formation. It was obvious that the amountsof stable supercomplexes are diminished as a result ofLIL3 deficiency (Fig. 7B). However, the levels ofLHCPs (Figs. 2B and 7A) do not correlate directly withthe decreased accumulation of PSII-LHCII super-complexes. Indeed, in the relative absence of LIL3, theloss of PSII-LHCII supercomplexes is associated withan elevated accumulation of LHC monomers (Fig. 7B).LIL3 function was linked previously with the associa-tion of peripheral antennae with PSII core complexes

(Reisinger et al., 2008; Lohscheider et al., 2015), but wecurrently see fewer indications for a specific function ofLIL3 in PSII assembly. Based on our results, we favor arole of LIL3 in the management of the metabolic path-way. The binding of tetrapyrrole intermediates to LIL3and its interaction with enzymes involved in Chl andterpenoid synthesis might reduce the risk of phototox-icity by freely diffusing light-absorbing tetrapyrrolesand enhanced photosensitivity of Chl during exposureto high-intensity light. We propose that the organiza-tion of Chl metabolism implies the action of auxiliaryfactors, such as LIL3, which contribute to the fine-tunedchanneling of tetrapyrrole intermediates. It is suggestedthat further studies of LIL3 and other members of theLIL family will be required before the roles of LIL3 inthe biogenesis of photosynthetic complexes and inother enzymatic steps in Chl biosynthesis can be de-fined more precisely.

Altered Levels of Chl Synthesis Proteins in PORB-Deficient Mutants

Since the stability of CHLP and POR depends onLIL3 expression, we examined whether the conversealso is true. We selected the porB mutants because theyrepresent the isoform that is constantly expressed indarkness and light, while PORC is light-dependentlyexpressed and PORA is expressed only in etiolatedtissue (Masuda et al., 2003). Interestingly, porBmutantsactually express higher levels of LIL3 than the wildtype. In addition, it is remarkable that the content ofsome enzymes of late Chl synthesis, such as CHLG andCHLP, also was elevated (Fig. 6B).

In conclusion, we report a more far-reaching role forLIL3 in tetrapyrrole metabolite binding and interactionwith POR. This mutual interaction of LIL3 with en-zymes associated in potential protein complexes (Fig. 9)in photosynthetic organisms is proposed to be crucialfor the maintenance and productivity of the metabolicpathway.

MATERIALS AND METHODS

Plant Material and Growth Conditions

Arabidopsis plants (Arabidopsis thaliana ecotypes Columbia and Nossen)were grown on soil in a growth chamber (100 mmol photons m22 s21, 8-hphotoperiod, 20°C, and 70% relative humidity). The single mutant T-DNA in-sertion lines lil3.1 (DS13-3953) and lil3.2 (Ds13-0193) and the double mutantlil3.1/lil3.2 (DS13-39533Ds13-0193) were kindly provided by R. Tanaka. Seedsof the porB mutant (SALK_001829, SALK_043887, SALK_060191, and SAIL429_A06) were obtained from the Nottingham Arabidopsis Stock Centre.Confirmation of the homozygous T-DNA insertion lines was performed byPCR using the primers listed in Supplemental Table S1.

VIGS Assay

Arabidopsis plants (ecotype Columbia) applied for the VIGS assay weregrown on soil in a growth chamber (100 mmol photons m22 s21, 16-h pho-toperiod, 20°C, and 70% relative humidity). The plasmids pTRV1 and pTRV2,based on Tobacco rattle virus, were used as described by Liu et al. (2002). LIL3.1and LIL3.2 cDNAs were amplified with specific primers (listed in

Plant Physiol. Vol. 174, 2017 1047

LIL3-POR Interaction

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

Supplemental Table S1) from Arabidopsis total cDNAs generated by re-verse transcription-PCR and cloned into pTRV2. As the negative control, apTRV2-GFP vector was used. Infiltration of Agrobacterium tumefaciens cellstransformedwith the pTRV plasmids was done as described previously (Burch-Smith et al., 2006).

Pigment Extraction and HPLC

Leafmaterial for pigment extractionwasharvested2 to 3hafter the transitionfrom dark to light and ground in liquid nitrogen. Subsequently, the pigmentswere extracted from the powder with ice-cold acetone:0.2 N NH4OH (9:1), withor without prior freeze drying, separated by HPLC, and quantified using purestandards.

