Elongator subunit 3 positively regulates plant immunity through its histone acetyltransferase and radical S-adenosylmethionine domains

Background Pathogen infection triggers a large-scale transcriptional reprogramming in plants, and the speed of this reprogramming affects the outcome of the infection. Our understanding of this process has significantly benefited from mutants that display either delayed or accelerated defense gene induction. In our previous work we demonstrated that the Arabidopsis Elongator complex subunit 2 (AtELP2) plays an important role in both basal immunity and effector-triggered immunity (ETI), and more recently showed that AtELP2 is involved in dynamic changes in histone acetylation and DNA methylation at several defense genes. However, the function of other Elongator subunits in plant immunity has not been characterized. Results In the same genetic screen used to identify Atelp2, we found another Elongator mutant, Atelp3-10, which mimics Atelp2 in that it exhibits a delay in defense gene induction following salicylic acid treatment or pathogen infection. Similarly to AtELP2, AtELP3 is required for basal immunity and ETI, but not for systemic acquired resistance (SAR). Furthermore, we demonstrate that both the histone acetyltransferase and radical S-adenosylmethionine domains of AtELP3 are essential for its function in plant immunity. Conclusion Our results indicate that the entire Elongator complex is involved in basal immunity and ETI, but not in SAR, and support that Elongator may play a role in facilitating the transcriptional induction of defense genes through alterations to their chromatin.


Background
Plants possess an innate immune system that protects them from microbial pathogens, but lack the adaptive immunity of mammals [1,2]. Each plant cell is capable of sensing non-self entities and mounting immune responses, demonstrating remarkable functional plasticity. Recognition of pathogen-associated molecular patterns (PAMPs) results in PAMP-triggered immunity (PTI) that prevents pathogen colonization. In turn, pathogens have evolved effectors to dampen PAMP-triggered signals and thereby attenuate PTI. The plants can only activate a weak response known as basal immunity. Some host plants have evolved resistance (R) proteins to detect the presence of pathogen effectors, inducing effector-triggered immunity (ETI) [3]. Activation of PTI or ETI leads to the generation of a blend of signal molecules, which move to distal tissues for the establishment of systemic acquired resistance (SAR). SAR is a long-lasting immunity against a broad spectrum of pathogens [4].
Salicylic acid (SA) is a key signal molecule for plant immunity against biotrophic and hemibiotrophic pathogens. It is not only required for the activation of SAR [5,6], but also plays an important role in plant basal immunity and ETI [7][8][9][10][11]. In plants, SA can be made through two metabolic pathways, the phenylalanine ammonia lyase (PAL)mediated phenylalanine pathway and the isochorismate synthase (ICS)-mediated isochorismate pathway. Although knockout of PAL genes significantly reduces SA production [12], the isochorismate pathway is thought to be more important during plant defense [13]. Increasing cellular SA levels induces profound transcriptional changes that are largely governed by the transcription coactivator NPR1 (nonexpressor of pathogenesis-related (PR) genes). Similarly to SA, NPR1 is not only required for SAR activation, but also plays a significant role in plant basal immunity and ETI [7,[14][15][16][17]. Interestingly, NPR1 is also a feedback inhibitor of SA biosynthesis. After pathogen infection, npr1 plants accumulate significantly higher levels of SA [13,17]. Hyperaccumulation of SA causes chlorosis in juvenile leaves and inflorescences of npr1 plants [18]. When grown on media containing high concentrations of SA, npr1 seedlings fail to develop beyond the cotyledon stage, while wild type displays tolerance to SA cytotoxicity [19,20].
In eukaryotic cells, RNA Polymerase II (RNAPII) catalyzes the transcription of protein-encoding genes. The Elongator complex was first identified as an interactor of hyperphosphorylated RNAPII in yeast [21,22], and subsequently purified from mammalian and plant cells [23,24]. Elongator consists of six subunits (ELP1-6) with ELP1-3 forming the core subcomplex and ELP4-6 the accessory subcomplex [25,26]. Among the six subunits, ELP3 is the catalytic subunit, harboring a C-terminal histone acetyltranferase (HAT) domain and an Nterminal cysteine-rich motif that resembles an ironsulfur (Fe-S) radical S-adenosylmethionine (SAM) domain [27,28]. ELP3 alone has intrinsic HAT activity and is capable of acetylating all four histones, whereas the six-subunit holo-Elongator predominantly acetylates lysine-14 of histone H3 and lysine-8 of histone H4 [22,27]. Consistently, the levels of acetylated histones H3 and H4 are reduced in yeast, mammalian, and plant elp mutants [24,27,29]. The radical SAM domain of yeast ELP3 is a structural motif required for the integrity of the complex [30], whereas the archaea Methanocaldococcus jannaschii ELP3 binds and cleaves SAM, a co-substrate involved in methyl group transfers, suggesting that M. jannaschii ELP3 may have another catalytic function other than HAT activity [31]. Indeed, a recent study indicated that the radical SAM domain of mouse ELP3, but not the HAT domain, is required for zygotic paternal genome demethylation [32].
In order to identify new components in SA signaling, we performed a genetic screen for suppressors of the npr1 mutation based on restoration of SA tolerance on half-strength MS medium supplemented with 0.5 mM SA. A total of 20 gns (green npr1 seedling on SA medium) mutants showing restored SA tolerance have been isolated. We have previously described the gns1 mutant, which harbors a mutation in AtELP2 [42]. Here we report the isolation and characterization of the gns2 mutant, in which a frameshift mutation was identified in the Arabidopsis Elongator complex subunit 3 (AtELP3).
Our results indicate that, like AtELP2, AtELP3 is required for plant basal immunity and ETI but not for SAR, and demonstrate that the HAT and radical SAM domains of AtELP3 are essential for its function in plant immunity.

Results
The gns2 mutation suppresses hyperaccumulation of SA in npr1 Similarly to the previously characterized gns1 npr1 [42], the gns2 npr1 mutant not only exhibited partially restored SA tolerance ( Figure 1A), but also accumulated significantly less SA than npr1 after infection by the virulent bacterial pathogen Pseudomonas syringae pv. maculicola (Psm) ES4326 ( Figure 1B and C), suggesting that gns2 suppresses SA hyperaccumulation in npr1. To test whether gns2 affects pathogen susceptibility, the growth of Psm ES4326 was determined in gns2 npr1 plants. As shown in Figure 1D, while Psm ES4326 grew 32-fold more in npr1 than in the wild type, its growth was further increased by~10-fold in gns2 npr1 plants, indicating that the gns2 mutation compromises NPR1independent disease resistance.
The leaves of gns2 npr1 plants were a lighter shade of green than npr1 ( Figure 1E). The gns2 npr1 mutant was backcrossed to npr1 three times before further characterization. The F 1 progeny were not SA tolerant and displayed npr1 morphology. Of 72 F 2 plants, 22 were gns2-like. The 1:3 gns2-like to npr1-like ratio (χ 2 = 1.1852; P > 0.1), together with the F 1 phenotypes fit the expectation that gns2 is a single, recessive mutation. To determine the co-segregation of gns2 morphology and SA tolerance, progeny from F 2 plants with npr1 or gns2 morphology were examined. Progeny from gns2-like parents were nearly all SA tolerant, while none or only a fraction of the progeny from npr1-like plants were SA tolerant, suggesting that the morphology and SA tolerance of gns2 co-segregate and are caused by the same mutation or two closely linked mutations.
The GNS2 locus encodes the Elongator subunit 3 To map the GNS2 locus, homozygous gns2 npr1, which is in Columbia (Col-0) background, was crossed with the polymorphic ecotype Landsberg erecta (Ler). Linkage analysis of 100 F 2 plants with gns2 morphology placed the GNS2 locus in the middle of the lower arm of chromosome 5, between the molecular markers CIW9 and CIW10 (Figure 2A). Recombination analysis of 1352 F 2 light green plants located the GNS2 locus in between markers at the genetic loci At5g50180 and At5g50360 (Figure 2A). One gene within this interval was AtELP3/ELONGATA3(ELO3) (At5g50320) [39], which encodes the third subunit of the Elongator complex. Since gns2 resembled the previously characterized gns1/Atelp2-5 mutants in SA responses [42], GNS2 might be AtELP3. The AtELP3 coding region was therefore amplified from gns2 npr1 and sequenced. A deletion of a guanine (39 bp from ATG) was detected in the first exon of AtELP3, which allowed the development of a dCAPS marker to genotypically distinguish the mutant from the wild type ( Figure 2C and B and Additional file 1: Table S1). This mutation caused a frameshift, likely resulting in a truncated and non-functional protein.
To confirm that GNS2 is AtELP3, the AtELP3 cDNA under the control of the constitutive 35S promoter of cauliflower mosaic virus was introduced into gns2 npr1. The 35S::AtELP3 transgene complemented all the phenotypes caused by the gns2 mutation in gns2 npr1, including partially restored SA tolerance, light green coloration, reduced SA accumulation after Psm ES4326 infection, and enhanced susceptibility to Psm ES4326 ( Figure 2D to H). These results indicate that the mutation in AtELP3 is responsible for the morphological and defense phenotypes of the gns2 npr1 plants. The gns2 mutant was therefore renamed Atelp3-10.

AtELP3 positively contributes to salicylic acid responsiveness
Compared with npr1, Atelp3-10 npr1 exhibited delayed and reduced expression of PR2 and PR5 during Psm ES4326 infection ( Figure 3A). To test the function of AtELP3 in the presence of NPR1, Atelp3-10 npr1 was   Table S1. (B) A dCAPS marker distinguishing gns2 and WT. A deletion of a guanine (39 bp from ATG) was detected in the first exon of AtELP3, which allowed the development of a dCAPS marker to genotypically distinguish the mutant from the wild type (Additional file 1: Table S1). (C) Structure of the GNS2/ELO3/AtELP3 gene and the positions of the gns2 and elo3-1 mutations [39]. Boxes denote the translated regions and lines between boxes denote introns. (D) Ten-day-old seedlings of WT, npr1, Atelp3-10 npr1, and two complementation lines Com(n) # 1 and Com(n) #  crossed with wild type, and Atelp3-10 single mutants were identified in the segregating F 2 population based on their morphology (same as the Atelp3-10 npr1 double mutant) and confirmed by the Atelp3-10 and npr1 dCAPS markers (Additional file 1: Table S1). The Atelp3-10 single mutant displayed decreased and/or delayed PR gene expression and enhanced susceptibility to Psm ES4326, and these phenotypes were all completely complemented by a 35S:: AtELP3 transgene ( Figure 3A and B), confirming that Atelp3-10 is a loss-of-function mutation.
When treated with SA, Atelp3-10 exhibited delayed PR gene expression ( Figure 4A), supporting the conclusion that Elongator plays a role downstream of SA in plant immune responses [42]. On the other hand, SA treatment substantially protected both Atelp3-10 and wild type against Psm ES4326 ( Figure 4B). We further analyzed the interaction between SA treatment and genotype with a linear mixed-effects model [44], and found that the Atelp3-10 mutation did not significantly affect SA-induced resistance against Psm ES4326 (p = 0.8285).

AtELP3 positively regulates effector-triggered immunity
To test whether AtELP3, like AtELP2, plays a role in ETI, we examined the expression profiles of nine defense genes that are rapidly induced during ETI [8,45]. As shown in Figure 5A, after infection with the ETIinducing bacterial pathogen P. s. pv. tomato (Pst) DC3000/avrRpt2, induction of all nine genes was delayed in Atelp3-10 compared with that in the wild type, though expression of some genes eventually reached wild-type levels. Furthermore, Pst DC3000/avrRpt2 growth was~10-fold more in Atelp3-10 than in the wild type but lower than in the fully susceptible rps2 mutant ( Figure 5B) [46]. In the double mutant Atelp3-10 npr1, however, Pst DC3000/avrRpt2 growth was~35-fold higher than in either Atelp3-10 or npr1, and about sevenfold higher than in rps2 ( Figure 5B). These results are consistent with the conclusion that both Elongator and NPR1 positively regulate ETI [42]. Using a linear mixed-effects model to statistically analyze the data in Figure 5B, we found that the Atelp3-10 and npr1 mutations did not have statistically significant genetic interaction in resistance against Pst DC3000/avrRpt2 (p = 0.064); therefore, AtELP3 and NPR1 appear to function largely independently of each other in ETImediated resistance to Pst DC3000/avrRpt2. However, since the p value is very close to the significance threshold (0.05), we cannot rule out the possibility of genetic interaction between AtELP3 and NPR1. Indeed, our recent work indicated that AtELP2 regulates both NPR1dependent and -independent defense gene induction [47].

AtELP3 is not required for biological induction of systemic acquired resistance
We next analyzed the induction of SAR in Atelp3-10 plants. Three lower leaves were injected with 10 mM MgCl 2 (mock treatment) or Psm ES4326 (SAR treatment). After two days, SAR treatment-induced gene expression in the upper, untreated leaves was examined. Six genes that are induced during SAR [48] were strongly NPR1-dependent, but only required AtELP3 for their full expression ( Figure 6A). Expression levels of these genes were significantly higher in Atelp3-10 plants than in npr1 plants ( Figure 6A). When the upper, untreated systemic leaves were challenge-inoculated with Psm ES4326, Atelp3-10 exhibited similar levels of resistance as the wild type ( Figure 6B). Using a linear mixed-effects model to statistically analyze the data in Figure 6B, we found that the Atelp3-10 mutation did not have statistically significant effect on SAR treatmentinduced resistance to Psm ES4326 (p = 0.279). Therefore, SAR induction restores pathogen resistance to Atelp3-10 plants, which supports the conclusion that Elongator is not required for SAR activation [42].
Both HAT and radical SAM domains of AtELP3 are required for its function in plant immunity ELP3 is the catalytic subunit of the Elongator complex, containing a HAT domain and a radical SAM domain. Both domains are required for Elongator's function in yeast, whereas only the radical SAM domain is essential for mouse paternal genome demethylation in zygotes [32]. To test whether the HAT and the radical SAM domain of AtELP3 are required for Elongator's function in plant immunity, we introduced point mutations into the AtELP3 cDNA to generate a HAT domain mutant, in which two conserved tyrosine residues (Y547 and Y548) were changed to alanine, and a radical SAM domain mutant, in which two conserved cysteine residues (C127 and C130) were changed to serine ( Figure 7A) [32]. Mutated Atelp3 cDNAs were put under the control of the constitutive 35S promoter and transformed into Atelp3-10 and Atelp3-10 npr1 to test the functionality of the HAT and the SAM mutant. Multiple transgenic lines were generated and characterized for both constructs, and one representative transgenic line for each construct was shown in Figure 7. Neither the HAT nor the SAM mutant complemented any of the Atelp3-10 phenotypes, including partially restored SA tolerance ( Figure 7B), reduced SA accumulation after Psm ES4326 infection ( Figure 7C and D), delayed/decreased defense gene expression ( Figure 7E), and enhanced susceptibility to Psm ES4326 ( Figure 7F), indicating that both the HAT and radical SAM domains of AtELP3 are essential for its function in plant immune responses.

Discussion
The multiprotein complex Elongator functions in diverse biological processes in yeast, animals, and plants likely by accelerating the induction of genes required for development or adaptation to changing environmental conditions. In Arabidopsis, Elongator regulates transcriptional changes induced during pathogen infection [42,47]. These transcriptional changes are accompanied by changes in DNA methylation, a hallmark of transcriptional suppression [49], and histone acetylation, which is generally associated with transcriptional activation [47,50]. Since the Elongator complex subunit AtELP3 contains both HAT and radical SAM domains, it seems possible that the enzymatic activities of Elongator are directly involved in altering the chromatin status of defense genes to facilitate their induction. Fortuitously, we were able to isolate an Atelp3 mutant in our genetic screen for suppressors of the npr1 mutation based on restoration of SA tolerance (Figures 1 and 2) [42]. The Atelp3-10 mutant displays similar deficiencies in basal immunity and ETI as Atelp2 (Figures 1D, 2H, and 5B), and like Atelp2 has normal SAR ( Figure 6B), suggesting that the Elongator complex itself, but not AtELP2 individually, is the functional unit of action for its role in Figure 5 Characterization of ETI in Atelp3-10 Plants. (A) Expression of nine ETI-inducible genes in Pst DC3000/avrRpt2-infected WT and Atelp3-10 plants. Plants were inoculated with Pst DC3000/avrRpt2 (OD 600 = 0.001). Leaf tissues were collected at the indicated time points. Total RNA was extracted and analyzed for the expression of indicated genes using real-time qPCR. The y-axes indicate relative expression levels monitored by qPCR. Expression was normalized against constitutively expressed UBQ5. The x-axes indicate hours after Pst DC3000/avrRpt2 infection. Data represent the mean of three independent samples with SD. An asterisk (*) indicates that the expression level of the gene in the WT was significantly different from that in Atelp3-10 (p < 0.05, Student's t-test). (B) Growth of Pst DC3000/avrRpt2 in WT, npr1, Atelp3-10, Atelp3-10 npr1, and rps2 plants. Plants were inoculated with Pst DC3000/avrRpt2 (OD 600 = 0.0001). The in planta bacterial titers were determined immediately and three days postinoculation. Cfu, colony-forming units. Data represent the mean of eight independent samples with SD. Different letters above the bars indicate significant differences (p < 0.05, Student's t-test). plant immunity. Indeed, T-DNA insertion mutants of other subunits of the Arabidopsis Elongator complex exhibit Atelp2-and Atelp3-like morphology [39,41]. The function of these subunits in plant immunity will be a subject of further inquiry.
Transgenic complementation of Atelp3-10 with the AtELP3 gene lacking conserved residues in the HAT domain did not restore wild-type levels of SA tolerance, SA accumulation, defense gene induction, and pathogen resistance ( Figure 7). Although Elongator's HAT activity has not been demonstrated in plants, it has been proven in yeast [27]. Consistent with this, the levels of acetylated histones are reduced in yeast, mammalian, and plant elp mutants [24,27,29]. Previously we found reduced histone acetylation levels at several defense genes in Atelp2, which was correlated with delayed and/or decreased induction of these genes [47]. Taken together, these results suggest that Elongator-dependent histone acetylation may maintain an inducible state at defense loci in Arabidopsis.
Conserved residues in the radical SAM domain proved equally important for AtELP3's function in plant immunity (Figure 7). Although these data demonstrate the importance of the domain, since the enzymatic activity of this domain is currently unknown, the reason for this importance remains elusive. One possibility is that the radical SAM domain is important for the structural integrity of the protein complex, and in yeast, this has been shown to be the case [30]. Alternatively, ELP3 may bind and cleave SAM, as seen in the archaea M. jannaschii [31]. Indeed, a recent study indicated that the radical SAM domain of mouse ELP3, but not the HAT domain, is required for zygotic paternal genome demethylation [32]. Two days later, total RNA was extracted from the upper uninfected/ untreated leaves and analyzed for the expression of indicated genes by real-time qPCR. Expression was normalized against constitutively expressed UBQ5. Data represent the mean of three independent samples with SD. Different letters above the bars indicate significant differences (p < 0.05, Student's t-test). Note that the comparison was made among WT, npr1, Atelp3-10, and Com(C) for each gene/treatment. (B) SAR induction in WT, npr1, Atelp3-10, and Com(C) plants. Three lower leaves on each plant were inoculated with Psm ES4326 (OD 600 = 0.002) (+SAR) or mock-treated with 10 mM MgCl 2 (−SAR). Two days later, two upper uninfected/untreated leaves were challenge-inoculated with Psm ES4326 (OD 600 = 0.001). The in planta bacterial titers were determined immediately and three days after challenge inoculation. Cfu, colony-forming units. Data represent the mean of ten independent samples with SD. Psm ES4326 grew significantly less in the SAR-induced WT, Atelp3-10, and Com(C) plants than in the mock-treated plants (*All p < 0.05, Student's t-test).
Previously we found altered dynamic DNA methylation changes induced by pathogen infection at some defense loci in Atelp2, which was also correlated with delayed and/or decreased induction [47]. Therefore, Elongator may also modulate chromatin at defense loci in Arabidopsis through DNA methylation.
Another, non-mutually exclusive possibility is that the role of Elongator in plant immunity depends on its function in the modification of wobble uridine tRNA. In yeast and plants, Elongator is essential for these modifications [51,52]. Using overexpression of hypomodified uridine-containing tRNAs, all phenotypes of yeast elp mutants were shown to stem from reduced formation of fully modified tRNAs [53]. Most yeast mutations in the radical SAM domain result in complete loss of the modified nucleoside 5-methoxycarbonylmethyl-2-thiouridine (mcm 5 s 2 U), while mutations of the HAT domain result in the loss of the majority of mcm 5 s 2 U. Loss of mcm 5 s 2 U is closely correlated with the elp mutant phenotypes [53]. In yeast, mutations of conserved amino acids in the SAM domain do not abrogate HAT or RNA binding activity [30]; therefore, for all Elongator roles examined including plant immunity, with the exception of zygotic paternal genome demethylation [32], both the HAT and radical SAM domains of ELP3 are essential, but neither is sufficient. Since plant Elongator is also essential for wobble uridine tRNA modifications [52], it is possible that the immune phenotypes, as well as other elp phenotypes, including altered histone modification and DNA methylation profiles, could be indirect effects of tRNA modification defects in these mutants. Future investigations in plants should address this important question.
Accumulating evidence from diverse organisms is starting to paint a picture where the transcriptional role of Elongator is to facilitate the induction of genes in general. Indeed, previous work performed in our lab and elsewhere has shown that Elongator influences the inducibility of thousands of genes [35,47,54]. In this study we established that the HAT and radical SAM domains of the Arabidopsis Elongator subunit AtELP3 are essential for Elongator's immune function. Elongator has also been implicated in diverse stress responses including oxidative and drought stress in Arabidopsis [41]. It would be interesting to determine the role of the HAT and radical SAM domains of AtELP3 in these processes. In this regard, the HAT and radical SAM domain mutants generated in this study provide a useful tool for dissecting the role of these domains throughout development and under diverse stress conditions.

Conclusions
In this study, we isolated the gns2 mutant and cloned the GNS2 gene using a map-based cloning approach. We found that GNS2 encodes the Elongator subunit AtELP3. Characterization of gns2/Atelp3-10 indicates that AtELP3 is required for basal immunity and ETI, but not for SAR, suggesting that Elongator as a whole is likely involved in basal immunity and ETI, but not in SAR. Furthermore, we demonstrate that AtELP3's function in plant immunity requires both the HAT and radical SAM domains, which is consistent with a role of Elongator in facilitating the transcriptional induction of defense genes through alterations to their chromatin.
Inoculation of plants with Psm ES4326 and Pst DC3000/avrRpt2 was performed by pressure-infiltration with a 1 mL needleless syringe as described previously [18]. After inoculation, eight infected leaves, one from each plant, were collected for each genotype, treatment or time point to determine in planta growth of the pathogen. For SAR induction, three lower leaves on each plant were inoculated with the virulent bacterial pathogen Psm ES4326 (OD 600 = 0.002). Two days later, the upper uninfected systemic leaves were either collected for gene expression analysis or challenge-inoculated with Psm ES4326 (OD 600 = 0.001) for resistance test. Ten leaves were collected three days after challenge inoculation to examine the pathogen growth.

RNA analysis
RNA extraction and RNA gel blot analysis were carried out as described by Cao et al. [14] and Glazbrook et al. [55]. For reverse transcription (RT), total RNA was treated with DNase I (Gibco, BRL) at 37°C for 30 min. After inactivation of the DNase, RT was performed using SUPERSCRIPT First-Strand Synthesis System (Gibco, BRL) and 2 μg of the DNase-treated RNA in a 20 μL reaction. Aliquots of the resulting RT reaction product were used for real-time quantitative PCR (qPCR), which was performed using SYBR Green protocol (Applied Biosystems, Foster City, CA) with 1 μM primers and 0.2 μL aliquot of RT product in a total of 12.5 μL per reaction. Reactions were run and analyzed on a MX3000P real-time PCR machine (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The relative quantity of the tested gene is expressed in relation to ubiquitin (UBQ5) using the formula 2^(Ct{UBQ5}-Ct {GENE}), where 2 represents perfect PCR efficiency. We chose UBQ5 as the reference gene for qPCR normalization, because it is one of the most stably expressed genes [56]. All primers used for qPCR in this study have been reported in DeFraia et al. [42].

SA measurement
SA measurement was done with HPLC as described by Verberne et al. [57].

Plasmid construction and plant transformation
A pair of primers BamH I-AtELP3F1 (5′-CGGGAT CCATGGCGACGGCGGTAGTGATG-3′) and Sac I-At ELP2R (5′-CCGAGCTCTCAAAGAAGATGCTTCACC ATGTAAG-3') was used to amplify the coding region of AtELP3 from total cDNA generated by RT. The PCR products were digested with BamH I and Sac I and then ligated into the corresponding sites of the vector pBI1.4 T, resulting in the plamsid pBI1.4 T-35S::AtELP3. Site-directed mutagenesis of the conserved amino acid residues in the HAT and radical SM domains of AtELP3 was performed in the pBI1.4 T-35S::AtELP3 construct using a PCR-based Quick-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA). The primers used for the site-directed mutagenesis of the HAT domain were AtELP3-HATMuF (5′-GTAGGAACCAGACATG CCGCCAGAAAGTTGGGTTATG-3′) and AtELP3-HA TMuR (5′-CATAACCCAACTTTCTGGCGGCATGTC TGGTTCCTAC-3′), and for the radical SAM domain mutagenesis, the primers were AtELP3-SAMMuF (5′-CGACGGGGAATATATCCGTTTATTCTCCCGGTGG-ACCTGAC-3′) and AtELP3-SAMMuR (5′-GTCAGGT CCACCGGGAGAATAAACGGATATATTCCCCGTCG-3′). The presence of the expected mutations in the resulting construct was verified by DNA sequencing. The plasmids were introduced into the Agrobacterium strain GV3101(pMP90) by electroporation [58], and transformation was performed following the floral dip method [59].

Statistical methods
Statistical analyses were performed with the data analysis tools (Student's t-test: Two Samples Assuming Unequal Variances) in Microsoft Excel of Microsoft Office 2004 for Macintosh. Linear mixed-effects model analysis was performed with the software SAS 9.3. All experiments were repeated at least three times with similar results.