Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

GCN5 contributes to stem cuticular wax biosynthesis by histone acetylation of CER3 in Arabidopsis

GCN5 contributes to stem cuticular wax biosynthesis by histone acetylation of CER3 in Arabidopsis Cuticular wax is a major component of the surface cuticle of plants, which performs crucial functions in optimizing plant growth. Histone acetylation regulates gene expression in diverse biological processes, but its role in cuticular wax synthesis is not well understood. In this study, we observed that mutations of the Arabidopsis thaliana histone acetyltransferase GENERAL CONTROL NON-REPRESSED PROTEIN5 (GCN5) impaired the accumulation of stem cuticular wax. Three target genes of GCN5, ECERIFERUM3 (CER3), CER26, and CER1-LIKE1 (CER1-L1), were identi- fied by RNA-seq and ChIP assays. H3K9/14 acetylation levels at the promoter regions of CER3, CER26, and CER1-L1 were consistently and significantly decreased in the gcn5-2 mutant as compared to the wild-type. Notably, overex- pression of CER3 in the gcn5-2 mutant rescued the defect in stem cuticular wax biosynthesis. Collectively, these data demonstrate that GCN5 is involved in stem cuticular wax accumulation by modulating CER3 expression via H3K9/14 acetylation, which underlines the important role of histone acetylation in cuticular wax biosynthesis. Keywords: Arabidopsis thaliana, CER3, GCN5, histone acetylation, stem, wax biosynthesis. Introduction Plant cuticular wax is a complex mixture of very-long-chain termed the alcohol-forming and the alkane-forming pathways, fatty acids (VLCFAs) and aldehydes, alcohols, alkanes, ketones, which yield 17~18% and 80% of the total amount of wax, and esters, with predominant carbon chain-lengths ranging respectively (Bernard and Joubès, 2013). As a crucial adap- from C22 to C36 (Samuels et al., 2008; Li et al., 2016), and it tive characteristic, cuticular wax protects plants against biotic forms one of the major lipid components of the cuticle that and abiotic stresses, such as pathogen attacks and water loss covers the outer surface of aerial plant tissues. The biosynthesis (Aharoni et  al., 2004; Samuels et  al., 2008). Therefore, eluci- of cuticular wax is processed through two distinct pathways, dating the regulatory mechanisms controlling cuticular wax Abbreviations: GCN5, general control non-repressed protein5; CER3, eceriferum3; VLCFA, very-long-chain fatty acid; HAT, histone acetyltransferase; HDAC, histone deacetylase; GO, gene ontology; GUS, β-glucuronidase; GC-FID, gas chromatography–flame ionization detector. © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Experimental Biology. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2912 | Wang et al. accumulation is of great interest for genetic engineering of However, to the best of our knowledge, histone-acetylating agricultural crops. events have not been reported in cuticular wax biosynthesis. To date, numerous candidate genes in the cuticular The reversible modulations of histone acetylation and de- wax pathway have been identified, most of which encode acetylation are catalysed by histone acetyltransferases (HATs) enzymes, or work with enzymes, in the VLCFA biosynthesis and histone de-acetylases (HDACs), respectively (Pandey and derivatization pathways (Koornneef et al., 1989; Bernard et  al., 2002). As in yeast and mammals, the Arabidopsis and Joubès, 2013). For example, CER2, and its homologues HATs are grouped into four classes: GNAT (GCN5-related CER2-LIKE1/CER26 and CER2-LIKE2 have biological N-acetyltransferase), MYST (for ‘MOZ, Ybf2/Sas2 and functions in two-carbon elongation of VLCFAs, from C28 Tip60’), p300/CBP (p300/CREB-binding protein), and TAF1 to C30, C30 to C32, and C32 to C34, respectively (Haslam [for ‘TATA-binding protein (TBP)-associated factor’] (Servet et  al., 2012, 2015; Pascal et  al., 2013; Lee and Suh, 2015). et al., 2010). GCN5, a GNAT-type HAT that harbors a HAT A  member of the bifunctional wax ester synthase/diacylg- domain and a bromodomain, has been well studied in exten- lycerol acyltransferase family, WSD1, is a key enzyme in wax sive research (Benhamed et  al., 2008; Servet et  al., 2010). It ester synthesis of Arabidopsis stems (Li et  al., 2008). As for has been observed that mutations of GCN5 result in various the essential alkane-forming pathway, CER3/WAX2 plays an growth impairments, such as dwarfism, defects in terminal important role in synthesis of major wax components (Aarts flower production, deformed seed development, and poor fer- et al., 1995; Chen et al., 2003; Rowland et al., 2007; Lee and tility (Bertrand et al., 2003; Vlachonasios et al., 2003). Moreover, Suh, 2015). The amount of wax found in the cer3 mutant was GCN5 regulates a variety of biological processes in Arabidopsis, severely reduced compared with the wild-type in Arabidopsis including cell differentiation, shoot and floral meristem forma- stems, especially with regards to aldehydes, alkanes, second- tion, light and abiotic stress (e.g. heat and cold) responses, iron ary alcohols, and ketones (Rowland et  al., 2007). Although homeostasis, and fatty acid biosynthesis, by catalysing histone the exact reactions catalysed by the CER3 enzyme remain acetylation levels of target promoters at certain sites, including unknown, it has been reported that CER3 may physically H3K14, H3K9, and H3K27 (Laux et al., 1996; Stockinger et al., interact with CER1 and CYTOCHROME B5 ISOFORM 2001; Bertrand et al., 2003; Vlachonasios et al., 2003; Benhamed (CYTB5) for the biosynthesis of very-long-chain alkanes et  al., 2006; Earley et  al., 2007; Hu et  al., 2015; Xing et  al., (Bernard et al., 2012). 2015; Wang et  al., 2016). In this current study, we found that An increasing focus on the transcriptional regulation of the the gcn5 mutants had significantly glossy (wax-deficient) stems genes in cuticular wax biosynthesis has led to some transcrip- compared with the wild-type. Analyses of chemical compo- tion factors being reported in recent studies (Lee and Suh, nents coupled with scanning electron microscopy showed an 2015). In Arabidopsis, WAX INDUCER1/SHINE1 (WIN1/ obvious reduction in amounts of total wax and changes in SHN1), an AP2-EREBP-type transcription factor, was the first their composition. Moreover, we found that GCN5 bound to to be identified and is a representative regulator of wax biosyn- the promoter of CER3 and this interaction was impaired in thesis, regulating the CER1, CER2, and 3-KETOACYL-COA the gcn5-2 mutant. Taken together, we conclude that GCN5- SYNTHASE1 (KCS1) genes (Aharoni et al., 2004; Broun et al., mediated histone acetylation of CER3 regulates stem cuticular 2004). WIN1 overexpression lines exhibited enhanced drought wax biosynthesis in Arabidopsis. tolerance compared with the wild-type (Aharoni et al., 2004). In addition, MYB transcription factors are important for wax Materials and methods biosynthesis under both biotic and abiotic stresses (Suh et al., 2005; Lee and Suh, 2015). For example, MYB96 controls wax Plant material and growth conditions biosynthesis by regulating the KCS1, KCS2, KCS6, BETA- The Arabidopsis thaliana wild-types Col-0 and Ws, together with T-DNA KETOACYL REDUCTASE1 (KCR1), and CER3 genes insertion mutants involved in histone modifications were used in this under drought stress (Lee et  al., 2016b), and KCS1, KCR1, study, as follows. (1) Histone acetylation: gcn5-1 and gcn5-2 (Ws back- ground), hda2, hda2c, hda5, hda7, hda9, hda13, hda18, hda19, and srt2 CER2, and CER3 are the targets of MYB30 in response to (Col-0 background); and (2) histone methylation: ashh1, ashh2, ashh3, pathogen attack (Raffaele et al., 2008). ashh4, ashr2, ashr3, atx1, atx2, atx4, atx5, atxr2, atxr3, and atxr4 (Col-0 With our increasing understanding of epigenetic mecha- background). The gcn5-1 and gcn5-2 mutants were both T-DNA inser- nisms, recent reports have demonstrated that several epige- tion mutants in the bromodomain-coding region (Bertrand et al., 2003; netic events are involved in wax biosynthesis (Lee and Suh, Vlachonasios et  al., 2003; Supplementary Fig.  S3A at JXB online). Notably, the gcn5-2 mutation removes the entire bromodomain, which is 2013, 2015). Recently, two RING E3 ligases, HISTONE required for binding to 11% of the GCN5 promoter targets (Servet et al., MONOUBIQUITINATION1 (HUB1) and HUB2, were 2010). The other mutants were obtained in the homozygous state from demonstrated to be involved in wax biosynthesis by mon- ABRC (https://abrc.osu.edu/) or from individual donors. For germina- oubiquitinating histone H2B proteins, which in turn activates tion, sterilized seeds were incubated at 4  °C for 3 d, and subsequently the transcriptional levels of the wax biosynthetic genes LONG- sown on Murashige and Skoog (MS) plates containing 1% sucrose and 0.6% agar. The seedlings were grown under 16/8 h light/dark conditions CHAIN ACYL-COA SYNTHETASE2 (LACS2) and CER1 at 22 °C in a growth room. (Ménard et  al., 2014). The Arabidopsis histone methyl trans- ferases SET DOMAIN GROUP8 (SDG8) and SDG25 have been reported to contribute to wax accumulation through his- RNA isolation and RNA-seq tone lysine methylation and/or H2B ubiquitination by target- Total RNA was extracted using TRIzol reagent (Invitrogen), according ing the key wax biosynthetic gene CER3 (Lee et  al., 2016a). to the manufacturer’s instructions. RNA concentrations were measured Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2913 using a NanoDrop 2000 spectrophotometer (ND-2000, ThermoFisher described by Fiil et  al. (2008). Six-week-old stems of Ws and gcn5-2 Scientific, Inc., MA, USA). RNA integrity was assessed using an Agilent mutants were harvested and fixed in 1% formaldehyde for 15 min in a 2100 Bioanalyser (Agilent Technologies, Inc., CA, USA). Paired-end vacuum and subsequently neutralized using 0.125 M glycine for 5 min. sequencing libraries with an average insert size of 200 bp were prepared After washing with sterilized water, the samples were dried with towels, using the TruSeq RNA Sample Preparation Kit v2 (Illumina, San Diego, and ground in liquid nitrogen. The resulting powders were resuspended in USA) and sequenced using a HiSeq2500 platform (Illumina, San Diego, the Nuclei Extraction Buffer 1, which contained 0.4 M sucrose, 10 mM USA) according to the manufacturers’ standard protocols. Raw data Tris-HCl, pH 8.0, 10 mM MgCl , 5 mM β-mercaptoethanol, 0.1 mM obtained from Illumina sequencing were processed and filtered using PMSF (Sigma, P7626), and protease inhibitors (Roche, 11873580001), the Illumina pipeline (http://www.illumina.com) to generate FastQ and mixed immediately. After incubation for 20 min at 4 °C with a rota- files. Approximately 12 G of high-quality 125-bp paired-end reads were tor, the solutions were filtered through four layers of Miracloth into generated from six libraries (Supplementary Table S1). The FastQC pro- new tubes, and the filtrate was then centrifuged for 20 min at 3000 g at gram (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was 4  °C. The nuclei pellets were resuspended in Nuclei Extraction Buffer used to evaluate the overall quality of the RNA-seq reads. Poor-quality 2, which contained 0.25 M sucrose, 10  mM Tris-HCl, pH8.0, 10  mM bases were filtered out using Sickle (https://github.com/najoshi/sickle). MgCl , 1% Triton X-100, 5  mM β-mercaptoethanol, 0.1  mM PMSF, High-quality RNA-seq reads from each library were mapped to The and protease inhibitors. The suspensions were transferred to microfuge Arabidopsis Information Resource (TAIR10) version of the Arabidopsis tubes and centrifuged at 12 000 g for 10 min at 4 °C. The pellets were genome using the splice-junction-aware short-read alignment suite resuspended in Nuclei Extraction Buffer 2 and centrifuged at 12 000 g TOPHAT v2.09 with default settings (Kim and Salzberg, 2011). The through a layer of Nuclei Extraction Buffer 3, which contained 1.7 M reads displaying unique alignment and not more than two nucleotide sucrose, 10 mM Tris-HCl, pH 8.0, 2 mM MgCl , 0.15% Triton X-100, mismatches were kept for further analysis. The differentially expressed 5  mM β-mercaptoethanol, 0.1  mM PMSF, and protease inhibitors, in genes were identified by using the edgeR package (ver. 3.2.3) with an microfuge tubes for 60  min at 4  °C. The nuclear pellets were lysed in absolute value of log -fold change ≥2 and a false-discovery rate <0.05 Nuclei Lysis Buffer, which contained 50 mM Tris-HCl, pH8.0, 10 mM as cut-off (Robinson et al., 2010). The groups of differentially expressed EDTA, 1% SDS, and protease inhibitors. The lysed nuclei were sonicated genes identified by RNA-seq in this study are shown in Supplementary four times with a Bioruptor (UCD-200) in a water bath at 4 °C, with Table  S2. Gene Ontology (GO) analysis was performed using agriGO each sonication period consisting of 15 s on and 15 s off for 5 min and v2.0 with a cut-off of P-value <0.05 (Tian et  al., 2017), and the total followed by centrifugation. The clear supernatants, which contained the enrichment categories are identified in Supplementary Table S3. sonicated chromatin, were transferred to new tubes. Immunoprecipitation (IP) was performed using 5 μl of chromatin, and the following IP steps were conducted using the Magna ChIP™ HiSens Kit. Aliquots of the Quantitative real-time PCR dilution were used for the IP assays. The anti-GCN5 antibody was gen- erated using two synthetic peptides (H2N-CARGADTDSDPDESED Real-time PCR was performed as previously described (Livak and and H2N-SSRNTKLKTESSTVKLC); both peptide epitopes are located Schmittgen, 2001) and ACTIN8 was used as the control gene, i.e. the between amino acids 85 to 99 and between amino acids 136 to 150, expression levels of each gene were normalized to that of ACTIN8. which is the N-terminal region of the protein, provided by Prof. D-X The primer pairs used for real-time PCR are listed in Supplementary Zhou, and the specificity of this GCN5 antibody was confirmed by Table S4. The PCR analysis was performed using a CFX96 System (Bio- protein gel blots (Benhamed et al., 2006, 2008). The anti-H3K14ac and Rad) with SYBR Green. The following program was used for the real- anti-H3K9ac antibodies were purchased from Upstate Biotechnology. time PCR: 95 °C for 3 min and 40 cycles of 95 °C for 30 s, 58 °C for CHALCONE SYNTHASE (CHS) and AT4G03800 (gypsy-like retro- 30 s, and 72 °C for 30 s. transposon family gene) were amplified as endogenous controls for the anti-GCN5 and anti-H3K9/H3K14 antibodies, respectively. The immu- Plasmid construction and plant transformation noprecipitated DNA was analysed by quantitative PCR in three bio- logical replicates using the primer sets listed in Supplementary Table S4. A DNA fragment containing a 2.0-kb fragment upstream of the CER3 Amplified DNA from the chromatin fractions prior to antibody incuba- coding sequence and full-length ORFs of GCN5 and CER3 were tion were used as the controls (inputs). The fold-enrichment was normal- amplified by PCR-directed cloning based on the annotation from TAIR ized to the chromatin inputs. using the following primer pairs: CER3-P-F and CER3-P-R, GCN5-F and GCN5-R, CER3-F and CER3-R, respectively (Supplementary Table  S4). The sequence-confirmed clones containing the ORFs of GUS histochemical and fluorometric assays GCN5 and CER3 were then respectively cloned into the binary expres- sion vector pCAMBIA1300 (driven by the CaMV35S promoter). The Three homologous transgenic T3 lines of ProCER3::GUS/Ws and the promoter region of CER3 was fused to the reporter gene encoding corresponding homologous ProCER3::GUS/gcn5-2 lines were used for β-glucuronidase (GUS). The chimeric gene was then cloned into the GUS histochemical analysis. The seedlings were grown under strictly iden- binary expression vector pCAMBIA1300 to generate ProCER3::GUS. tical conditions. After pollination, Arabidopsis stems, siliques, flowers, and These vectors were transferred into the Agrobacterium tumefaciens strain young leaves were vacuum-infiltrated with staining buffer (2 mM potas- GV3101. Transgenic plants were generated using the floral dip method sium ferricyanide, 10 mM phosphate buffer, 0.5% Triton X-100, and 1 mg –1 −1 and subsequently screened on solid plates containing 25 mg l hygromy- ml X-Gluc) and then incubated overnight. The tissue was then incubated cin (Clough and Bent, 1998). The hygromycin-resistant seedlings were in ethanol and acetic acid (1:1) for 4–8 h and cleared in 80% ethanol. The then transferred to a mixture of soil and vermiculite (2:1). At least three samples were observed with a stereomicroscope (Olympus SEX16). independent T3 homozygous lines with a single T-DNA insertion were For the quantification of GUS activity, we used the fluorometric assay subjected to a detailed analysis. Because the gcn5-2 mutant exhibits low- based on the method of Jefferson et al. (1987). Total protein extracts from fertility pollen that hinders the direct acquisition of transgenic plants in stems of three independent lines for each construct (ProCER3::GUS/ the gcn5-2 background (Bertrand et al., 2003; Xing et al., 2015), we ini- Ws and ProCER3::GUS/gcn5-2) were determined using bovine serum tially generated transgenic plants in the Ws background, and three inde- albumin (BSA) as a standard according to the Bradford assay (Bradford, pendent transgenic T3 lines were selected for crossing into the gcn5-2 1976). Fluorescence was measured using 4-methylumbelliferone (4-MU) mutant. The homologous transgenic lines in the gcn5-2 background were as a substrate, with an excitation wavelength of 365 nm and an emission selected using the same method described above. wavelength of 455 nm in a BioTek Synergy HT Multi-Mode Microplate Reader (BioTek, Vermont, USA). GUS activities of the extracts were cal- culated as nanomole 4-MU per minute per milligram protein. ChIP assay analysis Both GUS histochemical and fluorometric assays were conducted at ChIP assays were performed using the Magna ChIP™ HiSens Kit least three times, and only the transgenic lines with stable GUS signals (Catalogue No. 17-10460) combined with the ChIP method as previously throughout different generations were selected for further analysis. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2914 | Wang et al. Scanning electron microscopy ashr2, and hda9 showed more cuticular wax crystals than the Six-week-old stems were attached to double-sided carbon sealing tape. wild-type (Col-0), whereas less abundant wax was observed The specimens were examined under a Hitachi TM3000 Tabletop in atx5, ashh2, and hda18. Remarkably, fewer stem wax crys- Scanning Electron Microscope at 15 kV, and the digital recordings were tals were observed on the surface of the gcn5-2 mutant rela- saved as TIFF files. tive to the wild-type (Ws), which led to the appearance of a glossy stem (Fig.  1A, B). The compositions of the cuticular Cuticular wax analysis wax of the gcn5-2 mutant and wild-type were quantified by Cuticular wax analysis was conducted as previously described by Chen GC-FID and GC-MS analysis. The total cuticular wax content et al. (2011). Because the environmental conditions could affect configu- in the gcn5-2 mutant was approximately 63% of the control ration and distribution of the surface wax structures, all the plants we used (Fig.  1C), which was attributable to notable decreases in the in this study were cultivated in strictly controlled temperature and humid- major wax constituents in the mutant stem, including alkanes ity. Six-week-old stems were collected and pictures were taken (with (C29), ketone (C29), primary alcohols (C26, C28, and C30), an adjacent ruler) for later determination of area using ImageJ (https:// imagej.nih.gov/ij/). Cuticular waxes were extracted by immersing the secondary alcohols (C29), aldehydes (C28 and C30), and esters samples for 30 s in 1 ml of chloroform containing 10 μg of tetracosane (C42 and C44; Fig. 1D and Table 1). (Fluka) as an internal standard. Three biological replicates per genotype To confirm that the observed defects in stem wax were were performed. Five individual stems were used for each replicate. The indeed caused by the disruption of GCN5, another well- extracts were transferred to reactive vials, dried under nitrogen gas, and established T-DNA insertion mutant allele of GCN5 (gcn5- derivatized by adding 20 μl of N, N-bis-trimethylsilyltrifluoroacetamide (Macherey-Nagel) and 20  μl of pyridine, and incubated for 40  min at 1) was used for further investigation. As expected, the gcn5-1 70 °C. These derivatized samples were then analysed using a gas chro- mutant also exhibited wax-deficient phenotypes (Fig.  1 and matography–flame ionization detector (GC-FID, Agilent, Technologies) Table  1). Moreover, a genetic complementation experiment and GC-MS (Agilent gas chromatograph coupled to an Agilent 5973N was performed by introducing a full-length GCN5 coding quadrupole mass-selective detector). sequence driven by the CaMV35S promoter into the gcn5-2 Consistent with previous studies (Haslam et  al., 2015; Lee and Suh, 2015; Li et al., 2016), the content of the unidentified components (which mutant. Given the low-fertility pollen of the gcn5-2 mutant showed no significant differences among genotypes) was excluded from found in previous studies (Bertrand et  al., 2003; Xing et  al., the total wax load. 2015), we first generated three independent homologous 35S::GCN5/Ws transgenic plants harboring significantly Statistical analysis high GCN5 expression levels, and crossed them with gcn5-2 Statistical analyses of the phenotypic data and expression levels were per- mutants (Supplementary Fig.  S3B). After three generations, formed using Student’s t-test in Excel. To assess the overall differences in three independent homozygous 35S::GCN5/gcn5-2 (#1, the stem cuticular wax composition among genotypes, we compared the #5 and #6) transgenic lines were obtained (Supplementary means using one-way ANOVA together with a Bonferroni adjustment Fig.  S3B). No obvious phenotype differences were observed test in R. between 35S::GCN5/Ws transgenic plants and the wild-type (Supplementary Fig.  S3C). However, 35S::GCN5/gcn5-2 Accession numbers transgenic lines exhibited similar phenotypes to the wild-type Sequence data from this article can be found in the Arabidopsis Genome (Fig. 1, Table 1, and Supplementary Fig. S3C), indicating that Initiative or GenBank/EMBL databases under the following accession constitutive expression of GCN5 could rescue the gcn5-2 wax numbers: CER3, AT5G57800; GCN5, AT3G54610; CER26, AT4G13840; deficiency. These results indicated that GCN5 is essential for CER1-L1, AT1G02190; WSD1, AT5G37300; AT2, AT5G55370; FAR3, the normal accumulation of cuticular wax on the stem surface AT4G33790, ACTIN8, AT1G49240; CHS, AT5G13930; AT4G03800. The RNA-seq reads used for this study are deposited at the National of Arabidopsis. Center for Biotechnology Information Short Read Archive under the accession number SRP093334. RNA-seq analysis reveals significant alteration of lipid-related gene expression in the gcn5-2 mutant Results To investigate whether the reduction of cuticular wax con- tent in the gcn5-2 mutant was caused by decreased expression Mutation of Arabidopsis GCN5 impairs stem cuticular of wax-related genes, total RNA of 6-week-old stems of the wax deposition gcn5-2 mutant and the wild-type Ws were isolated for high- To investigate the potential roles of histone modification in throughput RNA sequencing and transcriptomic comparison. cuticular wax accumulation, 23 T-DNA insertion mutants that Three biological replicates per genotype were performed and harbor disruption in histone acetylation or methylation genes the correlation coefficients of each genotype showed favor- were selected for analysis, namely ashh1, ashh2, ashh3, ashh4, able reproducibility (Supplementary Fig.  S4). The high reli- ashr2, ashr3, atx1, atx2, atx4, atx5, atxr2, atxr3, atxr4, gcn5-2, ability of the RNA-seq data was verified by qPCR of 10 hda2, hda2c, hda5, hda7, hda9, hda13, hda18, hda19, and srt2. randomly selected genes (Supplementary Fig.  S5). For each The stem cuticular wax of these mutants and the wild-type sample, values for reads per kilobase of exon model per million controls were observed and compared using SEM. As shown mapped reads (RPKM) were calculated, and the genes with in Supplementary Figs  S1 and S2, we found wide variations at least 2-fold change and a false discovery rate value P≤0.05 for cuticular wax crystal morphology and crystallization pat- were selected. Compared with the control, we found that 54% terns in the mutants and wild-types. For example, atx4, atxr4, (2616 genes) of the total of differentially expressed genes were Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2915 Fig. 1. Mutations in GCN5 are responsible for defects in the stem cuticular wax in Arabidopsis. (A) Light reflectance and (B) SEM observations of 6-week-old stems from wild-type Ws, and gcn5-2, 35S::GCN5/gcn5-2 transgenic lines (#1, #5, #6), and gcn5-1 mutants. Scale bars: (A) 1 mm, (B) 30 μm. (C) Total cuticular wax content, calculated per unit area of 6-week-old stems from the four different genotypes. (D) Cuticular wax composition of 6-week-old stems for wild-type Ws, and gcn5-2 and gcn5-1 mutants, and (E) for Ws and 35S::GCN5/gcn5-2 transgenic lines (#1, #5, #6). The chemical classes and main chain-lengths of each constituent are indicated. Data are means (±SD) of three biological replicates. *P<0.05, **P<0.01; Student’s t-test. (This figure is available in colour at JXB online.) down-regulated in gcn5-2 mutant stems. Because GCN5 usu- cellular lipid metabolic process, and lipid biosynthetic pro- –6 ally positively regulates transcriptional processes (Servet et al., cess (P≤6.2  ×  10 ). Moreover, a significant fraction of genes 2010) and mutation of GCN5 decreased the total wax load in involved in metabolic and/or biosynthetic processes (including Arabidopsis stems, these 2616 genes were expected to be dir- glycerolipid, neutral lipid, fatty acid, and unsaturated fatty acid), ect or indirect GCN5 target genes and might be involved in lipid catabolic process, and regulation of lipid metabolic and cuticular wax accumulation. biosynthetic processes were enriched (Table 2), suggesting that We then used GO analysis of the 2616 candidate genes using GCN5 is involved in wax accumulation by modulating the tran- agriGO v2.0 (Tian et  al., 2017), and the categories showed scription of lipid-related genes. For further screening, 145 non- considerably high enrichments in lipid metabolic process, redundant genes from three most-abundant lipid-related GO Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2916 | Wang et al. Table 1. Cuticular wax composition of stems in the wild-type Ws, gcn5-2, gcn5-1, and complementary transgenic lines Total Alkanes Ketones 1-Alcohols 2-Alcohols Fatty acids Aldehydes Esters Ws 1347.99 ± 54.38 650.81 ± 34.27 226.39 ± 12.94 140.44 ± 2.34 122.16 ± 6.91 10.39 ± 1.13 109.86 ± 4.70 87.93 ± 4.82 gcn5-2 847.49 ± 35.42** 375.94 ± 7.98** 133.64 ± 2.16** 87.01 ± 22.68* 83.59 ± 4.50** 7.03 ± 2.08 66.84 ± 14.77** 93.45 ± 38.77 #1 1357.28 ± 67.64 659.13 ± 34.75 230.33 ± 20.01 138.11 ± 17.45 122.64 ± 8.57 9.75 ± 3.34 111.81 ± 14.64 85.52 ± 7.39 #5 1360.93 ± 61.58 666.21 ± 27.19 230.35 ± 19.85 141.23 ± 15.52 116.23 ± 12.49 11.59 ± 2.21 110.60 ± 8.14 84.72 ± 17.21 #6 1344.06 ± 57.09 655.99 ± 27.72 228.46 ± 8.61 134.99 ± 11.51 117.54 ± 8.16 9.87 ± 1.85 111.51 ± 8.10 85.72 ± 14.12 gcn5-1 875.47 ± 90.61** 387.21 ± 41.12** 136.54 ± 13.30** 96.61 ± 17.24* 96.91 ± 16.31* 7.38 ± 1.03 72.00 ± 10.66** 78.82 ± 7.08 –2 Data are means (±SD; n=3) for the total wax load and each constituent class in μg dm . Five individual stems were used for each replicate. #1, #5, #6 represent three complementary transgenic lines (35S::GCN5/gcn5-2#1, #5, #6). *P<0.05, **P<0.01; Student’s t-test. 1-Alcohols, primary alcohols; 2-Alcohols, secondary alcohols. Table 2. Lipid-related GO categories for the 2616 candidate genes down-regulated in the gcn5-2 mutant GO ID Term Query item Query total Bg item Bg total P-value –7 GO:0006629 Lipid metabolic process 145 2598 994 28362 2.7 × 10 –6 GO:0044255 Cellular lipid metabolic process 100 2598 649 28362 2.5 × 10 –6 GO:0008610 Lipid biosynthetic process 83 2598 522 28362 6.20 × 10 GO:0046486 Glycerolipid metabolic process 24 2598 122 28362 0.0011 GO:0045017 Glycerolipid biosynthetic process 15 2598 79 28362 0.011 GO:0046460 Neutral lipid biosynthetic process 6 2598 18 28362 0.012 GO:0006638 Neutral lipid metabolic process 7 2598 24 28362 0.013 GO:0019216 Regulation of lipid metabolic process 11 2598 54 28362 0.019 GO:0016042 Lipid catabolic process 31 2598 224 28362 0.024 GO:0046890 Regulation of lipid biosynthetic process 9 2598 43 28362 0.028 GO:0006636 Unsaturated fatty acid biosynthetic process 6 2598 24 28362 0.036 GO:0006631 Fatty acid metabolic process 33 2598 252 28362 0.037 Query item: number of down-regulated genes in the gcn5-2 mutant annotated as the corresponding GO term. Query total: number of down-regulated genes in the gcn5-2 mutant. Bg item: number of genes in Arabidopsis whole genome annotated as the corresponding GO term. Bg total: number of genes in Arabidopsis whole genome. and Ws plants and GCN5-specific antibodies (Benhamed et al., Table 3. Potential GCN5-regulated genes involved in cuticular 2006, 2008). ChIP-qPCRs with three primer pairs spanning wax synthesis the promoter regions and gene body regions of each gene Gene Name Annotation (WSD1, CER3, CER26, CER1-L1, and AT2) were conducted for binding tests (Fig. 2B). CHS was used as the negative con- AT5G37300 WSD1 Bifunctional wax synthase/acyl-CoA:diacylglycerol acyltransferase trol as its expression is not affected by GCN5 (Benhamed et al., AT5G57800 CER3 ECERIFERUM 3 2006). As shown in Fig. 2C, significant decreases in enrichment AT4G13840 CER26 ECERIFERUM 26 in the gcn5-2 mutant were observed for most of the examined AT1G02190 CER1-L1 Protein CER1-like 1 regions of the CER3, CER26, and CER1-L1 genes relative to AT5G55370 AT2 Long-chain-alcohol O-fatty-acyltransferase 2 the wild-type, especially in the transcription start-site region (P1) of CER3. By contrast, no significant changes in GCN5 terms (GO:0006629, GO:0044255, and GO:0008610) were enrichment were detected in any tested regions for the other analysed together with their biological properties and functions two genes (WSD1 and AT2). in cuticular wax development as described in previous reports Previous studies have reported that GCN5 is specifically (Bernard and Joubès, 2013; Lee and Suh, 2015). Finally, we fil- responsible for H3K14 acetylation (H3K14ac) and that it tered five cuticular wax genes that are potential target genes of influences the H3K9ac and H3K27ac at the promoters of their GCN5, namely WSD1, CER3, CER26, CER1-L1, and Long- targets, which are positively correlated with gene expression chain-alcohol O-fatty-acyltransferase 2 (AT2; Table 3). (Bertrand et al., 2003; Bhat et al., 2003; Earley et al., 2007). Thus, we analysed the acetylation levels of H3K14 and H3K9 at the CER3, CER26, and CER1-L1 loci using the primer sets indi- ChIP assays identify target genes of GCN5 involved in cated in Fig. 2B. Consistently, both the H3K14ac and H3K9ac cuticular wax biosynthesis levels of these candidate genes were significantly decreased in the gcn5-2 mutant compared to the wild-type, especially at the To confirm the GCN5-regulated target genes in stem cuticular wax biosynthesis, we first detected the transcript levels of the promoter regions (Fig. 3), which was in accordance with the expression profiles (Fig. 2A). Collectively, these data indicated five candidate genes derived from RNA-seq data by qRT-PCR. that CER3, CER26, and CER1-L1 are the targets of GCN5, As expected, their expression levels were significantly reduced in the gcn5-2 mutant (Fig. 2A). The candidate genes were then and their expression can be regulated by GCN5 by modulating analysed by ChIP assays, using 6-week-old stems of the gcn5-2 their H3K14 and H3K9 acetylation. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2917 Fig. 2. Identification of GCN5 target genes. (A) Expression levels of the GCN5 target genes as determined by qRT-PCR in 6-week-old stems of the wild-type Ws and the gcn5-2 mutant. (B) Diagram representing the genomic structure and primer sets (indicated by P1–P5) analysed for ChIP-qPCR in the AT2, CER1-L1, CER26, WSD1, and CER3 genes. The exact distance (in bp) of the primers from to the ATG start codon sites (indicated by triangles) are labeled. Black boxes represent exons and gray boxes represent untranslated regions (UTRs). (C) ChIP analysis with nuclei extracted from cross- linked, 6-week-old stems of Ws and the gcn5-2 mutant and antibody-specific for GCN5. The CHS gene was used as a negative control, which provided background level for the ChIP samples. Data are means (±SD) from at least three biological replicates. *P<0.05, **P<0.01; Student’s t-test. Fig. 3. H3K14 and H3K9 acetylation levels on GCN5 target genes. ChIP analysis of H3K14 and H3K9 acetylation on (A, B) CER1-L1, (C, D) CER26, and (E, F) CER3 genes. Nuclei extracted from cross-linked, 6-week-old stems of the wild-type Ws and gcn5-2 mutant and antibodies specific for H3K14ac and H3K9ac. AT4G03800 (gypsy-like retrotransposon family gene) was used as a negative control. Data are means (±SD) from at least three biological replicates. *P<0.05, **P<0.01; Student’s t-test. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2918 | Wang et al. CER3 is a critical enzyme for cuticular wax synthesis composition of the CER3 overexpression transgenic lines in (Aarts et  al., 1995; Chen et  al., 2003; Rowland et  al., 2007). both the Ws and gcn5-2 mutant backgrounds. Transcript lev- Interestingly, the cuticular wax composition of cer3 mutant els of the CER3 in 35S::CER3/Ws and 35S::CER3/gcn5-2 stems was quite similar to that of the gcn5-2 mutant stem, espe- transgenic lines increased significantly compared with the cially for aldehydes, alkanes, (secondary alcohols, and ketone controls (Fig.  5A). SEM imaging showed that the deposition (Lai et al., 2007; Rowland et al., 2007). Therefore, we hypoth- of wax crystals on the 35S::CER3/gcn5-2 stem was obviously esized that CER3 might be a critical target of GCN5 and increased compared with that of the gcn5-2 mutant and that decided to investigate the gene expression patterns of CER3 it basically recovered to the level of Ws, but no significant in ceriferous (i.e. wax-producing) tissues of gcn5-2 and Ws. differences were observed between 35S::CER3/Ws and Ws Isogenic ProCER3::GUS/gcn5-2 and ProCER3::GUS/Ws (Fig. 5B). lines were obtained in which each transgene was homozy- In addition, the total load and composition of stem cuticu- gous and correspondingly inserted into a single genomic locus. lar wax of 35S::CER3/Ws, 35S::CER3/gcn5, Ws, and gcn5- GUS staining revealed weaker signals in stems, siliques, flowers, 2 plants were measured by GC-FID and GC-MS analysis. and young leaves of the gcn5-2 mutant compared to that of the Although the wax crystal loading showed little increase under control (Fig.  4A). Quantification of GUS activity by fluoro- SEM observation (Fig.  5B), the total amount and individ- metric assay consistently revealed significantly lower activity in ual components of stem cuticular wax in two of the three three independent homozygous ProCER3::GUS/gcn5-2 lines 35S::CER3/Ws transgenic lines (#5 and #10) were signifi- than in the corresponding ProCER3::GUS/Ws lines (Fig. 4B). cantly increased as compared to Ws, especially for the amount qRT-PCR analysis also showed that the CER3 expression lev- of alkanes, primary alcohols, and ketone, which were consistent els were decreased in different tissues of gcn5-2 mutant as com- with the expression levels of the CER3 gene (Figs.  5C, 6A). pared to the wild-type (Fig. 4C). These results provided further Notably, the total wax load in 35S::CER3/gcn5-2 transgenic evidence that the expression of CER3 is positively regulated plants was obviously increased compared with that of the gcn5- by GCN5 in ceriferous tissues. 2 mutant, and in the 35S::CER3/gcn5-2 #10 transgenic line it was even restored to the wild-type level (Fig. 5C). Analysis of the cuticular wax composition showed that the contents of Overexpression of CER3 rescues the stem cuticular C29 alkane, C29 ketone, C26 and C28 primary alcohols, C29 wax-deficient phenotype in the gcn5-2 mutant secondary alcohols, and C30 aldehyde significantly increased To assess the role of CER3 in GCN5-modulated biosynthe- in the 35S::CER3/gcn5-2 transgenic lines compared with the sis of stem cuticular wax, we studied the accumulation and control, thus providing genetic evidence that CER3 plays an Fig. 4. Expression patterns of CER3 in ceriferous tissues of the wild-type Ws and gcn5-2 mutant. (A) Spatial expression patterns of the CER3 gene in transgenic Ws and gcn5-2 plants harboring the CER3 promoters fused to the GUS gene. Promoter activity was visualized through histochemical GUS-staining in stems, siliques, flowers, and young leaves of 6-week-old plants. Scale bars: stems and siliques, 1.5 mm; flowers and leaves, 2 mm. (B) Quantification of the GUS activity using 4-methylumbelliferone (4-MU) as a substrate and (C) CER3 expression levels in stems, siliques, flowers, and young leaves of 6-week-old plants. Data are means (±SD) from three biological replicates. **P<0.01; Student’s t-test. (This figure is available in colour at JXB online.) Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2919 important role in GCN5-regulated cuticular wax biosynthesis Post-translational modification of histone tails plays an important (Fig.  6B). It should be noted that overexpression of CER3 role in epigenetic regulation of gene expression, and this includes did not completely restore all wax component defects in the histone acetylation, methylation, phosphorylation, and ubiqui- gcn5-2 mutant, for example C30 primary alcohol and some tination (Pfluger and Wagner, 2007). As a member of the HAT esters, and this may be attributed to other GCN5-regulated enzymes, GCN5 is a versatile regulator of Arabidopsis devel- target genes in the wax biosynthetic pathway. FATTY ACID opment and stress responses (Bertrand et  al., 2003; Benhamed REDUCTASE3 (FAR3, also known as CER4), encodes an et  al., 2006; Kornet and Scheres, 2009; Hu et  al., 2015; Xing alcohol-forming fatty acyl-coenzyme A  reductase, and is et al., 2015). Recently, we found that GCN5 is involved in FA involved in the synthesis of primary alcohols (Rowland et al., biosynthesis by affecting the acetylation levels of FAD3 ( Wang 2006; Wang et al., 2015). Thus, we analysed the transcript levels et al., 2016). Here, we observed that GCN5 is essential for lipid of FAR3 in gcn5-2, Ws, and 35S::CER3 transgenic plants using metabolism in Arabidopsis stems. Firstly, mutation of GCN5 in qRT-PCR. However, the expression levels of FAR3 in these Arabidopsis compromised the content of multiple lipid com- plants were not obviously changed (Supplementary Fig. S6). pounds (including very-long-chain alkanes, aldehydes, ketones, and alcohols), which resulted in a complete deficiency in stem cuticular wax accumulation. This wax deficiency could be fully Discussion rescued by complementation with 35S::GCN5. Secondly, GO Lipids are an essential constituent of all plant cells, including fatty analysis indicated that down-regulated genes in the gcn5-2 acids (FAs), waxes, sterols, and others (Li-Beisson et  al., 2013). mutant were enriched in categories related to lipid synthesis, Fig. 5. Overexpression of CER3 in the gcn5-2 mutant increases the cuticular wax component and restores it to the wild-type Ws level. (A) Relative expression levels of CER3 in plants of Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10). Total RNA was isolated from 6-week-old stems of Ws and the gcn5-2 mutant. ACTIN8 was used as an endogenous control. Data are means (±SD) from at least three biological replicates. (B) Scanning electron microscopy of the stems of Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10). Scale bars: 20 μm. (C) Total cuticular wax content was calculated over the unit area of 6-week-old stems in plants of Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10). The mean expression levels and total wax load were compared using one-way ANOVA together with a Bonferroni adjustment test in R. Different letters indicate significant differences among genotypes (P<0.05). Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2920 | Wang et al. Fig. 6. Cuticular wax composition of the stems of wild-type Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10) from 6-week-old plants. The chemical classes and main chain-lengths of each constituent are indicated. Data are means (±SD) of three biological replicates. (A) Comparison between Ws and 35S::CER3/Ws (#2, #5, #10) and (B) among Ws, gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10) transgenic plants. The mean values were compared using one-way ANOVA together with a Bonferroni adjustment test in R. Different letters indicate significant differences among genotypes (P<0.05). (This figure is available in colour at JXB online.) including the lipid biosynthetic process, neutral lipid biosyn- Rowland et al., 2007). The wax components of the Arabidopsis thetic process, glycerolipid biosynthetic process, and unsaturated cer3 mutant are lacking in aldehydes, alkanes, secondary alco- fatty acid biosynthetic process. Thirdly, ChIP assays demonstrated hols, and ketones in the stem compared with the wild-type that CER1-L1, CER26, and CER3, which encode proteins (Rowland et al., 2007). Interestingly, our GC-FID and GC-MS involved in VLCFA production and alkane-forming pathways analyses found significant reductions in total wax in the gcn5- of wax synthesis, are target genes of GCN5. Finally, enrich- 2 mutant stem, especially for C30 aldehyde, C29 alkane, C29 ment of H3K9ac and H3K14ac at the promoters of CER1-L1, ketone, and C29 secondary alcohols. Moreover, overexpression CER26, and CER3 was significantly decreased in the gcn5-2 of CER3 in the gcn5-2 background significantly increased the mutant compared with the wild-type. Collectively, our previous levels of C29 ketone, C30 aldehydes, C29 alkanes, and C29 sec- data (Wang et  al., 2016) and that from the present study have ondary alcohols, indicating that CER3 played a pivotal role in demonstrated that histone acetyltransferase GCN5 is involved GCN5-regulated biosynthesis of stem cuticular wax. However, in multiple lipid metabolic processes, from upstream de novo FA we cannot rule out the possibility that other unknown genes synthesis to subsequent wax production. may also contribute, which may prove an interesting area for To determine the underlying mechanisms of GCN5- further investigation. For example, the ester component, which regulated biosynthesis of stem cuticular wax, three GCN5 cannot be catalysed by CER3, also changed significantly in target genes, CER3, CER26, and CER1-L1 were identified. the gcn5-2 mutant stem. The gcn5-2 mutant deficiency of C30 CER26 is involved in the elongation of VLCFAs (from 30 primary alcohols was not rescued by overexpression of CER3. C to 32 C) and has high specificity of tissue and substrate In addition, the expression of FAR3, an important enzyme in (Haslam et  al., 2015; Pascal et  al., 2013). Although little is primary alcohol formation, was not regulated by GCN5. known about CER1-L1 other than that it is a homolog of Many reports have demonstrated that acetylation of histone CER1 (Bernard et  al., 2012), there is the possibility that, like tails induces the accessibility of transcription factors to the CER1, it can physically interact with CER3 during the very- nucleosomal DNA, which subsequently influences the gene long-chain alkane biosynthesis process and contribute to the expression (e.g. Lee et  al., 1993). As a key wax biosynthetic total wax load. Previous studies have reported that CER3, enzyme gene, CER3 is regulated by both transcription factors which catalyses redox-dependent alkane formation, is the key and epigenetic modulators (Lee and Suh, 2015). CER3 is posi- wax biosynthetic enzyme (Aarts et al., 1995; Chen et al., 2003; tively regulated by two MYB transcription factors, MYB96 Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2921 and MYB30, in response to drought and pathogen attack, Supplementary data respectively (Lee and Suh, 2013; Lee et  al., 2016b). Based on Supplementary data are available at JXB online. our RNA-seq data (see Supplementary Table S2), the expres- Table S1. Summary of the RNA-seq data and read mapping. sion of MYB96 was up-regulated in the gcn5-2 mutant, sug- Table S2. Differently expressed genes in the RNA-seq data. gesting that it does not contribute much to the changed Table  S3. GO analysis of the 2616 down-regulated genes expression of CER3 in the stem of this mutant. However, we in the stems of the gcn5-2 mutant compared with the cannot rule out the possibility that CER3 could be regulated wild-type Ws. by other unknown transcription factors that are deregulated Table S4. Gene-specific primer pairs in this study. in the gcn5-2 mutant. Recent studies showed that epigenetic Fig. S1. Stem cuticular wax phenotype of the histone-acet- modulators, namely the histone methyl transferases SDG8 ylation mutants. (ASHH2) and SDG25 (ATXR7), were involved in wax accu- Fig. S2. Stem cuticular wax phenotype of the histone-meth- mulation through histone lysine methylation and/or indirectly ylation mutants. through H2B ubiquitination by targeting CER3, and this was Fig.  S3. GCN5 expression levels and phenotypes of the associated with diminished accumulation of lipids (Lee et  al., independent homologous 35S::GCN5 transgenic lines. 2016a). Here, our observations also showed that the ashh2 Fig.  S4. Reproducibility of the RNA-seq biological mutant exhibited mildly reduced wax crystal accumulation replicates. compared with Col-0. It might be of interest to further exam- Fig. S5. Expression levels of 10 genes randomly selected to ine the relationships between GCN5, SDG8 (ASHH2), and validate the accuracy of the RNA-seq data using qRT-PCR. SDG25 (ATXR7) in cuticular wax biosynthesis. Fig. S6. FAR3 expression levels in wild-type Ws, gcn5-2, and In conclusion, our results have demonstrated that, like his- 35S::CER3 transgenic lines. tone ubiquitination and methylation, histone acetylation is also involved in the regulation of biosynthesis of stem cuticular wax, and we propose a working model to explain this process in Acknowledgements Arabidopsis (Fig. 7). Briefly, the histone acetyltransferase GCN5 The authors thank Prof. Jianxin Shi (Shanghai Jiao Tong University) for regulates the biosynthesis of stem cuticular wax by regulating the the GC-FID and GC-MS analyses in his laboratory. This work was sup- expression of CER3, CER1-L1, and CER26 via histone acety- ported by the Ministry of Agriculture of China for Transgenic Research lation at the H3K9/14 sites. Thus, interruption of GCN5 dra- (2016ZX08009002), the State Key laboratory of Agrobiotechnology matically reduces the total amount of cuticular wax and changes Open Grant (2018SKLAB6-25) and the Fundamental Research Funds for the Central Universities (2412017QD016). its composition, especially with regards to alkanes, aldehydes, and ketone, which are mainly synthesized in the alkane-forming pro- cess. Remarkably, overexpression of CER3 in the gcn5-2 mutant References could rescue the cuticular wax deficiency, suggesting that it has Aarts MG, Keijzer CJ, Stiekema WJ, Pereira A. 1995. Molecular an important role in GCN5-mediated cuticular wax biosynthe- characterization of the CER1 gene of Arabidopsis involved in epicuticular sis. Our findings provide an insight into the epigenetic regula- wax biosynthesis and pollen fertility. The Plant Cell 7, 2115–2127. tion of cuticular wax development through histone acetylation, Aharoni A, Dixit S, Jetter R, Thoenes E, van Arkel G, Pereira A. which may contribute to wax-related stress responses in plants. 2004. The SHINE clade of AP2 domain transcription factors activates wax biosynthesis, alters cuticle properties, and confers drought tolerance when overexpressed in Arabidopsis. The Plant Cell 16, 2463–2480. Benhamed M, Bertrand C, Servet C, Zhou DX. 2006. Arabidopsis GCN5, HD1, and TAF1/HAF2 interact to regulate histone acetylation required for light-responsive gene expression. The Plant Cell 18, 2893–2903. Benhamed M, Martin-Magniette ML, Taconnat L, et  al. 2008. Genome-scale Arabidopsis promoter array identifies targets of the histone acetyltransferase GCN5. The Plant Journal 56, 493–504. Bernard A, Domergue F, Pascal S, Jetter R, Renne C, Faure JD, Haslam RP, Napier JA, Lessire R, Joubès J. 2012. Reconstitution of plant alkane biosynthesis in yeast demonstrates that Arabidopsis ECERIFERUM1 and ECERIFERUM3 are core components of a very-long- chain alkane synthesis complex. The Plant Cell 24, 3106–3118. Bernard A, Joubès J. 2013. Arabidopsis cuticular waxes: advances in Fig. 7. A model for the regulation of stem cuticular wax synthesis by synthesis, export and regulation. Progress in Lipid Research 52, 110–129. GCN5-associated acetylation in Arabidopsis. As the wax precursors, Bertrand C, Bergounioux C, Domenichini S, Delarue M, Zhou DX. the very-long-chain acyl-CoAs (VLC-CoA) can be processed through 2003. Arabidopsis histone acetyltransferase AtGCN5 regulates the floral the alcohol-forming pathway and alkane-forming pathway, which yield meristem activity through the WUSCHEL/AGAMOUS pathway. The Journal 17~18% and 80% of the total cuticular wax mixture, respectively. Solid of Biological Chemistry 278, 28246–28251. black arrows represent the cuticular wax biosynthesis pathway. The Bhat RA, Riehl M, Santandrea G, Velasco R, Slocombe S, Donn arrows from GCN5 indicate positive transcriptional regulation by GCN5 G, Steinbiss HH, Thompson RD, Becker HA. 2003. Alteration of via H3K9/14ac modifications. GCN5 targets are marked at the positions GCN5 levels in maize reveals dynamic responses to manipulating histone acetylation. The Plant Journal 33, 455–469. where the enzymes they encode are required. CER3 is a key cuticular wax biosynthetic enzyme that catalyses the alkane-forming pathway in Bradford MM. 1976. A rapid and sensitive method for the quantitation of Arabidopsis stems (solid arrow), whereas CER26 and CER1-L1 might microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248–254. partially contribute to the total wax load but they were not verified functionally in this study (dashed arrows). (This figure is available in colour Broun P, Poindexter P, Osborne E, Jiang CZ, Riechmann JL. 2004. at JXB online.) WIN1, a transcriptional activator of epidermal wax accumulation in Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2922 | Wang et al. Arabidopsis. Proceedings of the National Academy of Sciences, USA 101, Li-Beisson Y, Shorrosh B, Beisson F, et al. 2013. Acyl-lipid metabolism. 4706–4711. The Arabidopsis Book 11, e0161. Chen W, Yu XH, Zhang K, Shi J, De Oliveira S, Schreiber L, Shanklin J, Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression –ΔΔ t Zhang D. 2011. Male Sterile2 encodes a plastid-localized fatty acyl carrier data using real-time quantitative PCR and the 2 C method. Methods 25, protein reductase required for pollen exine development in Arabidopsis. 402–408. Plant Physiology 157, 842–853. Ménard R, Verdier G, Ors M, Erhardt M, Beisson F, Shen WH. 2014. Chen X, Goodwin SM, Boroff VL, Liu X, Jenks MA. 2003. Cloning Histone H2B monoubiquitination is involved in the regulation of cutin and and characterization of the WAX2 gene of Arabidopsis involved in cuticle wax composition in Arabidopsis thaliana. Plant & Cell Physiology 55, membrane and wax production. The Plant Cell 15, 1170–1185. 455–466. Clough SJ, Bent AF. 1998. Floral dip: a simplified method for Pandey R, Müller A, Napoli CA, Selinger Da, Pickaard CS, Richards Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant EJ, Bender J, Mount DW, Jorgensen RA. 2002. Analysis of histone Journal 16, 735–743. acetyltransferase and histone deacetylase families of Arabidopsis thaliana suggests functional diversification of chromatin modification among Earley KW, Shook MS, Brower-Toland B, Hicks L, Pikaard CS. 2007. multicellular eukaryotes. Nucleic Acids Research 30, 5036–5055. In vitro specificities of Arabidopsis co-activator histone acetyltransferases: implications for histone hyperacetylation in gene activation. The Plant Pascal S, Bernard A, Sorel M, Pervent M, Vile D, Haslam RP, Napier Journal 52, 615–626. JA, Lessire R, Domergue F, Joubès J. 2013. The Arabidopsis cer26 mutant, like the cer2 mutant, is specifically affected in the very long chain Fiil BK, Qiu JL, Petersen K, Petersen M, Mundy J. 2008. fatty acid elongation process. The Plant Journal 73, 733–746. Coimmunoprecipitation (co-IP) of nuclear proteins and chromatin immunoprecipitation (ChIP) from Arabidopsis. CSH Protocols 2008, pdb. Pfluger J, Wagner D. 2007. Histone modifications and dynamic regulation prot5049. of genome accessibility in plants. Current Opinion in Plant Biology 10, 645–652. Haslam TM, Haslam R, Thoraval D, et  al. 2015. ECERIFERUM2-LIKE proteins have unique biochemical and physiological functions in very-long- Raffaele S, Vailleau F, Léger A, Joubès J, Miersch O, Huard C, Blée chain fatty acid elongation. Plant Physiology 167, 682–692. E, Mongrand S, Domergue F, Roby D. 2008. A MYB transcription factor regulates very-long-chain fatty acid biosynthesis for activation of Haslam TM, Mañas-Fernández A, Zhao L, Kunst L. 2012. Arabidopsis the hypersensitive cell death response in Arabidopsis. The Plant Cell 20, ECERIFERUM2 is a component of the fatty acid elongation machinery 752–767. required for fatty acid extension to exceptional lengths. Plant Physiology 160, 1164–1174. Robinson MD, McCarthy DJ, Smyth GK. 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Hu Z, Song N, Zheng M, et al. 2015. Histone acetyltransferase GCN5 is Bioinformatics 26, 139–140. essential for heat stress-responsive gene activation and thermotolerance in Arabidopsis. The Plant Journal 84, 1178–1191. Rowland O, Lee R, Franke R, Schreiber L, Kunst L. 2007. The CER3 wax biosynthetic gene from Arabidopsis thaliana is allelic to WAX2/YRE/ Jefferson RA, Bevan M, Kavanagh T. 1987. The use of the Escherichia FLP1. FEBS Letters 581, 3538–3544. coli beta-glucuronidase as a gene fusion marker for studies of gene expression in higher plants. Biochemical Society Transactions 15, 17–18. Rowland O, Zheng H, Hepworth SR, Lam P, Jetter R, Kunst L. 2006. CER4 encodes an alcohol-forming fatty acyl-coenzyme A  reductase Kim D, Salzberg SL. 2011. TopHat-Fusion: an algorithm for discovery of involved in cuticular wax production in Arabidopsis. Plant Physiology 142, novel fusion transcripts. Genome Biology 12, R72. 866–877. Koornneef M, Hanhart CJ, Thiel F. 1989. A genetic and phenotypic Samuels L, Kunst L, Jetter R. 2008. Sealing plant surfaces: cuticular wax description of eceriferum mutants in Arabidopsis thaliana. The Journal of formation by epidermal cells. Annual Review of Plant Biology 59, 683–707. Heredity 80, 118–122. Servet C, Conde e Silva N, Zhou DX. 2010. Histone acetyltransferase Kornet N, Scheres B. 2009. Members of the GCN5 histone AtGCN5/HAG1 is a versatile regulator of developmental and inducible gene acetyltransferase complex regulate PLETHORA-mediated root stem cell expression in Arabidopsis. Molecular Plant 3, 670–677. niche maintenance and transit amplifying cell proliferation in Arabidopsis. The Plant Cell 21, 1070–1079. Stockinger EJ, Mao Y, Regier MK, Triezenberg SJ, Thomashow MF. 2001. Transcriptional adaptor and histone acetyltransferase proteins Lai C, Kunst L, Jetter R. 2007. Composition of alkyl esters in the cuticular in Arabidopsis and their interactions with CBF1, a transcriptional activator wax on inflorescence stems of Arabidopsis thaliana cer mutants. The Plant involved in cold-regulated gene expression. Nucleic Acids Research 29, Journal 50, 189–196. 1524–1533. Laux T, Mayer KF, Berger J, Jürgens G. 1996. The WUSCHEL gene is Suh MC, Samuels AL, Jetter R, Kunst L, Pollard M, Ohlrogge J, required for shoot and floral meristem integrity in Arabidopsis. Development Beisson F. 2005. Cuticular lipid composition, surface structure, and 122, 87–96. gene expression in Arabidopsis stem epidermis. Plant Physiology 139, Lee DY, Hayes JJ, Pruss D, Wolffe AP. 1993. A positive role for histone 1649–1665. acetylation in transcription factor access to nucleosomal DNA. Cell 72, Tian T, Liu Y, Yan H, You Q, Yi X, Du Z, Xu W, Su Z. 2017. agriGO v2.0: 73–84. a GO analysis toolkit for the agricultural community, 2017 update. Nucleic Lee S, Fu F, Xu S, Lee SY, Yun DJ, Mengiste T. 2016a. Global regulation Acids Research 45, W122–W129. of plant immunity by histone lysine methyl transferases. The Plant Cell 28, Vlachonasios KE, Thomashow MF, Triezenberg SJ. 2003. Disruption 1640–1661. mutations of ADA2b and GCN5 transcriptional adaptor genes dramatically Lee SB, Kim HU, Suh MC. 2016b. MYB94 and MYB96 additively activate affect Arabidopsis growth, development, and gene expression. The Plant cuticular wax biosynthesis in Arabidopsis. Plant & Cell Physiology 57, Cell 15, 626–638. 2300–2311. Wang T, Xing J, Liu X, et  al. 2016. Histone acetyltransferase general Lee SB, Suh MC. 2013. Recent advances in cuticular wax biosynthesis control non-repressed protein 5 (GCN5) affects the fatty acid composition and its regulation in Arabidopsis. Molecular Plant 6, 246–249. of Arabidopsis thaliana seeds by acetylating fatty acid desaturase3 (FAD3). Lee SB, Suh MC. 2015. Advances in the understanding of cuticular waxes The Plant Journal 88, 794–808. in Arabidopsis thaliana and crop species. Plant Cell Reports 34, 557–572. Wang Y, Wang M, Sun Y, Hegebarth D, Li T, Jetter R, Wang Z. Li F, Wu X, Lam P, Bird D, Zheng H, Samuels L, Jetter R, Kunst L. 2008. 2015. Molecular characterization of TaFAR1 involved in primary alcohol Identification of the wax ester synthase/acyl-coenzyme A: diacylglycerol biosynthesis of cuticular wax in hexaploid wheat. Plant & Cell Physiology acyltransferase WSD1 required for stem wax ester biosynthesis in 56, 1944–1961. Arabidopsis. Plant Physiology 148, 97–107. Xing J, Wang T, Liu Z, et  al. 2015. GENERAL CONTROL Li S, Wang X, He S, et  al. 2016. CFLAP1 and CFLAP2 are two bHLH NONREPRESSED PROTEIN5-mediated histone acetylation of FERRIC transcription factors participating in synergistic regulation of AtCFL1- REDUCTASE DEFECTIVE3 contributes to iron homeostasis in Arabidopsis. mediated cuticle development in Arabidopsis. PLoS Genetics 12, 1–27. Plant Physiology 168, 1309–1320. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Experimental Botany Oxford University Press

GCN5 contributes to stem cuticular wax biosynthesis by histone acetylation of CER3 in Arabidopsis

Loading next page...
 
/lp/ou_press/gcn5-contributes-to-stem-cuticular-wax-biosynthesis-by-histone-lEk0kAtIXQ

References (57)

Publisher
Oxford University Press
Copyright
Copyright © 2022 Society for Experimental Biology
ISSN
0022-0957
eISSN
1460-2431
DOI
10.1093/jxb/ery077
Publisher site
See Article on Publisher Site

Abstract

Cuticular wax is a major component of the surface cuticle of plants, which performs crucial functions in optimizing plant growth. Histone acetylation regulates gene expression in diverse biological processes, but its role in cuticular wax synthesis is not well understood. In this study, we observed that mutations of the Arabidopsis thaliana histone acetyltransferase GENERAL CONTROL NON-REPRESSED PROTEIN5 (GCN5) impaired the accumulation of stem cuticular wax. Three target genes of GCN5, ECERIFERUM3 (CER3), CER26, and CER1-LIKE1 (CER1-L1), were identi- fied by RNA-seq and ChIP assays. H3K9/14 acetylation levels at the promoter regions of CER3, CER26, and CER1-L1 were consistently and significantly decreased in the gcn5-2 mutant as compared to the wild-type. Notably, overex- pression of CER3 in the gcn5-2 mutant rescued the defect in stem cuticular wax biosynthesis. Collectively, these data demonstrate that GCN5 is involved in stem cuticular wax accumulation by modulating CER3 expression via H3K9/14 acetylation, which underlines the important role of histone acetylation in cuticular wax biosynthesis. Keywords: Arabidopsis thaliana, CER3, GCN5, histone acetylation, stem, wax biosynthesis. Introduction Plant cuticular wax is a complex mixture of very-long-chain termed the alcohol-forming and the alkane-forming pathways, fatty acids (VLCFAs) and aldehydes, alcohols, alkanes, ketones, which yield 17~18% and 80% of the total amount of wax, and esters, with predominant carbon chain-lengths ranging respectively (Bernard and Joubès, 2013). As a crucial adap- from C22 to C36 (Samuels et al., 2008; Li et al., 2016), and it tive characteristic, cuticular wax protects plants against biotic forms one of the major lipid components of the cuticle that and abiotic stresses, such as pathogen attacks and water loss covers the outer surface of aerial plant tissues. The biosynthesis (Aharoni et  al., 2004; Samuels et  al., 2008). Therefore, eluci- of cuticular wax is processed through two distinct pathways, dating the regulatory mechanisms controlling cuticular wax Abbreviations: GCN5, general control non-repressed protein5; CER3, eceriferum3; VLCFA, very-long-chain fatty acid; HAT, histone acetyltransferase; HDAC, histone deacetylase; GO, gene ontology; GUS, β-glucuronidase; GC-FID, gas chromatography–flame ionization detector. © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Experimental Biology. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2912 | Wang et al. accumulation is of great interest for genetic engineering of However, to the best of our knowledge, histone-acetylating agricultural crops. events have not been reported in cuticular wax biosynthesis. To date, numerous candidate genes in the cuticular The reversible modulations of histone acetylation and de- wax pathway have been identified, most of which encode acetylation are catalysed by histone acetyltransferases (HATs) enzymes, or work with enzymes, in the VLCFA biosynthesis and histone de-acetylases (HDACs), respectively (Pandey and derivatization pathways (Koornneef et al., 1989; Bernard et  al., 2002). As in yeast and mammals, the Arabidopsis and Joubès, 2013). For example, CER2, and its homologues HATs are grouped into four classes: GNAT (GCN5-related CER2-LIKE1/CER26 and CER2-LIKE2 have biological N-acetyltransferase), MYST (for ‘MOZ, Ybf2/Sas2 and functions in two-carbon elongation of VLCFAs, from C28 Tip60’), p300/CBP (p300/CREB-binding protein), and TAF1 to C30, C30 to C32, and C32 to C34, respectively (Haslam [for ‘TATA-binding protein (TBP)-associated factor’] (Servet et  al., 2012, 2015; Pascal et  al., 2013; Lee and Suh, 2015). et al., 2010). GCN5, a GNAT-type HAT that harbors a HAT A  member of the bifunctional wax ester synthase/diacylg- domain and a bromodomain, has been well studied in exten- lycerol acyltransferase family, WSD1, is a key enzyme in wax sive research (Benhamed et  al., 2008; Servet et  al., 2010). It ester synthesis of Arabidopsis stems (Li et  al., 2008). As for has been observed that mutations of GCN5 result in various the essential alkane-forming pathway, CER3/WAX2 plays an growth impairments, such as dwarfism, defects in terminal important role in synthesis of major wax components (Aarts flower production, deformed seed development, and poor fer- et al., 1995; Chen et al., 2003; Rowland et al., 2007; Lee and tility (Bertrand et al., 2003; Vlachonasios et al., 2003). Moreover, Suh, 2015). The amount of wax found in the cer3 mutant was GCN5 regulates a variety of biological processes in Arabidopsis, severely reduced compared with the wild-type in Arabidopsis including cell differentiation, shoot and floral meristem forma- stems, especially with regards to aldehydes, alkanes, second- tion, light and abiotic stress (e.g. heat and cold) responses, iron ary alcohols, and ketones (Rowland et  al., 2007). Although homeostasis, and fatty acid biosynthesis, by catalysing histone the exact reactions catalysed by the CER3 enzyme remain acetylation levels of target promoters at certain sites, including unknown, it has been reported that CER3 may physically H3K14, H3K9, and H3K27 (Laux et al., 1996; Stockinger et al., interact with CER1 and CYTOCHROME B5 ISOFORM 2001; Bertrand et al., 2003; Vlachonasios et al., 2003; Benhamed (CYTB5) for the biosynthesis of very-long-chain alkanes et  al., 2006; Earley et  al., 2007; Hu et  al., 2015; Xing et  al., (Bernard et al., 2012). 2015; Wang et  al., 2016). In this current study, we found that An increasing focus on the transcriptional regulation of the the gcn5 mutants had significantly glossy (wax-deficient) stems genes in cuticular wax biosynthesis has led to some transcrip- compared with the wild-type. Analyses of chemical compo- tion factors being reported in recent studies (Lee and Suh, nents coupled with scanning electron microscopy showed an 2015). In Arabidopsis, WAX INDUCER1/SHINE1 (WIN1/ obvious reduction in amounts of total wax and changes in SHN1), an AP2-EREBP-type transcription factor, was the first their composition. Moreover, we found that GCN5 bound to to be identified and is a representative regulator of wax biosyn- the promoter of CER3 and this interaction was impaired in thesis, regulating the CER1, CER2, and 3-KETOACYL-COA the gcn5-2 mutant. Taken together, we conclude that GCN5- SYNTHASE1 (KCS1) genes (Aharoni et al., 2004; Broun et al., mediated histone acetylation of CER3 regulates stem cuticular 2004). WIN1 overexpression lines exhibited enhanced drought wax biosynthesis in Arabidopsis. tolerance compared with the wild-type (Aharoni et al., 2004). In addition, MYB transcription factors are important for wax Materials and methods biosynthesis under both biotic and abiotic stresses (Suh et al., 2005; Lee and Suh, 2015). For example, MYB96 controls wax Plant material and growth conditions biosynthesis by regulating the KCS1, KCS2, KCS6, BETA- The Arabidopsis thaliana wild-types Col-0 and Ws, together with T-DNA KETOACYL REDUCTASE1 (KCR1), and CER3 genes insertion mutants involved in histone modifications were used in this under drought stress (Lee et  al., 2016b), and KCS1, KCR1, study, as follows. (1) Histone acetylation: gcn5-1 and gcn5-2 (Ws back- ground), hda2, hda2c, hda5, hda7, hda9, hda13, hda18, hda19, and srt2 CER2, and CER3 are the targets of MYB30 in response to (Col-0 background); and (2) histone methylation: ashh1, ashh2, ashh3, pathogen attack (Raffaele et al., 2008). ashh4, ashr2, ashr3, atx1, atx2, atx4, atx5, atxr2, atxr3, and atxr4 (Col-0 With our increasing understanding of epigenetic mecha- background). The gcn5-1 and gcn5-2 mutants were both T-DNA inser- nisms, recent reports have demonstrated that several epige- tion mutants in the bromodomain-coding region (Bertrand et al., 2003; netic events are involved in wax biosynthesis (Lee and Suh, Vlachonasios et  al., 2003; Supplementary Fig.  S3A at JXB online). Notably, the gcn5-2 mutation removes the entire bromodomain, which is 2013, 2015). Recently, two RING E3 ligases, HISTONE required for binding to 11% of the GCN5 promoter targets (Servet et al., MONOUBIQUITINATION1 (HUB1) and HUB2, were 2010). The other mutants were obtained in the homozygous state from demonstrated to be involved in wax biosynthesis by mon- ABRC (https://abrc.osu.edu/) or from individual donors. For germina- oubiquitinating histone H2B proteins, which in turn activates tion, sterilized seeds were incubated at 4  °C for 3 d, and subsequently the transcriptional levels of the wax biosynthetic genes LONG- sown on Murashige and Skoog (MS) plates containing 1% sucrose and 0.6% agar. The seedlings were grown under 16/8 h light/dark conditions CHAIN ACYL-COA SYNTHETASE2 (LACS2) and CER1 at 22 °C in a growth room. (Ménard et  al., 2014). The Arabidopsis histone methyl trans- ferases SET DOMAIN GROUP8 (SDG8) and SDG25 have been reported to contribute to wax accumulation through his- RNA isolation and RNA-seq tone lysine methylation and/or H2B ubiquitination by target- Total RNA was extracted using TRIzol reagent (Invitrogen), according ing the key wax biosynthetic gene CER3 (Lee et  al., 2016a). to the manufacturer’s instructions. RNA concentrations were measured Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2913 using a NanoDrop 2000 spectrophotometer (ND-2000, ThermoFisher described by Fiil et  al. (2008). Six-week-old stems of Ws and gcn5-2 Scientific, Inc., MA, USA). RNA integrity was assessed using an Agilent mutants were harvested and fixed in 1% formaldehyde for 15 min in a 2100 Bioanalyser (Agilent Technologies, Inc., CA, USA). Paired-end vacuum and subsequently neutralized using 0.125 M glycine for 5 min. sequencing libraries with an average insert size of 200 bp were prepared After washing with sterilized water, the samples were dried with towels, using the TruSeq RNA Sample Preparation Kit v2 (Illumina, San Diego, and ground in liquid nitrogen. The resulting powders were resuspended in USA) and sequenced using a HiSeq2500 platform (Illumina, San Diego, the Nuclei Extraction Buffer 1, which contained 0.4 M sucrose, 10 mM USA) according to the manufacturers’ standard protocols. Raw data Tris-HCl, pH 8.0, 10 mM MgCl , 5 mM β-mercaptoethanol, 0.1 mM obtained from Illumina sequencing were processed and filtered using PMSF (Sigma, P7626), and protease inhibitors (Roche, 11873580001), the Illumina pipeline (http://www.illumina.com) to generate FastQ and mixed immediately. After incubation for 20 min at 4 °C with a rota- files. Approximately 12 G of high-quality 125-bp paired-end reads were tor, the solutions were filtered through four layers of Miracloth into generated from six libraries (Supplementary Table S1). The FastQC pro- new tubes, and the filtrate was then centrifuged for 20 min at 3000 g at gram (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was 4  °C. The nuclei pellets were resuspended in Nuclei Extraction Buffer used to evaluate the overall quality of the RNA-seq reads. Poor-quality 2, which contained 0.25 M sucrose, 10  mM Tris-HCl, pH8.0, 10  mM bases were filtered out using Sickle (https://github.com/najoshi/sickle). MgCl , 1% Triton X-100, 5  mM β-mercaptoethanol, 0.1  mM PMSF, High-quality RNA-seq reads from each library were mapped to The and protease inhibitors. The suspensions were transferred to microfuge Arabidopsis Information Resource (TAIR10) version of the Arabidopsis tubes and centrifuged at 12 000 g for 10 min at 4 °C. The pellets were genome using the splice-junction-aware short-read alignment suite resuspended in Nuclei Extraction Buffer 2 and centrifuged at 12 000 g TOPHAT v2.09 with default settings (Kim and Salzberg, 2011). The through a layer of Nuclei Extraction Buffer 3, which contained 1.7 M reads displaying unique alignment and not more than two nucleotide sucrose, 10 mM Tris-HCl, pH 8.0, 2 mM MgCl , 0.15% Triton X-100, mismatches were kept for further analysis. The differentially expressed 5  mM β-mercaptoethanol, 0.1  mM PMSF, and protease inhibitors, in genes were identified by using the edgeR package (ver. 3.2.3) with an microfuge tubes for 60  min at 4  °C. The nuclear pellets were lysed in absolute value of log -fold change ≥2 and a false-discovery rate <0.05 Nuclei Lysis Buffer, which contained 50 mM Tris-HCl, pH8.0, 10 mM as cut-off (Robinson et al., 2010). The groups of differentially expressed EDTA, 1% SDS, and protease inhibitors. The lysed nuclei were sonicated genes identified by RNA-seq in this study are shown in Supplementary four times with a Bioruptor (UCD-200) in a water bath at 4 °C, with Table  S2. Gene Ontology (GO) analysis was performed using agriGO each sonication period consisting of 15 s on and 15 s off for 5 min and v2.0 with a cut-off of P-value <0.05 (Tian et  al., 2017), and the total followed by centrifugation. The clear supernatants, which contained the enrichment categories are identified in Supplementary Table S3. sonicated chromatin, were transferred to new tubes. Immunoprecipitation (IP) was performed using 5 μl of chromatin, and the following IP steps were conducted using the Magna ChIP™ HiSens Kit. Aliquots of the Quantitative real-time PCR dilution were used for the IP assays. The anti-GCN5 antibody was gen- erated using two synthetic peptides (H2N-CARGADTDSDPDESED Real-time PCR was performed as previously described (Livak and and H2N-SSRNTKLKTESSTVKLC); both peptide epitopes are located Schmittgen, 2001) and ACTIN8 was used as the control gene, i.e. the between amino acids 85 to 99 and between amino acids 136 to 150, expression levels of each gene were normalized to that of ACTIN8. which is the N-terminal region of the protein, provided by Prof. D-X The primer pairs used for real-time PCR are listed in Supplementary Zhou, and the specificity of this GCN5 antibody was confirmed by Table S4. The PCR analysis was performed using a CFX96 System (Bio- protein gel blots (Benhamed et al., 2006, 2008). The anti-H3K14ac and Rad) with SYBR Green. The following program was used for the real- anti-H3K9ac antibodies were purchased from Upstate Biotechnology. time PCR: 95 °C for 3 min and 40 cycles of 95 °C for 30 s, 58 °C for CHALCONE SYNTHASE (CHS) and AT4G03800 (gypsy-like retro- 30 s, and 72 °C for 30 s. transposon family gene) were amplified as endogenous controls for the anti-GCN5 and anti-H3K9/H3K14 antibodies, respectively. The immu- Plasmid construction and plant transformation noprecipitated DNA was analysed by quantitative PCR in three bio- logical replicates using the primer sets listed in Supplementary Table S4. A DNA fragment containing a 2.0-kb fragment upstream of the CER3 Amplified DNA from the chromatin fractions prior to antibody incuba- coding sequence and full-length ORFs of GCN5 and CER3 were tion were used as the controls (inputs). The fold-enrichment was normal- amplified by PCR-directed cloning based on the annotation from TAIR ized to the chromatin inputs. using the following primer pairs: CER3-P-F and CER3-P-R, GCN5-F and GCN5-R, CER3-F and CER3-R, respectively (Supplementary Table  S4). The sequence-confirmed clones containing the ORFs of GUS histochemical and fluorometric assays GCN5 and CER3 were then respectively cloned into the binary expres- sion vector pCAMBIA1300 (driven by the CaMV35S promoter). The Three homologous transgenic T3 lines of ProCER3::GUS/Ws and the promoter region of CER3 was fused to the reporter gene encoding corresponding homologous ProCER3::GUS/gcn5-2 lines were used for β-glucuronidase (GUS). The chimeric gene was then cloned into the GUS histochemical analysis. The seedlings were grown under strictly iden- binary expression vector pCAMBIA1300 to generate ProCER3::GUS. tical conditions. After pollination, Arabidopsis stems, siliques, flowers, and These vectors were transferred into the Agrobacterium tumefaciens strain young leaves were vacuum-infiltrated with staining buffer (2 mM potas- GV3101. Transgenic plants were generated using the floral dip method sium ferricyanide, 10 mM phosphate buffer, 0.5% Triton X-100, and 1 mg –1 −1 and subsequently screened on solid plates containing 25 mg l hygromy- ml X-Gluc) and then incubated overnight. The tissue was then incubated cin (Clough and Bent, 1998). The hygromycin-resistant seedlings were in ethanol and acetic acid (1:1) for 4–8 h and cleared in 80% ethanol. The then transferred to a mixture of soil and vermiculite (2:1). At least three samples were observed with a stereomicroscope (Olympus SEX16). independent T3 homozygous lines with a single T-DNA insertion were For the quantification of GUS activity, we used the fluorometric assay subjected to a detailed analysis. Because the gcn5-2 mutant exhibits low- based on the method of Jefferson et al. (1987). Total protein extracts from fertility pollen that hinders the direct acquisition of transgenic plants in stems of three independent lines for each construct (ProCER3::GUS/ the gcn5-2 background (Bertrand et al., 2003; Xing et al., 2015), we ini- Ws and ProCER3::GUS/gcn5-2) were determined using bovine serum tially generated transgenic plants in the Ws background, and three inde- albumin (BSA) as a standard according to the Bradford assay (Bradford, pendent transgenic T3 lines were selected for crossing into the gcn5-2 1976). Fluorescence was measured using 4-methylumbelliferone (4-MU) mutant. The homologous transgenic lines in the gcn5-2 background were as a substrate, with an excitation wavelength of 365 nm and an emission selected using the same method described above. wavelength of 455 nm in a BioTek Synergy HT Multi-Mode Microplate Reader (BioTek, Vermont, USA). GUS activities of the extracts were cal- culated as nanomole 4-MU per minute per milligram protein. ChIP assay analysis Both GUS histochemical and fluorometric assays were conducted at ChIP assays were performed using the Magna ChIP™ HiSens Kit least three times, and only the transgenic lines with stable GUS signals (Catalogue No. 17-10460) combined with the ChIP method as previously throughout different generations were selected for further analysis. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2914 | Wang et al. Scanning electron microscopy ashr2, and hda9 showed more cuticular wax crystals than the Six-week-old stems were attached to double-sided carbon sealing tape. wild-type (Col-0), whereas less abundant wax was observed The specimens were examined under a Hitachi TM3000 Tabletop in atx5, ashh2, and hda18. Remarkably, fewer stem wax crys- Scanning Electron Microscope at 15 kV, and the digital recordings were tals were observed on the surface of the gcn5-2 mutant rela- saved as TIFF files. tive to the wild-type (Ws), which led to the appearance of a glossy stem (Fig.  1A, B). The compositions of the cuticular Cuticular wax analysis wax of the gcn5-2 mutant and wild-type were quantified by Cuticular wax analysis was conducted as previously described by Chen GC-FID and GC-MS analysis. The total cuticular wax content et al. (2011). Because the environmental conditions could affect configu- in the gcn5-2 mutant was approximately 63% of the control ration and distribution of the surface wax structures, all the plants we used (Fig.  1C), which was attributable to notable decreases in the in this study were cultivated in strictly controlled temperature and humid- major wax constituents in the mutant stem, including alkanes ity. Six-week-old stems were collected and pictures were taken (with (C29), ketone (C29), primary alcohols (C26, C28, and C30), an adjacent ruler) for later determination of area using ImageJ (https:// imagej.nih.gov/ij/). Cuticular waxes were extracted by immersing the secondary alcohols (C29), aldehydes (C28 and C30), and esters samples for 30 s in 1 ml of chloroform containing 10 μg of tetracosane (C42 and C44; Fig. 1D and Table 1). (Fluka) as an internal standard. Three biological replicates per genotype To confirm that the observed defects in stem wax were were performed. Five individual stems were used for each replicate. The indeed caused by the disruption of GCN5, another well- extracts were transferred to reactive vials, dried under nitrogen gas, and established T-DNA insertion mutant allele of GCN5 (gcn5- derivatized by adding 20 μl of N, N-bis-trimethylsilyltrifluoroacetamide (Macherey-Nagel) and 20  μl of pyridine, and incubated for 40  min at 1) was used for further investigation. As expected, the gcn5-1 70 °C. These derivatized samples were then analysed using a gas chro- mutant also exhibited wax-deficient phenotypes (Fig.  1 and matography–flame ionization detector (GC-FID, Agilent, Technologies) Table  1). Moreover, a genetic complementation experiment and GC-MS (Agilent gas chromatograph coupled to an Agilent 5973N was performed by introducing a full-length GCN5 coding quadrupole mass-selective detector). sequence driven by the CaMV35S promoter into the gcn5-2 Consistent with previous studies (Haslam et  al., 2015; Lee and Suh, 2015; Li et al., 2016), the content of the unidentified components (which mutant. Given the low-fertility pollen of the gcn5-2 mutant showed no significant differences among genotypes) was excluded from found in previous studies (Bertrand et  al., 2003; Xing et  al., the total wax load. 2015), we first generated three independent homologous 35S::GCN5/Ws transgenic plants harboring significantly Statistical analysis high GCN5 expression levels, and crossed them with gcn5-2 Statistical analyses of the phenotypic data and expression levels were per- mutants (Supplementary Fig.  S3B). After three generations, formed using Student’s t-test in Excel. To assess the overall differences in three independent homozygous 35S::GCN5/gcn5-2 (#1, the stem cuticular wax composition among genotypes, we compared the #5 and #6) transgenic lines were obtained (Supplementary means using one-way ANOVA together with a Bonferroni adjustment Fig.  S3B). No obvious phenotype differences were observed test in R. between 35S::GCN5/Ws transgenic plants and the wild-type (Supplementary Fig.  S3C). However, 35S::GCN5/gcn5-2 Accession numbers transgenic lines exhibited similar phenotypes to the wild-type Sequence data from this article can be found in the Arabidopsis Genome (Fig. 1, Table 1, and Supplementary Fig. S3C), indicating that Initiative or GenBank/EMBL databases under the following accession constitutive expression of GCN5 could rescue the gcn5-2 wax numbers: CER3, AT5G57800; GCN5, AT3G54610; CER26, AT4G13840; deficiency. These results indicated that GCN5 is essential for CER1-L1, AT1G02190; WSD1, AT5G37300; AT2, AT5G55370; FAR3, the normal accumulation of cuticular wax on the stem surface AT4G33790, ACTIN8, AT1G49240; CHS, AT5G13930; AT4G03800. The RNA-seq reads used for this study are deposited at the National of Arabidopsis. Center for Biotechnology Information Short Read Archive under the accession number SRP093334. RNA-seq analysis reveals significant alteration of lipid-related gene expression in the gcn5-2 mutant Results To investigate whether the reduction of cuticular wax con- tent in the gcn5-2 mutant was caused by decreased expression Mutation of Arabidopsis GCN5 impairs stem cuticular of wax-related genes, total RNA of 6-week-old stems of the wax deposition gcn5-2 mutant and the wild-type Ws were isolated for high- To investigate the potential roles of histone modification in throughput RNA sequencing and transcriptomic comparison. cuticular wax accumulation, 23 T-DNA insertion mutants that Three biological replicates per genotype were performed and harbor disruption in histone acetylation or methylation genes the correlation coefficients of each genotype showed favor- were selected for analysis, namely ashh1, ashh2, ashh3, ashh4, able reproducibility (Supplementary Fig.  S4). The high reli- ashr2, ashr3, atx1, atx2, atx4, atx5, atxr2, atxr3, atxr4, gcn5-2, ability of the RNA-seq data was verified by qPCR of 10 hda2, hda2c, hda5, hda7, hda9, hda13, hda18, hda19, and srt2. randomly selected genes (Supplementary Fig.  S5). For each The stem cuticular wax of these mutants and the wild-type sample, values for reads per kilobase of exon model per million controls were observed and compared using SEM. As shown mapped reads (RPKM) were calculated, and the genes with in Supplementary Figs  S1 and S2, we found wide variations at least 2-fold change and a false discovery rate value P≤0.05 for cuticular wax crystal morphology and crystallization pat- were selected. Compared with the control, we found that 54% terns in the mutants and wild-types. For example, atx4, atxr4, (2616 genes) of the total of differentially expressed genes were Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2915 Fig. 1. Mutations in GCN5 are responsible for defects in the stem cuticular wax in Arabidopsis. (A) Light reflectance and (B) SEM observations of 6-week-old stems from wild-type Ws, and gcn5-2, 35S::GCN5/gcn5-2 transgenic lines (#1, #5, #6), and gcn5-1 mutants. Scale bars: (A) 1 mm, (B) 30 μm. (C) Total cuticular wax content, calculated per unit area of 6-week-old stems from the four different genotypes. (D) Cuticular wax composition of 6-week-old stems for wild-type Ws, and gcn5-2 and gcn5-1 mutants, and (E) for Ws and 35S::GCN5/gcn5-2 transgenic lines (#1, #5, #6). The chemical classes and main chain-lengths of each constituent are indicated. Data are means (±SD) of three biological replicates. *P<0.05, **P<0.01; Student’s t-test. (This figure is available in colour at JXB online.) down-regulated in gcn5-2 mutant stems. Because GCN5 usu- cellular lipid metabolic process, and lipid biosynthetic pro- –6 ally positively regulates transcriptional processes (Servet et al., cess (P≤6.2  ×  10 ). Moreover, a significant fraction of genes 2010) and mutation of GCN5 decreased the total wax load in involved in metabolic and/or biosynthetic processes (including Arabidopsis stems, these 2616 genes were expected to be dir- glycerolipid, neutral lipid, fatty acid, and unsaturated fatty acid), ect or indirect GCN5 target genes and might be involved in lipid catabolic process, and regulation of lipid metabolic and cuticular wax accumulation. biosynthetic processes were enriched (Table 2), suggesting that We then used GO analysis of the 2616 candidate genes using GCN5 is involved in wax accumulation by modulating the tran- agriGO v2.0 (Tian et  al., 2017), and the categories showed scription of lipid-related genes. For further screening, 145 non- considerably high enrichments in lipid metabolic process, redundant genes from three most-abundant lipid-related GO Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2916 | Wang et al. Table 1. Cuticular wax composition of stems in the wild-type Ws, gcn5-2, gcn5-1, and complementary transgenic lines Total Alkanes Ketones 1-Alcohols 2-Alcohols Fatty acids Aldehydes Esters Ws 1347.99 ± 54.38 650.81 ± 34.27 226.39 ± 12.94 140.44 ± 2.34 122.16 ± 6.91 10.39 ± 1.13 109.86 ± 4.70 87.93 ± 4.82 gcn5-2 847.49 ± 35.42** 375.94 ± 7.98** 133.64 ± 2.16** 87.01 ± 22.68* 83.59 ± 4.50** 7.03 ± 2.08 66.84 ± 14.77** 93.45 ± 38.77 #1 1357.28 ± 67.64 659.13 ± 34.75 230.33 ± 20.01 138.11 ± 17.45 122.64 ± 8.57 9.75 ± 3.34 111.81 ± 14.64 85.52 ± 7.39 #5 1360.93 ± 61.58 666.21 ± 27.19 230.35 ± 19.85 141.23 ± 15.52 116.23 ± 12.49 11.59 ± 2.21 110.60 ± 8.14 84.72 ± 17.21 #6 1344.06 ± 57.09 655.99 ± 27.72 228.46 ± 8.61 134.99 ± 11.51 117.54 ± 8.16 9.87 ± 1.85 111.51 ± 8.10 85.72 ± 14.12 gcn5-1 875.47 ± 90.61** 387.21 ± 41.12** 136.54 ± 13.30** 96.61 ± 17.24* 96.91 ± 16.31* 7.38 ± 1.03 72.00 ± 10.66** 78.82 ± 7.08 –2 Data are means (±SD; n=3) for the total wax load and each constituent class in μg dm . Five individual stems were used for each replicate. #1, #5, #6 represent three complementary transgenic lines (35S::GCN5/gcn5-2#1, #5, #6). *P<0.05, **P<0.01; Student’s t-test. 1-Alcohols, primary alcohols; 2-Alcohols, secondary alcohols. Table 2. Lipid-related GO categories for the 2616 candidate genes down-regulated in the gcn5-2 mutant GO ID Term Query item Query total Bg item Bg total P-value –7 GO:0006629 Lipid metabolic process 145 2598 994 28362 2.7 × 10 –6 GO:0044255 Cellular lipid metabolic process 100 2598 649 28362 2.5 × 10 –6 GO:0008610 Lipid biosynthetic process 83 2598 522 28362 6.20 × 10 GO:0046486 Glycerolipid metabolic process 24 2598 122 28362 0.0011 GO:0045017 Glycerolipid biosynthetic process 15 2598 79 28362 0.011 GO:0046460 Neutral lipid biosynthetic process 6 2598 18 28362 0.012 GO:0006638 Neutral lipid metabolic process 7 2598 24 28362 0.013 GO:0019216 Regulation of lipid metabolic process 11 2598 54 28362 0.019 GO:0016042 Lipid catabolic process 31 2598 224 28362 0.024 GO:0046890 Regulation of lipid biosynthetic process 9 2598 43 28362 0.028 GO:0006636 Unsaturated fatty acid biosynthetic process 6 2598 24 28362 0.036 GO:0006631 Fatty acid metabolic process 33 2598 252 28362 0.037 Query item: number of down-regulated genes in the gcn5-2 mutant annotated as the corresponding GO term. Query total: number of down-regulated genes in the gcn5-2 mutant. Bg item: number of genes in Arabidopsis whole genome annotated as the corresponding GO term. Bg total: number of genes in Arabidopsis whole genome. and Ws plants and GCN5-specific antibodies (Benhamed et al., Table 3. Potential GCN5-regulated genes involved in cuticular 2006, 2008). ChIP-qPCRs with three primer pairs spanning wax synthesis the promoter regions and gene body regions of each gene Gene Name Annotation (WSD1, CER3, CER26, CER1-L1, and AT2) were conducted for binding tests (Fig. 2B). CHS was used as the negative con- AT5G37300 WSD1 Bifunctional wax synthase/acyl-CoA:diacylglycerol acyltransferase trol as its expression is not affected by GCN5 (Benhamed et al., AT5G57800 CER3 ECERIFERUM 3 2006). As shown in Fig. 2C, significant decreases in enrichment AT4G13840 CER26 ECERIFERUM 26 in the gcn5-2 mutant were observed for most of the examined AT1G02190 CER1-L1 Protein CER1-like 1 regions of the CER3, CER26, and CER1-L1 genes relative to AT5G55370 AT2 Long-chain-alcohol O-fatty-acyltransferase 2 the wild-type, especially in the transcription start-site region (P1) of CER3. By contrast, no significant changes in GCN5 terms (GO:0006629, GO:0044255, and GO:0008610) were enrichment were detected in any tested regions for the other analysed together with their biological properties and functions two genes (WSD1 and AT2). in cuticular wax development as described in previous reports Previous studies have reported that GCN5 is specifically (Bernard and Joubès, 2013; Lee and Suh, 2015). Finally, we fil- responsible for H3K14 acetylation (H3K14ac) and that it tered five cuticular wax genes that are potential target genes of influences the H3K9ac and H3K27ac at the promoters of their GCN5, namely WSD1, CER3, CER26, CER1-L1, and Long- targets, which are positively correlated with gene expression chain-alcohol O-fatty-acyltransferase 2 (AT2; Table 3). (Bertrand et al., 2003; Bhat et al., 2003; Earley et al., 2007). Thus, we analysed the acetylation levels of H3K14 and H3K9 at the CER3, CER26, and CER1-L1 loci using the primer sets indi- ChIP assays identify target genes of GCN5 involved in cated in Fig. 2B. Consistently, both the H3K14ac and H3K9ac cuticular wax biosynthesis levels of these candidate genes were significantly decreased in the gcn5-2 mutant compared to the wild-type, especially at the To confirm the GCN5-regulated target genes in stem cuticular wax biosynthesis, we first detected the transcript levels of the promoter regions (Fig. 3), which was in accordance with the expression profiles (Fig. 2A). Collectively, these data indicated five candidate genes derived from RNA-seq data by qRT-PCR. that CER3, CER26, and CER1-L1 are the targets of GCN5, As expected, their expression levels were significantly reduced in the gcn5-2 mutant (Fig. 2A). The candidate genes were then and their expression can be regulated by GCN5 by modulating analysed by ChIP assays, using 6-week-old stems of the gcn5-2 their H3K14 and H3K9 acetylation. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2917 Fig. 2. Identification of GCN5 target genes. (A) Expression levels of the GCN5 target genes as determined by qRT-PCR in 6-week-old stems of the wild-type Ws and the gcn5-2 mutant. (B) Diagram representing the genomic structure and primer sets (indicated by P1–P5) analysed for ChIP-qPCR in the AT2, CER1-L1, CER26, WSD1, and CER3 genes. The exact distance (in bp) of the primers from to the ATG start codon sites (indicated by triangles) are labeled. Black boxes represent exons and gray boxes represent untranslated regions (UTRs). (C) ChIP analysis with nuclei extracted from cross- linked, 6-week-old stems of Ws and the gcn5-2 mutant and antibody-specific for GCN5. The CHS gene was used as a negative control, which provided background level for the ChIP samples. Data are means (±SD) from at least three biological replicates. *P<0.05, **P<0.01; Student’s t-test. Fig. 3. H3K14 and H3K9 acetylation levels on GCN5 target genes. ChIP analysis of H3K14 and H3K9 acetylation on (A, B) CER1-L1, (C, D) CER26, and (E, F) CER3 genes. Nuclei extracted from cross-linked, 6-week-old stems of the wild-type Ws and gcn5-2 mutant and antibodies specific for H3K14ac and H3K9ac. AT4G03800 (gypsy-like retrotransposon family gene) was used as a negative control. Data are means (±SD) from at least three biological replicates. *P<0.05, **P<0.01; Student’s t-test. Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2918 | Wang et al. CER3 is a critical enzyme for cuticular wax synthesis composition of the CER3 overexpression transgenic lines in (Aarts et  al., 1995; Chen et  al., 2003; Rowland et  al., 2007). both the Ws and gcn5-2 mutant backgrounds. Transcript lev- Interestingly, the cuticular wax composition of cer3 mutant els of the CER3 in 35S::CER3/Ws and 35S::CER3/gcn5-2 stems was quite similar to that of the gcn5-2 mutant stem, espe- transgenic lines increased significantly compared with the cially for aldehydes, alkanes, (secondary alcohols, and ketone controls (Fig.  5A). SEM imaging showed that the deposition (Lai et al., 2007; Rowland et al., 2007). Therefore, we hypoth- of wax crystals on the 35S::CER3/gcn5-2 stem was obviously esized that CER3 might be a critical target of GCN5 and increased compared with that of the gcn5-2 mutant and that decided to investigate the gene expression patterns of CER3 it basically recovered to the level of Ws, but no significant in ceriferous (i.e. wax-producing) tissues of gcn5-2 and Ws. differences were observed between 35S::CER3/Ws and Ws Isogenic ProCER3::GUS/gcn5-2 and ProCER3::GUS/Ws (Fig. 5B). lines were obtained in which each transgene was homozy- In addition, the total load and composition of stem cuticu- gous and correspondingly inserted into a single genomic locus. lar wax of 35S::CER3/Ws, 35S::CER3/gcn5, Ws, and gcn5- GUS staining revealed weaker signals in stems, siliques, flowers, 2 plants were measured by GC-FID and GC-MS analysis. and young leaves of the gcn5-2 mutant compared to that of the Although the wax crystal loading showed little increase under control (Fig.  4A). Quantification of GUS activity by fluoro- SEM observation (Fig.  5B), the total amount and individ- metric assay consistently revealed significantly lower activity in ual components of stem cuticular wax in two of the three three independent homozygous ProCER3::GUS/gcn5-2 lines 35S::CER3/Ws transgenic lines (#5 and #10) were signifi- than in the corresponding ProCER3::GUS/Ws lines (Fig. 4B). cantly increased as compared to Ws, especially for the amount qRT-PCR analysis also showed that the CER3 expression lev- of alkanes, primary alcohols, and ketone, which were consistent els were decreased in different tissues of gcn5-2 mutant as com- with the expression levels of the CER3 gene (Figs.  5C, 6A). pared to the wild-type (Fig. 4C). These results provided further Notably, the total wax load in 35S::CER3/gcn5-2 transgenic evidence that the expression of CER3 is positively regulated plants was obviously increased compared with that of the gcn5- by GCN5 in ceriferous tissues. 2 mutant, and in the 35S::CER3/gcn5-2 #10 transgenic line it was even restored to the wild-type level (Fig. 5C). Analysis of the cuticular wax composition showed that the contents of Overexpression of CER3 rescues the stem cuticular C29 alkane, C29 ketone, C26 and C28 primary alcohols, C29 wax-deficient phenotype in the gcn5-2 mutant secondary alcohols, and C30 aldehyde significantly increased To assess the role of CER3 in GCN5-modulated biosynthe- in the 35S::CER3/gcn5-2 transgenic lines compared with the sis of stem cuticular wax, we studied the accumulation and control, thus providing genetic evidence that CER3 plays an Fig. 4. Expression patterns of CER3 in ceriferous tissues of the wild-type Ws and gcn5-2 mutant. (A) Spatial expression patterns of the CER3 gene in transgenic Ws and gcn5-2 plants harboring the CER3 promoters fused to the GUS gene. Promoter activity was visualized through histochemical GUS-staining in stems, siliques, flowers, and young leaves of 6-week-old plants. Scale bars: stems and siliques, 1.5 mm; flowers and leaves, 2 mm. (B) Quantification of the GUS activity using 4-methylumbelliferone (4-MU) as a substrate and (C) CER3 expression levels in stems, siliques, flowers, and young leaves of 6-week-old plants. Data are means (±SD) from three biological replicates. **P<0.01; Student’s t-test. (This figure is available in colour at JXB online.) Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2919 important role in GCN5-regulated cuticular wax biosynthesis Post-translational modification of histone tails plays an important (Fig.  6B). It should be noted that overexpression of CER3 role in epigenetic regulation of gene expression, and this includes did not completely restore all wax component defects in the histone acetylation, methylation, phosphorylation, and ubiqui- gcn5-2 mutant, for example C30 primary alcohol and some tination (Pfluger and Wagner, 2007). As a member of the HAT esters, and this may be attributed to other GCN5-regulated enzymes, GCN5 is a versatile regulator of Arabidopsis devel- target genes in the wax biosynthetic pathway. FATTY ACID opment and stress responses (Bertrand et  al., 2003; Benhamed REDUCTASE3 (FAR3, also known as CER4), encodes an et  al., 2006; Kornet and Scheres, 2009; Hu et  al., 2015; Xing alcohol-forming fatty acyl-coenzyme A  reductase, and is et al., 2015). Recently, we found that GCN5 is involved in FA involved in the synthesis of primary alcohols (Rowland et al., biosynthesis by affecting the acetylation levels of FAD3 ( Wang 2006; Wang et al., 2015). Thus, we analysed the transcript levels et al., 2016). Here, we observed that GCN5 is essential for lipid of FAR3 in gcn5-2, Ws, and 35S::CER3 transgenic plants using metabolism in Arabidopsis stems. Firstly, mutation of GCN5 in qRT-PCR. However, the expression levels of FAR3 in these Arabidopsis compromised the content of multiple lipid com- plants were not obviously changed (Supplementary Fig. S6). pounds (including very-long-chain alkanes, aldehydes, ketones, and alcohols), which resulted in a complete deficiency in stem cuticular wax accumulation. This wax deficiency could be fully Discussion rescued by complementation with 35S::GCN5. Secondly, GO Lipids are an essential constituent of all plant cells, including fatty analysis indicated that down-regulated genes in the gcn5-2 acids (FAs), waxes, sterols, and others (Li-Beisson et  al., 2013). mutant were enriched in categories related to lipid synthesis, Fig. 5. Overexpression of CER3 in the gcn5-2 mutant increases the cuticular wax component and restores it to the wild-type Ws level. (A) Relative expression levels of CER3 in plants of Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10). Total RNA was isolated from 6-week-old stems of Ws and the gcn5-2 mutant. ACTIN8 was used as an endogenous control. Data are means (±SD) from at least three biological replicates. (B) Scanning electron microscopy of the stems of Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10). Scale bars: 20 μm. (C) Total cuticular wax content was calculated over the unit area of 6-week-old stems in plants of Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10). The mean expression levels and total wax load were compared using one-way ANOVA together with a Bonferroni adjustment test in R. Different letters indicate significant differences among genotypes (P<0.05). Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2920 | Wang et al. Fig. 6. Cuticular wax composition of the stems of wild-type Ws, 35S::CER3/Ws (#2, #5, #10), gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10) from 6-week-old plants. The chemical classes and main chain-lengths of each constituent are indicated. Data are means (±SD) of three biological replicates. (A) Comparison between Ws and 35S::CER3/Ws (#2, #5, #10) and (B) among Ws, gcn5-2, and 35S::CER3/gcn5-2 (#2, #5, #10) transgenic plants. The mean values were compared using one-way ANOVA together with a Bonferroni adjustment test in R. Different letters indicate significant differences among genotypes (P<0.05). (This figure is available in colour at JXB online.) including the lipid biosynthetic process, neutral lipid biosyn- Rowland et al., 2007). The wax components of the Arabidopsis thetic process, glycerolipid biosynthetic process, and unsaturated cer3 mutant are lacking in aldehydes, alkanes, secondary alco- fatty acid biosynthetic process. Thirdly, ChIP assays demonstrated hols, and ketones in the stem compared with the wild-type that CER1-L1, CER26, and CER3, which encode proteins (Rowland et al., 2007). Interestingly, our GC-FID and GC-MS involved in VLCFA production and alkane-forming pathways analyses found significant reductions in total wax in the gcn5- of wax synthesis, are target genes of GCN5. Finally, enrich- 2 mutant stem, especially for C30 aldehyde, C29 alkane, C29 ment of H3K9ac and H3K14ac at the promoters of CER1-L1, ketone, and C29 secondary alcohols. Moreover, overexpression CER26, and CER3 was significantly decreased in the gcn5-2 of CER3 in the gcn5-2 background significantly increased the mutant compared with the wild-type. Collectively, our previous levels of C29 ketone, C30 aldehydes, C29 alkanes, and C29 sec- data (Wang et  al., 2016) and that from the present study have ondary alcohols, indicating that CER3 played a pivotal role in demonstrated that histone acetyltransferase GCN5 is involved GCN5-regulated biosynthesis of stem cuticular wax. However, in multiple lipid metabolic processes, from upstream de novo FA we cannot rule out the possibility that other unknown genes synthesis to subsequent wax production. may also contribute, which may prove an interesting area for To determine the underlying mechanisms of GCN5- further investigation. For example, the ester component, which regulated biosynthesis of stem cuticular wax, three GCN5 cannot be catalysed by CER3, also changed significantly in target genes, CER3, CER26, and CER1-L1 were identified. the gcn5-2 mutant stem. The gcn5-2 mutant deficiency of C30 CER26 is involved in the elongation of VLCFAs (from 30 primary alcohols was not rescued by overexpression of CER3. C to 32 C) and has high specificity of tissue and substrate In addition, the expression of FAR3, an important enzyme in (Haslam et  al., 2015; Pascal et  al., 2013). Although little is primary alcohol formation, was not regulated by GCN5. known about CER1-L1 other than that it is a homolog of Many reports have demonstrated that acetylation of histone CER1 (Bernard et  al., 2012), there is the possibility that, like tails induces the accessibility of transcription factors to the CER1, it can physically interact with CER3 during the very- nucleosomal DNA, which subsequently influences the gene long-chain alkane biosynthesis process and contribute to the expression (e.g. Lee et  al., 1993). As a key wax biosynthetic total wax load. Previous studies have reported that CER3, enzyme gene, CER3 is regulated by both transcription factors which catalyses redox-dependent alkane formation, is the key and epigenetic modulators (Lee and Suh, 2015). CER3 is posi- wax biosynthetic enzyme (Aarts et al., 1995; Chen et al., 2003; tively regulated by two MYB transcription factors, MYB96 Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 GCN5 affects stem cuticular wax synthesis via CER3 | 2921 and MYB30, in response to drought and pathogen attack, Supplementary data respectively (Lee and Suh, 2013; Lee et  al., 2016b). Based on Supplementary data are available at JXB online. our RNA-seq data (see Supplementary Table S2), the expres- Table S1. Summary of the RNA-seq data and read mapping. sion of MYB96 was up-regulated in the gcn5-2 mutant, sug- Table S2. Differently expressed genes in the RNA-seq data. gesting that it does not contribute much to the changed Table  S3. GO analysis of the 2616 down-regulated genes expression of CER3 in the stem of this mutant. However, we in the stems of the gcn5-2 mutant compared with the cannot rule out the possibility that CER3 could be regulated wild-type Ws. by other unknown transcription factors that are deregulated Table S4. Gene-specific primer pairs in this study. in the gcn5-2 mutant. Recent studies showed that epigenetic Fig. S1. Stem cuticular wax phenotype of the histone-acet- modulators, namely the histone methyl transferases SDG8 ylation mutants. (ASHH2) and SDG25 (ATXR7), were involved in wax accu- Fig. S2. Stem cuticular wax phenotype of the histone-meth- mulation through histone lysine methylation and/or indirectly ylation mutants. through H2B ubiquitination by targeting CER3, and this was Fig.  S3. GCN5 expression levels and phenotypes of the associated with diminished accumulation of lipids (Lee et  al., independent homologous 35S::GCN5 transgenic lines. 2016a). Here, our observations also showed that the ashh2 Fig.  S4. Reproducibility of the RNA-seq biological mutant exhibited mildly reduced wax crystal accumulation replicates. compared with Col-0. It might be of interest to further exam- Fig. S5. Expression levels of 10 genes randomly selected to ine the relationships between GCN5, SDG8 (ASHH2), and validate the accuracy of the RNA-seq data using qRT-PCR. SDG25 (ATXR7) in cuticular wax biosynthesis. Fig. S6. FAR3 expression levels in wild-type Ws, gcn5-2, and In conclusion, our results have demonstrated that, like his- 35S::CER3 transgenic lines. tone ubiquitination and methylation, histone acetylation is also involved in the regulation of biosynthesis of stem cuticular wax, and we propose a working model to explain this process in Acknowledgements Arabidopsis (Fig. 7). Briefly, the histone acetyltransferase GCN5 The authors thank Prof. Jianxin Shi (Shanghai Jiao Tong University) for regulates the biosynthesis of stem cuticular wax by regulating the the GC-FID and GC-MS analyses in his laboratory. This work was sup- expression of CER3, CER1-L1, and CER26 via histone acety- ported by the Ministry of Agriculture of China for Transgenic Research lation at the H3K9/14 sites. Thus, interruption of GCN5 dra- (2016ZX08009002), the State Key laboratory of Agrobiotechnology matically reduces the total amount of cuticular wax and changes Open Grant (2018SKLAB6-25) and the Fundamental Research Funds for the Central Universities (2412017QD016). its composition, especially with regards to alkanes, aldehydes, and ketone, which are mainly synthesized in the alkane-forming pro- cess. Remarkably, overexpression of CER3 in the gcn5-2 mutant References could rescue the cuticular wax deficiency, suggesting that it has Aarts MG, Keijzer CJ, Stiekema WJ, Pereira A. 1995. Molecular an important role in GCN5-mediated cuticular wax biosynthe- characterization of the CER1 gene of Arabidopsis involved in epicuticular sis. Our findings provide an insight into the epigenetic regula- wax biosynthesis and pollen fertility. The Plant Cell 7, 2115–2127. tion of cuticular wax development through histone acetylation, Aharoni A, Dixit S, Jetter R, Thoenes E, van Arkel G, Pereira A. which may contribute to wax-related stress responses in plants. 2004. The SHINE clade of AP2 domain transcription factors activates wax biosynthesis, alters cuticle properties, and confers drought tolerance when overexpressed in Arabidopsis. The Plant Cell 16, 2463–2480. Benhamed M, Bertrand C, Servet C, Zhou DX. 2006. Arabidopsis GCN5, HD1, and TAF1/HAF2 interact to regulate histone acetylation required for light-responsive gene expression. The Plant Cell 18, 2893–2903. Benhamed M, Martin-Magniette ML, Taconnat L, et  al. 2008. Genome-scale Arabidopsis promoter array identifies targets of the histone acetyltransferase GCN5. The Plant Journal 56, 493–504. Bernard A, Domergue F, Pascal S, Jetter R, Renne C, Faure JD, Haslam RP, Napier JA, Lessire R, