functional evidence for the critical amino-terminal conserved

Functional Evidence for the Critical Amino-TerminalConserved Domain and Key Amino Acids of Arabidopsis4-HYDROXY-3-METHYLBUT-2-ENYLDIPHOSPHATE REDUCTASE1[W][OPEN]

Wei-Yu Hsieh, Tzu-Ying Sung, Hsin-Tzu Wang, and Ming-Hsiun Hsieh*

Institute of Plant and Microbial Biology, Academia Sinica, Taipei 11529, Taiwan

The plant 4-HYDROXY-3-METHYLBUT-2-ENYL DIPHOSPHATE REDUCTASE (HDR) catalyzes the last step of themethylerythritol phosphate pathway to synthesize isopentenyl diphosphate and its allyl isomer dimethylallyl diphosphate,which are common precursors for the synthesis of plastid isoprenoids. The Arabidopsis (Arabidopsis thaliana) genomic HDRtransgene-induced gene-silencing lines are albino, variegated, or pale green, confirming that HDR is essential for plants. We usedEscherichia coli isoprenoid synthesis H (Protein Data Bank code 3F7T) as a template for homology modeling to identify key aminoacids of Arabidopsis HDR. The predicted model reveals that cysteine (Cys)-122, Cys-213, and Cys-350 are involved in iron-sulfurcluster formation and that histidine (His)-152, His-241, glutamate (Glu)-242, Glu-243, threonine (Thr)-244, Thr-312, serine-379,and asparagine-381 are related to substrate binding or catalysis. Glu-242 and Thr-244 are conserved only in cyanobacteria, greenalgae, and land plants, whereas the other key amino acids are absolutely conserved from bacteria to plants. We used site-directed mutagenesis and complementation assay to confirm that these amino acids, except His-152 and His-241, were critical forArabidopsis HDR function. Furthermore, the Arabidopsis HDR contains an extra amino-terminal domain following the transitpeptide that is highly conserved from cyanobacteria, and green algae to land plants but not existing in the other bacteria. Wedemonstrated that the amino-terminal conserved domain was essential for Arabidopsis and cyanobacterial HDR function.Further analysis of conserved amino acids in the amino-terminal conserved domain revealed that the tyrosine-72 residue wascritical for Arabidopsis HDR. These results suggest that the structure and reaction mechanism of HDR evolution have becomespecific for oxygen-evolving photosynthesis organisms and that HDR probably evolved independently in cyanobacteria versusother prokaryotes.

Isoprenoids are the largest group of natural productsfound in living organisms (Sacchettini and Poulter, 1997).All isoprenoids are derived from a basic five-carbon unit,isopentenyl diphosphate (IPP), and its allyl isomerdimethylallyl diphosphate (DMAPP). In animals, IPPand DMAPP are synthesized via the mevalonate (MVA)pathway using acetyl-CoA as the precursor. By contrast,in most eubacteria, including many pathogenic bacteria,IPP and DMAPP are synthesized via the methylerythritolphosphate (MEP) pathway, which uses pyruvate andglyceraldehyde 3-phosphate as precursors. In plants,IPP and DMAPP are synthesized via two independentpathways, the cytosolic MVA pathway and the chloro-plastic MEP pathway. The common products IPP andDMAPP synthesized in distinct subcellular compartmentsare used for the biosynthesis of different isoprenoids.

For instance, sesquiterpenes, sterols (triterpenes), andpolyterpenes are mainly synthesized from the cytosolicMVA pathway. Isoprene, monoterpenes, phytol, plas-toquinones, tocopherols, carotenoids, and the planthormones GA and abscisic acid are derived fromthe chloroplastic MEP pathway (Lichtenthaler, 1999;Vranová et al., 2013).

The MEP pathway is essential for plant survival, asmutants defective in any of the MEP pathway genesare lethal (Mandel et al., 1996; de la Luz Gutiérrez-Nava et al., 2004; Guevara-García et al., 2005; Hsiehand Goodman, 2005, 2006; Hsieh et al., 2008; Tsenget al., 2013). Moreover, the antibiotic fosmidomycin,an inhibitor of the MEP pathway enzyme DEOXY-XYLULOSE PHOSPHATE REDUCTOISOMERASE(DXR), can efficiently kill plants (Rodríguez-Concepciónand Boronat, 2002). It is not surprising that the field ofMEP pathway research has advanced rapidly sinceits discovery two decades ago (Rohmer et al., 1993;Rodríguez-Concepción and Boronat, 2002; Oldfield, 2010).

The LytB (for lysis tolerant B) gene of Escherichia coliwas initially identified to be involved in penicillin tol-erance and control of the stringent response (Gustafsonet al., 1993). LytB/isoprenoid synthesis H (IspH) IspHwas later shown to be involved in isoprenoid biosyn-thesis (Cunningham et al., 2000; Altincicek et al., 2001;

1 This work was supported by the Ministry of Science and Tech-nology and Academia Sinica (grant no. 98–CDA–L04).

* Address correspondence to [email protected] 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:Ming-Hsiun Hsieh ([email protected]).

[W] The online version of this article contains Web-only data.[OPEN] Articles can be viewed online without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.114.243642

Plant Physiology�, September 2014, Vol. 166, pp. 57–69, www.plantphysiol.org � 2014 American Society of Plant Biologists. All Rights Reserved. 57 www.plantphysiol.orgon March 27, 2018 - Published by Downloaded from

Copyright © 2014 American Society of Plant Biologists. All rights reserved.

mailto:[email protected]

http://www.plantphysiol.org

mailto:[email protected]

http://www.plantphysiol.org/cgi/doi/10.1104/pp.114.243642


McAteer et al., 2001). In eubacteria, the final step of theMEP pathway involves the 2H+/2e2 reduction of4-hydroxy-3-methylbut-2-enyl diphosphate (HMBPP) toform a 6:1 mixture of IPP and DMAPP by the iron-sulfurprotein LytB/IspH or 4-HYDROXY-3-METHYLBUT-2-ENYL DIPHOSPHATE REDUCTASE (HDR; Adamet al., 2002; Altincicek et al., 2002; Rohdich et al., 2002,2003; Wolff et al., 2003). In addition to 2H+/2e2 reduction,IspH was recently shown to have acetylene hydrataseactivity (Span et al., 2012b). The physiological significanceof this unexpected activity is unknown.

The structure and reaction mechanism of IspH havebeen actively investigated since its discovery. Studiesfrom Aquifex aeolicus and E. coli indicate that IspHhas a trefoil-like structure with a central Fe4S4 cluster,which is coordinated by the Cys residues located ineach of the three folding domains (Rekittke et al., 2008;Gräwert et al., 2009). The central cavity is the activesite, and the iron-sulfur cluster serves as an electron-transfer cofactor in the active site (Gräwert et al., 2004,2009, 2010; Rekittke et al., 2008; Wang et al., 2010). TheIspH mechanism of action is still under debate, butthere is considerable evidence in support of the orga-nometallic hypothesis (Wang et al., 2010, 2012; Spanet al., 2012a; Xu et al., 2012; Li et al., 2013). The sub-strate HMBPP initially forms a hydroxy complexwith the iron-sulfur cluster, and a subsequent series ofcomplexes with rare organometallic bonds leads to thesynthesis of IPP and DMAPP (Wang et al., 2010, 2012;Li et al., 2013).

Similar to bacteria, HDR is an essential gene in plants.Virus-induced gene silencing was shown to knock downthe expression ofHDR in tobacco (Nicotiana tabacum), andthe resulting plants had albino leaves (Page et al., 2004).The Arabidopsis (Arabidopsis thaliana) HDR knockoutmutants are lethal (Guevara-García et al., 2005; Hsieh andGoodman, 2005). In maize (Zea mays), 35S:HDR antisensetransgene-induced gene-silencing plants are albino (Luet al., 2012). Compared with the bacterial enzyme, thestructure and reaction mechanism of plant HDR arelargely unknown. Nonetheless, the plant enzyme cansuccessfully rescue the E. coli ispHmutant, indicating thatplant HDR and bacterial IspH may have similar enzy-matic mechanisms (Hsieh and Goodman, 2005).

