N-Terminal Structure of Maize Ferredoxin:NADP+ ReductaseDetermines Recruitment into Different ThylakoidMembrane Complexes W

Manuel Twachtmann,a Bianca Altmann,a Norifumi Muraki,b Ingo Voss,a Satoshi Okutani,c Genji Kurisu,b

Toshiharu Hase,c and Guy T. Hankea,c,1

a Department of Plant Physiology, University of Osnabrück, Osnabruck 49076, Germanyb Laboratory of Protein Crystalography, Institute for Protein Research, Osaka University, Suita, Osaka 565-0871, Japanc Laboratory for the Regulation of Biological Reactions, Institute for Protein Research, Osaka University, Suita, Osaka 565-0871,Japan

To adapt to different light intensities, photosynthetic organisms manipulate the flow of electrons through several alternativepathways at the thylakoid membrane. The enzyme ferredoxin:NADP+ reductase (FNR) has the potential to regulate thiselectron partitioning because it is integral to most of these electron cascades and can associate with several differentmembrane complexes. However, the factors controlling relative localization of FNR to different membrane complexes havenot yet been established. Maize (Zea mays) contains three chloroplast FNR proteins with totally different membraneassociation, and we found that these proteins have variable distribution between cells conducting predominantly cyclicelectron transport (bundle sheath) and linear electron transport (mesophyll). Here, the crystal structures of all three enzymeswere solved, revealing major structural differences at the N-terminal domain and dimer interface. Expression in Arabidopsisthaliana of maize FNRs as chimeras and truncated proteins showed the N-terminal determines recruitment of FNR to differentmembrane complexes. In addition, the different maize FNR proteins localized to different thylakoid membrane complexes onexpression in Arabidopsis, and analysis of chlorophyll fluorescence and photosystem I absorbance demonstrates the impactof FNR location on photosynthetic electron flow.

INTRODUCTION

The ATP and NADPH that support photoautotrophic growth areprovided by photoexcitation and electron transport throughprotein complexes at the thylakoid membrane. Under naturalgrowth conditions, light intensity does not remain constant, andit is a challenge to maximize energy conversion under low lightconditions while minimizing damage due to overexcitation athigh light. Photosynthetic organisms can respond to these en-vironmental changes by altering photosynthetic electron trans-port (PET) at the thylakoid membrane. In the main pathway oflinear electron transport (LET), water is split to O2 and H+, andNADP+ is the final electron acceptor (Arnon et al., 1963; Shinand Arnon, 1965). The final transfer of electrons to NADP+ iscatalyzed by the enzyme ferredoxin:NADP+ reductase (FNR),which is both bound to the thylakoid as a peripheral protein andpresent as a soluble protein in the chloroplast stroma. In thealternative pathway of cyclic electron transport (CET), electronsare passed back to the plastoquinone pool from either ferre-doxin or NADPH, which generates a proton gradient resulting inATP synthesis, with minimal water splitting and no net reduction

of NADP+ (Arnon et al., 1963; Hosler and Yocum, 1985; Munekageet al., 2004). A role for FNR in CET has been proposed due toreported interactions between FNR and characterized compo-nents of the CET apparatus in higher plants and algae (Zhanget al., 2001; DalCorso et al., 2008; Iwai et al., 2010), perturbation ofCET in cyanobacteria that lack membrane-bound FNR (van Thoret al., 2000), electron donation from FNR to quinones (Bojko et al.,2003), and the disruption of CET in spinach (Spinacia oleracea)thylakoids by FNR inhibitors (Shahak et al., 1981). However, theprecise function of FNR in the CET pathway remains an openquestion.These roles in different electron transport pathways have led

to intense research interest in the mechanism and location ofFNR binding to the thylakoid membrane (Andersen et al., 1992;Jose Quiles and Cuello, 1998; Zhang et al., 2001; Peng et al.,2008; Benz et al., 2009; Juri�c et al., 2009; Alte et al., 2010;Baniulis et al., 2011). Two thylakoid FNR binding proteins namedTROL and Tic62 have been identified in Arabidopsis thaliana.TROL mutants are impaired in photosynthesis (Juri�c et al., 2009),and Tic62 has no direct photosynthetic role but seems rather tostabilize FNR (Benz et al., 2009). An FNR binding domain con-taining a Ser/Pro-rich motif has been identified in both TROLand Tic62, and its interaction with FNR has recently beenthoroughly characterized (Alte et al., 2010). In the crystal struc-ture, FNR dimerizes around this Ser/Pro-rich motif. This in-teraction increases the surface area available for FNR:FNRbinding from 380 to 2070 Å2, and the strength of this interactionis pH dependent. In addition to these specific binding proteins,

1 Address correspondence to guy.hanke@biologie.uni-osnabrueck.de.The author responsible for distribution of materials integral to the findingspresented in this article in accordance with the policy described in theInstructions for Authors (www.plantcell.org) is: Guy T. Hanke (guy.hanke@biologie.uni-osnabrueck.de).WOnline version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.111.094532

The Plant Cell, Vol. 24: 2979–2991, July 2012, www.plantcell.org ã 2012 American Society of Plant Biologists. All rights reserved.

FNR copurifies with the cytochrome b6f complex (Cyt b6f) ofseveral higher plants and cyanobacteria (Zhang et al., 2001;Okutani et al., 2005; Baniulis et al., 2011), although the area onthe complex to which FNR binds is not yet clear. Isolation fromChlamydomonas reinhardtii of a supercomplex capable of CET,and composed of only FNR, photosystem I (PSI), Cyt b6f, andPgrL1 (Iwai et al., 2010), supports the involvement of FNR boundto Cyt b6f in CET.

In all higher plants with extensive sequence availability, two dif-ferent FNR isoproteins are present. The best studied example isArabidopsis, where both proteins (At-FNR1 and At-FNR2) wereidentified in membrane-bound and unbound states and were foundto have similar electron transfer properties (Hanke et al., 2005).Reciprocal knockouts are quite similar in phenotype but have re-vealed that At-FNR2 is dependent on At-FNR1 for membrane as-sociation (Lintala et al., 2007; Hanke et al., 2008), and At-FNR2 mayhave a role in stress signaling (Lintala et al., 2009). By contrast,maize (Zea mays) chloroplasts contain three separate FNR iso-proteins (Zm-FNR1, Zm-FNR2, and Zm-FNR3) (Okutani et al., 2005).

These three FNRs have distinct membrane association pat-terns (Okutani et al., 2005). Zm-FNR1 is restricted to the thyla-koid, Zm-FNR3 is only recovered from chloroplasts as a solubleprotein, and Zm-FNR2 is found in both membrane-bound andunbound states. The physiological role of the uniquely solubleFNR3 in maize chloroplasts is not clear, although it has a slightlyhigher efficiency than maize FNR2 or FNR1 in nonphotosyntheticelectron transfer, from NADPH to ferredoxin, and transcripts areupregulated in the presence of NH4 and NO3, leading to thesuggestion that it might have an alternative, heterotrophic func-tion (Okutani et al., 2005). Unlike other plants in which FNR hasbeen studied, maize is a C4 plant and therefore contains twodifferentiated photosynthetic cell types (Schuster et al., 1985). Inmaize, mesophyll cells perform predominantly LET, while bundlesheath cells have very low photosystem II (PSII) activity, and soconduct mainly CET. Therefore, using maize as a model providesan opportunity to discover what determines the different mem-brane association behavior of FNR proteins and how they con-tribute to different PET pathways.

In this article, we show the FNR isoprotein composition of cellsconducting mainly LET or CET varies and that the membranebound:free ratio of FNR is completely different between these celltypes. In addition, we identified the N terminus of FNR as criticalin determining membrane association by solving the crystalstructures of all three maize FNRs. Transformation of Arabidopsiswith maize FNRs, truncated FNR sequences, and a chimeric FNRconstruct with exchanged N termini established that the N-terminalcontribution to membrane binding could be a universal mechanismin higher plants and that N-terminal structure also contributes torelative recruitment into Tic62 or TROL coordinated thylakoidmembrane complexes.