ALA Synthesis Rate

Rates of ALA synthesis were measured as described by Mauzerall andGranick (1956). Briefly, detached leaves or leaf discs were incubated in 40 mM

levulinic acid in 50 mM Tris-HCl, pH 7.2, for 3 to 4 h under growth light con-ditions. Afterward, the leaf material was ground in liquid nitrogen and resus-pended in 50mMKPO4 buffer, pH 6.8. After centrifugation, the supernatant wasmixed 5:1 with ethyl acetoacetate, boiled for 10 min, and mixed 1:1 withmodified Ehrlich reagent. The ALA concentration was determined photomet-rically and compared with a standard as described.

Preparation of Total Leaf Protein, SDS-PAGE, andImmunoblot Analyses

Total leafproteinswere extracted from leafmaterial ground in liquidnitrogenby resuspending the powder in an equal volume of protein extraction buffer(55 mM Na2CO3, 5 mM DTT, 350 mM Suc, 2% SDS, and 2 mM EDTA) and incu-bating at 70°C for 10 min. After precipitation of 25 mL of protein extract withcold TCA and resuspension in 100 mL of 0.1 N NaOH, the protein concentrationwas determined with the BCA reagent (Thermo Fisher). Equal amounts ofprotein were fractionated by SDS-PAGE, transferred onto nitrocellulosemembranes, and probed with specific antibodies. For signal detection, ClarityWestern ECL Blotting Substrate (Bio-Rad) was used.

Thylakoid Extraction and BN-PAGE

Thylakoid membranes were extracted as described previously (Järvi et al.,2011), except that leaves were homogenized in buffer containing 450 mM sor-bitol, 20 mM Tricine-KOH (pH 8.4), 10 mM EDTA, and 0.1% BSA. Finally, thy-lakoids were resuspended in 25BTH20G buffer (25 mM Bis-Tris, pH 7, and 20%glycerol) and stored at 280°C. BN-PAGE was performed as described earlier(Peng et al., 2008).

Gene Expression Analyses

Total RNAs were extracted from leaf material using Trisure (Bioline) fol-lowing the manufacturer’s instructions. Subsequently, 1-mg aliquots of RNAwere digested with DNase, and cDNA synthesis was performed with theRevertAid RT Enzyme (Thermo) according to the manufacturer’s instructions.Gene expression was analyzed with SensiMixSYBR (Bioline) using the CFX96 real-time system (Bio-Rad) with the primers listed in Supplemental Table S2.ACT2 (At3g18780) was used as the reference gene. Data analysis was performedaccording to the DDCt method (Pfaffl, 2001).

Expression and Purification of Recombinant LIL3.1 Protein

For the expression of recombinant LIL3.1 protein, an Escherichia coli-optimized cDNA sequence was ordered from Integrated DNA Technologiesand cloned into the pET28a vector (Novagen) modifiedwith an N-terminal TEVcleavage site. Protein expression in the E. coli BL21-pRIL strain (DE3; Stra-tagene) was performed at 26°C for 3 h and was induced routinely with 0.25 mM

isopropylthio-b-galactoside at an OD600 of 0.6 to 0.8. Purification of recombi-nant 6xHis-tagged proteins was performed using a 50% (v/v) suspension ofNi-NTA agarose (Thermo Scientific) according to the manufacturer’s protocol.Proteins were concentrated with Amicon Ultra-15 (molecular mass cutoff,30 kD) centrifugal filter units (Millipore).

Pull-Down Assay

Purified 6xHis-LIL3.1 protein (100mg) was immobilized onNi-NTA agarose(100 mL of a 50% [v/v] suspension of Ni-NTA agarose [Thermo Scientific] inequilibration buffer) for 60 min at 4°C. Total chloroplast proteins (100 mg ofChl) solubilized in 1% (w/v) DDM were incubated with immobilized 6xHis-LIL3.1 protein overnight at 4°C. The empty agarose and agarose bound to6xHis-RibA1 were used as negative controls. The agarose was washed exten-sively four times with phosphate buffer containing 150 mM NaCl. Bound pro-teins were eluted with elution buffer containing 250 mM imidazole and 150 mM

NaCl and analyzed by immunoblotting.