We previously reported that some of the Arabidopsistransgenic plants harboring 35S:HDR complementaryDNA (cDNA) had various albino patterns caused bygene silencing (Hsieh and Goodman, 2005). In this study,we usedHDR genomic DNA including its own promoterto generate transgenic Arabidopsis. Interestingly, some ofthe transgenic plants showed albino, variegated, or pale-green phenotype, indicating that the Arabidopsis HDRgene was very effective in triggering transgene-inducedgene silencing. To understand the reaction mechanism ofArabidopsis HDR, we used the crystallized IspH fromE. coli as a template to predict the structure of Arabi-dopsis HDR. Furthermore, the hydroxy complex formedby the substrate HMBPP and the iron-sulfur cluster wasdocked into the protein model. This molecular modelingrevealed key amino acids around the active site of

Arabidopsis HDR, which may be involved in substratebinding or catalysis. The importance of these key aminoacids was verified by a complementation test of theE. coli ispH mutant with mutated Arabidopsis HDRproteins generated by site-directed mutagenesis. Inaddition, we identified an extra N-terminal domain thatis highly conserved in cyanobacteria, green algae, andland plants but not in E. coli and the other bacteria. Weprovide in vivo evidence to show that the N-terminalconserved domain (NCD) is essential for Arabidopsisand cyanobacterial HDR. We further identified that theconserved Tyr-72 residue in the NCD was critical forArabidopsis HDR function. These results suggest thatthe structure and reaction mechanism of cyanobacterialand plant HDR are different from those of bacteria.

RESULTS

The Arabidopsis HDR Transgene Is Effective in InducingGene Silencing

The enzyme HDR catalyzes the last step of the MEPpathway that converts HMBPP into IPP and DMAPP.In Arabidopsis, HDR is encoded by a single-copy gene,and knockout mutants are albino lethal (Guevara-Garcíaet al., 2005; Hsieh and Goodman, 2005). We previouslyshowed that the expression of HDR was silenced inmany of the 35S:HDR cDNA transgenic plants, leadingto various albino phenotypes (Hsieh and Goodman,2005). Interestingly, when we introduced an approxi-mately 3.5-kb HDR genomic clone containing its ownpromoter into Arabidopsis wild-type plants, some T1transgenic plants exhibited a pale-green or variegatedphenotype (Fig. 1, A and B). The morphology of chlo-roplasts in the pale-green leaves of an HDR gene-silencing line ranged from lens shape with very thingrana to oval or round shape without any thylakoids(Fig. 1C). T2 and T3 progeny of these plants were palegreen, variegated, or albino (Supplemental Fig. S1). Inmost seedlings, the cotyledons were green or pale greenand their first leaves were variegated, light yellow, oralbino (Supplemental Fig. S1). The whitening of seed-lings progressed along the newly developed tissues, andthus very few seeds were recovered from these HDRtransgenic plants. Two independent transgenic lines, G16and G50, had progeny 100% showing the pale-green,variegated, or albino phenotype and were carried to theT4 generation for further analysis. In both G16 and G50lines, the cotyledons of T4 seedlings were pale green, lightyellow, or albino, indicating that the bleaching phenotypeof these lines occurred earlier and was more severe thanin previous generations (Fig. 1D, left). The PSII maximumquantum yield of G16 and G50 was significantly lowerthan that of the wild type (Fig. 1D, right). The amounts ofchlorophylls and carotenoids in G16 and G50 were aboutone-quarter of the wild-type levels (Fig. 1E). RNA gel-blotanalysis revealed that the expression of HDR was dra-matically reduced in G16 and G50 lines (Fig. 1F). Theseresults suggested that the expression of HDR was si-lenced in the HDR genomic DNA transgenic lines.

58 Plant Physiol. Vol. 166, 2014

Hsieh et al.

www.plantphysiol.orgon March 27, 2018 - Published by Downloaded from Copyright © 2014 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org/cgi/content/full/pp.114.243642/DC1



Sequence Analysis of Plant HDR and Bacterial IspH

The Arabidopsis HDR was able to rescue the lethalphenotype of the E. coli ispH mutant (Hsieh andGoodman, 2005), indicating that the function of this en-zyme is conserved between plants and bacteria. Wealigned IspH/HDR from E. coli, A. aeolicus, Rhodobactercapsulatus (purple bacterium), Synechocystis sp. PCC 6803(cyanobacterium), Chlamydomonas reinhardtii (green alga),Physcomitrella patens (bryophyte), Selaginella moellendorffii(lycophyte), Pinus densiflora (gymnosperm), tobacco (di-cotyledon), Arabidopsis (dicotyledon), rice (Oryza sativa;monocotyledon), and maize (monocotyledon) to identifythe invariant amino acid residues between plants andbacteria. The sequence alignment revealed that the cya-nobacterial HDR had an extra stretch of 53 amino acidsin the N terminus that were missing in E. coli, A. aeolicus,and R. capsulatus IspH. Interestingly, these amino acidsare highly conserved in cyanobacteria, green algae, andland plants (Fig. 2). Beyond the NCD, the HDR of green

algae and land plants had an extended N-terminal se-quence with various lengths that was not found in cya-nobacteria. These sequences were not highly conserved,which may serve as transit peptides to target plant HDRsto the chloroplast. The Arabidopsis IspH domain (e.g.amino acid residues 111 to 466, encompassing the bac-terial IspH) shares approximately 25% identity and ap-proximately 42% similarity with the E. coli protein. Manyamino acid residues found to be critical for E. coli andA. aeolicus IspH (Gräwert et al., 2004, 2009, 2010; Rekittkeet al., 2008, Wang et al., 2010) were also conserved inpurple bacteria, cyanobacteria, green algae, and landplants (Fig. 2). These invariant amino acids may playimportant roles such as iron-sulfur cluster formation,substrate binding, and catalysis in plant HDRs.

Homology Modeling of Arabidopsis HDR

We used the E. coli IspH (Protein Data Bank [PDB]code 3F7T) as a template for Arabidopsis HDR structure

Figure 1. HDR-induced gene silencing in transgenic Arabidopsis. A, Basta-resistant T1 transgenic plants harboring the ap-proximately 3.5-kb HDR genomic DNA transgene. The arrow indicates a pale-green transgenic plant. B, Three representativeT1 transgenic plants showing pale-green or variegated rosette leaves. C, Transmission electron micrographs of chloroplasts fromwild-type (WT) and pale-green rosette leaves of the G16 transgenic line. Bars = 500 nm. D, Light (left) and chlorophyll fluo-rescence (right) images of the wild type and T4 progeny of G16 and G50 HDR gene-silencing lines. An imaging chlorophyllfluorometer was used to measure PSII activity (maximum photochemical efficiency of PSII in the dark-adapted state [Fv/Fm]),and the values are indicated by the pseudocolor scale at the bottom. E, Contents of chlorophylls and carotenoids in G16 andG50 HDR gene-silencing lines. FW, Fresh weight. F, RNA gel-blot analysis of Arabidopsis HDR genomic DNA-induced gene-silencing lines G16 and G50.

Plant Physiol. Vol. 166, 2014 59

Critical Domain and Amino Acids for Plant HDR



Figure 2. Amino acid sequence alignment of plant HDR and bacterial IspH. At, Arabidopsis; Nt, tobacco; Os, rice; Zm, maize;Pd, Pinus densiflora; Sm, S. moellendorffii; Pp, Physcomitrella patens; Cr, C. reinhardtii; Sy, Synechocystis sp. PCC 6803; Rc,R. capsulatus; Ec, E. coli; Aa, A. aeolicus. The NCD among the oxygenic photosynthesis organisms is indicated at the top of thealignment. The arrow indicates the predicted cleavage site of the Arabidopsis HDR transit peptide. Arrowheads indicate thecritical Cys residues that are involved in iron-sulfur cluster formation. Asterisks indicate the conserved amino acids located inthe proximity of the substrate-binding site.

60 Plant Physiol. Vol. 166, 2014 www.plantphysiol.orgon March 27, 2018 - Published by Downloaded from

Copyright © 2014 American Society of Plant Biologists. All rights reserved.


modeling. The E. coli IspH is monomeric, which foldsinto three similar domains forming a trefoil-like struc-ture (Gräwert et al., 2009). Each folding domain consistsof four strands of parallel b-sheets in the center flankedby three a-helices (Fig. 3A, left). Although these do-mains have similar folding patterns, the primary se-quences are not similar. We used the software AccelrysDiscovery Studio 2.5.5 to construct a three-dimensionalmodel of Arabidopsis HDR. The N-terminal region (e.g.amino acids 1–110) of Arabidopsis HDR does not existin E. coli IspH. So we only used the IspH domain(amino acids 111–466) of Arabidopsis HDR to performhomology modeling. The predicted IspH domain ofArabidopsis HDR showed a trefoil-like structure similarto that of E. coli (Fig. 3A, right). The three domainsprotrude into a cavity at the center of the protein, whereit possibly coordinates an iron-sulfur cluster and HMBPPsubstrate as E. coli IspH (Gräwert et al., 2009). The mostapparent differences were at the tip of the three domains,where the Arabidopsis protein had extended structures(Fig. 3A). These extensions correspond to the mismatchgaps between the Arabidopsis and E. coli proteins(Supplemental Fig. S2).