RESULTS

Variable FNR Composition of Maize Bundle Sheath andMesophyll Cells

Evidence from maize thylakoid proteomic studies suggests thatFNR2 is enriched in the mesophyll, and FNR1 is equally abundant

in mesophyll and bundle sheath cells (Majeran et al., 2008). Noinformation is available for FNR3. To clarify the contribution ofFNR isoproteins to the different electron transfer processes thatoccur in these cell types, we compared their abundance inseparated mesophyll and bundle sheath cells by immunoblot-ting. Figure 1A shows that mesophyll cells contain a higherabundance of FNR than the bundle sheath, and Figure 1Bdemonstrates the different FNR profiles of these cell types.Membrane-bound forms were detected in both the bundlesheath and mesophyll cells, with the ratio of FNR1 to FNR2higher in the bundle sheath than the mesophyll. By comparison,soluble proteins were only detected in the mesophyll cells,where FNR2 is the most abundant isoprotein. The higher pro-portion of FNR at the membrane in the bundle sheath thereforecorrelates with a higher demand for CET, while the greaterabundance of soluble FNR occurs in cell types with a greaterdemand for LET. It should be noted that, in our experiments,soluble proteins were extracted from leaves under growth lightconditions in 100 mM NaCl ionic strength buffer; thus, tran-siently or weakly membrane-bound FNR might be recovered inthe soluble fraction.

Differences in Maize FNR Recruitment into ArabidopsisThylakoid Complexes

The lack of soluble FNR in bundle sheath chloroplasts, incomparison to the abundant soluble FNR in mesophyll chlor-oplasts, could either be due to variation between the chloroplastenvironment in these cell types or FNR composition. Variablemembrane association of Zm-FNR1, Zm-FNR2, and Zm-FNR3suggests that their intrinsic properties might determine the ratiobetween soluble and membrane-associated FNR. We thereforefurther tested their membrane association by introducing themaize (Zm) FNR1, FNR2, and FNR3 coding sequences intoArabidopsis and examining the location of the proteins. Figure2A shows higher FNR activity than the wild type in two in-dependent lines expressing each Zm-FNR gene. The phenotypeof the transgenics varied with the Zm-FNR gene introduced.Figure 2B shows that lines overexpressing Zm-FNR2 and Zm-FNR3 were larger than the wild type. By contrast, lines over-expressing Zm-FNR1 were smaller, and increasing FNR activityin the Zm-FNR1 lines correlated with decreasing size. The maizeFNRs have similar molecular mass (;35 kD) and can only beclearly separated from native Arabidopsis FNR isoproteins bynative-PAGE followed by immunoblotting, as shown in Figure2C. On introduction to Arabidopsis, the maize FNRs appear toassociate with the membrane in a comparable pattern tothat seen in maize chloroplasts. In Arabidopsis, transgenicallyintroduced Zm-FNR1 was found principally at the membrane,Zm-FNR2 was equally distributed between membrane and sol-uble fractions, while Zm-FNR3 was mostly recovered in thesoluble fraction, indicating that differences between the proteinscontribute to their membrane association. Apparent differencesin Zm-FNR abundance in the transgenic lines are probably dueto variable antigenicity of the isoproteins to the antisera, whichwas raised against maize FNR2.Two proteins have recently been reported to bind FNR at the

thylakoid membrane in Arabidopsis: TROL (Juri�c et al., 2009) and

2980 The Plant Cell

Tic62 (Benz et al., 2009). Blue native-PAGE analysis has shownthat TROL binds FNR to complexes of ;140 to 190 kD, andTic62 binds FNR into several complexes at higher molecularmasses. To identify the membrane complexes into which maizeFNR proteins are recruited at the Arabidopsis thylakoid, weperformed blue native-PAGE and protein gel blotting on the

transgenic Zm-FNR plants. Figure 2D shows the distribution ofFNR in thylakoids from representative lines expressing Zm-FNR1, Zm-FNR2, or Zm-FNR3. In comparison to the wild type,Zm-FNR1–expressing plants showed enrichment in FNR boundto the 140- to 190-kD complexes coordinated by TROL. Incomparison, thylakoids from Zm-FNR2–expressing plants werespecifically enriched in the higher molecular mass complexescoordinated by Tic62. Membranes from Zm-FNR3 lines were notnoticeably enriched, and relative staining of TROL and Tic62complexes was similar to that seen in untransformed Arabidopsis.The kinetics of FNR binding to a motif conserved between TROLand Tic62 are well described (Alte et al., 2010), and the data inFigure 2D now indicate there may be differences between themechanism of FNR interaction with TROL and Tic62.

N-Terminal Structure Varies between Maize FNR Proteins

The recruitment of maize FNR1 and FNR2 into different mem-brane complexes and the weakened membrane association ofmaize FNR3 offer an opportunity to identify the structuraldeterminants of FNR membrane recruitment. To investigate this,we crystallized the purified, mature maize FNR proteins andsolved their structures. Crystal data and refinement statistics aregiven in Supplemental Table 1 online. The maize FNR iso-proteins closely resemble each other (Figure 3A), with the onlymajor difference between them in the orientation of the N ter-minus. In the Zm-FNR1 structure, resolution of N-terminal aminoacids was excellent, showing a rigid structure, perpendicular tothe surface of the molecule. In Zm-FNR2, the N-terminal waspoorly resolved, presumably due to a high degree of flexibility.The N-terminal region in the Zm-FNR3 structure was resolvedclearly enough to identify its orientation, lying flat across themolecule.All three maize FNRs were crystallized as homodimers (Figure

3B). The orientation of Zm-FNR1 and Zm-FNR2 subunits in thehomodimer is very similar to that in the cocrystal of pea (Pisumsativum) FNR with a 27–amino acid fragment of Tic62 (Alte et al.,2010). By contrast, the structure of the Zm-FNR3 homodimer ismuch more open, with the potential Tic62 binding surfacesfurther apart. This is consistent with the poor membrane asso-ciation behavior of Zm-FNR3. Interestingly, in the Zm-FNR1homodimer, the N termini of two adjacent Zm-FNR1 moleculesare inserted at the interface of the homodimer (Figure 3C), oc-cupying the space reported to be filled by the Tic62 peptide(Figure 3E; coordinates taken from Alte et al., 2010). In additionto N-terminal differences, there is variation between the chargedistribution at the homodimer-forming surface of FNR proteins(Figures 3C and 3D). Maize Zm-FNR1 is similar to pea FNR(Figure 3E), with a basically charged area to one side of themostly acidic region to which the Tic62 peptide binds. By con-trast, Zm-FNR2 has a much more acidic surface in this region(Figure 3D). Zm-FNR3 shows almost identical surface charge toZm-FNR2 and so is not shown. Although the FNR binding do-mains on Tic62 and TROL are very similar, the increased TROLbinding in Arabidopsis transformed with Zm-FNR1 indicates thatthese differences in N-terminal and homodimer interface maydetermine preferential recruitment of Zm-FNR1 into TROL andZm-FNR2 into Tic62-mediated complexes.

Figure 1. Variable Cellular and Subcellular Locations of Maize FNRProteins.

(A) Cellular distribution of FNR isoproteins. Chlorophyll and proteinswere extracted from total leaves (T) or mesophyll cells (MC) and bundlesheath cells (BSC) isolated from mature maize leaves. Chlorophyll a/bratios 6 SD of three independent measurements were calculated, andprotein extracts were separated by SDS-PAGE before immunoblotting todetect the indicated proteins. Lanes were loaded with equal protein, ei-ther 4 µg for detection of FNR and NADP-MDH or 0.5 µg for detection ofthe large subunit of ribulose-1,5-bis-phosphate carboxylase/oxygenase(RuBisCO). Antisera used are indicated to the left of the blots. Migrationposition of the different maize FNR isoproteins, as established previously(Okutani et al., 2005), is indicated to the right of the aFNR blot.(B) Subchloroplast location of FNR isoproteins. Proteins from bundlesheath and mesophyll cells were separated into membrane (M) andsoluble (S) fractions and loaded on the gel to give equivalent ratios ofmembrane to soluble extract. Twenty micrograms of protein from bundlesheath cell extracts and 4 µg protein from mesophyll cell extracts wereloaded on the gel and immunoblotted to detect FNR. Migration positionof the different maize FNR isoproteins, as established previously (Okutaniet al., 2005), is indicated to the right of the blot.