In Planta Interaction Studies by BiFC

LIL3.1 and POR cDNAs were amplified from total cDNAs with primerscarrying attB sites and cloned into the G1 and G3 vectors (pDEST-GW-VYNE/-VYCE) via pDON207 using the Gateway system (Invitrogen) accord-ing to the manufacturer’s instructions. A. tumefaciens (GV2260) cells transformedwith the G1/G3 vectors were resuspended in transformation buffer to an OD of0.2, mixed in the desired combinations, and infiltrated intoNicotiana benthamianaleaves. Plants were kept in the dark overnight and subsequently incubated at120 mE PAR for 2 d. YFP fluorescence was recorded on a confocal laser-scanningmicroscope (CLSM TCS SP2 AOBS; Leica).

Yeast Two-Hybrid Assays

cDNA sequences of all genes analyzed were cloned either by restriction sitecloning via pJET2.1 into pDHB1MCS2 (bait) or by Gateway cloning viapDONR221 into met25pXCgate (pNub; prey). The yeast strain L40ccA wascotransformed with bait and prey constructs in all variations. Transformedcloneswere selected on agar plateswith synthetic dextrose (SD)medium lackingLeu and Trp, and liquid overnight cultures of these selected colonies weregrown in 3 mL of SD medium lacking Leu and Trp. OD600 was measured andadjusted to 1. For the selection of positive interaction, the cultures were spottedon agar plates containing SD medium lacking His, Leu, Trp, and uracil, sup-plementedwith 10mM 3-amino-1,2,4-triazole, and incubated for 3 to 4 d at 30°C.

Chl Fluorescence Spectroscopy

Leafmaterial fromArabidopsis plantswashomogenized inLHCbuffer (0.4M

sorbitol in 50mM Tricine, pH 7.8) andmixedwith 1 volume of 80% glycerol. Thesuspension was filled into a capillary, frozen in liquid nitrogen, before its Chlfluorescence was measured with a fluorescence spectrophotometer (HitachiF-7000). Pulse amplitude-modulated room temperature fluorescence wasmeasured with an FMS2 Portable Pulse Modulated Chlorophyll Fluorometer(Hansatech). In general, a protocol described by Meurer et al. (1996) was fol-lowed, except that the actinic light was adjusted to growth light conditions(100 mmol photons m22 s21).

Trp Fluorescence Quenching

An aliquot of the 6xHis-tagged LIL3.1 protein (360 nM) was diluted inquenching buffer (50 mM Tricine, pH 7.9, 300 mM glycerol, and 1 mM DTT),mixedwith increasing amounts of various porphyrins (final concentrations, 0–8mM), dissolved in DMSO, and incubated for 5 min. Trp fluorescence was mea-sured with a fluorescence spectrophotometer (Hitachi F-7000) after excitation at280 nm.

Data Analysis

Generally, experiments were performed three times, and a representativeresult is shown. Data were analyzed using Microsoft Excel except for the curvefitting of the Trp fluorescence quenching data, for which SigmaPlot 11 (SystatSoftware) was used.

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. qPCR analysis of tetrapyrrole biosynthesis genesin lil3 single and double mutants.

1048 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

Supplemental Figure S2. Additional analyses confirming LIL3-POR interaction.

Supplemental Figure S3. qPCR of tetrapyrrole biosynthesis genes in porBmutants and T-DNA insertion sites in the PORB gene.

Supplemental Figure S4. Predicted structure of LIL3.1.

Supplemental Table S1. Sequences of primers used for genotyping andcloning.

Supplemental Table S2. Sequences of primers used for quantitative re-verse transcription-PCR.

Received April 12, 2017; accepted April 19, 2017; published April 21, 2017.