Molecular Docking of HMBPP in Arabidopsis HDR

It has been proposed that the substrate of IspH,HMBPP, initially forms a hydroxy complex with theFe4S4 center (Wang et al., 2010). In Arabidopsis HDR,the iron-sulfur cluster appeared in close approxima-tion with three Cys residues, Cys-122, Cys-213, andCys-350, as it was in the bacterial protein (e.g. Cys-12,Cys-96, and Cys-197 of E. coli IspH; Fig. 3B). Dockingof HMBPP to the central active site of ArabidopsisHDR revealed that the amino acid residues His-152,His-241, Glu-242, Glu-243, Thr-244, Thr-312, Ser-379, andAsn-381 were located around the substrate (Fig. 3B). Mostof these amino acids are absolutely conserved betweenbacteria and plants, indicating that these residues may beinvolved in substrate binding or the catalysis of HMBPPto IPP and DMAPP.

Cys-122, Cys-213, and Cys-350 Are Essential forArabidopsis HDR

We used site-directed mutagenesis to generate threeindependent constructs that would produce Arabi-dopsis HDR mutant proteins harboring Cys-122Gly,Cys-213Gly, and Cys-350Gly. These constructs wereseparately transformed into the E. coli ispH mutant totest if they could complement the mutant. In the E. coliispH mutant, the endogenous IspH gene was knockedout and an engineered IspH gene under the control ofthe arabinose-inducible PBAD promoter was present onthe chromosome (McAteer et al., 2001). Because theIspH gene is essential for survival, the growth of theE. coli ispH mutant can be controlled by the addition ofAra or Glc in the medium. In the presence of 0.2% (w/v)

Ara, the engineered IspH gene located on the bacterialchromosome would be induced, and all E. coli ispHmutant strains transformed with plasmids containingwild-type or mutated Arabidopsis HDR cDNAs wouldgrow and form colonies (Fig. 4, left). In the presence of0.2% (w/v) Glc, the expression of wild-type ArabidopsisHDR could restore the growth of the E. coli ispH mutant,but none of the Cys-122Gly, Cys-213Gly, and Cys-350Gly mutant proteins was able to complement themutant (Fig. 4, right). These results indicate that the Cys-122, Cys-213, and Cys-350 residues are essential forArabidopsis HDR function.

Some Key Amino Acids Are Specific to Plant HDR

In addition to Cys-122, Cys-213, and Cys-350, wealso examined the functions of His-152, His-241, Glu-242,

Figure 3. Homology modeling of Arabidopsis HDR. A, The E. coliIspH (PDB code 3F7T) was used as a template (left) for homologymodeling of the IspH domain (amino acid residues 111–466) ofArabidopsis HDR (right). Red and blue ribbons represent a-helices andb-strands, respectively. The center of the three lobes is the active site forHMBPP substrate binding. N, N terminus. B, Closeup view of active-siteresidues in Arabidopsis HDR determined by homology modeling inrelation to modeled Fe4S3 cluster and HMBPP (shown in color sticks)substrate molecules. Amino acid residues Cys-122, His-152, Cys-213,His-241, Glu-242, Glu-243, Thr-244, Cys-350, Thr-312, Ser-379,and Asp-381 surrounding the substrate-binding site were analyzedin this study. This figure was created with Accelrys Discovery Studio2.5.5.






Glu-243, Thr-244, Thr-312, Ser-379, and Asn-381, whichwere predicted to locate in the central active site ofArabidopsis HDR (Fig. 3B). These amino acid residues,except Glu-242 and Thr-244, are absolutely conservedfrom bacteria to plants (Fig. 2). We used site-directedmutagenesis to generate constructs encoding ArabidopsisHDR His-152Asn, His-241Asn, Glu-242Lys, Glu-243Lys,Thr-244Pro, Thr-312Pro, Ser-379Pro, Asn-381His, andHis-152Asn/His-241Asn mutants. The Glu-242Lys,Glu-243Lys, Thr-244Pro, Thr-312Pro, Ser-379Pro, andAsn-381His mutant constructs were unable to com-plement the E. coli ispHmutant (Supplemental Fig. S3,A and B), indicating that Glu-242, Glu-243, Thr-244,Thr-312, Ser-379, and Asn-381 are critical for Arabi-dopsis HDR function.

Unexpectedly, Arabidopsis HDR His-152Asn andHis-241Asn single mutants but not the His-152Asn/His-241Asn double mutant were able to complementthe E. coli ispHmutant (Supplemental Fig. S3C). Althoughthe Arabidopsis HDR His-152Asn mutant was able torescue the E. coli ispH mutant, the complemented straingrew poorly. We used liquid culture supplemented with0.2% (w/v) Glc to measure the growth rate of E. coli ispHmutant strains containing Arabidopsis wild-type, His-152Asn, or His-241Asn HDR protein. Consistent withthe plate colony-forming assay, the growth rate of theHis-241Asn mutant complemented strain was similar tothat of the wild type, whereas the His-152Asn mutantcomplemented strain grew poorly in liquid culture(Supplemental Fig. S3D). These results suggest that His-152 is more important than His-241 and that mutations inboth His-152 and His-241 will abolish the function ofArabidopsis HDR. This is in contrast to the reactionmechanism of E. coli IspH, in which the correspondingHis-41 and His-124 are critical residues (Wang et al.,2010). The results of these complementation assays aresummarized in Table I.

Arabidopsis HDRD110 Failed to Complement the E. coliispH Mutant

The Arabidopsis HDR protein consisted of a puta-tive transit peptide, the NCD (amino acids 56–110),and the IspH domain (amino acids 111–466; Fig. 5A).Similarly, transit peptides also exist in the otherArabidopsis MEP pathway enzymes, as revealed byamino acid sequence alignments (Supplemental Fig. S4).These analyses indicated that the cyanobacterial MEPpathway enzymes aligned well with those of E. coliexcept in the N-terminal region of HDR (SupplementalFig. S4). Thus, the NCD of plant and cyanobacterialHDR is unique among the MEP pathway enzymes,which may have specific functions in the oxygenicphotosynthetic lineage.

We generated a series of Arabidopsis HDR N-terminaldeletion constructs, HDRD24, HDRD55, HDRD110,and HDRD122, to examine the function of the NCD.The predicted cleavage site of the Arabidopsis HDRtransit peptide is located at the 38th amino acid resi-due. In planta, the mature Arabidopsis HDR in thechloroplast will be without the transit peptide. Weconstructed a cDNA clone encoding the ArabidopsisHDR protein without the first 24 amino acid residues(e.g. HDRD24), which was able to complement theE. coli ispH mutant (Hsieh and Goodman, 2005). Ara-bidopsis HDRD24, considered as a wild-type protein,was used as a positive control in this study. In thepresence of 0.2% (w/v) Glc, Arabidopsis HDRD55 thatretained the complete NCD and the IspH domain wasable to rescue the E. coli ispH mutant (Fig. 5B). As ex-pected, Arabidopsis HDRD122 without the criticalCys-122 residue failed to complement the E. coli ispHmutant (Fig. 5B). Surprisingly, Arabidopsis HDRD110,corresponding to the full-length E. coli IspH, also failedto complement the E. coli mutant (Fig. 5B). These re-sults indicated that amino acid residues 56 to 110 (e.g.the NCD) were critical for Arabidopsis HDR function.

The NCD Is Critical for Cyanobacterial HDR Function

To examine if the NCD is critical for cyanobacterialHDR function, we cloned a full-length HDR cDNA fromSynechocystis sp. PCC 6803 (SyHDR) and generatedtwo deletion constructs, SyHDRD54 and SyHDRD66(Fig. 6A). These clones were used for a complementationassay to test if they could rescue the E. coli ispH mutant.The full-length SyHDR was able to complement theE. coli mutant (Fig. 6B). However, the NCD deletionconstruct (SyHDRD54) and the NCD plus the conservedCys-66 deletion construct (SyHDRD66) failed to rescuethe E. coli ispH mutant (Fig. 6B). These results confirmedthat the NCD is also critical for HDR function in Syne-chocystis sp. PCC 6803.

Essential Role of the Arabidopsis HDR NCD

To further confirm the NCD function in planta, wetested if the Arabidopsis HDRDNCD could complement

Figure 4. Arabidopsis HDR C122G, C213G, and C350G mutantproteins failed to complement the E. coli ispHmutant. The E. coli ispHmutant strain complemented with the Arabidopsis HDR wild type(WT) or the C122G, C213G, or C350G mutant was able to grow onmedium containing 0.2% (w/v) Ara (left), but only the wild typecomplemented strain was able to grow on medium containing 0.2%(w/v) Glc (right). Cys-122, Cys-213, and Cys-350 of Arabidopsis HDRcorrespond to Cys-12, Cys-96, and Cys-197 of E. coli IspH, which areinvolved in iron-sulfur cluster formation.