FNR N Terminus Controls Membrane Association 2981

The Maize FNR1 N Terminus Confers Membrane BindingActivity on Maize FNR3

To determine the contribution of the different N-terminal struc-tures to membrane association, we investigated whether ex-change of N-terminal regions between FNR proteins couldchange their membrane association (Figure 4). Chimeric con-structs were generated, encoding the transit peptide and matureN-terminal of Zm-FNR1 and the remaining part of the Zm-FNR3molecule (Figure 4A). Transformation of Arabidopsis with theseZm-FNR1-3 constructs yielded several independent lines. Theintroduced Zm-FNR chimeras could not be distinguished fromindigenous Arabidopsis FNRs by SDS-PAGE (data not shown),so native PAGE was used to determine whether the chimeraswere membrane bound or not. The typical distribution of chi-meric Zm-FNR1-3 between membrane and soluble fractions isshown in Figure 4B, and a single additional band migratingfaster than the endogenous At-FNRs is visible. In addition, theseplants appear enriched in native At-FNR1, presumably due tosome pleiotropic effect, as described in previous studies of FNRoverexpression (Higuchi-Takeuchi et al., 2011). In comparison totransgenic Arabidopsis plants expressing the parent molecules,in which Zm-FNR3 was predominantly recovered in the solublefraction, the additional Zm-FNR1-3 band is more abundant atthe membrane. When thylakoids from Zm-FNR1-3 lines wereseparated by blue native-PAGE and immunoblotted, enrichmentof TROL coordinated FNR complexes was clearly seen (Figure4C). This result demonstrates that introduction of the maizeFNR1 N-terminal to maize FNR3 increases membrane associa-tion and specifically enhances interaction with TROL.

Figure 2. Transformation of Arabidopsis with Maize FNR Isoproteins.

(A) Arabidopsis transformed with maize FNR proteins. Genes encodingmaize FNR1 (Zm-FNR1), FNR2 (Zm-FNR2), and FNR3 (Zm-FNR3) werecloned and used to stably transform wild-type (wt) Arabidopsis undercontrol of the Arabidopsis FNR1 promoter. Crude protein extracts weremade from leaves of the transformants and assayed for FNR activity,

measured as ferredoxin-dependent cytochrome c reduction in crudeprotein extracts from leaves of the wild type and two independentArabidopsis lines overexpressing Zm-FNR1, Zm-FNR2, and Zm-FNR3.Values are means 6 SE of three independent measurements.(B) Phenotype of Zm-FNR transgenics. Typical representative plants oftwo independent lines expressing either Zm-FNR1, Zm-FNR2, or Zm-FNR3 are shown following 10 weeks of growth.(C) Subcellular location of Zm-FNR proteins in Arabidopsis. Crude ex-tracts from the wild type and representative Zm-FNR1–, Zm-FNR2–, andZm-FNR3–expressing lines were separated into soluble (S) and mem-brane (M) protein fractions. For each sample, 20 µg of total protein wasdivided into S and M fractions with equal volume, before native-PAGE,and immunoblotting with an antibody raised against Zm-FNR2. Positionof native Arabidopsis FNR1 (At-FNR1) and FNR2 (At-FNR2) proteins isindicated to the left of the blot. Positions of introduced Zm-FNR proteinsare indicated to the right of the blot.(D) Recruitment of maize FNR proteins into Arabidopsis thylakoidmembrane complexes. Chloroplasts were isolated from the wild type andrepresentative Arabidopsis lines expressing the maize FNR genes. Thy-lakoid membranes were extracted, digested, and subjected to bluenative-PAGE (BNP) before immunoblotting and detection of FNR-containing complexes. Samples were loaded on an equal chlorophyllbasis, with 4 µg per lane. Left panel shows the blue native-PAGE gelfollowing the run, with major native complexes indicated to the left of thegel. Right panel shows an immunoblot of the gel, detected with an an-tibody raised against maize FNR2, with FNR-containing complexes in-dicated to the right of the blot. Position of molecular mass markers isindicated between the gel and the blot in kilodaltons.

2982 The Plant Cell

Figure 3. Structural Variation at the N-Terminal and Homodimer Interface of Maize FNR Proteins.

Maize FNR1, FNR2, and FNR3 were expressed recombinantly as mature proteins, purified, and crystallized.(A) Crystal structures of single FNR molecules. Structures are shown as ribbon with a-helices in red and b-sheets in yellow. The FAD moiety is shown inblue. Lower resolution of structure is indicated by the green and red dotted lines representing parts of the Zm-FNR2 and Zm-FNR3 N termini.(B) Arrangement of Zm-FNR homodimers. Structures are shown as a ribbon with a-helices in red and b-sheets in yellow. In the Zm-FNR1 structure, Ntermini of two adjacent Zm-FNR1 molecules are shown in either white or blue balls and sticks.(C) to (E) Charge distribution at the interface of Zm-FNR homodimers (red, area of acidic charge; blue, area of basic charge). Charge distribution on space-filled homodimers of Zm-FNR1 (C), Zm-FNR2 (D), and pea FNR (E) (taken from Alte et al., 2010). The dimer structure is shown in the center, composed of twomonomers ([A] and [B]). To the left and right of each dimer, the constituent FNRA and FNRB monomers of each dimer are shown rotated through 90° todisplay the surface area of interaction. In the structure of Zm-FNR1, the N termini of other Zm-FNR1 molecules adjacent to the dimer in the crystal matrixinsert into the dimer interface. These adjacent FNRmolecules are labeled FNR1A9 and FNR1B9 and are shown as ball and stick models in white (FNR1 A9) andblue (FNR1 B9). The location of the 27–amino acid fragment of Tic62 in the structure of the pea FNR structure is shown in a yellow ball and stick model.

FNR N Terminus Controls Membrane Association 2983

Truncated Maize FNR1 Retains Some TROL Binding Ability

Because the addition of the Zm-FNR1 N-terminal enhanced Zm-FNR3 binding to the membrane in general, and to TROL medi-ated complexes in particular, we tested whether truncation ofthis domain would affect the properties of Zm-FNR1. We con-structed a truncated Zm-FNR1 coding sequence (Zm-FNR1t),with the transit peptide still intact (Figure 5). The truncatedconstruct lacked the N-terminal 14 amino acids of the matureprotein that correspond to those raised from the globularstructure of Zm-FNR1 in Figure 3A. To confirm correct pro-cessing of the transit peptide of Zm-FNR1t and investigatemembrane association, we used SDS-PAGE, rather than native-PAGE, because the smaller, truncated protein can easily be dis-tinguished from the native At-FNR isoproteins. The smaller size ofthe additional protein confirms that the transit peptide was cor-rectly processed (Figure 5B). The relative proportion of Zm-FNR1tbound to the membrane decreased in comparison to Zm-FNR1,but some membrane binding capacity was retained. Blue native-PAGE analysis demonstrated that Zm-FNR1t–expressing plantswere still enriched in TROL binding FNR relative to the wild type,although this was reduced in comparison to Zm-FNR1–expressingplants. This suggests that there are additional factors that con-tribute toward specific recruitment of maize FNR1 to TROL. Themost likely of these is the Zm-FNR1 homodimer interface surface,which is more basic than that seen in Zm-FNR2 or Zm-FNR3(Figure 3), and this would also provide some explanation for theenriched Tic62 binding in one line expressing the Zm-FNR1-3chimera (Figure 4C), as Zm-FNR3 is also more acidic in thisregion.

FNR Location Influences Electron Transport aroundthe Photosystems

Since the discovery that two different proteins, Tic62 and TROL,are capable of mediating FNR association with the thylakoid

Figure 4. Impact of Chimeric N-Terminal Exchange between Maize FNRProteins.

(A) Construction of chimeric Zm-FNR1-3. Amino acid sequence align-ment, comparing Zm-FNR1 (black on white) and Zm-FNR3 (black ongray) with chimeric Zm-FNR1-3, in which the Zm-FNR1 N-terminal se-quence was exchanged for that of Zm-FNR3. Chimeras were created to

exchange the transit peptides and amino acids perpendicular to thesurface of Zm-FNR1 in the structure shown in Figure 3. The first aminoacids of the mature proteins are indicated as white lettering on black.(B) Detection of chimeric Zm-FNR1-3 proteins expressed in transgenicArabidopsis. Plants were stably transformed with the chimeric Zm-FNR1-3construct under control of the At-FNR1 promoter. Protein extracts weremade from Arabidopsis wild-type (wt), Zm-FNR1, and Zm-FNR3–expressingplants, and a representative line expressing Zm-FNR1-3. For each sample,20 µg of total protein was divided into S and M fractions with equal volume,before native-PAGE and immunoblotting with an antibody raised againstZm-FNR2. Positions of native Arabidopsis FNR1 (At-FNR1) and FNR2 (At-FNR2) proteins are indicated to the left of the blot. Positions of introducedZm-FNR proteins are indicated to the right of the blot.(C) Assembly of Zm-FNR1-3 into thylakoid membrane complexes. Thy-lakoid membrane protein complexes from Arabidopsis wild type, trans-genics expressing Zm-FNR1 or Zm-FNR3, and three independent linesoverexpressing Zm-FNR1-3 were solubilized and subjected to blue native-PAGE (BNP) before immunoblotting and detection of FNR-containingcomplexes. Samples loaded contained 4 µg chlorophyll per lane, exceptthe wild type, which was loaded as 2 µg per lane. Approximate molecularmasses are given in kilodaltons to the left of the blot, and FNR-containingcomplexes coordinated by Tic62 and TROL are indicated to the right ofthe blot.

2984 The Plant Cell

(Benz et al., 2009; Juri�c et al., 2009), there has been muchspeculation that FNR bound to different protein complexes, andin the soluble phase, might coordinate different electron transferprocesses (Benz et al., 2010; Joliot and Johnson, 2011). As wefound Arabidopsis plants expressing maize FNR1, FNR2, andFNR3 are respectively enriched in FNR bound to TROL, Tic62,and in soluble FNR, we compared their PET capacity, to gainfurther insight into this question. There are early reports thatFNR is the likely point of light regulation at the reducing side ofPSI on light induction (Satoh and Katoh, 1980; Satoh, 1982),and more recent articles have provided further evidence thatFNR may regulate PET (Ilík et al., 2006; Joliot and Johnson,2011) and that this might be dependent on the NADP+/NADPHratio (Rajagopal et al., 2003; Hald et al., 2008). In addition, FNRmembrane association is known to change during dark-to-lighttransitions: FNR bound to Tic62 is released at basic pH valuesthat could be generated in PET (Alte et al., 2010), and in-teraction with thylakoid Tic62- and TROL-mediated complexesis reported to be disrupted in the light (Benz et al., 2009). Wetherefore examined electron transport processes at PSII andPSI over the transition from dark to moderate light. Figure 6shows parameters calculated from chlorophyll fluorescence atPSII and P700 absorption at PSI. As expected, differencesfrom the wild type were mainly seen during the initial stages oflight induction. Electron transport rates at PSI and PSII in themaize FNR-expressing plants show relatively small deviationsfrom the wild type, but there is a more rapid increase in ФI/ФIIfor Zm-FNR1–expressing plants when the light is switched on,reflecting higher activity at PSI relative to PSII. This rise ismirrored in a more rapid increase in nonphotochemical quenching(NPQ). By contrast, Zm-FNR2– and Zm-FNR3–expressing plantsshow slower increases in both these parameters, and this is mostdramatic for Zm-FNR3–expressing plants. The differences of Zm-FNR2– and Zm-FNR3–expressing plants disappear over the lightinduction, but Zm-FNR1–expressing plants retain a higher ФI/ФIIfor the duration of the measurement.

Figure 5. Impact of N-Terminal Truncation on Maize FNR1 Assemblyinto Thylakoid Membrane Complexes.

(A) Truncation of maize FNR1 sequence. Alignment comparing theN-terminal portion of the native (Zm-FNR1) and truncated (Zm-FNR1t)nucleotide and amino acid sequences. The Zm-FNR1 sequence wastruncated to retain the transit peptide, while shortening the mature pro-tein sequence by 14 amino acids that are perpendicular to the surface ofthe protein in the structure shown in Figure 3. The first amino acid of themature Zm-FNR1 protein is indicated as white lettering on black.(B) Detection of the mature, truncated Zm-FNR1 protein expressed intransgenic Arabidopsis. Plants were stably transformed with the truncated

Zm-FNR1t construct under control of the At-FNR1 promoter. Proteinextracts were made from Arabidopsis wild type (wt), one representativeline expressing Zm-FNR1, and one representative line expressing Zm-FNR1t (Zm-FNR1t-7). For each sample, 20 µg of total protein was dividedinto soluble (S) and membrane-bound (M) fractions, before SDS-PAGE,and immunoblotting with an antibody raised against maize FNR2. Posi-tion of native Arabidopsis FNR proteins (At-FNR1 + At-FNR2) is indicatedto the left of the blots, and positions of introduced Zm-FNR1 and Zm-FNR1t are indicated to the right of the blots.(C) Assembly of Zm-FNR1t into thylakoid membrane complexes. Thy-lakoid membrane protein complexes from Arabidopsis wild type, trans-genics expressing Zm-FNR1, and two independent lines overexpressingZm-FNR1t were solubilized and subjected to blue native-PAGE (BNP)before immunoblotting and detection of FNR-containing complexes.Samples loaded contained 4 µg chlorophyll per lane. Approximate mo-lecular masses are given in kilodaltons to the left of the blot, and FNR-containing complexes coordinated by Tic62 and TROL are indicated tothe right of the blot.

FNR N Terminus Controls Membrane Association 2985

DISCUSSION

N-Terminal Structure Promotes FNR Dimerization,Enhancing Recruitment to the Membrane

The three FNR proteins in maize chloroplasts are all unique intheir association with the membrane. The aim of our study wasto identify structural variation between these proteins that coulddetermine membrane interaction. In Zm-FNR1 (membranebound), the N terminus sticks out from the surface of the protein,while in Zm-FNR2 (dual location), this region is not well resolved,and in Zm-FNR3 (soluble), it is folded back across the surface ofthe protein. N termini from adjacent Zm-FNR1 molecules fill thespace between the dimerized Zm-FNR1 molecules in the crystal.By contrast, the relative orientation of the crystallized Zm-FNR3in the homodimer is markedly different, with a very openstructure. Assuming the orientation of the Zm-FNR1 N terminusin the crystal structure is not an artifact, this suggests that theZm-FNR1 N terminus may promote dimer formation and thatdimerization is crucial for recruitment to the thylakoid. Excludingthe N terminus, maize FNR2 and FNR3 monomer structures arevery similar (Figure 3), and amino acid sequences are 93%identical (see Supplemental Figure 1 online). We thereforespeculate that the more efficient dimerization of Zm-FNR2 isalso due to the N terminus. Although it is not resolved in thecrystal structure, it may adopt an orientation favorable to dimerformation, such as that shown by Zm-FNR1. Because Zm-FNR2is the only maize FNR capable of both membrane and solublelocation, structural rearrangement of the flexible Zm-FNR2 Nterminus could thus regulate its recruitment and release from thethylakoid by influencing dimer formation. Interestingly, regula-tion of wheat (Triticum aestivum) FNR association with the thy-lakoid is also influenced by N-terminal structure, with truncationof the FNR1 N terminus by only three amino acids, resulting inan increase from 33 to 76% of the enzyme found associatedwith the membrane (Moolna and Bowsher, 2010).In contrast with their native environment in the chloroplast, the

membrane and soluble localization of maize FNR1 and FNR3,respectively, is not as exclusive when the genes are expressedin Arabidopsis; a small amount of Zm-FNR3 is bound to themembrane, and a small amount of Zm-FNR1 is soluble (Figure2C). We speculate that this is due to evolutionary refinement ofFNR–binding protein interactions in maize and that the evolu-tionary distance from Arabidopsis means relative affinities arenot completely conserved. FNR binding to complexes mediated

Figure 6. PET in Arabidopsis with FNR Enriched in Soluble, TROL-Bound, or Tic62-Bound FNR.

Electron transport parameters of Arabidopsis wild type (closed triangles)and representative lines expressing either Zm-FNR1 (Zm-FNR1-5, openboxes), Zm-FNR2 (Zm-FNR2-2, open triangles), or Zm-FNR3 (Zm-FNR3-4,open circles) over light induction. Plants were dark adapted before expo-sure to growth light intensity actinic light. Saturation pulses were given atthe indicated times for calculation of the indicated parameters. Values aremeans 6 SE of measurements on mature leaves from five to six differentindividuals, and the experiment was repeated three times with basically thesame result. The experiment was also repeated with other lines expressingthe same genes to confirm tendencies.

2986 The Plant Cell

by TROL is necessary for optimal photosynthetic electron flow(Juri�c et al., 2009), while binding to Tic62 is not (Benz et al.,2009). It is therefore of interest to speculate on how maize FNR1and FNR2 are differentially recruited to these two differentcomplexes (Figure 2). Addition of the Zm-FNR1 N terminus re-sults in recruitment of the normally soluble Zm-FNR3 to TROL,so the N terminus has a clear role in this specific association.Tic62:FNR association is weaker than TROL:FNR association(Juri�c et al., 2009), and because the N terminus of Zm-FNR1appears to sterically hinder Tic62-type binding (Figure 3), it maybe that this prevents binding to Tic62. The stronger affinity ofTROL for FNR could allow it to displace the N terminus, enablingFNR dimerization around TROL. However, truncation of the Zm-FNR1 N terminus could not completely eliminate recruitmentinto TROL complexes; therefore, other structural factors mustcontribute. We propose a model (Figure 7) in which a combina-tion of N-terminal structure and charge distribution on the FNR:FNR dimer interface determines membrane association. Here,the N termini of adjacent Zm-FNR1 molecules promote thepreformation of an FNR dimer, by occupying the space normallyfilled by an FNR binding domain on Tic62 or TROL (Alte et al.,2010). This preformed dimer has a greater tendency to dimerizein turn around the TROL binding domain, an association that isenhanced by the basic charges on the Zm-FNR1 dimer interface.