LITERATURE CITED

Adhikari ND, Froehlich JE, Strand DD, Buck SM, Kramer DM, LarkinRM (2011) GUN4-porphyrin complexes bind the ChlH/GUN5 subunitof Mg-chelatase and promote chlorophyll biosynthesis in Arabidopsis.Plant Cell 23: 1449–1467

Albus CA, Salinas A, Czarnecki O, Kahlau S, Rothbart M, Thiele W, LeinW, Bock R, Grimm B, Schöttler MA (2012) LCAA, a novel factor re-quired for magnesium protoporphyrin monomethylester cyclase accu-mulation and feedback control of aminolevulinic acid biosynthesis intobacco. Plant Physiol 160: 1923–1939

Allahverdiyeva Y, Suorsa M, Tikkanen M, Aro EM (2015) Photoprotectionof photosystems in fluctuating light intensities. J Exp Bot 66: 2427–2436

Andersson U, Heddad M, Adamska I (2003) Light stress-induced one-helixprotein of the chlorophyll a/b-binding family associated with photo-system I. Plant Physiol 132: 811–820

Armstrong GA, Runge S, Frick G, Sperling U, Apel K (1995) Identificationof NADPH:protochlorophyllide oxidoreductases A and B: a branchedpathway for light-dependent chlorophyll biosynthesis in Arabidopsisthaliana. Plant Physiol 108: 1505–1517

Brzezowski P, Richter AS, Grimm B (2015) Regulation and function oftetrapyrrole biosynthesis in plants and algae. Biochim Biophys Acta1847: 968–985

Burch-Smith TM, Schiff M, Liu Y, Dinesh-Kumar SP (2006) Efficientvirus-induced gene silencing in Arabidopsis. Plant Physiol 142: 21–27

Chidgey JW, Linhartová M, Komenda J, Jackson PJ, Dickman MJ,Canniffe DP, Koník P, Pilný J, Hunter CN, Sobotka R (2014) A cya-nobacterial chlorophyll synthase-HliD complex associates with theYcf39 protein and the YidC/Alb3 insertase. Plant Cell 26: 1267–1279

Dolganov NA, Bhaya D, Grossman AR (1995) Cyanobacterial protein withsimilarity to the chlorophyll a/b binding proteins of higher plants: evo-lution and regulation. Proc Natl Acad Sci USA 92: 636–640

Green BR, Pichersky E, Kloppstech K (1991) Chlorophyll a/b-bindingproteins: an extended family. Trends Biochem Sci 16: 181–186

Grimm B, Kloppstech K (1987) The early light-inducible proteins of barley:characterization of two families of 2-h-specific nuclear-coded chloro-plast proteins. Eur J Biochem 167: 493–499

Hayami N, Sakai Y, Kimura M, Saito T, Tokizawa M, Iuchi S, Kurihara Y,Matsui M, Nomoto M, Tada Y, et al (2015) The responses of Arabi-dopsis Early Light-Induced Protein2 to ultraviolet B, high light, and coldstress are regulated by a transcriptional regulatory unit composed oftwo elements. Plant Physiol 169: 840–855

Heddad M, Adamska I (2000) Light stress-regulated two-helix proteins inArabidopsis thaliana related to the chlorophyll a/b-binding gene family.Proc Natl Acad Sci USA 97: 3741–3746

Hernandez-Prieto MA, Tibiletti T, Abasova L, Kirilovsky D, Vass I, FunkC (2011) The small CAB-like proteins of the cyanobacterium Synecho-cystis sp. PCC 6803: their involvement in chlorophyll biogenesis forphotosystem II. Biochim Biophys Acta 1807: 1143–1151

Hiltunen HM, Illarionov B, Hedtke B, Fischer M, Grimm B (2012) ArabidopsisRIBA proteins: two out of three isoforms have lost their bifunctional activityin riboflavin biosynthesis. Int J Mol Sci 13: 14086–14105

Horton P (2012) Optimization of light harvesting and photoprotection:molecular mechanisms and physiological consequences. Philos Trans RSoc Lond B Biol Sci 367: 3455–3465

Hutin C, Nussaume L, Moise N, Moya I, Kloppstech K, Havaux M (2003)Early light-induced proteins protect Arabidopsis from photooxidativestress. Proc Natl Acad Sci USA 100: 4921–4926

Jansson S (1994) The light-harvesting chlorophyll a/b-binding proteins.Biochim Biophys Acta 1184: 1–19

Jansson S, Andersson J, Kim SJ, Jackowski G (2000) An Arabidopsisthaliana protein homologous to cyanobacterial high-light-inducibleproteins. Plant Mol Biol 42: 345–351

Järvi S, Suorsa M, Paakkarinen V, Aro EM (2011) Optimized native gelsystems for separation of thylakoid protein complexes: novel super- andmega-complexes. Biochem J 439: 207–214