Hsieh et al.










the hdr-1 albino mutant. The 35S:HDRD(562110) con-struct, which encodes a recombinant protein con-taining the putative transit peptide (amino acids 1–55)and the IspH domain (amino acids 111–466), was trans-formed into Arabidopsis hdr-1 heterozygous plants. TheArabidopsis hdr-1 mutant is resistant to kanamycin, andthe 35S:HDRD(562110) construct carries a hygromycinselection marker. So T1 seeds were germinated andgrown on hygromycin plus kanamycin selective mediumto screen for successful transformants in the hdr-1mutant(2/2 or +/2) background. If the 35S:HDRD(562110)transgene was able to complement the hdr-1 albino mu-tant, all the hygromycin-resistant (HygR)/kanamycin-resistant (KanR) transformants should be green. Whilethe majority of T1 transformants were green, some HygR/KanR albino seedlings did appear during the screening(data not shown). This result indicated that the 35S:HDRD(562110) transgene failed to complement the hdr-1mutant. Some of the 35S:HDRD(562110) transgenicplants were carried to T3 homozygosity. Progeny of arepresentative transgenic line, which is hdr (+/2) hetero-zygous and 35S:HDRD(562110) transgene homozy-gous, segregated mainly green and albino seedlings on ahygromycin selective medium (Fig. 7A). The presenceof homozygous HygR albino seedlings implied that theHDRDΝCD protein, if expressed and localized to thechloroplast, is not functional in Arabidopsis. Interest-ingly, a few HygR variegated or pale-green seedlings alsoappeared in the progeny (Fig. 7A). We examined twoadditional 35S:HDRD(562110) homozygous trans-genic lines and found that green, pale-green, variegated,and albino seedlings all appeared in the progeny(Supplemental Fig. S5). The phenotypes of these trans-genic plants suggested that the 35S:HDRD(562110)

transgene was effective in triggering gene silencing.Thus, the NCD is not required for HDR transgene-induced gene silencing in Arabidopsis.

To confirm whether the 35S:HDRD(562110) trans-gene was properly expressed, we extracted total RNAand proteins from the HygR albino seedlings for furtheranalysis. The original hdr-1 (2/2) albino mutants donot have detectable HDR transcripts or proteins (Hsiehand Goodman, 2005). Reverse transcription (RT)-PCRand sequencing results confirmed that these HygR

albino seedlings contained HDRD(562110) transcripts,which were approximately 150 bp shorter than those ofthe wild type (Fig. 7B). Immunoblot analysis revealedthat the HDRD(562110) but not the wild-type proteinwas detected in the HygR albino seedlings (Fig. 7C).Moreover, the abundance of HDRD(562110) protein inthe HygR albino seedlings was lower than that of HDRin the wild type (Fig. 7C).

Although the HDRD(562110) protein was expressedin the transgenic plants, it was not clear if the NCD-deleted protein was properly localized to the chloroplast.We previously showed that the first 52 amino acids ofArabidopsis HDRwere sufficient to target the fused GFPto the chloroplast (Hsieh et al., 2008). So the expressedHDRD(562110) protein in the HygR albino seedlingsthat still contains the first 55 amino acids should have afunctional transit peptide to target the protein to the

Table I. Complementation of the E. coli ispH mutant by ArabidopsisHDR

+, Complemented; 2, not complemented; +*, complemented butgrew slowly; n.a., not available.

Arabidopsis

HDR Protein

Complementation of the

E. coli ispH Mutant

Corresponding E. coli

IspH Residue

Wild type +Cys-122Gly 2 Cys-12Cys-213Gly 2 Cys-96Cys-350Gly 2 Cys-197His-152Asn +* His-41His-241Asn + His-124His-152Asn/

His-241Asn2 His-41/His-124

Glu-242Lys 2 Pro-125Glu-243Lys 2 Glu-126Thr-244Pro 2 Val-127Thr-312Pro 2 Thr-167Ser-379Pro 2 Ser-225Asn-381His 2 Asn-227Asp-58His + n.a.Leu-66Phe + n.a.Tyr-72Ser 2 n.a.Tyr-91Ser + n.a.Gly-109Arg +* n.a.

Figure 5. Functional analysis of the NCD of Arabidopsis HDR in E. coli.A, Schematic diagram of the full-length Arabidopsis HDR. TP, Transitpeptide; IspH, amino acid (a.a.) residues 111 to 466 that encompass theE. coli IspH protein. B, The NCD is required for Arabidopsis HDR tocomplement the E. coli ispH mutant. The E. coli ispH mutant straincomplemented with the wild-type Arabidopsis HDR (HDRD24) orHDRD55 was able to grow on medium containing 0.2% (w/v) Ara (left)or 0.2% (w/v) Glc (right). The Arabidopsis HDR deletion mutantHDRD110 and HDRD122 failed to rescue the E. coli ispH mutant in thepresence of 0.2% (w/v) Glc (right).






chloroplast. Nonetheless, we generated a 35S:HDRD(562110)-GFP construct and used an Arabidopsis pro-toplast transient expression assay to examine the sub-cellular localization of HDRD(562110)-GFP. The greenfluorescent signals colocalized with the autofluorescentsignals of chlorophyll in the transformed protoplast, in-dicating that the HDRD(562110)-GFP fusion proteinwas localized to the chloroplast (Fig. 7D). Together,these results suggested that the Arabidopsis HDRD(562110) protein was expressed and properly localized to thechloroplast but failed to rescue the hdr-1 mutant.

Important Amino Acid Residues in the NCD

Alignment of HDRs from Arabidopsis and sevenrepresentative cyanobacterial species belonging todifferent orders revealed that the NCD exists in allcyanobacteria, including Gloeobacter violaceus (Fig. 8A).The HDR of G. violaceus is slightly shorter in the NCDthan that of the other cyanobacteria, which is consis-tent with the notion that G. violaceus is an early-branching cyanobacterium that diverged from othercyanobacteria before the emergence of plant plastids(Nelissen et al., 1995). The sequence alignment alsorevealed that Asp-58, Leu-66, Tyr-72, Tyr-91, and Gly-109 located in the NCD of Arabidopsis HDR are ab-solutely conserved among cyanobacteria, green algae,and land plants (Figs. 2 and 8A). We further used site-directed mutagenesis and an E. coli complementation

assay to examine the importance of these conserved re-sidues. Interestingly, Arabidopsis HDR Tyr-72Ser failedto rescue the E. coli ispH mutant (Fig. 8B). ArabidopsisHDR Gly-109Arg was able to complement the E. colimutant, but it only formed small colonies (Fig. 8B). Bycontrast, the Asp-58His, Leu-66Phe, and Tyr-91Ser mu-tants did not affect the function of Arabidopsis HDR(Fig. 8B). These results suggest that Tyr-72 is critical andGly-109 is important for Arabidopsis HDR function.

DISCUSSION

The Arabidopsis HDR Transgene Is Effective in InducingGene Silencing

We previously showed that the Arabidopsis 35S:HDR cDNA transgene was very effective in triggeringgene silencing (Hsieh and Goodman, 2005). Here, wedemonstrated that transformation of an approximately3.5-kb HDR genomic clone into wild-type plants couldalso induce gene silencing in Arabidopsis. Some trans-genic plants harboring this genomic clone are albino,pale green, or variegated, and the silencing effect wasmore profound in younger tissues. The clonal sectors ofalbino/green tissues in the variegated plants indicatedthat some of the silencing effects were stochastic andcell autonomous. These phenotypes suggested that bothtranscriptional and posttranscriptional gene-silencingmechanisms could be involved in these lines.

In addition to HDR, the DXR gene encoding thesecond enzyme of the MEP pathway was able to triggertransgene-induced gene silencing in Arabidopsis(Carretero-Paulet et al., 2006). Nonetheless, in HDRtransgene-induced gene-silencing lines, the degree ofbleaching correlates with the reduction of HDR tran-scripts. Thus, it is important for HDR to maintain suffi-cient amounts to fulfill its function. The expression andactivity of HDR should be tightly regulated during plantgrowth and development. It will be interesting to iden-tify those components that are involved in the regulationof HDR gene expression and enzyme activity.

Structure and Enzymatic Mechanism of HDR: Similarityand Difference between Plants and Bacteria

The E. coli IspH is an iron-sulfur protein, and theconserved Cys-12, Cys-96, and Cys-197 residues areinvolved in iron-sulfur cluster formation (Wolff et al.,2003; Gräwert et al., 2004, 2009; Rekittke et al., 2008).These Cys residues are also conserved in cyanobacte-ria, green algae, and land plants. Homology modelingof Arabidopsis HDR revealed that the correspondingCys-122, Cys-213, and Cys-350 residues were locatedin the central active site, which was critical for itsfunction. It is likely that the Arabidopsis HDR is alsoan iron-sulfur protein. Similar to the reaction mecha-nism of E. coli IspH, the conserved Cys-122, Cys-213,and Cys-350 residues are involved in the formation ofthe iron-sulfur cluster in Arabidopsis HDR.

Figure 6. The NCD is critical for cyanobacterial HDR. A, Schematicdiagrams of Synechocystis sp. PCC 6803 HDR. IspH, Amino acid (a.a.)residues 55 to 406 corresponding to the full-length E. coli IspH; C66,C157, and C288 are critical Cys residues involved in iron-sulfur clusterformation. B, The wild-type cyanobacterial HDR (SyHDR), but not theNCD deletion constructs SyHDRD54 and SyHDRD66, was able tocomplement the E. coli ispH mutant. The E. coli ispH mutant com-plemented with the wild-type Arabidopsis HDR (AtHDR) was used as apositive control.


Hsieh et al.



In E. coli and A. aeolicus IspH, Glu-126 is a key catalyticresidue that delivers H+ to the active site, and Thr-167,Ser-225, and Asn-227 are involved in substrate binding(Gräwert et al., 2009, 2010; Wang et al., 2010). Theseamino acids are conserved in cyanobacteria, green algae,and land plants. We confirmed that these conservedresidues are critical for Arabidopsis HDR function, asmutations in any of these residues (e.g. Glu-243Lys,Thr-312Pro, Ser-379Pro, and Asn-381His) failed tocomplement the E. coli ispH mutant. It is possible thatthe conserved Glu-243 in Arabidopsis HDR may alsofunction as a key catalytic residue involved in H+ de-livery. In homology modeling of Arabidopsis HDR,Thr-312, Ser-379, and Asn-381 were located in thecentral active site surrounding the HMBPP (Fig. 3B),which was consistent with the predicted role of theseamino acids in substrate binding.In addition to the above amino acids, His-41 and

His-124 also play important roles in E. coli andA. aeolicus IspH. These two residues are absolutely

conserved, but their functions may have evolvedin plants. In bacterial IspH, His-41 is involved in sub-strate binding, whereas His-124 is required for deliv-ering H+ to the key residue Glu-126 and the substrateHMBPP during catalysis (Rekittke et al., 2008; Gräwertet al., 2009, 2010; Wang et al., 2010). In E. coli IspH, theHis-41Asn mutant did not have detrimental effects,but the activity of IspH was undetectable in the His-124Asn mutant (Gräwert et al., 2009). We made thesame amino acid substitutions (e.g. His-152Asn andHis-241Asn) in Arabidopsis HDR. If the functions ofthese two residues were conserved, Arabidopsis HDRHis-152Asn but not His-241Asn should be able to res-cue the E. coli ispH mutant. Surprisingly, the Arabi-dopsis HDR His-241Asn mutant was fully functional inthe E. coli complementation assay. While the function ofArabidopsis HDR was partially lost in the His-152Asnsingle mutant, it was completely lost in the His-152Asn/His-241Asn double mutant. Still, these resultsare consistent with the notion that His-152 and His-241

Figure 7. The NCD is required for Arabidopsis HDR to complement the hdr-1 albino mutant. A, Progeny of a representative hdr(+/2) heterozygous line harboring the homozygous 35S:HDRD(562110) transgene segregated green and albino HygR seedlingson a selective medium. The 35S:HDRD(562110) transgene encodes the Arabidopsis HDR without the NCD (amino acidresidues 56–110). B, RT-PCR analysis of HDR in the wild type (WT) and a 35S:HDRD(562110) albino transgenic line. Trun-cated, but not full-length, HDR transcripts were detected in HygR albino seedlings, indicating that the 35S:HDRD(562110)transgene failed to complement the hdr-1 mutant. Elongation factor 1 a (EF1a) was used as a control for RT-PCR analysis.C, Immunoblot analysis of HDR in the wild type and a 35S:HDRD(562110) albino transgenic line. The arrowhead and arrowindicate wild-type HDR and HDRD(562110), respectively. Immunoblot analysis of actin (ACT) and a Coomassie Brilliant Blue-stained (CBB) gel of the same samples are shown as loading controls. D, HDRD(562110)-GFP is localized to the chloroplast.DIC, Differential interference contrast. Bars = 10 mm.





are important for Arabidopsis HDR function. However,in contrast to the critical role of His-124 in E. coli IspH,the corresponding His-241 residue is not essential forArabidopsis HDR. If His-241 is involved in H+ delivery,its role may be limited or replaceable by other aminoacids in Arabidopsis HDR.

Although His-241 is not essential, the neighboringGlu-242, Glu-243, and Thr-244 residues are critical forArabidopsis HDR function. Amino acids Glu-242 andThr-244 flanking the key Glu-243 residue of Arabi-dopsis HDR are conserved in cyanobacteria, green al-gae, and land plants but not in E. coli, A. aeolicus, andR. capsulatus (Fig. 2). The Arabidopsis HDR may useGlu-242 and Thr-244, instead of His-241, to deliver H+

to the key residue Glu-243 and the substrate HMBPPduring catalysis. Homology modeling of ArabidopsisHDR revealed that His-152, His-241, Glu-242, and Thr-244 were located around the substrate HMBPP in thecentral active site (Fig. 3B). Alternatively, these res-idues, together with Thr-312, Ser-379, and Asn-381, mayform a bonding network with the substrate HMBPP.The HEET motif in the central active site is specificallyconserved in oxygenic photosynthesis organisms, in-dicating that the catalytic mechanism of HDR mayhave evolved independently in the green lineage oflife. It will be interesting to further examine the func-tions of these amino acids and the reaction mechanismof plant HDR.

A Critical Role of the NCD for Cyanobacterial andArabidopsis HDR

In addition to key amino acids in the central activesite, we have identified that the NCD and the Tyr-72residue in the NCD are critical for Arabidopsis HDR.The NCD is present in cyanobacteria but not in otherprokaryotes. So the occurrence of this domain shouldbe early in evolution. It is possible that the cyano-bacterial HDR may evolve independently from thecommon ancestor of prokaryotes to obtain the NCD,which remains an essential part of the enzyme alongthe evolution of green algae and land plants. The NCDis present in all cyanobacterial HDRs (Fig. 8), sug-gesting that this domain is not part of the transitpeptide in plants. Furthermore, we showed that Ara-bidopsis HDR(D56–110)-GFP was localized to thechloroplast (Fig. 7D). So the NCD is not required fortargeting HDR to the chloroplast in plants. This is incontrast to the Pro-rich motif located in the N-terminalregion of plant DXR, which may be required for tar-geting the protein to the chloroplast and to the thyla-koid lumen (Carretero-Paulet et al., 2002; Fung et al.,2010). Moreover, the Pro-rich motif is present in DXRsof green algae and land plants but not in the cyano-bacterial protein (Supplemental Fig. S4).

The exact function of the NCD has yet to be estab-lished. Analysis of 35S:HDR(D56–110) transgenic plants

Figure 8. The Tyr-72 residue located in the NCD is critical for Arabidopsis HDR function. A, Alignment of the NCD of HDRfrom Arabidopsis (At) and seven representative cyanobacteria: Sy, Synechocystis sp. PCC 6803; Pl, Pleurocapsa sp. PCC 7319;Ac, Anabaena cylindrica PCC 7122; Fm, Fischerella muscicola; Oc, Oscillatoria sp. PCC 6506; Gv, G. violaceus; Pm, Pro-chlorococcus marinus strain MIT 9303. Asterisks indicate the conserved amino acids in cyanobacteria, green algae, and landplants. B, E. coli complementation assay of Arabidopsis HDR D58H, L66F, Y72S, Y91S, and G109R mutants. The ArabidopsisHDR Y72S construct failed to complement the E. coli ispH mutant. The Arabidopsis HDR G109R complemented strain grewslowly and only formed small colonies. WT, Wild type.