Differential Roles of Membrane-Bound andSoluble FNR in CET

There is strong evidence that membrane-bound FNR is neces-sary for optimal electron transport. In higher plants, it has beenshown that membrane-bound FNR has a much higher activity inNADP+ photoreduction than that of soluble FNR (Forti andBracale, 1984) and even addition of 5 µM soluble FNR could notrestore wild-type NADP+ photoreduction rates to mutant thyla-koids lacking native FNR (Hanke et al., 2008). In addition, mu-tants lacking TROL were perturbed in LET at high light intensities(Juri�c et al., 2009), Finally, work on cyanobacteria (Korn et al.,2009) showed that membrane-bound FNR is predominantly in-volved in NADP+ photoreduction, while soluble FNR reducesferredoxin using NADPH under heterotrophic conditions. In-terestingly, not all membrane-bound FNR is directly involved inelectron transport, as FNR recruited to Tic62-mediated com-plexes was stabilized against degradation rather than involved inPET (Benz et al., 2009).

The differential distribution of membrane-bound and solubleFNR isoproteins in maize bundle sheath and mesophyll cellsis therefore enlightening. Homologous Tic62 and TROL cDNAsequences have been identified in maize (see SupplementalFigure 2 online). The maize TROL protein contains two of theFNR binding domains defined previously (Stengel et al., 2008),but maize Tic62 cDNAs lack any sequence encoding an FNRbinding motif. The FNR binding motif is also missing from Tic62sequences of sorghum (Sorghum bicolor), indicating evolution-ary loss in some C4 type grasses. Proteomic work has shownthat maize TROL is equally abundant in both mesophyll andbundle sheath cells (Majeran et al., 2008), so membrane-boundFNR is presumably recruited to TROL or Cyt b6f in both celltypes in maize. When genes were expressed in Arabidopsis, we

found that maize FNR1 and FNR2 were bound to the recentlyidentified FNR binding proteins TROL and Tic62, respectively,but that maize FNR3 was not (Figure 2D). Both maize FNR1 andFNR2 were also copurified from maize leaves with Cyt b6f, butmaize FNR3 was not (Okutani et al., 2005), indicating that FNRassociation with the Cyt b6f might be either similar to, or de-pendent on, Tic62 or TROL.Bundle sheath cells perform predominantly CET, with little

need for regulatory adjustment of PET. The fixed membranelocalization of FNR in bundle sheath chloroplasts thereforepoints to its permanent localization at specific complexes in-volved in CET, such as that reported by Iwai et al. (2010). Bycontrast, mesophyll cells must have the capacity to adjustelectron flow between LET and CET, depending on environ-mental conditions (Munekage et al., 2004). There is evidencefor FNR recruitment to several different thylakoid complexes(Andersen et al., 1992; Jose Quiles and Cuello, 1998; Zhanget al., 2001; Peng et al., 2008; Benz et al., 2009; Juri�c et al.,

Figure 7. The Contribution of Rigid N-Terminal Structure and TROLBinding Domain Amino Acid Composition to Membrane Recruitment ofMaize FNR.

This model summarizes structural and transgenic results indicating thatboth FNR N-terminal structure and amino acid composition of the in-teraction site with TROL contribute to membrane recruitment of FNR inmaize and is based on the assumption that FNR:FNR binding proteininteractions are dependent on FNR dimerization. The structure of Zm-FNR1 shows N termini of adjacent FNR molecules helping to coordinatedimer formation. The basic charge (blue shading) distribution at the Zm-FNR1 homodimer interface causes a strong interaction with TROL.Preformation of a dimer in solution promotes dimerization around theTROL peptide at the membrane, with the strongly interacting FNRbinding domain of TROL displacing the N termini of other Zm-FNR1molecules. In the Zm-FNR3 structure, the more flexible N terminus doesnot contribute to preformation of a dimer. In addition, the TROL bindingsite is surrounded by more acidic amino acids (red shading), and thesefactors preclude Zm-FNR3 recruitment to the membrane. When the Nterminus of Zm-FNR1 is added to Zm-FNR3, this favors preformation ofa dimer, promoting recruitment of Zm-FNR1-3 to TROL. When the Nterminus of Zm-FNR1 is truncated, dimer formation is less likely, andbinding to TROL is therefore weakened. However, the basic amino acidsin the TROL binding domain enable Zm-FNR1t to retain some TROLbinding.

FNR N Terminus Controls Membrane Association 2987

2009; Baniulis et al., 2011), and the greater abundance of solubleFNR in maize mesophyll cells may provide capacity for dynamicrelease and rapid recruitment of FNR to different complexes, ashas recently been suggested (Joliot and Johnson, 2011).

Because of the differences in FNR binding protein composi-tion and PET, between the chloroplasts of Arabidopsis (a C3plant) and the bundle sheath and mesophyll cells of maize, wecan only tentatively interpret the physiological roles of FNRisoproteins using the Arabidopsis maize FNR transgenics.However, the data in Figure 6 support a model where tightbinding to TROL enhances CET, as the Zm-FNR1–expressingplants showed an increased ratio of PSI to PSII activity, partic-ularly during light induction. This is reflected in the stimulation ofNPQ, which is dependent on CET enhancement of DpH for-mation. Such activity is consistent with the higher abundance ofZm-FNR1 in bundle sheath cells (Figure 1), where CET pre-dominantly occurs. Arabidopsis expressing maize FNR2 showedalmost no differences from the wild type in the PET parametersmeasured. As these plants are enriched in FNR bound to Tic62,this is consistent with reports that Tic62-associated FNR is notinvolved in PET (Benz et al., 2009).

In contrast with the Zm-FNR1 plants, Arabidopsis expressingmaize FNR3 shows a slightly slower rise in FI/FII and a slowerinduction of NPQ on light induction (Figure 6). As Zm-FNR3 ispredominantly soluble when expressed in Arabidopsis (Figure2B), one interpretation of these results is that CET is enhancedby FNR recruitment to a specific complex, as previously pro-posed (Joliot and Johnson, 2011) and that this is facilitated ifFNR is already localized at the membrane. The greater abundanceof soluble FNR in the Zm-FNR3 plants may therefore competewith CET-recruited FNR for electrons carried by ferredoxin, slow-ing the onset of DpH and NPQ. The physiological role of the FNR3in maize remains unclear, although it appears to be restricted tochloroplasts capable of performing LET (Figure 1), and one pos-sibility is that it channels electrons from NADPH to support fer-redoxin dependent reactions, such as those involved in N and Sassimilation, when the PET chain is not operating, in a systemanalogous to that seen in cyanobacteria (Korn et al., 2009).

Previous transformation of tobacco (Nicotiana tabacum) withpea FNR did not yield a phenotypic difference (Rodriguez et al.,2007), indicating that factors other than FNR limit photosyn-thesis in tobacco. It was recently reported that expression of rice(Oryza sativa) FNR2 in Arabidopsis perturbed both LET and CET,while rice FNR1 disturbed N assimilation (Higuchi-Takeuchiet al., 2011). No information about association with differentFNR binding proteins is available for this study, and inter-pretation of the phenotypes is complicated by the large de-creases in native Arabidopsis FNRs, which were not seen in ourstudy, where the native At-FNR1 promoter was used to driveZm-FNR expression. The phenotypes of transgenics with vari-able FNR localization indicate that FNR microcompartmentationto different locations in the chloroplast has a big impact on thephysiology of Arabidopsis. The more abundant FNR at the TROLcomplex in Zm-FNR1 plants seems to have a negative impact,whereas enrichment in Tic62 or soluble FNR results in plantslarger than the wild type under these conditions (Figure 2B).Taken together with the analysis of PET in these plants, theseresults could indicate that under the growth conditions used in

this work, FNR has a high control coefficient for PET but thatdynamic association and dissociation from different membranecomplexes is necessary for optimal efficiency. When associationto thylakoid membrane complexes (specifically TROL) is en-hanced, as in the maize FNR1-expressing Arabidopsis, dynamicdissociation is retarded and the balance of excitation betweenthe photosystems disturbed, with deleterious consequences forPET and photosynthesis.In summary, it has long been proposed that some structural

change in FNR may regulate PET at the acceptor side of PSI(Satoh and Katoh, 1980; Satoh, 1982). The data presented inthis article demonstrate that this regulation is probably related tothe N-terminal structure of FNR, which determines the re-cruitment of the enzyme to thylakoid membrane complexes inmaize, in a mechanism that can also operate in Arabidopsis.Analysis of PSI and PSII activity in Arabidopsis with FNR en-riched either at TROL, Tic62, or in the soluble phase shows thatPET activity over light induction is influenced by FNR location.Therefore, changes in N-terminal structure may be involved incontrolling relative recruitment into different membrane com-plexes, allowing for crucial regulation of electron flow around thethylakoid membrane through changes in FNR location.

METHODS

Plant Material and Sample Preparation

Maize plants (Zea mays cv Golden Cross Bantam T51) were grown on soilin a glasshouse, with supplementary light from Son-T Agro lamps (400 W;Philips). Arabidopsis thaliana plants were all Columbia ecotype and grownunder long-day conditions, basically as described previously (Voss et al.,2011). Membrane and soluble protein fractions were prepared from wholeplants and isolated chloroplasts and subjected to native-PAGE or SDS-PAGE and immunoblotting as described previously (Hanke et al., 2008).Mesophyll protoplasts and bundle sheath strands were isolated frommature maize leaves by physical disruption, basically as described bySheen (1995). Antibodies used in immunoblotting detection were raisedagainst maize FNR2 (1:50,000), spinach (Spinacia oleracea) NADP-MDH(1:10,000), and pea (Pisum sativum) ribulose-1,5-bis-phosphate car-boxylase/oxygenase (1:20,000).

Cloning, Plasmid Construction, and Plant Transformation

Mature maize FNR coding sequences and the Arabidopsis FNR1 pro-moter (At-FNR1pro) were amplified from maize cDNA and Arabidopsisgenomic DNA, respectively, with the primers listed in Supplemental Table2 online as described previously (Okutani et al., 2005), and cloned usingthe Mighty TA-cloning kit (Takara). The At-FNR1pro sequence wassubcloned into pSMAB704 plant transformation vector (kind gift of HiroakiIchikawa) using a native HindIII site and an XbaI site introduced duringcloning to give pSAtFNR1pro-GUS (for b-glucuronidase). Zm-FNR se-quences were then separately subcloned into pSAtFNR1pro-GUS, usingNcoI (native including the start codon following At-FNR1pro and the startcodon of Zm-FNR2 and introduced during cloning at the start codon forZm-FNR1 and Zm-FNR3) and SmaI sites (Zm-FNR sequences) or SacIdigestion followed by blunting with T4 polymerase (pSAt-FNR1pro-GUS),yielding pSAt-FNR1pro-Zm-FNR1, pSAt-FNR1pro-Zm-FNR2, and pSAt-FNR1pro-Zm-FNR3. Plasmids were introduced into Arabidopsis viaAgrobacterium tumefaciens using the floral dip method as described pre-viously (Hanke and Hase, 2008) and transformants selected on 100 µgmL21

glyphosinate.

2988 The Plant Cell

FNR Activity Assay

FNR proteins were extracted from Arabidopsis leaves in the presence of50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 2 mM EDTA, 1 mM PMSF, 0.1%Triton X-100, and 0.1 mg mL21 polyvinyl-polypyrrolidone. Extracts weredesalted anddetergent removedover Sephadex-G50. Ferredoxin-dependentelectron transfer fromNADPH to cytochrome cwasmeasured in 100 µg totalprotein, basically as described previously (Hanke et al., 2004), in 50 mMHEPES-NaOH, pH 7.5, 100 mM NaCl, 200 µM NADPH, and 200 µM cyto-chrome c with 20 µM Arabidopsis ferredoxin 2.

Blue Native-PAGE

All steps were performed at 4°C. Washed chloroplasts were resuspendedin 20% (v/v) glycerol, 25 mM Bis-Tris HCl, pH 7, and 250 µg mL21 Pe-fabloc and centrifuged at 10,000g for 30 min. Thylakoids were then re-suspended in the same buffer at a chlorophyll concentration of 0.2 µg/mLand solubilized with n-dodecyl-b-D-maltoside. After a 3-min incubation,samples were centrifuged at 18,000g for 15 min and the supernatantadjusted to 5% (w/v) Coomassie Brilliant Blue G, 100 mM Bis-Tris-HCl,pH 7, 30% (w/v) Suc, 500 mM e-amino-n-caproic acid, and 10% (v/v)glycerol. Samples were analyzed by blue native-PAGE as described(Reisinger and Eichacker, 2007) using 6 to 12% polyacrylamide gradientgels, before immunoblotting to detect FNR.

Crystallization and Structural Analysis of Recombinant MaizeFNR Proteins

Maize FNR proteins were expressed and purified as described previously(Okutani et al., 2005), and Zm-FNR1 was crystallized as described pre-viously (Kurisu et al., 2001). The crystals of Zm-FNR2 were obtained usinga solution containing 30% (w/v) polyethylene glycol 4000, 0.1 M am-monium acetate, and 0.1 M sodium acetate buffer, pH 5.5, as precipitantat 293K. The crystals of Zm-FNR3 were obtained using a solution con-taining 18% (w/v) polyethylene glycol 8000, 0.2 M sodium chloride, and0.1 M sodium acetate buffer, pH 5.5, as precipitant at 293K.

For data collection under cryogenic conditions, both crystals werebriefly soaked in reservoir solution containing 15% (v/v) glycerol. Crystalswere mounted in a nylon loop and flash-cooled in a stream of gaseousnitrogen at 100K. Diffraction data from crystals of Zm-FNR2 and Zm-FNR3 were collected with oscillation method (Df = 1.0°) using syn-chrotron radiation at beamline NW-12 of Photon Factory. Diffractionimages were collected at 100K using a Quantum4R charge-coupleddevice detector (ADSC) and GN2 cryosystem (Rigaku). The data wereprocessed and scaled using the HKL2000 program package.

Initial phase angles of Zm-FNR2 and Zm-FNR3 were both calculatedby the molecular replacement method, using MOLREP in the CCP4 suite.The crystal structure of Zm-FNR1 (PDB ID: 1GAW) was used as the searchmodel in both cases. Iterative manual model buildings were performedusing the program COOT. Throughout the crystallographic refinement,the programs CNS and Refmac5 in the CCP4 suite were used. The finalmodels of Zm-FNR2 and Zm-FNR3 were refined at 2.00 and 1.39 Å,respectively. The coordinates and structure factors for Zm-FNR2 and Zm-FNR3 have been deposited at the Protein Data Bank under accessioncodes 3VO1 and 3VO2, respectively.

The crystallographic data and the statistics of the refinement aresummarized in Supplemental Table 1 online.

Construction of Chimeric and Truncated Maize FNR Constructs

Plasmids for expression of chimeric and truncated maize FNR sequenceswere constructed by extracting At-FNR1pro-Zm-FNR1 and At-FNR1pro-Zm-FNR3 constructs from plant expression vectors by HindIII and XbaIdigestion, followed by subcloning into pASK IBA3 (Cayman Chemical

Company) to give pIBA3-At-FNR1pro-Zm-FNR1 and pIBA3-At-FNR1pro-Zm-FNR3. Restriction sites for BstBI were inserted by mutagenizing thecommon sequence coding for amino acids Ser61-Lys62 (Zm-FNR1 no-menclature) in both Zm-FNR1 and Zm-FNR3 (primers in SupplementalTable 2 online) using the QuikChange site-directed mutagenesis kit(Stratagene) to give pIBA3-At-FNR1pro-Zm-FNR1B and pIBA3-At-FNR1pro-Zm-FNR3B. Mutagenized plasmids were digested with BstBIand HindIII, and the excised At-FNR1pro-59-FNR regions ligated into theappropriate vectors containing 39-FNR regions, to create chimeras. Thetruncated Zm-FNR1 sequence (Zm-FNR1t) was created by mutagenizingthe codons for amino acids Arg43-Ala44 (primers in Supplemental Table 2online) in the pIBA3-At-FNR1pro -Zm-FNR1B vector to introduce a re-striction site forBssHII, giving the plasmid pIBA3-At-FNR1pro -Zm-FNR1BB.This plasmidwasdigestedwithBstBI andBssHII to excise the region codingfor the N terminus in the mature protein. The oligonucleotides 59-CGCGCAGGCGT-39 and 59-GTCCGCAGC-39 were then annealed and li-gated into the linearized pIBA3-At-FNR1pro-ZmFNR1BB, yielding a Zm-FNR1sequence with an intact transit peptide, but truncated mature sequence.Chimeric and truncated At-FNR1pro-Zm-FNR constructs were subcloned intopSMAB704 following digestionwith XbaI (At-FNR1pro-Zm-FNR constructs) orSacI (pSMAB704) followed by blunting with T4 polymerase and then di-gestion with HindIII.