Kauss D, Bischof S, Steiner S, Apel K, Meskauskiene R (2012) FLU, anegative feedback regulator of tetrapyrrole biosynthesis, is physicallylinked to the final steps of the Mg++-branch of this pathway. FEBS Lett586: 211–216

Kelley LA, Mezulis S, Yates CM, Wass MN, Sternberg MJE (2015) ThePhyre2 web portal for protein modeling, prediction and analysis. NatProtoc 10: 845–858

Klimmek F, Sjödin A, Noutsos C, Leister D, Jansson S (2006) Abundantlyand rarely expressed Lhc protein genes exhibit distinct regulation pat-terns in plants. Plant Physiol 140: 793–804

Komenda J, Sobotka R (2016) Cyanobacterial high-light-inducible proteins:protectors of chlorophyll-protein synthesis and assembly. Biochim Bio-phys Acta 1857: 288–295

Larkin RM, Alonso JM, Ecker JR, Chory J (2003) GUN4, a regulator ofchlorophyll synthesis and intracellular signaling. Science 299: 902–906

Li XP, Björkman O, Shih C, Grossman AR, Rosenquist M, Jansson S,Niyogi KK (2000) A pigment-binding protein essential for regulation ofphotosynthetic light harvesting. Nature 403: 391–395

Liu Y, Schiff M, Dinesh-Kumar SP (2002) Virus-induced gene silencing intomato. Plant J 31: 777–786

Lohscheider JN, Rojas-Stütz MC, Rothbart M, Andersson U, Funck D,Mendgen K, Grimm B, Adamska I (2015) Altered levels of LIL3 iso-forms in Arabidopsis lead to disturbed pigment-protein assembly andchlorophyll synthesis, chlorotic phenotype and impaired photosyntheticperformance. Plant Cell Environ 38: 2115–2127

Masuda T, Fusada N, Oosawa N, Takamatsu K, Yamamoto YY, Ohto M,Nakamura K, Goto K, Shibata D, Shirano Y, et al (2003) Functionalanalysis of isoforms of NADPH:protochlorophyllide oxidoreductase(POR), PORB and PORC, in Arabidopsis thaliana. Plant Cell Physiol 44:963–974

Mauzerall D, Granick S (1956) The occurrence and determination of delta-amino-levulinic acid and porphobilinogen in urine. J Biol Chem 219:435–446

Meurer J, Meierhoff K, Westhoff P (1996) Isolation of high-chlorophyll-fluorescence mutants of Arabidopsis thaliana and their characterisation byspectroscopy, immunoblotting and northern hybridisation. Planta 198:385–396

Montané MH, Kloppstech K (2000) The family of light-harvesting-relatedproteins (LHCs, ELIPs, HLIPs): was the harvesting of light their primaryfunction? Gene 258: 1–8

Mork-Jansson A, Bue AK, Gargano D, Furnes C, Reisinger V, Arnold J,Kmiec K, Eichacker LA (2015a) Lil3 assembles with proteins regulatingchlorophyll synthesis in barley. PLoS ONE 10: e0133145

Mork-Jansson AE, Gargano D, Kmiec K, Furnes C, Shevela D, EichackerLA (2015b) Lil3 dimerization and chlorophyll binding in Arabidopsisthaliana. FEBS Lett 589: 3064–3070

Niyogi KK, Li XP, Rosenberg V, Jung HS (2005) Is PsbS the site of non-photochemical quenching in photosynthesis? J Exp Bot 56: 375–382

Peng L, Shimizu H, Shikanai T (2008) The chloroplast NAD(P)H dehy-drogenase complex interacts with photosystem I in Arabidopsis. J BiolChem 283: 34873–34879

Peter E, Grimm B (2009) GUN4 is required for posttranslational control ofplant tetrapyrrole biosynthesis. Mol Plant 2: 1198–1210

Pfaffl MW (2001) A new mathematical model for relative quantification inreal-time RT-PCR. Nucleic Acids Res 29: e45

Pribil M, Labs M, Leister D (2014) Structure and dynamics of thylakoids inland plants. J Exp Bot 65: 1955–1972

Reisinger V, Plöscher M, Eichacker LA (2008) Lil3 assembles as chlorophyll-binding protein complex during deetiolation. FEBS Lett 582: 1547–1551

