Structural Features Determining Flower-Promoting Activity ofArabidopsis FLOWERING LOCUS TW OPEN

William Wing Ho Ho and Detlef Weigel1

Department of Molecular Biology, Max Planck Institute for Developmental Biology, D-72076 Tuebingen, Germany

ORCID IDs: 0000-0002-7360-2243 (W.W.H.H.); 0000-0002-2114-7963 (D.W.)

In Arabidopsis thaliana, the genes FLOWERING LOCUS T (FT) and TERMINAL FLOWER1 (TFL1) have antagonistic rolesin regulating the onset of flowering: FT activates and TFL1 represses flowering. Both encode small, closely relatedtranscription cofactors of ;175 amino acids. Previous studies identified a potential ligand binding residue as well asa divergent external loop as critical for the differences in activity of FT and TFL1, but the mechanisms for the differentialaction of FT and TFL1 remain unclear. Here, we took an unbiased approach to probe the importance of residues throughoutFT protein, testing the effects of hundreds of mutations in vivo. FT is surprisingly robust to a wide range of mutations, evenin highly conserved residues. However, specific mutations in at least four different residues are sufficient to convert FT intoa complete TFL1 mimic, even when expressed from TFL1 regulatory sequences. Modeling the effects of these mutations onthe surface charge of FT protein suggests that the affected residues regulate the docking of an unknown ligand. Theseresidues do not seem to alter the interaction with FD or 14-3-3 proteins, known FT interactors. Potential candidates fordifferential mediators of FT and TFL1 activities belong to the TCP (for TEOSINTE BRANCHED1, CYCLOIDEA, PCF) family oftranscription factors.

INTRODUCTION

The onset of flowering is a crucial event in the plant life cycle. Amultitude of external and endogenous signals are integrated onthe promoters of a small set of genes, among which homologsof FLOWERING LOCUS T (FT) play a particularly important andhighly conserved role (Kardailsky et al., 1999; Kobayashi et al.,1999; Kojima et al., 2002; Lifschitz et al., 2006; Kobayashi andWeigel, 2007; Turck et al., 2008; Hecht et al., 2011). While FTRNA is produced exclusively in leaves, the protein acts primarilyat the shoot apex, apparently by associating with the FD tran-scription factor to directly activate downstream genes (Abeet al., 2005; Wigge et al., 2005). A large body of circumstantialevidence suggests that FT protein moves from the site of itsproduction, the leaves, through the plant to the site of its action,the shoot apex (Lifschitz et al., 2006; Corbesier et al., 2007;Jaeger and Wigge, 2007; Lin et al., 2007; Mathieu et al., 2007;Tamaki et al., 2007; Notaguchi et al., 2008). FT thus qualifies asan essential component of the mobile flower-promoting signal,florigen (Kobayashi and Weigel, 2007; Turck et al., 2008). FTencodes a small protein of only 175 amino acids, and its mobilityis thus not surprising. However, recent work indicates that FTtrafficking is regulated (Liu et al., 2012).

In Arabidopsis thaliana, FT belongs to a small gene familythat includes TERMINAL FLOWER1 (TFL1), an inhibitor offlowering (Bradley et al., 1997; Ohshima et al., 1997). The

structures of FT, TFL1, and the TFL1 orthologous proteinCENTRORADIALIS from snapdragon (Antirrhinum majus)have all been solved (Banfield and Brady, 2000; Ahn et al.,2006). As expected from the primary sequence similarity, thefold is similar to that of the phosphatidylethanolamine bindingprotein (PEBP) family from animals, with some notable dif-ferences in and around the anion binding pocket of the animalproteins (Banfield et al., 1998; Serre et al., 1998). In plants,there is no evidence for in vivo binding of phospholipids orother anions.There are only 39 nonconservative substitutions that distin-

guish FT and TFL1, yet their in vivo activities are opposite, evenwhen expressed from the same promoter. This intriguing obser-vation has led several groups to investigate the sequence differ-ences underlying antagonistic function. An early report showedthat a Y85H substitution conferred partial TFL1-like activity on FT,while the converse change, H88Y, conferred weak FT-like activityon TFL1 (Hanzawa et al., 2005). Subsequent experiments dem-onstrated that an external loop together with the adjacent seg-ment is largely sufficient to convert TFL1 into an FT-like activator.The effects are not entirely symmetrical, as both the external loopand the adjacent segment on their own can convert FT into aTFL1-like repressor (Ahn et al., 2006). Notably, the external loopis not the most different portion of the FT and TFL1 proteins, but itevolves very differently in the two proteins: While it is almostcompletely conserved in FT, it appears to be under relaxed se-lection in TFL1 (Ahn et al., 2006). Furthermore, a residue in theexternal loop of TFL1 interacts with the Y85 residue required forTFL1 function, while there is no equivalent interaction in FT(Hanzawa et al., 2005; Ahn et al., 2006). This difference has beenexploited during natural evolution in sugar beet (Beta vulgaris sspvulgaris), where mutations in the loop have imparted flower-repressing activity of FT paralogs (Pin et al., 2010).

1 Address correspondence to weigel@weigelworld.org.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: Detlef Weigel (weigel@weigelworld.org).W Online version contains Web-only data.OPENArticles can be viewed online without a subscription.www.plantcell.org/cgi/doi/10.1105/tpc.113.115220

The Plant Cell, Vol. 26: 552–564, February 2014, www.plantcell.org ã 2014 American Society of Plant Biologists. All rights reserved.

We used an unbiased approach in which we examined hun-dreds of mutations to map residues required for FT flower-promoting activity. We show that altering surface charge canconvert FT into a TFL1-like repressor. Based on these obser-vations, we propose a model for docking of an unknown ligand.This ligand does not appear to be the FD transcription factor(Abe et al., 2005; Wigge et al., 2005) or the 14-3-3 proteins thatlink FD and FT or TFL1 (Pnueli et al., 2001; Taoka et al., 2011).Instead, members of the TCP family of transcription factors arecandidates for factors that mediate differential activity of FT andTFL1 (Cubas et al., 1999; Mimida et al., 2011).

RESULTS

Assay System and Random Mutagenesis

The FT protein sequence is highly conserved among floweringplants (Supplemental Figure 1). To identify residues with strongeffects on FT activity, we chose to overexpress random pointmutations of the FT cDNA. Overexpression of wild-type FT leadsto very early flowering and the formation of terminal flowers, withtransformants having a narrow distribution of flowering times(Kardailsky et al., 1999; Kobayashi et al., 1999). The absence ofaccelerated flowering in transformants therefore indicates mu-tations that inactivate FT.

We used PCR to generate random mutants of FT. After eval-uating the PCR-based mutagenesis conditions (SupplementalFigure 2), we sequenced 1241 clones of the final FT muta-genesis, of which 1100 had a complete FT insert. Amongthese, 9% had no mutation and 31% had only synonymoussubstitutions. The remainder, 60%, was almost equally dividedbetween clones with one or with more than one nonsynonymousmutation.

The products from the mutagenesis reaction were cloned intoan Escherichia coli plasmid vector, plasmid DNA was preparedfrom a pool of ;36,000 bacterial colonies, and the inserts wereexcised and subcloned en masse into a plant transformationvector between a cauliflower mosaic virus 35S (CaMV35S)promoter and yellow fluorescent protein (YFP [mCitrine]) se-quences. The purpose of fusing the FT open reading frame tothat of YFP was to avoid false negatives in the overexpressionassay due to introduction of premature stop codons. Agro-bacterium tumefaciens transformants were pooled and used toinfect wild-type Arabidopsis plants.

The general strategy for assessing FT activity of mutants isdelineated in Figure 1. Because the effects of FT overexpressionare strongest in short days, we first screened 3320 primarytransformants in this condition. There was no overlap in theflowering times of the positive (overexpressing FT-YFP) andnegative (empty vector) controls (Figure 2A). Two hundred andfifty-two negative control plants transformed with an emptyvector flowered with 17 leaves or more (average, 35), and of 401positive controls transformed with an unmutagenized 35S:FT-YFP construct, all flowered with fewer than 13 leaves, and 398with fewer than nine leaves.