Hsieh et al.




revealed that the NCD was not required for HDRtransgene-induced gene silencing. The amount ofHDRDΝCD protein in 35S:HDR(D56–110) transgenicalbino plants was lower than that of HDR in wild-typeplants, suggesting that the NCD might play an impor-tant role in protein stability or that the HDRDΝCDtranscript is less efficient in translation (Fig. 7C). Inter-estingly, the maize zebra7 mutant, which has transversegreen/yellow striped leaves, is caused by an Arg-64Cysmutation in the NCD of maize HDR. The authors sug-gest that the mutant HDR is less effective or unstable inthe zebra7 mutant (Lu et al., 2012). Nonetheless, it willbe interesting to further examine the molecular mech-anism concerning the regulation of plant HDR stabilityby the NCD.Since the NCD is essential for Arabidopsis HDR, it

may be involved directly in catalytic reaction or sub-strate binding rather than functioning as a regulatorydomain. In E. coli IspH, the N-terminal strand plays akey role for structural cohesion (Gräwert et al., 2009).The addition of extended amino acids, including theNCD, to the N-terminal strand would create a differentstructure in Arabidopsis HDR. The NCD of Arabi-dopsis HDR was predicted to form two to threea-helices and one b-sheet (Supplemental Fig. S2). Thisstructure may interact with the trefoil-like domain andchange the conformation of the central active site.Therefore, the passage of substrate and product or thehydrogen bonding of substrate may be different be-tween plant HDR and E. coli IspH. Alternatively, theNCD may interact with the central active site or itselfmay be involved in the formation of the active site.Some of the amino acid residues in the NCD, suchas Tyr-72 and Gly-109, may be directly involved insubstrate binding or catalytic reactions in Arabi-dopsis HDR.Although the involvement of HDR in the last step of

the MEP pathway is conserved from bacteria to plants,the structure and reaction mechanism of this enzymemay be different between these two domains of life.The occurrence of the NCD and some critical aminoacids is specific to the green lineage of life. Furtherstudies, especially crystal structure analysis, may pro-vide insights into the enzymatic mechanism of plantand cyanobacterial HDR.

MATERIALS AND METHODS

Nomenclature

To conform to the nomenclature of HDR in plants (Phillips et al., 2008), theArabidopsis (Arabidopsis thaliana) IspH homolog and the ispH-1 mutant (Hsiehand Goodman, 2005) have been renamed HDR and hdr-1, respectively.

Plant Materials and Growth Conditions

Arabidopsis (ecotype Columbia-0) was grown on one-half-strength Mura-shige and Skoog plates (Murashige and Skoog salts [Phytotechnology Labo-ratories], pH adjusted to 5.7 with 1 N KOH, and 0.8% [w/v] phytoagar)containing 2% Suc in a growth chamber or in soil in the greenhouse on a 16-h-light/8-h-dark cycle at 23°C. Primers 59-AGAGAGGGAATGTACGGAAG-39

and 59-GGTAAGAACATTAAGTGGAG-39 were used to amplify the HDRgenomic DNA, and the 3,735-bp PCR product was cloned into the TOPO2.1 vector (Invitrogen). An approximately 3.5-kb SpeI fragment harboring theHDR genomic DNA containing 903 bp upstream of the ATG start codon and153 bp downstream of the stop codon was subcloned into the XbaI-digestedpCambia 3201 vector and transformed into Agrobacterium tumefaciens GV3101.The HDR genomic clone was transformed into Arabidopsis ecotype Columbia-0by floral dip. T1 seeds were sown in soil, and Basta (120 mg L21) wassprayed twice on 10- and 17-d-old seedlings. About one-third of Basta-resistant T1 plants showed the pale-green, variegated, or albino phenotype.T2 seeds harvested from 30 independent T1 pale-green lines were furtheranalyzed. Of these 30 lines, progeny of G16 and G50 all showed the pale-green, variegated, or albino phenotype.

Analysis of HDR Genomic DNA-InducedGene-Silencing Lines

Rosette leaves from the soil-grown wild type and a pale-green HDR ge-nomic DNA-induced gene-silencing line were used for transmission electronmicroscopy analysis according to methods described previously (Hsieh andGoodman, 2005). Twelve-day-old wild-type, G16, and G50 seedlings grown intissue culture medium were used for chlorophyll fluorescence imaging andphotosynthetic pigment analysis. Chlorophyll fluorescence imaging and analysiswere performed with the Maxi-Imaging-PAM Chlorophyll Fluorometer (HeinzWalz). Plants were dark adapted for 20 min before measuring the PSII maxi-mum quantum yield. Determination of total chlorophylls and carotenoids wasperformed as described (Lichtenthaler and Wellburn, 1983).

RNA Gel-Blot and RT-PCR Analyses

Total RNA extraction and RNA gel-blot analysis ofHDRwere conducted asdescribed (Hsieh and Goodman, 2005). Digoxigenin probe labeling, prehy-bridization, hybridization, wash conditions, and detection were according tothe Boehringer Mannheim Genius System User’s Guide, version 3.0. For RT-PCR analysis, total RNA was digested with DNase I, and RT was performedwith SuperScript III reverse transcriptase (Invitrogen) according to the manu-facturer’s instructions. Oligo(dT) was used in RT to synthesize the first-strandcDNA. Primers 59-ATGGCTGTTGCGCTCCAATTC-39 and 59-TCAAGCCAG-CTGCAATAACTC-39 were used to amplify wild-typeHDR (approximately 1.4 kb)and HDRD(56–110) transgene (approximately 1.25 kb) cDNAs in Figure 8B.

Sequence Alignment of HDR/IspH Proteins

HDR/IspH amino acid sequences from Arabidopsis (AAN87171), tobacco(Nicotiana tabacum; AF159699), rice (Oryza sativa; NM_001057702), maize (Zeamays; NM_001175829), Pinus densiflora (ACC54561), Selaginella moellendorffii(XP_002960319), Physcomitrella patens (XM_001758317), Chlamydomonas rein-hardtii (XP_001701302), Synechocystis sp. PCC 6803 (NP_442089), Rhodobactercapsulatus (ADE87147), Aquifex aeolicus (O67625), and Escherichia coli (NP_414570)were aligned with ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/)and BoxShade (http://www.ch.embnet.org/software/BOX_form.html; Fig. 2). Toconfirm that the NCD exists in all cyanobacteria, HDR sequences from Arabi-dopsis (AAN87171) and seven arbitrarily selected cyanobacterial species belong-ing to different orders (http://www.ncbi.nlm.nih.gov/taxonomy) were alignedwith ClustalW2 and BoxShade. The accession numbers of these cyanobacterialHDRs are as follows: WP_011143289, Gloeobacter violaceus (Gloeobacterales);AFZ57717, Anabaena cylindrica PCC 7122 (Nostocales); NP_442089, Synechocystissp. PCC 6803 (Chroococcales); CBN54263,Oscillatoria sp. PCC 6506 (Oscillatoriales);WP_019504311, Pleurocapsa sp. PCC 7319 (Pleurocapsales); YP_001018478,Prochlorococcus marinus strain MIT 9303 (Prochlorales); and WP_016869831,Fischerella muscicola (Stigonematales). Only the NCD of the resulting alignment isshown in Figure 8A.

Homology Modeling of Arabidopsis HDR

The crystal structure of E. coli IspH (PDB code 3F7T; Gräwert et al., 2009) wasused as a template for homology modeling of Arabidopsis HDR using thesoftware Discovery Studio version 2.5.5 (Accelrys). Arabidopsis HDR has anextra N-terminal sequence (e.g. amino acid residues 1–110) compared with thatof E. coli (Fig. 2), which was not included in molecular modeling. The predictedstructure of Arabidopsis HDR is shown in Figure 3A. A substrate-binding site





http://www.ebi.ac.uk/Tools/msa/clustalw2/

http://www.ch.embnet.org/software/BOX_form.html

http://www.ncbi.nlm.nih.gov/taxonomy


was also predicted using Discovery Studio version 2.5.5 with the Receptor-Ligand panel in the program. The cavity of the possible substrate-binding sitewas predicted, and then the orientation of the substrate, and HMBPP wasdocked into the predicted cavity using the CDocker panel in the program. Thebest predicted result is shown in Figure 3B.

Generation of Arabidopsis HDR Mutant Proteins

The Arabidopsis HDR cDNA digested with BamHI and SacI was clonedinto pQE-30 (Qiagen) to express a His-tagged HDR protein missing the first 24amino acid residues. The resulting clone, pQE-AtHDR, was able to comple-ment the E. coli ispH mutant (Hsieh and Goodman, 2005). pQE-AtHDR wasused as a template to generate Arabidopsis HDR mutant constructs by site-directed mutagenesis using the Quickchange II Site-Directed Mutagenesis Kit(Agilent Technologies) according to the manufacturer’s instructions. Only onenucleotide is mutated at a time for each amino acid residue of interest. TheHDR His-152Asn/His-241Asn double mutant was constructed with an addi-tional round of mutagenesis using the His-152Asn single mutant as a tem-plate. The mutated pQE-AtHDR clones were verified by sequencing. Primersused for site-directed mutagenesis to generate Arabidopsis HDR Asp-58His,Leu-66Phe, Tyr-72Ser, Tyr-91Ser, Gly-109Arg, Cys-122Gly, His-152Asn, Cys-213Gly, His-241Asn, Glu-242Lys, Glu-243Lys, Thr-244Pro, Thr-312Pro, Ser-379Pro, Asn-381His, and His-152Asn/His-241Asn mutant constructs are listedin Supplemental Table S1. pQE-AtHDR was used as a template to generate aseries of N-terminal deletion constructs by PCR. Primers containing BamHI andSacI restriction sites used to generate cDNAs encoding HDRD55, HDRD110, andHDRD122 mutant proteins are listed in Supplemental Table S1. PCR productsdigested with BamHI and SacI were cloned into the similarly cut pQE-30 vector(Qiagen) and verified by sequencing.

Generation of Cyanobacterial HDR Mutant Proteins

Primers 59-CCTTGGATCCATGGATACCAAAGCTT-39 and 59-CCTTGAGCTCCTATCCCGCAATTTC-39 were used to amplify the full-length HDR genefrom Synechocystis sp. PCC 6803 by PCR. The PCR products were digested withBamHI and SacI and ligated to the pQE-30 vector (Qiagen) that was cut by the samerestriction enzymes. The full-length SyHDR cDNA was verified by sequencing andwas used as a template to generate SyHDRD54 and SyHDRD66 mutant constructsby PCR with primers 59-CCTTGGATCCGTCACCATTCTGTTGGC-39/59-CCTTGAGCTCCTATCCCGCAATTT-39 and 59-CCTTGGATCCTGGGGCGTG-GAGCGAGCC-39/59-CCTTGAGCTCCTATCCCGCAATTTC-39, respectively. ThePCR products were digested with BamHI and SacI, cloned into pQE-30 vector, andverified by sequencing. The resulting SyHDRD54 and SyHDRD66 mutant cDNAs,and the wild type SyHDR full-length construct, were transformed into the E. coliispH mutant for complementation assays.

Complementation of the E. coli ispH Mutant

The E. coli ispH mutant strain MG1655 ara,.ispH was maintained onLuria-Bertani (LB) medium containing 50 mg mL21 kanamycin and 0.2% (w/v)Ara (McAteer et al., 2001). Arabidopsis and cyanobacterial HDR mutantconstructs were transformed into E. coli ispH mutant competent cells and se-lected on LB plates containing 50 mg mL21 kanamycin, 50 mg mL21 ampicillin,and 0.2% (w/v) Ara. The presence of mutant plasmids in surviving colonieswas verified. Transformants containing mutant plasmids were grown on LBplates containing 50 mg mL21 kanamycin, 50 mg mL21 ampicillin, 0.2% (w/v)Glc, and 0.5 mM isopropylthio-b-galactoside to test if the mutated HDR pro-teins were able to complement the E. coli ispH mutant. As a positive control,the wild-type pQE-AtHDR plasmid was transformed into the E. coli ispHmutant and grown on the same medium.

Growth Curves of E. coli ispH Mutants

The E. coli ispH mutants transformed with wild-type or mutant (e.g. His-152Asn or His-241Asn) pQE-AtHDR plasmids were grown in LB liquid culturecontaining 50 mg mL21 kanamycin, 50 mg mL21 ampicillin, and 0.2% (w/v) Araat 37°C overnight. Then, 1 mL of the overnight culture was added to 20 mL offresh LB liquid medium containing 50 mg mL21 kanamycin, 50 mg mL21 ampi-cillin, 0.2% (w/v) Glc, and 0.5 mM isopropylthio-b-galactoside, and the initialmeasurement of optical density at 600 nm was taken. The liquid cultures weregrown on a rotary shaker (250 rpm) at 37°C with the consecutive measurementsof optical density at 600 nm taken at 30-min intervals for a total of 450 min.

Complementation of the Arabidopsis hdr-1 Mutant by35S:HDRD(56–110)

To obtain the Arabidopsis HDRD(56–110) clone, full-length HDR cDNA wasused as a template to amplify the coding sequences of amino acids 1 to 55 and111 to 466 by PCR with primers 59-CACCATGGCTGTTGCGCTCCAATTC-39/59-CCTTGAGCTCGGAGTCCATCACCACCG-39 and 59-CCTTGAGCTCGTTACTGTGAAACTCGCT-39/59-TCAAGCCAGCTGCAATAACTC-39, respectively.These two PCR fragments were digested with SacI, ligated with T4 DNA ligase,and cloned into the Gateway pENTR/D-TOPO vector (Life Technologies). TheHDRD(56–110) clone was verified by DNA sequencing, subcloned into the plantgene expression vector pGWB502V containing the hygromycin selectablemarker, and transformed into A. tumefaciens GV3101. The resulting 35S:HDRD(56–110) clone was transformed into Arabidopsis hdr-1 heterozygous (+/2;KanR) plants by floral dip. T1 seeds were screened on selective medium con-taining kanamycin and hygromycin. Successful transformants (HygR and KanR)would be hdr-1 (2/2 or +/2) mutants harboring the 35S:HDRD(56–110)transgene. Total protein extraction and immunoblot analysis of ArabidopsisHDR were performed as described (Hsieh and Goodman, 2005). The monoclonalanti-actin antibody (A3853) was purchased from Sigma.

Subcellular Localization of ArabidopsisHDRD(56–110)-GFP

The Arabidopsis HDRD(56–110) clone was used as a template for PCR withprimers 59-CACCATGGCTGTTGCGCTCCAATTC-39 and 59-AGCCAGCTG-CAATAACTCTT-39. The PCR product was cloned into the Gateway pENTR/D-TOPO vector (Life Technologies), verified by DNA sequencing, and subcl-oned into the 35S:GFP expression vector pGWB505. The resulting clone, 35S:HDRD(56–110)-GFP, was transformed into Arabidopsis mesophyll protoplaststo examine the subcellular localization of the HDRD(56–110)-GFP fusionprotein. Protoplast isolation and polyethylene glycol-mediated transforma-tion were performed as described (Hsieh et al., 2008). The Zeiss LSM 510Meta confocal microscope was used to observe the protoplast after 16 h oftransformation.

Supplemental Data

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

Supplemental Figure S1. Phenotypes of T2 and T3 progeny of the Arabi-dopsis HDR genomic DNA-induced gene-silencing line G16.

Supplemental Figure S2. Amino acid sequence alignment and secondarystructure comparison between Arabidopsis HDR and E. coli IspH.

Supplemental Figure S3. Analysis of Arabidopsis HDR key amino acidsby site-directed mutagenesis and complementation test in the E. coli ispHmutant.

Supplemental Figure S4. Schematic diagrams of MEP pathway enzymes.

Supplemental Figure S5. Progeny of two homozygous 35S:HDRD(562110) transgenic lines segregated green, pale-green, variegated, and al-bino seedlings on a nonselective medium.

Supplemental Table S1. Primers used to generate Arabidopsis HDR mu-tant constructs.

ACKNOWLEDGMENTS

We thank Millicent Masters for the E. coli ispHmutant, Tsuyoshi Nakagawafor the pGWB vectors, Hsiu-An Chu for the Synechocystis sp. PCC 6803 culture,and Mei-Jane Fang for assistance in confocal microscopy.

Received May 26, 2014; accepted July 15, 2014; published July 18, 2014.

LITERATURE CITED

Adam P, Hecht S, Eisenreich W, Kaiser J, Grawert T, Arigoni D, Bacher A,Rohdich F (2002) Biosynthesis of terpenes: studies on 1-hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate reductase. Proc Natl Acad SciUSA 99: 12108–12113


Hsieh et al.











Altincicek B, Duin EC, Reichenberg A, Hedderich R, Kollas AK, Hintz M,Wagner S, Wiesner J, Beck E, Jomaa H (2002) LytB protein catalyzes theterminal step of the 2-C-methyl-D-erythritol-4-phosphate pathway ofisoprenoid biosynthesis. FEBS Lett 532: 437–440

Altincicek B, Kollas A, Eberl M, Wiesner J, Sanderbrand S, Hintz M,Beck E, Jomaa H (2001) LytB, a novel gene of the 2-C-methyl-D-erythritol 4-phosphate pathway of isoprenoid biosynthesis in Esche-richia coli. FEBS Lett 499: 37–40

Carretero-Paulet L, Ahumada I, Cunillera N, Rodríguez-Concepción M,Ferrer A, Boronat A, Campos N (2002) Expression and molecularanalysis of the Arabidopsis DXR gene encoding 1-deoxy-D-xylulose5-phosphate reductoisomerase, the first committed enzyme of the 2-C-methyl-D-erythritol 4-phosphate pathway. Plant Physiol 129: 1581–1591

Carretero-Paulet L, Cairó A, Botella-Pavía P, Besumbes O, Campos N,Boronat A, Rodríguez-Concepción M (2006) Enhanced flux through themethylerythritol 4-phosphate pathway in Arabidopsis plants over-expressing deoxyxylulose 5-phosphate reductoisomerase. Plant Mol Biol62: 683–695

Cunningham FX Jr, Lafond TP, Gantt E (2000) Evidence of a role for LytBin the nonmevalonate pathway of isoprenoid biosynthesis. J Bacteriol182: 5841–5848

de la Luz Gutiérrez-Nava M, Gillmor CS, Jiménez LF, Guevara-García A,León P (2004) CHLOROPLAST BIOGENESIS genes act cell and noncellautonomously in early chloroplast development. Plant Physiol 135:471–482

Fung PK, Krushkal J, Weathers PJ (2010) Computational analysis of theevolution of 1-deoxy-D-xylulose-5-phosphate reductoisomerase, an impor-tant enzyme in plant terpene biosynthesis. Chem Biodivers 7: 1098–1110

Gräwert T, Kaiser J, Zepeck F, Laupitz R, Hecht S, Amslinger S,Schramek N, Schleicher E, Weber S, Haslbeck M, et al (2004) IspHprotein of Escherichia coli: studies on iron-sulfur cluster implementationand catalysis. J Am Chem Soc 126: 12847–12855

Gräwert T, Rohdich F, Span I, Bacher A, Eisenreich W, Eppinger J, Groll M(2009) Structure of active IspH enzyme from Escherichia coli providesmechanistic insights into substrate reduction. Angew Chem Int Ed Engl 48:5756–5759

Gräwert T, Span I, Eisenreich W, Rohdich F, Eppinger J, Bacher A, Groll M(2010) Probing the reaction mechanism of IspH protein by x-ray structureanalysis. Proc Natl Acad Sci USA 107: 1077–1081

Guevara-García A, San Román C, Arroyo A, Cortés ME, de la LuzGutiérrez-Nava M, León P (2005) Characterization of the Arabidopsisclb6 mutant illustrates the importance of posttranscriptional regulationof the methyl-D-erythritol 4-phosphate pathway. Plant Cell 17: 628–643

Gustafson CE, Kaul S, Ishiguro EE (1993) Identification of the Escherichiacoli lytB gene, which is involved in penicillin tolerance and control of thestringent response. J Bacteriol 175: 1203–1205

Hsieh MH, Chang CY, Hsu SJ, Chen JJ (2008) Chloroplast localization ofmethylerythritol 4-phosphate pathway enzymes and regulation of mi-tochondrial genes in ispD and ispE albino mutants in Arabidopsis. PlantMol Biol 66: 663–673

Hsieh MH, Goodman HM (2005) The Arabidopsis IspH homolog is in-volved in the plastid nonmevalonate pathway of isoprenoid biosyn-thesis. Plant Physiol 138: 641–653

Hsieh MH, Goodman HM (2006) Functional evidence for the involvementof Arabidopsis IspF homolog in the nonmevalonate pathway of plastidisoprenoid biosynthesis. Planta 223: 779–784

Li J, Wang K, Smirnova TI, Khade RL, Zhang Y, Oldfield E (2013) Iso-prenoid biosynthesis: ferraoxetane or allyl anion mechanism for IspHcatalysis? Angew Chem Int Ed Engl 52: 6522–6525

Lichtenthaler HK (1999) The 1-deoxy-D-xylulose-5-phosphate pathway ofisoprenoid synthesis in plants. Annu Rev Plant Physiol Plant Mol Biol50: 47–65

Lichtenthaler HK, Wellburn AR (1983) Determinations of total carotenoidsand chlorophylls a and b of leaf extracts in different solvents. BiochemSoc Trans 11: 591–592

Lu XM, Hu XJ, Zhao YZ, Song WB, Zhang M, Chen ZL, Chen W, Dong YB,Wang ZH, Lai JS (2012) Map-based cloning of zb7 encoding an IPP andDMAPP synthase in the MEP pathway of maize. Mol Plant 5: 1100–1112

Mandel MA, Feldmann KA, Herrera-Estrella L, Rocha-Sosa M, León P(1996) CLA1, a novel gene required for chloroplast development, ishighly conserved in evolution. Plant J 9: 649–658

McAteer S, Coulson A, McLennan N, Masters M (2001) The lytB gene ofEscherichia coli is essential and specifies a product needed for isoprenoidbiosynthesis. J Bacteriol 183: 7403–7407

Nelissen B, Van de Peer Y, Wilmotte A, De Wachter R (1995) An earlyorigin of plastids within the cyanobacterial divergence is suggested byevolutionary trees based on complete 16S rRNA sequences. Mol BiolEvol 12: 1166–1173

Oldfield E (2010) Targeting isoprenoid biosynthesis for drug discovery:bench to bedside. Acc Chem Res 43: 1216–1226

Page JE, Hause G, Raschke M, Gao W, Schmidt J, Zenk MH, Kutchan TM(2004) Functional analysis of the final steps of the 1-deoxy-D-xylulose5-phosphate (DXP) pathway to isoprenoids in plants using virus-inducedgene silencing. Plant Physiol 134: 1401–1413

Phillips MA, León P, Boronat A, Rodríguez-Concepción M (2008) Theplastidial MEP pathway: unified nomenclature and resources. TrendsPlant Sci 13: 619–623

Rekittke I, Wiesner J, Röhrich R, Demmer U, Warkentin E, Xu W,Troschke K, Hintz M, No JH, Duin EC, et al (2008) Structure of (E)-4-hydroxy-3-methyl-but-2-enyl diphosphate reductase, the terminal en-zyme of the non-mevalonate pathway. J Am Chem Soc 130: 17206–17207

Rodríguez-Concepción M, Boronat A (2002) Elucidation of the methyl-erythritol phosphate pathway for isoprenoid biosynthesis in bacteriaand plastids: a metabolic milestone achieved through genomics. PlantPhysiol 130: 1079–1089

Rohdich F, Hecht S, Gärtner K, Adam P, Krieger C, Amslinger S, Arigoni D,Bacher A, Eisenreich W (2002) Studies on the nonmevalonate terpene bio-synthetic pathway: metabolic role of IspH (LytB) protein. Proc Natl Acad SciUSA 99: 1158–1163

Rohdich F, Zepeck F, Adam P, Hecht S, Kaiser J, Laupitz R, Gräwert T,Amslinger S, Eisenreich W, Bacher A, et al (2003) The deoxyxylulosephosphate pathway of isoprenoid biosynthesis: studies on the mecha-nisms of the reactions catalyzed by IspG and IspH protein. Proc NatlAcad Sci USA 100: 1586–1591

Rohmer M, Knani M, Simonin P, Sutter B, Sahm H (1993) Isoprenoidbiosynthesis in bacteria: a novel pathway for the early steps leading toisopentenyl diphosphate. Biochem J 295: 517–524

Sacchettini JC, Poulter CD (1997) Creating isoprenoid diversity. Science277: 1788–1789

Span I, Gräwert T, Bacher A, Eisenreich W, Groll M (2012a) Crystalstructures of mutant IspH proteins reveal a rotation of the substrate’shydroxymethyl group during catalysis. J Mol Biol 416: 1–9

Span I, Wang K, Wang W, Zhang Y, Bacher A, Eisenreich W, Li K, Schulz C,Oldfield E, Groll M (2012b) Discovery of acetylene hydratase activity of theiron-sulphur protein IspH. Nat Commun 3: 1042

Tseng CC, Lee CJ, Chung YT, Sung TY, Hsieh MH (2013) Differentialregulation of Arabidopsis plastid gene expression and RNA editing innon-photosynthetic tissues. Plant Mol Biol 82: 375–392

Vranová E, Coman D, Gruissem W (2013) Network analysis of the MVAand MEP pathways for isoprenoid synthesis. Annu Rev Plant Biol 64:665–700

Wang W, Wang K, Liu YL, No JH, Li J, Nilges MJ, Oldfield E (2010) Bi-oorganometallic mechanism of action, and inhibition, of IspH. Proc NatlAcad Sci USA 107: 4522–4527

Wang W, Wang K, Span I, Jauch J, Bacher A, Groll M, Oldfield E (2012)Are free radicals involved in IspH catalysis? An EPR and crystallo-graphic investigation. J Am Chem Soc 134: 11225–11234

Wolff M, Seemann M, Tse Sum Bui B, Frapart Y, Tritsch D, GarciaEstrabot A, Rodríguez-Concepción M, Boronat A, Marquet A, RohmerM (2003) Isoprenoid biosynthesis via the methylerythritol phosphatepathway: the (E)-4-hydroxy-3-methylbut-2-enyl diphosphate reductase(LytB/IspH) from Escherichia coli is a [4Fe-4S] protein. FEBS Lett 541:115–120

Xu W, Lees NS, Hall D, Welideniya D, Hoffman BM, Duin EC (2012) Acloser look at the spectroscopic properties of possible reaction inter-mediates in wild-type and mutant (E)-4-hydroxy-3-methylbut-2-enyldiphosphate reductase. Biochemistry 51: 4835–4849





functional evidence for the critical amino-terminal conserved

Documents