Rochaix JD (2014) Regulation and dynamics of the light-harvesting system.Annu Rev Plant Biol 65: 287–309

Ruban AV, Johnson MP, Duffy CD (2012) The photoprotective molecularswitch in the photosystem II antenna. Biochim Biophys Acta 1817: 167–181

Plant Physiol. Vol. 174, 2017 1049

LIL3-POR Interaction

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

Sobotka R, Tichy M, Wilde A, Hunter CN (2011) Functional assignmentsfor the carboxyl-terminal domains of the ferrochelatase from Synecho-cystis PCC 6803: the CAB domain plays a regulatory role, and region II isessential for catalysis. Plant Physiol 155: 1735–1747

Szenzenstein A, Kósa A, Solymosi K, Sárvári E, Böddi B (2010) Preferentialregeneration of the NADPH:protochlorophyllide oxidoreductase oligomercomplexes in pea epicotyls after bleaching. Physiol Plant 138: 102–112

Takahashi K, Takabayashi A, Tanaka A, Tanaka R (2014) Functionalanalysis of light-harvesting-like protein 3 (LIL3) and its light-harvestingchlorophyll-binding motif in Arabidopsis. J Biol Chem 289: 987–999

Tanaka R, Kobayashi K, Masuda T (2011) Tetrapyrrole metabolism inArabidopsis thaliana. The Arabidopsis Book 9: e0145, doi/0.1199/tab.0145

Tanaka R, Rothbart M, Oka S, Takabayashi A, Takahashi K, Shibata M,Myouga F, Motohashi R, Shinozaki K, Grimm B, et al (2010) LIL3, alight-harvesting-like protein, plays an essential role in chlorophyll andtocopherol biosynthesis. Proc Natl Acad Sci USA 107: 16721–16725

Tanaka R, Tanaka A (2007) Tetrapyrrole biosynthesis in higher plants.Annu Rev Plant Biol 58: 321–346

Tanaka R, Tanaka A (2011) Chlorophyll cycle regulates the constructionand destruction of the light-harvesting complexes. Biochim BiophysActa 1807: 968–976

Tottey S, Block MA, Allen M, Westergren T, Albrieux C, Scheller HV,Merchant S, Jensen PE (2003) Arabidopsis CHL27, located in both

envelope and thylakoid membranes, is required for the synthesis ofprotochlorophyllide. Proc Natl Acad Sci USA 100: 16119–16124

Tzvetkova-Chevolleau T, Franck F, Alawady AE, Dall’Osto L, Carrière F,Bassi R, Grimm B, Nussaume L, Havaux M (2007) The light stress-induced protein ELIP2 is a regulator of chlorophyll synthesis in Arabi-dopsis thaliana. Plant J 50: 795–809

Wang P, Grimm B (2015) Organization of chlorophyll biosynthesis andinsertion of chlorophyll into the chlorophyll-binding proteins in chlo-roplasts. Photosynth Res 126: 189–202

Waterhouse AM, Procter JB, Martin DMA, Clamp M, Barton GJ (2009)Jalview version 2: a multiple sequence alignment editor and analysisworkbench. Bioinformatics 25: 1189–1191

Yao DC, Brune DC, Vavilin D, Vermaas WF (2012) Photosystem II com-ponent lifetimes in the cyanobacterium Synechocystis sp. strain PCC6803: small Cab-like proteins stabilize biosynthesis intermediates andaffect early steps in chlorophyll synthesis. J Biol Chem 287: 682–692

Yuan M, Zhang DW, Zhang ZW, Chen YE, Yuan S, Guo YR, Lin HH(2012) Assembly of NADPH:protochlorophyllide oxidoreductase com-plex is needed for effective greening of barley seedlings. J Plant Physiol169: 1311–1316

Zhang M, Zhang F, Fang Y, Chen X, Chen Y, Zhang W, Dai HE, Lin R, LiuL (2015) The non-canonical Tetratricopeptide Repeat (TPR) domain ofFluorescent (FLU) mediates complex formation with glutamyl-tRNAreductase. J Biol Chem 290: 17559–17565

1050 Plant Physiol. Vol. 174, 2017

Hey et al.

https://plantphysiol.orgDownloaded on May 27, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.