Measurement of Chlorophyll Fluorescence and P700

All measurements were performed on 8-week-old plants between 1 and 6h into the light period. Plants were dark incubated for 20 min beforefollowing the light induction response of a single mature leaf usinga DUAL-PAM-100 (Walz). For the measurement, chlorophyll fluorescenceand P700 maxima were established followed by a further 2-min darkadaptation and then light induction at 126 µE red actinic light and 29 µEblue actinic light. Chlorophyll fluorescence parameters were calculatedbasically according to Genty et al. (1989) from dark-adapted maxima andmeasurements taken at saturating pulses, except that qP = (Fm92F)/(Fm92

Fo9), where Fo9 was estimated according to Oxborough and Baker (1997),andNPQwas calculated according to Kramer et al. (2004).ФI was calculatedbasically according to Klughammer and Schriebe (1994). ETRI and ETRIIwere calculated as µmol electrons/(m2$s) from the photosynthetically activeradiation and either ФI or ФII, respectively, based on the assumptions thatPSI and PSII absorbed equal quanta in all plants and that leaves of all plantsabsorbed 0.84 of quanta available.

Accession Numbers

Sequence data from this article can be found in the Arabidopsis GenomeInitiative or GenBank/EMBL databases under the following accessionnumbers: maize FNR1, BAA88236; maize FNR2, BAA88237; maize FNR3,ACF85815;Arabidopsis FNR1, AT5G66190; Arabidopsis FNR2, AT1G20020;maize Tic62, ACG28394.1; Arabidopsis Tic62, AT3G18890; maize TROL,ACF79627.1; Arabidopsis TROL, AT4G01050.1.

Supplemental Data

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

Supplemental Figure 1. Alignment of FNR Isoproteins from Arabi-dopsis and Maize.

Supplemental Figure 2. Comparison of FNR Binding Proteins fromArabidopsis and Maize.

Supplemental Table 1. Data Collection and Refinement Statistics forMaize FNR2 and FNR3.

Supplemental Table 2. Primers Used in Preparing ExperimentalMaterial.

FNR N Terminus Controls Membrane Association 2989

ACKNOWLEDGMENTS

The financial support of the Monbu Kagakusho Grant-in-Aid for YoungScientists 18770035 and the Deutsche Forschungsgemeinshaft Grant P2within Sonderforschungsbereich 944 is gratefully acknowledged. Wethank Renate Scheibe for helpful discussion.

AUTHOR CONTRIBUTIONS

M.T. performed research and analyzed data. B.A. performed research.N.M., I.V., and S.O. performed research. G.K. analyzed data. T.H.designed and performed research. G.T.H. designed and performedresearch, analyzed data, and wrote the article.

Received December 8, 2011; revised May 15, 2012; accepted June 27,2012; published July 17, 2012.

REFERENCES

Alte, F., Stengel, A., Benz, J.P., Petersen, E., Soll, J., Groll, M., andBölter, B. (2010). Ferredoxin:NADPH oxidoreductase is recruited tothylakoids by binding to a polyproline type II helix in a pH-dependentmanner. Proc. Natl. Acad. Sci. USA 107: 19260–19265.

Andersen, B., Scheller, H.V., and Møller, B.L. (1992). The PSI-Esubunit of photosystem I binds ferredoxin:NADP+ oxidoreductase.FEBS Lett. 311: 169–173.

Arnon, D.I., Tagawa, K., and Tsujimoto, H.Y. (1963). Role of ferre-doxin in the energy conversion process of photosynthesis. Science140: 378.

Baniulis, D., Zhang, H., Zakharova, T., Hasan, S.S., and Cramer,W.A. (2011). Purification and crystallization of the cyanobacterialcytochrome b6f complex. Methods Mol. Biol. 684: 65–77.

Benz, J.P., Lintala, M., Soll, J., Mulo, P., and Bölter, B. (2010). A newconcept for ferredoxin-NADP(H) oxidoreductase binding to plantthylakoids. Trends Plant Sci. 15: 608–613.

Benz, J.P., Stengel, A., Lintala, M., Lee, Y.H., Weber, A., Philippar,K., Gügel, I.L., Kaieda, S., Ikegami, T., Mulo, P., Soll, J., andBölter, B. (2009). Arabidopsis Tic62 and ferredoxin-NADP(H) oxi-doreductase form light-regulated complexes that are integrated intothe chloroplast redox poise. Plant Cell 21: 3965–3983.

Bojko, M., Kruk, J., and Wieckowski, S. (2003). Plastoquinones areeffectively reduced by ferredoxin:NADP+ oxidoreductase in thepresence of sodium cholate micelles. Significance for cyclic elec-tron transport and chlororespiration. Phytochemistry 64: 1055–1060.

DalCorso, G., Pesaresi, P., Masiero, S., Aseeva, E., Schünemann,D., Finazzi, G., Joliot, P., Barbato, R., and Leister, D. (2008). Acomplex containing PGRL1 and PGR5 is involved in the switchbetween linear and cyclic electron flow in Arabidopsis. Cell 132:273–285.

Forti, G., and Bracale, M. (1984). Ferredoxin-ferredoxin NADP re-ductase interaction, catalytic differences between the soluble andthylakoid bound complex. FEBS Lett. 166: 81–84.

Genty, B., Briantais, J.-M., and Baker, N.R. (1989). The relationshipbetween the quantum yield of photosynthetic electron transport andquenching of chlorophyll fluorescence. Biochim. Biophys. Acta 990:87–92.

Hald, S., Nandha, B., Gallois, P., and Johnson, G.N. (2008). Feed-back regulation of photosynthetic electron transport by NADP(H)redox poise. Biochim. Biophys. Acta 1777: 433–440.

Hanke, G.T., Endo, T., Satoh, F., and Hase, T. (2008). Altered pho-tosynthetic electron channelling into cyclic electron flow and nitriteassimilation in a mutant of ferredoxin:NADP(H) reductase. Plant CellEnviron. 31: 1017–1028.

Hanke, G.T., and Hase, T. (2008). Variable photosynthetic roles oftwo leaf-type ferredoxins in arabidopsis, as revealed by RNA in-terference. Photochem. Photobiol. 84: 1302–1309.

Hanke, G.T., Kimata-Ariga, Y., Taniguchi, I., and Hase, T. (2004). Apost genomic characterization of Arabidopsis ferredoxins. PlantPhysiol. 134: 255–264.

Hanke, G.T., Okutani, S., Satomi, Y., Takao, T., Suzuki, A., andHase, T. (2005). Multiple iso-proteins of FNR in Arabidopsis: Evi-dence for different contributions to chloroplast function and nitro-gen assimilation. Plant Cell Environ. 28: 1146–1157.

Higuchi-Takeuchi, M., Ichikawa, T., Kondou, Y., Matsui, K.,Hasegawa, Y., Kawashima, M., Sonoike, K., Mori, M., Hirochika,H., and Matsui, M. (2011). Functional analysis of two isoforms ofleaf-type ferredoxin-NADP(+)-oxidoreductase in rice using the heterol-ogous expression system of Arabidopsis. Plant Physiol. 157: 96–108.

Hosler, J.P., and Yocum, C.F. (1985). Evidence for two cyclic pho-tophosphorylation reactions concurrent with ferredoxin-catalyzednon-cyclic electron transport. Biochim. Biophys. Acta 808: 21–31.

Ilík, P., Schansker, G., Kotabová, E., Váczi, P., Strasser, R.J., andBarták, M. (2006). A dip in the chlorophyll fluorescence induction at0.2-2 s in Trebouxia-possessing lichens reflects a fast reoxidationof photosystem I. A comparison with higher plants. Biochim. Bio-phys. Acta 1757: 12–20.

