TERMINAL FLOWER2 (TFL2) is the plant homologue of metazoan HETEROCHROMATIN PROTEIN1 (HP1) protein family. It is known that, unlike most HP1 proteins, TFL2 does not primarily localize to heterochromatin; instead it functions in regulation of specific genes in euchromatic regions. We show that the tfl2 mutant has a lower rate of auxin biosynthesis, resulting in low levels of auxin. In line with this, tfl2 mutants have lower levels of expression of auxin response genes and retain an auxin response. The reduced rate of auxin biosynthesis in tfl2 is correlated to the down-regulation of specific genes in the tryptophan-dependent auxin biosynthesis pathway, a sub-set of the YUCCA genes. In vivo, TFL2 is targeted to a number of the YUCCA genes in an auxin-dependent fashion revealing a role of TFL2 in auxin regulation, probably as a component of protein complexes affecting transcriptional control.
TERMINAL FLOWER2 (TFL2, also know as LIKE HETEROCHROMATIN PROTEIN1, LHP1) encodes the only plant homologue of metazoan HETEROCHROMATIN PROTEIN1 (HP1) protein family (Gaudin et al., 2001; Kotake et al., 2003). HP1 proteins contain two highly conserved functional domains, a chromo domain and a chromo shadow domain (Hiragami and Festenstein, 2005). In animals, HP1 proteins act in general in heterochromatin formation and gene silencing. Mutations in TFL2 results in a pleiotropic phenotype, including early flowering, reduced photoperiod sensitivity, terminating inflorescences and dwarfing (Larsson et al., 1998). Further, TFL2 functions as a regulator of specific genes in euchromatic regions, e.g. in the maintenance of FLOWERING LOCUS C repression post vernalization, in repression of FLOWERING LOCUS T (FT) and in the regulation of AGAMOUS, PISTILLATA and APETALA1 (Kotake et al., 2003; Mylne et al., 2006; Sung et al., 2006; Liu et al., 2009). The model for TFL2 mediated repression suggests that TFL2 binds to polycomb target genes marked by H3K27 trimethylation to maintain repression (Exner et al., 2009).
In this paper, we show that the tfl2 mutation reduces the rate of auxin biosynthesis and leads to lower levels of free auxin in the plant. Furthermore, TFL2 directly regulates genes in the auxin biosynthetic pathway.
The phytohormone auxin affects several processes throughout the plant’s life cycle, including pattern formation during embryogenesis, lateral root formation, vascular patterning, and tropisms and, at the cellular level, cell division, expansion and differentiation (Lau et al., 2008). In Arabidopsis, the main sites of auxin (indole-3 acetic acid (IAA)) synthesis are found in young tissues with active cell division, such as the shoot apex, developing leaves and root tips (Ljung et al., 2001). When the rate of cell division decreases and cells start to expand, auxin levels drop, demonstrating that auxin production is reduced as tissues mature (Ljung et al., 2001, 2005). In addition, the production of IAA is regulated by feedback inhibition (Ljung et al., 2001). In higher plants, two major routes for auxin production have been identified. In one proposed, but yet uncharacterized route, IAA biosynthesis depends on indole or indole-3-glycerol phosphate that subsequently is converted to IAA without the amino acid tryptophan (Trp) as an intermediate molecule (Woodward and Bartel, 2005; Normanly, 2010). The second route to IAA is dependent on Trp as a precursor molecule. Trp can be converted to IAA following four converging pathways: the indole-3-acetaldoxime, indole-3-pyruvic acid, indole-3-acetamide and tryptamine pathways. None of these pathways has been fully characterized although several genes encoding key proteins have been identified (Ljung et al., 2002; Zazimalova and Napier, 2003; Woodward and Bartel, 2005; Stepanova et al., 2008; Zhao, 2008; Chandler, 2009; Normanly, 2010). Auxin-related developmental changes occur when genes encoding Trp aminotransferases (TAA) or flavin monooxygenases (YUCCA) are mutated (Cheng et al., 2006, 2007; Stepanova et al., 2008); indicating that certain pathways play a more central role in plant development (Sugawara et al., 2009). YUCCA genes have overlapping functions but also very distinct expression patterns throughout plant development, allowing local production of auxin required at distinct developmental stages (Cheng et al., 2006, 2007; Li et al., 2008).
Altered auxin levels rapidly induce changes in gene expression, mediated by three major classes of primary auxin-responsive genes, SMALL AUXIN UP RNA (SAUR), GRETCHEN HAGEN3 (GH3), and AUXIN/INDOLE-3-ACETIC ACID (Aux/IAA). In Arabidopsis over 70 SAUR genes have been identified, encoding short-lived proteins with unknown functions (Gil et al., 1994; Hagen and Guilfoyle, 2002). A sub-group of the GH3 proteins are involved in auxin conjugation, leading to reduced levels of free auxin (Staswick et al., 2005). Several of the genes in the Aux/IAA family are well characterized, and analyses of gain of function mutants reveal that they regulate diverse processes in several tissues, throughout the plant life cycle. Genetic analyses, combining up to three different loss-of-function mutants within this family do not show phenotypic alterations, suggesting a substantial functional redundancy among the genes (Overvoorde et al., 2005). In the current model of auxin signal transduction Aux/IAA proteins repress the function of ARFs through dimerization. ARFs are transcriptional regulators that bind to auxin response elements (AuxRE) in promoter regions of auxin-regulated genes, including Aux/IAA genes. Following elevated auxin concentrations, Aux/IAA proteins ubiquitinated by the SCFTIR1 complex, thereby releasing the inhibited ARFs. The F-box protein TRANSPORT INHIBITOR RESPONSE1 (TIR1), one of the components of the SCF complex, is an auxin receptor (Dharmasiri et al., 2005; Kepinski and Leyser, 2005), directly linking auxin to protein degradation.
Here we report that TFL2 is targeted to YUCCA genes active in Trp-dependent auxin biosynthesis. We also show that this binding is dependent on auxin concentration.
The induction of DR5::GUS is severely reduced in the tfl2 rosette
The tu8 mutant, an allele of tfl2, has been shown to have an altered glucosinolate profile as well as reduced auxin levels (Ludwig-Müller et al., 1999, 2000; Kim et al., 2004; Bennett et al., 2005). These findings prompted us to analyze the sites and levels of auxin activity in the tfl2 mutant, utilizing the synthetic auxin response reporter construct DR5::GUS (Ulmasov et al., 1997). We analysed homozygous DR5::GUS tfl2 and DR5::GUS wt plants 7, 14 and 21 days after germination (DAG). As shown in Figure 1a, GUS activity was detected in the root tip, in lateral root primordia, and in the tip and margins of the leaves. We did not detect any differences in GUS staining, comparing 7-day-old plants of both genotypes. However, after 2 and 3 weeks of growth, the GUS staining in the DR5::GUS tfl2 rosette was strongly reduced as compared with wt, although the GUS activity in the tfl2 roots did not deviate, in distribution or strength, from wt (Figure 1a). Next, we investigated the GUS staining in DR5::GUS tfl2 and wt plants exogenously supplied with the synthetic auxin naphthalene-1-acetic acid (NAA) or the polar auxin transport inhibitor naphthylphthalamic acid (NPA). We did not detect any differences in GUS staining intensity and distribution in the rosette between tfl2 and wt when NAA was supplied to the seedlings (Figure 1b). Following NPA treatment, auxin is trapped at the sites of synthesis (Ljung et al., 2001). tfl2 plants responded to NPA as lateral root development was inhibited in tfl2 as in wt (data not shown). As seen in Figure 1b, blue staining is faint but clearly visible in the rosette of DR5::GUS tfl2 plants, whereas in wt plants the staining is strong.
Our results indicate that the DR5 induction is weaker, and auxin levels are lower in the tfl2 mutant compared with wt, specifically in the rosette. The mutant can still respond to exogenously supplied auxin, and shoots of tfl2 plants do synthesize auxin but at lower levels compared with wt.
The tfl2 mutation reduces auxin levels and biosynthesis rates
To confirm the differences in auxin levels indicated by the DR5::GUS reporter experiment we analyzed the concentration of free indole-3-acetic acid (IAA) in tfl2 and wt. As shown in Figure 2a, the level of IAA in the roots of tfl2 did not statistically differ from that of wt, at any tested time point (7 and 14 DAG). However, in the rosette, the IAA content was significantly reduced to 70% in the tfl2 mutant already at 7 days. The shoot tissue collected after 14 days was sub-divided into rosette leaves 1–2, 3–4, and the remaining part of the rosette, excluding the inflorescence that is present only in tfl2 plants at 14 DAG. In these tissues auxin levels were between 30% and 45% of wt levels. The low amount of free auxin in tfl2 leaves could be due to reduced auxin biosynthesis in the aerial parts of the plant. To determine if that was the case, we performed a deuterated water (2H2O) feeding experiment, in which the rate of incorporation of deuterium into newly synthesized IAA molecules was measured. The plants were divided into root and shoot, incubated with medium containing 30%2H2O for 24 h and dissected as described above. Indeed, a difference was observed in auxin biosynthesis rate when comparing the aerial tissues (Figure 2b); tfl2 exhibiting a statistically lower IAA biosynthesis rate compared with wt. Taken together, TFL2 is required for proper auxin biosynthesis in the Arabidopsis shoot.
The expression of YUCCA genes is down-regulated in tfl2
The auxin biosynthetic pathway(s) are still under investigation. It’s a complex network of at least two major routes that lead to the synthesis of IAA (Figure 3a). The tryptophan-dependent pathways have been extensively studied and several genes involved in these pathways have been identified (Ljung et al., 2002; Woodward and Bartel, 2005; Zhao, 2008; Chandler, 2009; Normanly, 2010). To address how TFL2 affects auxin biosynthesis, we measured the relative expression of genes known to be involved in the tryptophan-dependent auxin biosynthesis pathway by qRT-PCR in 7- and 14-day-old whole seedlings of tfl2 and wt. We observed that the relative expression of YUCCA5, YUCCA8 and YUCCA9 was significantly reduced in tfl2 plants compared with wt (Figure 3b). The expression level of YUCCA5 was significantly reduced already at 7 DAG while the relative expression levels of YUCCA8 and YUCCA9 were significantly reduced only at 14 DAG (Figure 3b). No other auxin biosynthesis gene analyzed was affected (Figure 3a). As tfl2 has been identified as a glucosinolate mutant (Kim et al., 2004; Normanly, 2010), several genes involved in glucosinolate synthesis and breakdown were also included in the screen, e.g. CYP83A1 and CYP83B1 (Naur et al., 2003), IQD1, ST5a, UGT74B1, TGG1 and TGG2 (Grubb and Abel, 2006). The expression of these genes was not affected by the tfl2 mutation.
Chromatin immunoprecipitation (ChIP) experiments were performed on 14-day-old gTFL2::GFP whole seedlings using anti GFP antibodies to analyze if TFL2 localizes to the YUCCA genes in planta. It has been shown for several genes that TFL2 localizes throughout the promoter as well as to the transcribed regions of the genes it regulates (Turck et al., 2007; Zhang et al., 2007). We analyzed the transcribed region of all YUCCA genes and the promoter of YUCCA5. The qPCR analysis revealed that TFL2 binds to YUCCA5 and to a lesser extent to YUCCA8 and YUCCA9 (Figure 4a); the three YUCCA genes severely down-regulated in the tfl2 mutant. In addition, at this developmental stage, TFL2 was also found to localize to YUCCA1, YUCCA2, YUCCA4, YUCCA6 and YUCCA10, whose expression was not affected by the tfl2 mutation (Figure 4a).
To monitor the effect of exogenously applied auxin on the localization of TFL2 over a subset of YUCCA genes, we performed ChIP-qPCR on seedlings incubated 1h in the presence or absence of 10 μm IAA. In the presence of IAA, the binding of TFL2 to the YUCCA genes targeted in Figure 4a, was reduced by approximately 50% for all assayed YUCCA genes (Figure 4b). The non-targeted YUCCA11 (Figure 4a,b) included as a control was unaffected by the auxin treatment in its binding of TFL2. Furthermore, to assess the wt response of YUCCA5, YUCCA8 and YUCCA9 to the above mentioned auxin treatment we measured their relative expression at 7 and 14 DAG (Figure 4c). Auxin treatment reduced the expression levels of all three genes both at 7 and 14 DAG by at least 1.5- to two-fold.
In summary, at the analyzed stage tfl2 causes down-regulation of YUCCA5, 8 and 9 genes believed to be involved in auxin biosynthesis. Further, in wt plants, TFL2 is localized to a majority of the YUCCA genes, including YUCCA5, 8 and 9, in an auxin-dependent fashion.
Specific auxin response genes are misexpressed in tfl2 plants
To analyze the auxin response of tfl2, we assayed the expression of representatives of all three major classes of primary auxin response genes, as well as of the ARF family, in tfl2 in comparison with wt. Aerial tissues from 7- and 14-day-old tfl2 and wt plants were analyzed by qRT-PCR. The GH3 gene DWARF IN LIGHT1 (DFL1) in addition to ARF2, ARF7 and ARF19 did not show a difference in transcript levels between the genotypes (Figure S1a–d). In contrast, IAA5, IAA6 and IAA19 as well as SAUR-AC1 displayed reduced expression levels, by at least a factor 2, in tfl2 compared with wt, at both developmental stages (Figure S1e–h). No significant differences in expression levels between genotypes were detected analysing IAA12/BODENLOS (BDL) that is distantly related to the IAA5, IAA6, IAA19 genes (Figure S1i). Further, we investigated the ability of tfl2 to respond to exogenous auxin by examining the expression of IAA5, IAA6, IAA19 and SAUR-AC1 following application of IAA. Plants were grown for 7 or 14 days, incubated on media containing IAA of different concentrations for 1 h and thereafter expression levels were analyzed. Transcript levels were increased in both genotypes in a concentration-dependent manner (Figure 5a–d). The expression of IAA19 after IAA treatment was unaffected by the tfl2 mutation, IAA5 showed significantly reduced expression compared with wt only in the 7 DAG sample supplied with 10 μm IAA, while IAA6 and SAUR-AC1 transcript levels were significantly reduced in tfl2 compared with wt at both 7 and 14 DAG.
We also examined tfl2 for auxin responses at the morphological level. We analyzed auxin treated tfl2 and wt seedlings for root length inhibition and hypocotyl responses to IAA. Root response experiments were carried out on 3-day-old tfl2 seedlings that were transferred to media with IAA at different concentrations (Figure 6a). After an additional 6 days of growth we noted that the relative root length of tfl2 seedlings was less inhibited by auxin, compared with wt plants, on all the concentrations examined. The largest difference between the genotypes was found when the seedlings were grown on 1 μm IAA, here tfl2 displayed more than double the root length of wt seedlings. Exogenous auxin at low concentrations promotes cell expansion in the hypocotyl while at high concentrations auxin inhibits cell expansion. The hypocotyl elongation of plants grown on auxin at different concentrations was examined in darkness and under continuous light. At low levels of IAA (0.01 and 0.1 μm) in light, the tfl2 seedlings did not display auxin-induced hypocotyl elongation. However, at high levels of IAA (100 μm) the tfl2 seedlings displayed a significantly reduced hypocotyl length both in light and darkness, although this response was significantly weaker than in wt. (Figure 6b,c).
The deviations in auxin response between wt and tfl2 could be due to auxin regulation of TFL2 or to misexpression of the auxin receptor TIR. Comparing 3-, 7- and 14-day-old plants, exogenously supplied with 1 μm IAA for 1 h with untreated plants, the TFL2 expression level was unaltered, demonstrating that TFL2 expression is not regulated by auxin (Figure S1j). The expression of TIR1 did not differ between wt and tfl2 in any of the developmental stages analyzed, showing that its expression is not dependent on TFL2 (Figure S1k).
Taken together, our results show that tfl2 does respond to exogenous IAA both at a morphological and molecular level, although the response is reduced in tfl2 compared with wt. Also a specific subset of auxin response genes is affected by the tfl2 mutation.
The data presented in this paper suggest a direct role for TFL2 in auxin-regulated plant development, through its localization to a sub-set of YUCCA genes, positively regulating their expression.
We have shown that levels of free IAA, as well as the rate of auxin biosynthesis, are reduced in tfl2 as compared with wt plants and that the reduction is found in aerial tissues. In line with our results, altered levels of auxin have been reported for the tu8 mutant, the tfl2-6 allele previously described as a mutant with a developmentally altered leaf glucosinolate profile (Ludwig-Müller et al., 1999, 2000). The data we present show that TFL2 is involved in the regulation of auxin biosynthesis. First, the expression of YUCCA5, YUCCA8 and YUCCA9 is reduced in tfl2; the YUCCA gene family of flavin monooxygenases is reported to be a rate-limiting step in one of the tryptophan-dependent auxin biosynthetic pathways (Zhao et al., 2001). Second, TFL2 was found to localize to a subset of YUCCA genes. Third, we see a decrease of TFL2 enrichment over the transcribed region of several YUCCA genes upon auxin induction. For example YUCCA5 is down-regulated in tfl2 plants and in wt TFL2 is targeted to the promoter and transcribed regions of the gene in an auxin-dependent fashion. In line with this, auxin treatment of wt seedlings represses the expression of the YUCCA genes down-regulated in tfl2 suggesting that TFL2 has a positive role in transcriptional activation of the YUCCA genes assayed. The promoters of YUCCA genes have not been reported to contain any AuxRE and the transcription factors reported to regulate YUCCA genes (Li et al., 2008; Stone et al., 2008; Trigueros et al., 2009; Eklund et al., 2010) are not ARF transcription factors. Thus, we suggest that TFL2 is involved in positive regulation of the YUCCA genes by a mechanism-independent from the present auxin response gene regulation model involving Aux/IAA inhibition of ARFs (Dharmasiri and Estelle, 2004; Quint and Gray, 2006).
There is not a complete correlation between our datasets as TFL2 binds to YUCCA genes that are unaffected in their expression by the tfl2 mutation. We suggest that this finding reflects that different YUCCA genes are active in the regulation of auxin biosynthesis in different developmental stages and tissues. Alternatively, only the expression of the detected subset of YUCCA genes is dependent on TFL2 binding. We favour the first alternative as it has been shown that the expression of the YUCCA gene family is differentially regulated both spatially and temporally. The importance of local auxin produced by the YUCCA genes in different developmental stages has been described for several members of the gene family (Cheng et al., 2006, 2007; Li et al., 2008) and higher order YUCCA mutants confer severe developmental defects rescued only by local expression of the bacterial iaaM biosynthesis gene driven by YUCCA promoters (Cheng et al., 2006). YUCCA5, YUCCA8 and YUCCA9 are phylogenetically closely related (Cheng et al., 2006). Regarding their expression patterns YUCCA5 and YUCCA8 are expressed in roots and young rosette leaves and YUCCA5 expression is limited to non-reproductive tissues (Woodward et al., 2005; Rawat et al., 2009). Although the expression pattern of YUCCA9 has not been experimentally described it is plausible to believe, from the studies of other gene family members, that the expression domains are distinct throughout the plants life cycle. Furthermore the absolute levels of expression of the members of the YUCCA gene family are very low compared with all the other genes involved in the auxin biosynthesis pathway analyzed in this study (Arabidopsis eFP browser at bar.utoronto.ca; Winter et al., 2007) suggesting that small changes in YUCCA gene expression can have a pronounced effect on auxin biosynthesis.
Wild-type plants respond to auxin treatment both at the morphological and molecular level and our analyzes show that tfl2 is able to respond to auxin. At the morphological level we see reduced sensitivity to both high and low levels of auxin in tfl2 plants, while the relative expression data shows that some, but not all assayed early auxin responsive genes are misexpressed in tfl2. We find reduced levels of expression of the closely related IAA5, IAA6, IAA19 and of SAUR-AC1 in untreated tfl2 shoots, data were consistent with the low levels of auxin in tfl2. In line with this proposal, Vadassery et al. (2008) show that reduced levels of free auxin result in low expression of IAA6. In auxin-treated seedlings, on the other hand, we saw a clear effect of the tfl2 mutation on the transcript levels of IAA6 and SAUR-AC1, while IAA19 and IAA5 responded to auxin treatment in a wt fashion. Either TFL2 has additional functions during auxin regulation of plant development, or the reduced auxin response in tfl2 could be attributed to the severe down-regulation of YUCCA genes as exogenous auxin application is not able to rescue YUCCA mutant phenotypes (Cheng et al., 2006). This hypothesis is also supported by the fact that loss of TFL2 function does not have a general effect on all auxin response genes analyzed. Furthermore TFL2 is an epigenetic regulator shown to act cell type autonomously (Luo et al., 2009), enabling the required regulation of the YUCCA genes throughout plant development.
To summarize, TFL2 has a role in positive transcriptional regulation and feedback regulation of auxin biosynthesis. Such a role for TFL2 in positive transcriptional regulation has not been suggested previously in plants, but studies of two homologues of TFL2, the mammalian HP1γ and Drosophila melanogaster HP1a, show that both are involved in positive euchromatic gene regulation (Piacentini et al., 2003; Cryderman et al., 2005; Vakoc et al., 2005). The mechanisms for gene activation involving HP1 proteins are still unknown, but interaction with transcription factors as well as association with transcriptional elongation are indicated (Vermaak and Malik, 2009).
Our results widens the view showing that plant HETEROCHROMATIN PROTEIN1, although it is a single copy gene in Arabidopsis, is able to perform several of the functions carried out by the HETEROCHROMATIN PROTEIN1 family in animals.
Plant material and growth conditions
Plants were grown on soil as described by Larsson et al. (1998). Plant material grown in vitro was surface sterilized (15% commercial bleach, 0.1% SDS and 70% EtOH wash), sown on solid growth media (GM, 0.5 × MS salts, 1%sucrose), incubated in darkness at +4°C for 2 days and transferred to a constant light at 20–22°C. In all the experiments Columbia (Col-0) was used as wild type (wt) control. The tfl2-1 mutant has previously been described by Larsson et al. (1998) and DR5::GUS has been described by Ulmasov et al. (1997).
Auxin induction, hypocotyl elongation and root assays
Plants were grown in vitro on GM as described above for 7 or 14 days, and whole plants including the root were transferred to 0.5 × MS liquid media containing different concentrations of IAA (Sigma, http://www.sigmaaldrich.com), and incubated for 1 h with shaking.
For hypocotyl elongation assays, surface sterilized seeds plated on GM supplied with different IAA concentrations, were cold treated for 3 days, light induced for 3 h and kept in darkness for 21 h before exposure to continuous light (61 μE) or darkness for 6 days.
For root length assays seedlings were grown for 3 days on media without auxin, and transferred to dishes with or without IAA of different concentrations. At day 4 counting from when the seeds were transferred to light, the position of the root tip was marked, and after an additional 6 days the root growth was measured. For calculations, the 20 plants with the longest roots from each genotype and treatment were measured.
Gene expression analyses
Plant material was harvested at the time points indicated in the result section. Total RNA was extracted by RNeasy® Plant Mini Kit (Qiagen) and DNase treated with RNase-free DNase set (Qiagen, http://www.qiagen.com) according to the supplied protocols. cDNA synthesis and quantitative reverse transcript PCR (qRT-PCR) were conducted as described by Kuusk et al. (2006), cDNA was synthesized using the iscript cDNA synthesis kit (Bio-Rad, http://www.bio-rad.com) and iQ SYBR Green Super Mix (Bio-Rad) was used in the qRT-PCR reactions, that were conducted on MyiQ real-time PCR detection system (Bio-Rad). Significant relative expression changes were determined by a cut off value of 2.5. The primers used are reported in Table S1.
Auxin quantification and biosynthesis rate measurements
For IAA quantification, tissues from 7- and 14-day-old Arabidopsis wt and tfl2-2 seedlings were dissected into root and shoot as described in the result section and snap-frozen in liquid nitrogen. For IAA biosynthesis measurements, the seedlings from 7- and 14-day-old Arabidopsis wt and tfl2-2 were first divided into root and shoot and then incubated in liquid GM containing 30%2H2O for 24 h. The shoot was then dissected as described in the result section and all tissues were snap-frozen in liquid nitrogen. 500 pg 13C6-IAA internal standard (Cambridge Isotope Laboratories, Andover, MA, USA) was added to each sample and extraction and purification of the samples was performed as described in Andersen et al. (2008). IAA quantification and biosynthesis measurements were then performed by combined gas chromatography-selected reaction monitoring-mass spectrometry (GC-SRM-MS) (Edlund et al., 1995; Ljung et al., 2005). The IAA biosynthesis rates were calculated as the ratio between labeled (m/z (203 + 204 + 205)) and unlabelled tracer (m/z 202) as described in Ljung et al. (2005). All samples were analyzed in triplicate.
Histochemical localization of GUS activity
Histochemical GUS staining was performed using X-gluc as chromogenic substrate (Jefferson et al., 1987). Entire plants were incubated in GUS staining solution (50 mm sodium phosphate pH 7.0; 0.1% Triton X-100; K3/K4 FeCN 0.5 and 1 mm X-gluc) for 5 or 24 h, and subsequently destained and stored in 70% ethanol. Plants were analyzed and photographed using a stereomicroscope.
Chromatin immunoprecipitation and quantitative PCR
ChIP was carried out on 2 g of gTFL2::GFP (Kotake et al., 2003) whole seedlings grown vertically on GM. For auxin-induction experiments seeds were grown on GM with sterile nylon mesh and transferred to plates with or without 10 μm IAA and incubated for 1 h. ChIP was carried out as described by Gendrel et al. (2005) TFL2 was precipitated using anti GFP (Upstate) and anti-IgG (Upstate) was used as a mock antibody. The purified DNA was analyzed using qPCR, experiments were carried out on three independent ChIP replicates. 1 μl of ChIP and mock DNA per reaction were used and a serial input DNA dilution (1:5, 1:25 and 1:125) was run in parallel. qPCR conditions were 95°C for 3 min and 40 cycles of 95°C, 10 sec, 60°C, 15 sec and 72°C, 10 sec, followed by a melting curve analysis. Data were normalized against negative controls (Haring et al., 2007) WRKY DNA-binding protein 33, RNA-DEPENDENT RNA POLYMERASE2 and OVULE ABORTION8, described as negative controls by Turck et al. (2007), FLOWERING LOCUS T was used as a positive control. The fold enrichment of positive regions was calculated by dividing the absolute signals by the mean signal of the negative controls. The cut off value for determining a positive binding of TFL2 to the target genes was three-fold enrichment together with a ratio GFP enrichment/IgG enrichment >2.
We would like to thank Professor Koji Goto for kindly providing seeds of gTFL2::GFP, Carl-Johan Simola for preliminary screens of DR5::GUStfl2, Gun-Britt Berglund and Roger Granbom for technical assistance and Professor Peter Engström and Christina Roberts for helpful comments on the manuscript. Support from Nils and Dorthi Troedssons Foundation and Faculty of Science and Technology, Uppsala University is acknowledged.