Given the fraction of nonsynonymous mutations de-termined by sequencing, at least 40% of all transformantswere expected to flower similarly early as the positive control.

Indeed, the majority of T1 plants flowered with fewer than nineleaves, as seen in 99.3% of positive controls. These were as-sumed to be early-flowering and to have retained substantial FTactivity, even though some had only barely detectable YFPfluorescence (Figure 2A; Supplemental Figure 3).Only a minority of T1 plants, <4%, flowered in the same

range as the negative controls, with 17 leaves or more andwere classified as not early flowering. Fifty-two T1 plants withstrong or intermediate fluorescence, representing 33 uniquemutations, fell into this class. To distinguish mutations thatinactivated FT activity from those that interfered with normalFT function in a dominant-negative manner or that had gainedneomorphic activity, the T2 progeny from these transformantswas evaluated in long days (Figure 2B; Supplemental Table 1),where FT function is required for the acceleration of flowering

Figure 1. Random Mutagenesis Assay Strategy.

Around 36,000 randomly mutated FT cDNA copies were generated byPCR. These were cloned en masse into a plant transformation vector foroverexpression in wild-type plants as YFP fusions. Screening of 3320primary transformants in short days identified 52 late-flowering, YFP-positive lines with reduced FT activity, representing 33 unique mutants.The progeny of these 33 lines was reevaluated in long days to distinguishmutations that inactivated FT (14 lines) from those that introduced ne-omorphic activity and delayed rather than accelerated flowering whenoverexpressed (19 lines). See Supplemental Figure 3 for fluorescencelevels.

Structural Determinants of FT Activity 553

(Koornneef et al., 1991). The transgene in 14 of these T2 lineshad apparently lost FT activity, and plants flowered with fewerthan 13 leaves on average, while 19 transgenes had dominant-negative or neomorphic activity and caused plants to flowerwith on average 24 or more leaves, with some flowering evenlater than ft-10, a null allele (Yoo et al., 2005).

The FT coding regions from the mutated transgenes in allthree categories (wild type, loss of function, and gain of function)were amplified by PCR, and the products were individually se-quenced. Out of the 175 codons, 472 unique substitutions in166 codons were found (Supplemental Figure 4). One hundredand fifty mutants had nonsynonymous mutations in more than

Figure 2. Flowering Times of T1 and T2 Lines.

(A) Flowering times of T1 transformants and controls in short days. Light green and pink indicate distribution of positive and negative controls. The twocohorts on the right are a subset of the 35S:mFT-YFP population shown in the middle. Note that many plants with barely detectable fluorescence alsoflowered very early but were not considered further (see Figure 1 and Supplemental Figure 3).(B) Flowering time ranges of selected T2 lines in long days. Ochre indicates range between negative control and ft-10. Red type indicates residuessubsequently investigated in detail. Lines in dark gray are statistically significantly later than ft-10 in a two-tailed Student’s t test with Bonferronicorrection. Error bars indicate SD. See also Supplemental Table 1.

554 The Plant Cell

one codon. In the 245 mutants with retained FT activity, 155codons were affected, in the 14 mutants with reduced FT func-tion, 11 codons were affected, and in the 19 mutants with anti- orneomorphic activity, 12 codons were affected (SupplementalFigure 4). In some cases, which are discussed in more detailbelow, different substitutions at the same codon differed in theireffects on FT activity.

Targeted Mutagenesis

Twelve of fifteen unique anti- or neomorphic mutations werefound at or close to the potential ligand binding pocket and theadjacent domains encoded by segments B and C of the fourthexon (Ahn et al., 2006). The Adaptive Poisson-Boltzmann Solversoftware (Baker et al., 2001) predicted this surface to have mostlynegative charge (Supplemental Figure 5). Intriguingly, at least 11of the anti- or neomorphic activity mutations (Y85H, E109K,Q127R, L128H, W138R, Q140H, Q140K, E146K, N152K, L155R,and R173E) changed the charge of the corresponding residues,suggesting that they might also affect overall surface charge ofFT. Therefore, we modeled such changes by mutating in silico170 of the 175 residues of FT. Point mutations predicted tochange the surface charge were then engineered into the FT openreading frame by site-directed mutagenesis. In this series of ex-periments, we did not use the YFP tag, to avoid any potentialconfounding effects due to the more than doubling in size of FT-YFP relative to native FT. In addition, we used site-directed mu-tagenesis to confirm the effects of mutations in 108 residues thatare highly conserved in FT homologs (Supplemental Figure 1).Altogether, 228 mutations at 122 codons were introduced intowild-type plants. Some mutations at conserved positions did notaffect the ability of FT to induce early flowering but reduced theextent to which terminal flowers were formed, while others had nodetectable effects on FT function at all (Supplemental Table 2).Only a relatively small number, mutations at 17 out of 108 highlyconserved residues, reduced both the ability of FT to accelerateflowering in short days and to convert the indeterminate shootmeristem into a determinate floral meristem. Because plantstransformed with these variants did not have any of the featuresseen in TFL1 overexpressors (Ratcliffe et al., 1998), we infer thatthese were loss-of-function mutations.

Mutations Near the Entrance of the Potential LigandBinding Pocket

The entrance of the potential ligand binding pocket is impor-tant for FT- or TFL1-like activity, as a Y85H substitution par-tially confers TFL1 properties onto FT (Hanzawa et al., 2005).Both Tyr-85 in FT and the corresponding His-88 in TFL1 formhydrogen bonds with the orthologous residues Glu-109 andGlu-112, but only His-88 forms a second hydrogen bond withAsp-144 in TFL1, while Tyr-85 and Gln-140 (which corre-sponds to Asp-144) in FT do not (Ahn et al., 2006) (Figure 3).Importantly, both pairs of residues, Y85/Q140 and H88/D144,are diagnostic for FT- and TFL1-like proteins (with the excep-tion of Brassica TFL1 proteins, which have an Arg instead ofa His at position 88 (Mimida et al., 1999). The importance ofhydrogen bonds was confirmed by an Y85A substitution (Fig-ure 3), which largely inactivated FT (Supplemental Table 2). Wealso examined a Y85W substitution; similar to the side groupsof Tyr-85 and His-88, the indole side group of Trp can supportp-p stacking, but it should form a hydrogen bond only withGlu-109 (Figure 3). In contrast with Y85A, Y85W had partial FTactivity, supporting the hypothesis that the intramolecularbond with Glu-109 is important for FT function (SupplementalTable 2). We note that with regard to the Y85H mutation, the2.9-Å hydrogen bond between His-88 and Asp-144 in TFL1(Ahn et al., 2006) would be replaced by a weaker interactionbetween the equivalent His-85 and Gln-140 in FT, with a pre-dicted length of more than 3.5 Å (Figure 3). This may explainthe relatively weak TFL1-like activity of Y85H.Among the newly identified residues that could be mutated to

confer TFL1-like properties, we decided to focus on six posi-tions that are invariant or almost invariant among FT homologs.The first two are Glu-109, which shows similar interactions inFT and TFL1, and Gln-140, which does not. The surface sur-rounding these two residues is strongly negative in charge in FT,but positive in TFL1 (Figure 4). Notably, even though Glu-109 inFT corresponds to Glu-112 in TFL1, charge reversal througha E109K mutation caused FT to behave like TFL1 (Figure 5).Similar effects were seen when hydrophobic side chains wereintroduced (E109A and E109M), predicted to cause a localizedexpansion of positive charge (Figure 4). By contrast, E109D,

Figure 3. Intramolecular Hydrogen Bonds.

Effects of Tyr-85 mutations are shown; wild-type FT is at the far left, and wild-type TFL1 is at the far right. Hydrogen bonds are shown as dashed lines.Hot pink indicates FT residues, cyan TFL1 residues, and green other substitutions.

Structural Determinants of FT Activity 555

which preserves the negative charge, did not affect FT activity.Compared with Glu-109, Gln-140 tolerates both negative andneutral substitutions (Q140D and Q140A) but not positive ones(Q140K and Q140R) (Figures 4 and 5; Supplemental Table 3).Furthermore, charge reversion deep inside the ligand bindingpocket (such as D71N) did not convert FT into a TFL1-likemolecule (Supplemental Table 2). This is reminiscent of theobservation that binding of the human PEBP homolog Raf ki-nase inhibitor (RKIP) to Raf-1 relies on residues near but notinside the ligand binding pocket (Tavel et al., 2012).

Mutations Distant from the Potential Ligand Binding Pocket

The next two residues we investigated in detail were Leu-128and Asn-152, which are positioned in an external loop and anadjacent a-helix. The external loop including Leu-128 is understrong purifying selection and almost invariant in FT proteins,but it is under relaxed selection and very variable in TFL1 pro-teins (Ahn et al., 2006). The only tolerated mutation was a veryconservative substitution, L128I, while introduction of a charged

Lys, as in TFL1 (L128K), bestowed weak TFL1-like activity onthe mutant protein (Supplemental Table 3). Different from theexternal loops, the adjacent a-helices are very similar in FT andTFL1. However, although Asn-152 is strongly conserved in FT,but not TFL1 orthologs, mutations to negative (N152D), neutral(N152A), hydrophilic (N152S), or aromatic (N152Y) side chainswere all largely tolerated. The exception was N152K, which incontrast with the other mutations strongly affected surfacecharge and completely converted FT into a TFL1-like floral re-pressor (Figures 4 and 5).A beautiful example of Darwinian selection having changed

a floral activator in the FT/TFL1 family into a floral repressorcomes from sugar beet, where floral repressing activity of oneof the two FT orthologs, Bv-FT1, is associated with threemutations in the external loop corresponding to Y134N,G137Q, and W138Q in Arabidopsis FT. Simultaneous mutationof the three residues converts Bv-FT1 into a floral activator,while the opposite mutations in the activator Bv-FT2 impartfloral repressive activity (Pin et al., 2010), in line with our pre-vious demonstration of the external loop as being essential for

Figure 4. Predicted Effects of Point Mutations on Surface Electrostatic Potential of FT.

Residues tested in each structure are highlighted in yellow and the external loop and FT sequence signature from segment 4C (Ahn et al., 2006) inpurple in the ribbon representation on the far left. The ligand binding pocket is indicated by an asterisk where visible. TFL1 is shown for comparison onthe right. FT variants with FT-like activity are labeled red, those with TFL1-like activity blue, and reduced- or loss-of-function variants in black.

556 The Plant Cell

discriminating between FT- and TFL-like activities (Ahn et al.,2006). Based on heterologous overexpression in Arabidopsis, ithas been suggested that only a Y134N/W138Q double mutationcan turn Bv-FT2 into a floral repressor (Klintenäs et al., 2012).

To dissect the contributions made by individual residues inthe external loop, we introduced mutations at all three posi-tions. We found that neither G137A, G137W, G137E (which issimilar to G137Q), nor G137R conferred repressive activity onArabidopsis FT, while Y134A and Y134K as well as W138E,W138S, and W138A do (Supplemental Table 3). That mutatingeither Tyr-134 or Trp-138 was sufficient for converting FT intoa TFL1-like molecule is in agreement with the hypothesis thatthe external loop functions as a molecular interface with otherproteins only in FT, but not in TFL1 (Ahn et al., 2006). However,unlike the Glu-109 and Gln-140 residues near the ligandbinding pocket or Leu-128 and Asn-152 in and next to theexternal loop, the activities of Tyr-134 or Trp-138 wereindependent of the surface charge change (Figure 5;Supplemental Figure 6). They share with Y85 the characteristicof an aromatic side chain, and all three are close in the FTstructure. Different from Tyr-85, which is embedded in thebinding pocket entrance, the side groups of Tyr-134 and Trp-138 are protruding and easily accessible, suggesting roles inmediating interactions with other molecules.

Finally, to confirm that the TFL1-like mutants discussedabove, E109K, Y134K, W138E, Q140K, and N152K, had full

TFL1-like activity, we expressed them in a TFL1 genomiccontext with 2.2-kb upstream and 4.6-kb downstream se-quences. Even though it has not been shown directly thatthese sequences recapitulate the native TFL1 expressiondomain, they are sufficient for phenotypic complementationof tfl1 mutants without inducing gain-of-function phenotypes(Sohn et al., 2007; Kaufmann et al., 2010). Inactivation of FTand its paralog TWIN SISTER OF FT (TSF ) greatly delaysflowering but does not affect flower and shoot morphology,as overexpression of TFL1 does (Figure 6A), allowing dif-ferentiation between antimorphic, dominant-negative FTmutants and neomorphic FT mutants with TFL1-like activity.All five mutants behaved very differently from normal or loss-of-function variants of FT expressed from TFL1 regulatorysequences and complemented the tfl1-1 mutant defects(Figure 6B; Supplemental Table 4), confirming that TFL1-likeactivity did not require constitutive overexpression.

Correlation of TCP Interaction with FT Surface Charge

Several of the mutations that endow FT mutants with TFL1-like activity, especially E109K and Q140K, were predicted todrastically alter the charge of a single surface near the po-tential ligand binding pocket. So far, two types of proteinsthat interact with both FT and TFL1 orthologs have beendescribed in detail: 14-3-3/GENERAL REGULATORY FAC-TOR (GRF) proteins and the FD transcription factor (Pnueli

Figure 5. Effects of Side Chain Properties at Critical Positions on FTActivity.

Flowering time ranges of T1 transformants overexpressing FT variants andcontrols in short days. Far left, controls; negative control is empty vector withCaMV35S promoter. Middle, charge-dependent positions; right, charge-in-dependent positions. T1 plants were grown in short days. The mutants in-dicated in ochre were not significantly earlier than 35S:TFL1 using two-tailedStudent’s t test with Bonferroni correction. Error bars indicate 6 SD.

Figure 6. FT Mutants with TFL1-Like Activity.

(A) Comparison of inflorescence tips of plants overexpressing TFL1, FT,or FT mutants with TFL1-like activity. ft-10 tsf-1 is shown for comparisonon the far right. Asterisks indicate leafy shoots typical for 35S:TFL1plants (Ratcliffe et al., 1998). Arrow points to terminal flower in 35S:FT.(B) Complementation of tfl1-1 mutant with constructs using TFL1 ge-nomic sequences. The bottom panel shows top views of inflorescences.

Structural Determinants of FT Activity 557

et al., 2001; Abe et al., 2005; Wigge et al., 2005). Based onstudies with a small FT peptide, it has been suggested thatthe interaction of FT and TFL1 with FD is thought to be in-direct, mediated by the 14-3-3 adaptor protein, with yeast14-3-3 being able to substitute for the plant 14-3-3 in yeasttwo-hybrid assays (Taoka et al., 2011). We tested proteininteractions in quantitative yeast two-hybrid assays and inqualitative bimolecular luminescence complementation as-says after transient expression in Nicotiana benthamiana. Asexpected, FT, TFL1, the TFL-1–like FT mutant Q140K as wellas the FT mutant Q140D, which is predicted to have the samesurface charge as native FT, all interacted similarly well withseveral 14-3-3/GRF proteins and with FD (Figures 7 and 8;Supplemental Figure 7).

Recently, a diverse set of TCP proteins, from both miR319-targeted and non-miR319-targeted clades (Martín-Trillo andCubas, 2010), has been shown to interact with FT (Mimida et al.,2011; Niwa et al., 2013). At least one, TCP18/BRANCHED1

(BRC1), interacts specifically with FT but not TFL1 (Niwa et al.,2013). Furthermore, several of these affect flowering time (Efroniet al., 2008; Schommer et al., 2008; Niwa et al., 2013)(Supplemental Table 5). TCPs are thus candidates for discrimi-nating between FT and TFL1 proteins based on their surfacecharge. We tested all 24 TCP proteins encoded in the Arabi-dopsis genome for interaction with FT in yeast two-hybrid as-says; nine TCP proteins from class I and one from class IIinteracted with FT at a level that was comparable to that of FDand the majority of 14-3-3/GRF proteins (Figure 7). In sevencases, the interactions with TFL1 were significantly weaker, anda similar trend was seen with the TFL1 mimic FT (Q140K), butnot with the Q140D mutant, which retains FT activity. By con-trast, no statistically significant differences were seen for theinteraction with GRF proteins or FD. The same trends wereobserved in qualitative assays using bimolecular luminescencecomplementation assays inN. benthamiana (Figure 8; SupplementalFigure 7).

Figure 7. Yeast Two-Hybrid Interaction Assays.

Results from six independent replicates. The darker gray line is for comparison to GRF results. Symbols (‡) indicate significant differences betweenvertical comparisons using two-tailed Student’s t test with Bonferroni correction at P < 0.001. Error bars indicate 6SD.

558 The Plant Cell

Effect of Expression Domain on Mutant FT Activity

FT protein likely travels from its site expression in phloemcompanion cells via the sieve tubes to the shoot apex, where itforms an active complex with the FD transcription factor (Takadaand Goto, 2003; Abe et al., 2005; Wigge et al., 2005; Corbesieret al., 2007; Jaeger and Wigge, 2007; Mathieu et al., 2007;Notaguchi et al., 2008). Because the export of FT from thephloem companion cells has recently been shown to be regu-lated (Liu et al., 2012), we wanted to ascertain whether any of themutations were likely to specifically affect FT mobility. To thisend, we expressed the 228 variants from the targeted mu-tagenesis collection under the companion cell-specificSUCROSE-PROTON SYMPORTER2 (SUC2) promoter and shootapical meristem–specific FD promoter and compared the effectswith those resulting from constitutive expression from theCaMV35S promoter (Odell et al., 1985; Imlau et al., 1999; Mathieuet al., 2007). While in the vast majority there was very goodcorrelation between the three promoters, two mutations stoodout: D17K and V18A, where the SUC2 promoter driven mutantsseemed to be much less effective than the FD or CaMV35Spromoter–driven mutants (Supplemental Figure 8 and SupplementalTable 6). This was reminiscent of what has been observed forFT variants with limited mobility due to the addition of largeprotein tags (Jaeger and Wigge, 2007; Mathieu et al., 2007;

Notaguchi et al., 2008). The two adjacent conserved residuesAsp-17 and Val-18 are thus good candidates for interactionwith proteins such as FT-INTERACTING PROTEIN, whichregulates FT export from the phloem (Liu et al., 2012). In ad-dition, we also found five mutants where the FD promoterconstruct was less effective than the CaMV35S or SUC2 pro-moter constructs (N39A, N39D, R44F, G57H, G58E, and D60H;Supplemental Table 6). As there was a general tendency for theFD promoter constructs to have weaker effects, these fivemutants are more difficult to interpret.

DISCUSSION

The closely related FT and TFL1 proteins have oppositefunctions, with FT strongly promoting flowering and TFL1repressing flowering as well as suppressing floral identity(Shannon and Meeks-Wagner, 1991; Bradley et al., 1997;Ohshima et al., 1997; Ratcliffe et al., 1998; Kardailsky et al.,1999; Kobayashi et al., 1999). Previous work identified in-dividual residues that reduce the floral promoting activity ofFT as well as larger segments that impart floral repressingactivity on FT, with only the larger changes converting FTcompletely into a TFL1-like molecule (Hanzawa et al., 2005;Ahn et al., 2006; Pin et al., 2010). Here, we demonstrated that

Figure 8. Bimolecular Luciferase Complementation Assays.

Young but fully expanded 4-week-old leaves were strictly chosen for infiltration. Experiments were repeated twice. Complete set of assay results isshown in Supplemental Figure 7.

Structural Determinants of FT Activity 559

despite 39 nonconservative amino acid differences betweenFT and TFL1, specific mutations in each of four critical resi-dues, Glu-109, Trp-138, Gln-140, and Asn-152, can turn FTinto a TFL1-like floral repressor. The relevant mutations at allfour positions have strong effects on protein surface charge.It is striking that the effect of the E109K mutation could nothave been predicted based on a sequence comparison be-tween FT and TFL1, as the orthologous position in TFL1 isalso occupied by a Glu.

It has been noted before that the conformation of the po-tential ligand binding pocket does not readily explain differ-ences in FT and TFL1 activity (Ahn et al., 2006) but that theexternal loop, which is critical for FT and TFL1 function,could be involved in regulating access of a potential ligand tothe binding pocket. The surfaces adjacent to the ligandbinding pocket as well as the external loop have the samenegative charge as the ligand binding pocket only in FT,suggesting that FT, but not TFL1, would support access bya positively charged coactivator. The two aromatic residuesTyr-134 and Trp-138 found only on FT may further increaseselectivity for a coactivator with corresponding aromatic sidegroups that can engage in p-p stacking (Figure 9). This re-pulsion/lock model thus makes explicit predictions aboutthe molecular details of FT and TFL1 interaction with so farunknown coregulators. Whether TFL1 is merely unable toeffectively bind this coactivator, as proposed before (Ahnet al., 2006), or actively recruits a corepressor remainsunclear.

There has been conflicting evidence whether FD protein inter-acts similarly well with FT and TFL1 (Abe et al., 2005; Wiggeet al., 2005; Hanano and Goto, 2011). It has recently beenshown that FD is genetically required for the effects of TFL1overexpression (Jaeger et al., 2013). This may explain whyplants that lack activity of both FD and its paralog FDP flowersubstantially earlier than mutants in which FT and its closeparalog TSF are inactivated (Yoo et al., 2005; Jang et al., 2009;Jaeger et al., 2013), since ft tfl1 double mutants flower at anintermediate time relative to the ft and tfl1 single mutants(Hanzawa et al., 2005; Ahn et al., 2006). Thus, FD and FDP areunlikely candidates for discriminating between FT and TFL1effects, in agreement with our observation that changing thesurface charges of FT does not have a major impact on in-teraction with FD.

It is possible that FT directly recruits the basic transcriptionactivation machinery through interaction with a positivelycharged subunit, similar to fusions of FT or TFL1, which havehyper- and antimorphic activity, respectively (Wigge et al.,2005; Hanano and Goto, 2011). Alternatively, their oppositeactivities might be mediated by differential recruitment ofspecific transcription factors. However, increasing evidencesupports recruitment of bridging molecules, such as TCPproteins (Kosugi and Ohashi, 1997; Cubas et al., 1999). It hasbeen reported that TCP18/BRC1 interacts in yeast with FTand TSF, but not with TFL1, and genetic studies have shownthat BRC1 functions apparently as an inhibitor of FT/TSFactivity in lateral shoots (Niwa et al., 2013). We have foundadditional TCPs that interact more strongly with FT than withTFL1 in both yeast and in transient plant assays, with most

of them belonging to class I. Flowering defects have beendescribed for class II TCPs, although the molecular targetsare unknown (Palatnik et al., 2007; Efroni et al., 2008;Koyama et al., 2010; Niwa et al., 2013). Common to bothclass I and class I TCPs are effects on cell proliferation(Martín-Trillo and Cubas, 2010), and there is evidence for FTand TFL1 being able to affect more general growth pro-cesses (Pnueli et al., 1998, 2001; Lifschitz et al., 2006; Kumaret al., 2012). Further experiments are required both for identi-fying shared genomic targets of FT, TFL1 and TCPs, and forsubstantiating the generality of FT/TFL1 interactions with TCPsin vivo.In conclusion, we presented evidence for surface charge be-

ing a critical determinant of FT and TFL1 function, likely affectingthe recruitment of transcriptional coactivators and corepressors.In addition, our work confirms the power of large-scale randommutagenesis experiments coupled with functional tests intransgenic plants for the identification of structural activitydeterminants (Mateos et al., 2010).

Figure 9. Repulsion/Lock Models.

FT binds a coactivator, while TFL1 may either merely exclude a coac-tivator (middle) or attract a corepressor (bottom). Blue color representspositively charged, while red and pink represent strongly and moreweakly negatively charged regions. Gray arrows indicate electrostaticinteractions. p-p represents stacking between aromatic groups. AR,aromatic side chain.

560 The Plant Cell

METHODS

Plant Material and Growth Conditions

The wild type was Arabidopsis thaliana Columbia-0. T-DNA insertion lineswere obtained from the Nottingham Arabidopsis Stock Center unlessotherwise specified. tcp5 tcp13 tcp17 and tcp2 tcp4 tcp10 triple mutantshave been described (Efroni et al., 2008; Schommer et al., 2008). Plantswere cultivated under a 1:2 mixture of cool white and warm fluorescentlight, with a fluence rate of 125 to 175 µmol m22 s21. Plants were grown at65% humidity under long (16 h of light) or short days (8 h of light) at 23°C.Transgenic Arabidopsis plants were generated with the floral dip method(Clough and Bent, 1998). Transgenic seedlings were selected on soil with0.05% glufosinate (Basta).

Plasmid Construction

To construct 35S:HA:mFT:(GS)4:YFP:rbcs for random mutagenesis,a mCitrine version of YFP (Griesbeck et al., 2001) preceded by a multi-cloning site (MCS) was cloned into pGEM-T Easy (Promega) (pWH26). TheCaMV35S promoter and rbcs terminator were amplified and inserted intoApaI-SacII andSpeI-PstI sites of pWH26, and the resulting gene cassetteswere shuttled into pFK202, a derivative of pGREEN-IIS (Hellens et al.,2000), yielding pWH31. PCR was used to insert sequences coding for anHA tag at the N terminus and a (Gly-Ser)4 linker at the C terminus of FTprotein. This fragment was cloned into pGEM-T Easy (pWH25). Usingprimers in the HA and GS4 sequences, PCR-mediated random muta-genesis was performed with the GeneMorph II random mutagenesis kit(Clontech). PCR products were digested and ligated into pWH31 line-arized with HindIII-XhoI (pWH32). pWH31 was used as empty vectorcontrol, while the positive control was made in the same way as pWH32but using Pfu DNA Pol (Fermentas) for FT amplification (pWH33).

For site-directedmutagenesis of the central portion ofFT, overlap extensionPCR was performed (Ho et al., 1989). For site-directed mutagenesis of thetermini, single mutagenic primers were used. The PCR products were clonedinto entry vector pJLBlueF and recombined into pGREEN-IIS with CaMV35S,SUC2, or FD promoters (Mathieu et al., 2007) using LR Clonase II (Invitrogen).

For tfl1-1 complementation, the SUC2 promoter in an pGREEN-IISdestination vector (de Felippes et al., 2011) was released by XhoI-PstIexcision and replaced by 2.2-kb upstream sequences of TFL1 (pWH1352).The 4.6-kb downstream sequences essential for TFL1 expression (Kaufmannet al., 2010) and a MCS (PstI-NotI-SmaI-SalI-NcoI-BamHI) were inserted intoSpeI-SacI and PstI/BamHI sites of pWH1352 (pWH1354). Sequences of FT,mFT, and TFL1 from start to stop codon were inserted intoNotI-NcoI and PstI-SpeI sites of pWH1354.

To facilitate subsequent cloning, pairs of new entry and destinationvectors were made for yeast growth and quantitative ortho-nitrophenyl-b-galactoside assays. For the entry vector, an ApaI site was removedfrom a self-ligated pCR8/GW/TOPO vector (Invitrogen) by treatment ofS1 nuclease (Fermentas) (pEN1). A linker with MCS (EcoRI overhang-2bp-NotI-NcoI-KpnI-HindIII-SmaI-BglII-PstI-ApaI-SalI-SpeI-1 bp-EcoRIoverhang) was ligated to pEN1 linearized by EcoRI. The entry vector withforward-oriented MCS was named pEN2F. For destination vectors,MCS of pGBKT7:GAL4_DBD and pGADT7:GAL4_AD (Clontech) werereplaced by Gateway cloning sites (attR:ccdB-CmR) flanked by NdeI-NotI (pWH1016) and NdeI-XhoI (pWH1017) sites. cDNA was synthe-sized from wild-type RNA using the SuperScript II cDNA synthesis kit(Invitrogen). GRF1-15 and TCP1-24 (except TCP9, 11, and 21) wereamplified from the cDNA and cloned into pEN2F using NotI-XhoI:SalIsites. TCP9, 11, and 21 were cloned into pEN2F using NotI-SpeI sites.FT,mFT, FD, and TFL1were cloned into pEN2F using HindIII-XhoI sites.Entry clones of GRF1-15, TCP1-24, and FD were recombined topWH1017 and transformed to yeast strain AH109. FT, Q140K, Q140D,

and TFL1 were recombined into pWH1016 and transformed into yeaststrain Y187 for mating (Clontech).

For the bimolecular luciferase complementation assays, destinationvectors were constructed by replacing the MCS of JW771 (35S:LUCn)and JW772 (35S:LUCc) (Gou et al., 2011) with Gateway cloning sites (attR:ccdB-CmR) flanked by KpnI-SalI (pWH1354) and KpnI-PstI (pWH1353)sites. FT, TFL1,Q140K, andQ140Dwere cloned into pEN2F usingHindIII-XhoI sites. Entry clones of GRF1-15, TCP1-24, and FD (from the yeast two-hybrid set) were recombined into pWH1353, while FT, Q140K, Q140D, andTFL1were recombined into pWH1354. Expression clones were transformed toAgrobacterium tumefaciens GV3101 for infiltration of Nicotiana benthamiana.

Primers used for vector construction are shown inSupplemental Table 7.

Fluorescence Measurement

A Leica MZ FLIII stereomicroscope was calibrated each time using leavesof transgenic plants transformed with pWH33 (positive control) andpWH31 (negative control) to define maximal as well as backgroundfluorescence. Rosette leaves (at four- to six-leaves stage) from each plantwere freshly excised in dim light. Transgenic plants were examined for theexpression of YFP using the stereomicroscope fitted with wide- andband-pass YFP filters and an AxioCam HRc digital camera (Carl Zeiss)with AxioVison software (version 3.1).

In Silico FT Mutagenesis and Electrostatic Potential Calculation

Three-dimensional structures of wild-type FT (entry 1WKP) and TFL1(entry 1WKO) were obtained from the RCSB Protein Data Bank and werevisualized by PyMOL molecular graphics system (version 1.5.0.4;Schrödinger) (Ahn et al., 2006). During the in silico mutagenesis, first thebackbone-dependent rotamer representing the highest frequencies ofoccurrence in proteins was adopted, unless otherwise specified. For theelectrostatic potential calculation, the Adaptive Poisson-BoltzmannSolver plug-in was applied (Baker et al., 2001). Before calculation, anywater and solvent molecules that disturb the native potential of proteinsfound in the visualization were removed. PDB2PQR and default gridsettings were applied for the calculations.

Yeast Two-Hybrid Assays

b-Galactosidase activity was quantified using ortho-nitrophenyl-b-galactoside assays as described in the Yeast Protocols Handbook(Clontech), with at least six replicates.

Bimolecular Luciferase Complementation Assays

N. benthamiana was transiently transformed as described (Voinnet et al.,2003), and samples were analyzed 4 d after inoculation (Gou et al., 2011).Before measurement, leaves only infiltrated with luciferin (negative control)and leaves infiltrated with a 35S:LUC construct and luciferin (positivecontrol) were used for calibrating the Orca 2-BT cooled charge-coupleddevice camera (Hamamatsu Photonics). At least three replicates wereperformed on independent plants for each assay.

Accession Numbers

Sequence data from this article can be found in the Arabidopsis GenomeInitiative or GenBank/EMBL databases under the following accession numbers:FT (At1g65480), TFL1 (At5g03840), FD (At4g35900), GRF1 (At4g09000), GRF2(At1g78300), GRF3 (At5g38480), GRF4 (At1g35160), GRF5 (At5g16050), GRF6(At5g10450),GRF7 (At3g02520),GRF8 (At5g65430),GRF9 (At2g42590),GRF10(At1g22300), GRF11 (At1g34760), GRF12 (At1g26480), GRF13 (At1g78220),GRF14 (At1g22290), GRF15 (At2g10450), TCP1 (At1g67260), TCP2

Structural Determinants of FT Activity 561

(At4g18390), TCP3 (At1g53230), TCP4 (At3g15030), TCP5 (At5g60970),TCP6 (At5g41030), TCP7 (At5g23280), TCP8 (At1g58100), TCP9 (At2g45680),TCP10 (At2g31070), TCP11 (At2g37000), TCP12 (At1g68800), TCP13(At3g02150), TCP14 (At3g47620), TCP15 (At1g69690), TCP16 (At3g45150),TCP17 (At5g08070), TCP18 (At3g18550), TCP19 (At5g51910), TCP20(At3g27010), TCP21 (At5g08330), TCP22 (At1g72010), TCP23 (At1g35560),and TCP24 (At1g30210).

Supplemental Data

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

Supplemental Figure 1. Conservation of FT Amino Acid Sequence in88 FT Homologs.

Supplemental Figure 2. Mutation Rate Optimization and BiasEvaluation.

Supplemental Figure 3. Fluorescence Levels in T1 Transformants.

Supplemental Figure 4. Effects of Mutations in 35S:FT-YFP T1 Plantson Flowering in Short Days.

Supplemental Figure 5. Distribution of Residues Conferring TFL1-Like Properties on FT.

Supplemental Figure 6. Point Mutations Conferring TFL-Like ActivityIndependently of Surface Charge.

Supplemental Figure 7. Bimolecular Luciferase ComplementationAssays.

Supplemental Figure 8. Correlation between Flowering Times of Trans-genic Plants Expressing FTMutants underCaMV35S,SUC2, or FDPromoter.

Supplemental Table 1. Comparison of T1 and T2 Flowering Times forSelected Lines.

Supplemental Table 2. Effects of Mutating Ultraconserved Residuesin Short Days.

Supplemental Table 3. Surface Charge–Dependent Effects of FTMutants in Short Days.

Supplemental Table 4. tfl1-1 Complementation Tests.

Supplemental Table 5. Flowering Times of Plants with CompromisedTCP Activity.

Supplemental Table 6. Comparison of Effects of CaMV35S, SUC2,and FD Promoters.

Supplemental Table 7. Oligonucleotide Primers.

ACKNOWLEDGMENTS

We thank Pauline Ip and Rebecca Schwab for support, Robert Leo Bradyfor communicating unpublished findings, Markus Schmid for themCitrine clone, Yuval Eshed for seeds, and Yasushi Kobayashi for FTvariant P75L. We also thank Michael Christie, Yasushi Kobayashi, RoosaLaitinen, Sascha Laubinger, Ignacio Rubio, Markus Schmid, Fritz Schöffl,Rebecca Schwab, and Jiawei Wang for advice and comments on thearticle. W.W.H.H. was supported by a PhD scholarship from the CroucherFoundation. This work was supported by the Max Planck Society.

AUTHOR CONTRIBUTIONS

W.W.H.H. and D.W. designed research. W.W.H.H. performed research.W.W.H.H. and D.W. analyzed data. W.W.H.H. and D.W. wrote the article.

Received June 20, 2013; revised January 9, 2014; accepted January 18,2014; published February 14, 2014.

REFERENCES

Abe, M., Kobayashi, Y., Yamamoto, S., Daimon, Y., Yamaguchi, A.,Ikeda, Y., Ichinoki, H., Notaguchi, M., Goto, K., and Araki, T.(2005). FD, a bZIP protein mediating signals from the floral pathwayintegrator FT at the shoot apex. Science 309: 1052–1056.

Ahn, J.H., Miller, D., Winter, V.J., Banfield, M.J., Lee, J.H., Yoo,S.Y., Henz, S.R., Brady, R.L., and Weigel, D. (2006). A divergentexternal loop confers antagonistic activity on floral regulators FTand TFL1. EMBO J. 25: 605–614.

Baker, N.A., Sept, D., Joseph, S., Holst, M.J., and McCammon, J.A.(2001). Electrostatics of nanosystems: Application to microtubulesand the ribosome. Proc. Natl. Acad. Sci. USA 98: 10037–10041.

Banfield, M.J., Barker, J.J., Perry, A.C., and Brady, R.L. (1998). Functionfrom structure? The crystal structure of human phosphatidylethanolamine-binding protein suggests a role in membrane signal transduction.Structure 6: 1245–1254.

Banfield, M.J., and Brady, R.L. (2000). The structure of Antirrhinumcentroradialis protein (CEN) suggests a role as a kinase regulator. J.Mol. Biol. 297: 1159–1170.

Bradley, D.J., Ratcliffe, O.J., Vincent, C., Carpenter, R., and Coen,E.S. (1997). Inflorescence commitment and architecture in Arabidopsis.Science 275: 80–83.

Clough, S.J., and Bent, A.F. (1998). Floral dip: A simplified method forAgrobacterium-mediated transformation of Arabidopsis thaliana.Plant J. 16: 735–743.

Corbesier, L., Vincent, C., Jang, S., Fornara, F., Fan, Q., Searle, I.,Giakountis, A., Farrona, S., Gissot, L., Turnbull, C., and Coupland, G.(2007). FT protein movement contributes to long-distance signaling infloral induction of Arabidopsis. Science 316: 1030–1033.

Cubas, P., Lauter, N., Doebley, J., and Coen, E. (1999). The TCPdomain: A motif found in proteins regulating plant growth anddevelopment. Plant J. 18: 215–222.

de Felippes, F.F., Ott, F., and Weigel, D. (2011). Comparativeanalysis of non-autonomous effects of tasiRNAs and miRNAs inArabidopsis thaliana. Nucleic Acids Res. 39: 2880–2889.

Efroni, I., Blum, E., Goldshmidt, A., and Eshed, Y. (2008). Aprotracted and dynamic maturation schedule underlies Arabidopsisleaf development. Plant Cell 20: 2293–2306.

Gou, J.Y., Felippes, F.F., Liu, C.J., Weigel, D., and Wang, J.W. (2011).Negative regulation of anthocyanin biosynthesis in Arabidopsis by amiR156-targeted SPL transcription factor. Plant Cell 23: 1512–1522.

Griesbeck, O., Baird, G.S., Campbell, R.E., Zacharias, D.A., andTsien, R.Y. (2001). Reducing the environmental sensitivity of yellowfluorescent protein. Mechanism and applications. J. Biol. Chem.276: 29188–29194.

Hanano, S., and Goto, K. (2011). Arabidopsis TERMINAL FLOWER1 isinvolved in the regulation of flowering time and inflorescence developmentthrough transcriptional repression. Plant Cell 23: 3172–3184.

Hanzawa, Y., Money, T., and Bradley, D. (2005). A single amino acidconverts a repressor to an activator of flowering. Proc. Natl. Acad.Sci. USA 102: 7748–7753.

Hecht, V., Laurie, R.E., Vander Schoor, J.K., Ridge, S., Knowles, C.L.,Liew, L.C., Sussmilch, F.C., Murfet, I.C., Macknight, R.C., and Weller,J.L. (2011). The pea GIGAS gene is a FLOWERING LOCUS T homolognecessary for graft-transmissible specification of flowering but not forresponsiveness to photoperiod. Plant Cell 23: 147–161.

Hellens, R.P., Edwards, E.A., Leyland, N.R., Bean, S., andMullineaux, P.M. (2000). pGreen: A versatile and flexible binary Ti

562 The Plant Cell

vector for Agrobacterium-mediated plant transformation. Plant Mol.Biol. 42: 819–832.

Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K., and Pease, L.R.(1989). Site-directed mutagenesis by overlap extension using thepolymerase chain reaction. Gene 77: 51–59.

Imlau, A., Truernit, E., and Sauer, N. (1999). Cell-to-cell and long-distancetrafficking of the green fluorescent protein in the phloem and symplasticunloading of the protein into sink tissues. Plant Cell 11: 309–322.

Jaeger, K.E., Pullen, N., Lamzin, S., Morris, R.J., and Wigge, P.A.(2013). Interlocking feedback loops govern the dynamic behavior ofthe floral transition in Arabidopsis. Plant Cell 25: 820–833.

Jaeger, K.E., and Wigge, P.A. (2007). FT protein acts as a long-rangesignal in Arabidopsis. Curr. Biol. 17: 1050–1054.

Jang, S., Torti, S., and Coupland, G. (2009). Genetic and spatialinteractions between FT, TSF and SVP during the early stages offloral induction in Arabidopsis. Plant J. 60: 614–625.

Kardailsky, I., Shukla, V.K., Ahn, J.H., Dagenais, N., Christensen, S.K.,Nguyen, J.T., Chory, J., Harrison, M.J., and Weigel, D. (1999).Activation tagging of the floral inducer FT. Science 286: 1962–1965.

Kaufmann, K., Wellmer, F., Muiño, J.M., Ferrier, T., Wuest, S.E.,Kumar, V., Serrano-Mislata, A., Madueño, F., Krajewski, P.,Meyerowitz, E.M., Angenent, G.C., and Riechmann, J.L. (2010).Orchestration of floral initiation by APETALA1. Science 328: 85–89.

Klintenäs, M., Pin, P.A., Benlloch, R., Ingvarsson, P.K., and Nilsson, O.(2012). Analysis of conifer FLOWERING LOCUS T/TERMINAL FLOWER1-like genes provides evidence for dramatic biochemical evolution in theangiosperm FT lineage. New Phytol. 196: 1260–1273.

Kobayashi, Y., Kaya, H., Goto, K., Iwabuchi, M., and Araki, T.(1999). A pair of related genes with antagonistic roles in mediatingflowering signals. Science 286: 1960–1962.

Kobayashi, Y., and Weigel, D. (2007). Move on up, it’s time forchange—Mobile signals controlling photoperiod-dependentflowering. Genes Dev. 21: 2371–2384.

Kojima, S., Takahashi, Y., Kobayashi, Y., Monna, L., Sasaki, T., Araki,T., and Yano, M. (2002). Hd3a, a rice ortholog of the Arabidopsis FTgene, promotes transition to flowering downstream of Hd1 under short-day conditions. Plant Cell Physiol. 43: 1096–1105.

Koornneef, M., Hanhart, C.J., and van der Veen, J.H. (1991). Agenetic and physiological analysis of late flowering mutants inArabidopsis thaliana. Mol. Gen. Genet. 229: 57–66.

Kosugi, S., and Ohashi, Y. (1997). PCF1 and PCF2 specifically bindto cis elements in the rice proliferating cell nuclear antigen gene.Plant Cell 9: 1607–1619.

Koyama, T., Mitsuda, N., Seki, M., Shinozaki, K., and Ohme-Takagi, M.(2010). TCP transcription factors regulate the activities of ASYMMETRICLEAVES1 and miR164, as well as the auxin response, duringdifferentiation of leaves in Arabidopsis. Plant Cell 22: 3574–3588.

Kumar, S.V., Lucyshyn, D., Jaeger, K.E., Alós, E., Alvey, E., Harberd,N.P., and Wigge, P.A. (2012). Transcription factor PIF4 controls thethermosensory activation of flowering. Nature 484: 242–245.

Lifschitz, E., Eviatar, T., Rozman, A., Shalit, A., Goldshmidt, A.,Amsellem, Z., Alvarez, J.P., and Eshed, Y. (2006). The tomatoFT ortholog triggers systemic signals that regulate growth andflowering and substitute for diverse environmental stimuli. Proc.Natl. Acad. Sci. USA 103: 6398–6403.

Lin, M.K., Belanger, H., Lee, Y.J., Varkonyi-Gasic, E., Taoka, K.I.,Miura, E., Xoconostle-Cázares, B., Gendler, K., Jorgensen, R.A.,Phinney, B., Lough, T.J., and Lucas, W.J. (2007). FLOWERINGLOCUS T protein may act as the long-distance florigenic signal inthe cucurbits. Plant Cell 19: 1488–1506.

Liu, L., Liu, C., Hou, X., Xi, W., Shen, L., Tao, Z., Wang, Y., and Yu,H. (2012). FTIP1 is an essential regulator required for florigentransport. PLoS Biol. 10: e1001313.

Martín-Trillo, M., and Cubas, P. (2010). TCP genes: A familysnapshot ten years later. Trends Plant Sci. 15: 31–39.

Mateos, J.L., Bologna, N.G., Chorostecki, U., and Palatnik, J.F.(2010). Identification of microRNA processing determinants byrandom mutagenesis of Arabidopsis MIR172a precursor. Curr. Biol.20: 49–54.

Mathieu, J., Warthmann, N., Küttner, F., and Schmid, M. (2007).Export of FT protein from phloem companion cells is sufficient forfloral induction in Arabidopsis. Curr. Biol. 17: 1055–1060.

Mimida, N., Kidou, S., Iwanami, H., Moriya, S., Abe, K., Voogd,C., Varkonyi-Gasic, E., and Kotoda, N. (2011). AppleFLOWERING LOCUS T proteins interact with transcription factorsimplicated in cell growth and organ development. Tree Physiol.31: 555–566.

Mimida, N., Sakamoto, W., Murata, M., and Motoyoshi, F. (1999).TERMINAL FLOWER 1-like genes in Brassica species. Plant Sci.142: 155–162.

Niwa, M., Daimon, Y., Kurotani, K.I., Higo, A., Pruneda-Paz, J.L.,Breton, G., Mitsuda, N., Kay, S.A., Ohme-Takagi, M., Endo, M.,and Araki, T. (2013). BRANCHED1 interacts with FLOWERINGLOCUS T to repress the floral transition of the axillary meristems inArabidopsis. Plant Cell 25: 1228–1242.

Notaguchi, M., Abe, M., Kimura, T., Daimon, Y., Kobayashi, T.,Yamaguchi, A., Tomita, Y., Dohi, K., Mori, M., and Araki, T.(2008). Long-distance, graft-transmissible action of ArabidopsisFLOWERING LOCUS T protein to promote flowering. Plant CellPhysiol. 49: 1645–1658.

Odell, J.T., Nagy, F., and Chua, N.-H. (1985). Identification of DNAsequences required for activity of the cauliflower mosaic virus 35Spromoter. Nature 313: 810–812.

Ohshima, S., Murata, M., Sakamoto, W., Ogura, Y., and Motoyoshi,F. (1997). Cloning and molecular analysis of the Arabidopsis geneTerminal Flower 1. Mol. Gen. Genet. 254: 186–194.

Palatnik, J.F., Wollmann, H., Schommer, C., Schwab, R.,Boisbouvier, J., Rodriguez, R., Warthmann, N., Allen, E.,Dezulian, T., Huson, D., Carrington, J.C., and Weigel, D. (2007).Sequence and expression differences underlie functional specializationof Arabidopsis microRNAs miR159 and miR319. Dev. Cell 13:115–125.

Pin, P.A., Benlloch, R., Bonnet, D., Wremerth-Weich, E., Kraft, T.,Gielen, J.J., and Nilsson, O. (2010). An antagonistic pair of FThomologs mediates the control of flowering time in sugar beet.Science 330: 1397–1400.

Pnueli, L., Carmel-Goren, L., Hareven, D., Gutfinger, T., Alvarez, J.,Ganal, M., Zamir, D., and Lifschitz, E. (1998). The SELF-PRUNINGgene of tomato regulates vegetative to reproductive switching ofsympodial meristems and is the ortholog of CEN and TFL1.Development 125: 1979–1989.

Pnueli, L., Gutfinger, T., Hareven, D., Ben-Naim, O., Ron, N., Adir,N., and Lifschitz, E. (2001). Tomato SP-interacting proteins definea conserved signaling system that regulates shoot architecture andflowering. Plant Cell 13: 2687–2702.

Ratcliffe, O.J., Amaya, I., Vincent, C.A., Rothstein, S., Carpenter, R.,Coen, E.S., and Bradley, D.J. (1998). A common mechanism controlsthe life cycle and architecture of plants. Development 125: 1609–1615.

Schommer, C., Palatnik, J.F., Aggarwal, P., Chételat, A., Cubas, P.,Farmer, E.E., Nath, U., and Weigel, D. (2008). Control ofjasmonate biosynthesis and senescence by miR319 targets. PLoSBiol. 6: e230.

Serre, L., Vallée, B., Bureaud, N., Schoentgen, F., and Zelwer, C.(1998). Crystal structure of the phosphatidylethanolamine-bindingprotein from bovine brain: A novel structural class of phospholipid-binding proteins. Structure 6: 1255–1265.

Structural Determinants of FT Activity 563

Shannon, S., and Meeks-Wagner, D.R. (1991). A mutation in theArabidopsis TFL1 gene affects inflorescence meristem development.Plant Cell 3: 877–892.

Sohn, E.J., Rojas-Pierce, M., Pan, S., Carter, C., Serrano-Mislata,A., Madueño, F., Rojo, E., Surpin, M., and Raikhel, N.V. (2007).The shoot meristem identity gene TFL1 is involved in flowerdevelopment and trafficking to the protein storage vacuole. Proc.Natl. Acad. Sci. USA 104: 18801–18806.

Takada, S., and Goto, K. (2003). Terminal flower2, an Arabidopsishomolog of heterochromatin protein1, counteracts the activation offlowering locus T by constans in the vascular tissues of leaves toregulate flowering time. Plant Cell 15: 2856–2865.

Tamaki, S., Matsuo, S., Wong, H.L., Yokoi, S., and Shimamoto, K.(2007). Hd3a protein is a mobile flowering signal in rice. Science316: 1033–1036.

Taoka, K., et al. (2011). 14-3-3 proteins act as intracellular receptorsfor rice Hd3a florigen. Nature 476: 332–335.

Tavel, L., Jaquillard, L., Karsisiotis, A.I., Saab, F., Jouvensal, L.,Brans, A., Delmas, A.F., Schoentgen, F., Cadene, M., and

Damblon, C. (2012). Ligand binding study of human PEBP1/RKIP:Interaction with nucleotides and Raf-1 peptides evidenced by NMRand mass spectrometry. PLoS ONE 7: e36187.

Turck, F., Fornara, F., and Coupland, G. (2008). Regulation andidentity of florigen: FLOWERING LOCUS T moves center stage.Annu. Rev. Plant Biol. 59: 573–594.

Voinnet, O., Rivas, S., Mestre, P., and Baulcombe, D. (2003).An enhanced transient expression system in plants based onsuppression of gene silencing by the p19 protein of tomato bushystunt virus. Plant J. 33: 949–956.

Wigge, P.A., Kim, M.C., Jaeger, K.E., Busch, W., Schmid, M.,Lohmann, J.U., and Weigel, D. (2005). Integration of spatial andtemporal information during floral induction in Arabidopsis. Science309: 1056–1059.

Yoo, S.K., Chung, K.S., Kim, J., Lee, J.H., Hong, S.M., Yoo, S.J.,Yoo, S.Y., Lee, J.S., and Ahn, J.H. (2005). CONSTANS activatesSUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 throughFLOWERING LOCUS T to promote flowering in Arabidopsis. PlantPhysiol. 139: 770–778.

564 The Plant Cell

CORRECTION

Ho, W.W.H., Weigel, D. (2014). Structural Features Determining Flower-Promoting Activity of Arabidopsis FLOWERING LOCUS T.

Plant Cell 26: 552–564.

During the assembly of Supplemental Figure 7 and Figure 8, six luciferase images of individual leaves were inadvertently duplicated.

These images served to illustrate the strength of interactions between FT andTFL1 proteins and theirmutant variantswithGRFandTCP

proteins, asmeasured qualitatively in split luciferase assays and in support of independent, quantitative yeast-two-hybrid experiments.

The authors went back to the original photographs, and prepared a corrected version that contains the correct images for the six

duplicated leaves. Figure 8 shows a subset of the leaves shown in Supplemental Figure 7; corrections for both figures are shownbelow.

The figures are referred to twice in the text:

“As expected, FT, TFL1, the TFL-1–like FTmutant Q140K as well as the FTmutant Q140D, which is predicted to have the same surface

charge as native FT, all interacted similarly well with several 14-3-3/GRFproteins andwith FD (Figures 7 and 8; Supplemental Figure 7).”

“The same trends were observed in qualitative assays using bimolecular luminescence complementation assays in N. benthamiana

(Figure 8; Supplemental Figure 7).”

These descriptions of the experiments remain correct.

Versions of the original Supplemental Figure 7 and of the original Figure 8 in which the duplications are indicated are shown below.

Leaves surrounded by heavy lines and numbered in gray indicate images that were correctly assigned. Leaves surrounded by light

lines and numbered in red indicate images that were duplicated and which are replaced in the corrected versions of Supplemental

Figure 7 and Figure 8.

All the original photographs from which the leaves of all experiments were extracted, including a table with a key to the images, are

available at https://datadryad.org/stash/dataset/doi:10.5061/dryad.h9w0vt4fh. The experiments affected by the duplication are

highlighted in red in this table.

The authors apologize for these errors.

Note: This correction was reviewed by members of The Plant Cell editorial board. The authors are responsible for providing a complete

listing and accurate explanations for all known errors or instances of inappropriate data handling or image manipulation associated with

the original publication.

www.plantcell.org/cgi/doi/10.1105/tpc.20.00239

The Plant Cell, Vol. 32: 2043–2047, June 2020, www.plantcell.org ã 2020 ASPB.

Figure 8 (original). Bimolecular luciferase complementation assays.

Young but fully expanded 4-week-old N. benthamiana leaves were strictly chosen for infiltration. Experiments were repeated twice. Complete set of

assay results is shown in Supplemental Figure 7.

Note added in correction: Leaves surrounded by heavy lines and numbered in gray indicate images that were correctly assigned. Leaves surrounded by

light lines and numbered in red indicate images that were duplicated and which are replaced in the corrected version.

2044 The Plant Cell

Figure 8 (corrected). Bimolecular luciferase complementation assays.

Young but fully expanded 4-week-old N. benthamiana leaves were strictly chosen for infiltration. Experiments were repeated twice. Complete set of

assay results is shown in Supplemental Figure 7.

Leaves surrounded by light lines and numbered in red indicate the images that were corrected compared to the original.

June 2020 2045

Supplemental Figure 7 (original). Bimolecular luciferase complementation assays.

Young but fully expanded 4-week-old N. benthamiana leaves were strictly chosen for infiltration. Experiments were repeated twice. A selection of TCP

interactions is shown in Figure 8.

Note added in correction: Leaves surrounded by heavy lines and numbered in gray indicate images that were correctly assigned. Leaves surrounded by

light lines and numbered in red indicate images that were duplicated and which are replaced in the corrected version.

2046 The Plant Cell

Supplemental Figure 7 (corrected). Bimolecular luciferase complementation assays.

Young but fully expanded 4-week-old N. benthamiana leaves were strictly chosen for infiltration. Experiments were repeated twice. A selection of TCP

interactions is shown in Figure 8.

Leaves surrounded by light lines and numbered in red indicate the images that were corrected compared to the original.

June 2020 2047

DOI 10.1105/tpc.113.115220; originally published online February 14, 2014; 2014;26;552-564Plant Cell

William Wing Ho Ho and Detlef WeigelLOCUS T

FLOWERINGArabidopsisStructural Features Determining Flower-Promoting Activity of

 This information is current as of June 21, 2020

 

Supplemental Data /content/suppl/2014/02/05/tpc.113.115220.DC2.html /content/suppl/2014/01/27/tpc.113.115220.DC1.html

References /content/26/2/552.full.html#ref-list-1

This article cites 61 articles, 32 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

Errata

/content/32/6/2043.full.pdfbe accessed at:An erratum has been published regarding this article. It is appended to this PDF and can also

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