Iwai, M., Takizawa, K., Tokutsu, R., Okamuro, A., Takahashi, Y.,and Minagawa, J. (2010). Isolation of the elusive supercomplexthat drives cyclic electron flow in photosynthesis. Nature 464:1210–1213.

Joliot, P., and Johnson, G.N. (2011). Regulation of cyclic and linearelectron flow in higher plants. Proc. Natl. Acad. Sci. USA 108: 13317–13322.

Jose Quiles, M., and Cuello, J. (1998). Association of ferredoxin-NADP oxidoreductase with the chloroplastic pyridine nucleotidedehydrogenase complex in barley leaves. Plant Physiol. 117:235–244.

Juri�c, S., Hazler-Pilepi�c, K., Tomasi�c, A., Lepedus, H., Jelici�c, B.,Puthiyaveetil, S., Bionda, T., Vojta, L., Allen, J.F., Schleiff, E., andFulgosi, H. (2009). Tethering of ferredoxin:NADP+ oxidoreductaseto thylakoid membranes is mediated by novel chloroplast proteinTROL. Plant J. 60: 783–794.

Klughammer, C., and Schriebe, U. (1994). An improved method,using saturating light pulses, for the determination of photosystem Iquantum yield via P700+-absorbance changes at 830 nm. Planta192: 261–268.

Korn, A., Ajlani, G., Lagoutte, B., Gall, A., and Sétif, P. (2009).Ferredoxin:NADP+ oxidoreductase association with phycocyaninmodulates its properties. J. Biol. Chem. 284: 31789–31797.

Kramer, D.M., Johnson, G., Kiirats, O., and Edwards, G.E. (2004).New fluorescence parameters for the determination of q(a) redoxstate and excitation energy fluxes. Photosynth. Res. 79: 209–218.

Kurisu, G., Kusunoki, M., Katoh, E., Yamazaki, T., Teshima, K.,Onda, Y., Kimata-Ariga, Y., and Hase, T. (2001). Structure of theelectron transfer complex between ferredoxin and ferredoxin-NADP(+) reductase. Nat. Struct. Biol. 8: 117–121.

Lintala, M., Allahverdiyeva, Y., Kangasjärvi, S., Lehtimäki, N.,Keränen, M., Rintamäki, E., Aro, E.M., and Mulo, P. (2009).Comparative analysis of leaf-type ferredoxin-NADP oxidoreductaseisoforms in Arabidopsis thaliana. Plant J. 57: 1103–1115.

Lintala, M., Allahverdiyeva, Y., Kidron, H., Piippo, M., Battchikova,N., Suorsa, M., Rintamäki, E., Salminen, T.A., Aro, E.M., and

2990 The Plant Cell

Mulo, P. (2007). Structural and functional characterization of ferredoxin-NADP+-oxidoreductase using knock-out mutants of Arabidopsis. PlantJ. 49: 1041–1052.

Majeran, W., Zybailov, B., Ytterberg, A.J., Dunsmore, J., Sun, Q.,and van Wijk, K.J. (2008). Consequences of C4 differentiation forchloroplast membrane proteomes in maize mesophyll and bundlesheath cells. Mol. Cell. Proteomics 7: 1609–1638.

Moolna, A., and Bowsher, C.G. (2010). The physiological importanceof photosynthetic ferredoxin NADP+ oxidoreductase (FNR) isoformsin wheat. J. Exp. Bot. 61: 2669–2681.

Munekage, Y., Hashimoto, M., Miyake, C., Tomizawa, K., Endo, T.,Tasaka, M., and Shikanai, T. (2004). Cyclic electron flow aroundphotosystem I is essential for photosynthesis. Nature 429: 579–582.

Okutani, S., Hanke, G.T., Satomi, Y., Takao, T., Kurisu, G., Suzuki,A., and Hase, T. (2005). Three maize leaf ferredoxin:NADPH oxi-doreductases vary in subchloroplast location, expression, and in-teraction with ferredoxin. Plant Physiol. 139: 1451–1459.

Oxborough, K., and Baker, N.R. (1997). Resolving chlorophylla fluorescence images of photosynthetic efficiency into photo-chemical and non-photochemical components: Calculation of qPand F9v/F9m without measuring F9o. Photosynth. Res. 54: 135–142.

Peng, L., Shimizu, H., and Shikanai, T. (2008). The chloroplast NAD(P)H dehydrogenase complex interacts with photosystem I in Arabidopsis.J. Biol. Chem. 283: 34873–34879.

Rajagopal, S., Bukhov, N.G., Tajmir-Riahi, H.-A., and Carpentier,R. (2003). Control of energy dissipation and photochemical activityin photosystem I by NADP-dependent reversible conformationalchanges. Biochemistry 42: 11839–11845.

Reisinger, V., and Eichacker, L.A. (2007). How to analyze protein com-plexes by 2D blue native SDS-PAGE. Proteomics 7 (suppl. 1): 6–16.

Rodriguez, R.E., Lodeyro, A., Poli, H.O., Zurbriggen, M., Peisker, M.,Palatnik, J.F., Tognetti, V.B., Tschiersch, H., Hajirezaei, M.R., Valle,E.M., and Carrillo, N. (2007). Transgenic tobacco plants over-expressing chloroplastic ferredoxin-NADP(H) reductase display normalrates of photosynthesis and increased tolerance to oxidative stress.Plant Physiol. 143: 639–649.

Satoh, K. (1982). Mechanism of photoactivation of electron transportin intact bryopsis chloroplasts. Plant Physiol. 70: 1413–1416.

Satoh, K., and Katoh, S. (1980). Light-induced changes in chlorophylla fluorescence and cytochrome f in intact spinach chloroplasts: Thesite of light-dependent regulation of electron transport. Plant CellPhysiol. 21: 907–916.

Schuster, G., Ohad, I., Martineau, B., and Taylor, W.C. (1985). Dif-ferentiation and development of bundle sheath and mesophyllthylakoids in maize. Thylakoid polypeptide composition, phos-phorylation, and organization of photosystem II. J. Biol. Chem. 260:11866–11873.

Shahak, Y., Crowther, D., and Hind, G. (1981). The involvement offerredoxin-NADP+ reductase in cyclic electron transport in chlor-oplasts. Biochim. Biophys. Acta 636: 234–243.

Sheen, J. (1995). Methods for mesophyll and bundle sheath cellseparation. Methods Cell Biol. 49: 305–314.

Shin, M., and Arnon, D.I. (1965). Enzymic mechanisms of pyridinenucleotide reduction in chloroplasts. J. Biol. Chem. 240: 1405–1411.

Stengel, A., Benz, P., Balsera, M., Soll, J., and Bölter, B. (2008).TIC62 redox-regulated translocon composition and dynamics. J.Biol. Chem. 283: 6656–6667.

van Thor, J.J., Jeanjean, R., Havaux, M., Sjollema, K.A., Joset, F.,Hellingwerf, K.J., and Matthijs, H.C. (2000). Salt shock-induciblephotosystem I cyclic electron transfer in Synechocystis PCC6803 relieson binding of ferredoxin:NADP(+) reductase to the thylakoid membranesvia its CpcD phycobilisome-linker homologous N-terminal domain.Biochim. Biophys. Acta 1457: 129–144.

Voss, I., Goss, T., Murozuka, E., Altmann, B., McLean, K.J., Rigby,S.E., Munro, A.W., Scheibe, R., Hase, T., and Hanke, G.T. (2011).FdC1, a novel ferredoxin protein capable of alternative electronpartitioning, increases in conditions of acceptor limitation at pho-tosystem I. J. Biol. Chem. 286: 50–59.

Zhang, H., Whitelegge, J.P., and Cramer, W.A. (2001). Ferredoxin:NADP+ oxidoreductase is a subunit of the chloroplast cytochromeb6f complex. J. Biol. Chem. 276: 38159–38165.

FNR N Terminus Controls Membrane Association 2991

DOI 10.1105/tpc.111.094532; originally published online July 17, 2012; 2012;24;2979-2991Plant Cell

Toshiharu Hase and Guy T. HankeManuel Twachtmann, Bianca Altmann, Norifumi Muraki, Ingo Voss, Satoshi Okutani, Genji Kurisu,

Different Thylakoid Membrane Complexes Reductase Determines Recruitment into+N-Terminal Structure of Maize Ferredoxin:NADP

 This information is current as of June 2, 2018

 

Supplemental Data /content/suppl/2012/07/03/tpc.111.094532.DC1.html

References /content/24/7/2979.full.html#ref-list-1

This article cites 46 articles, 18 of which can be accessed free at:

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists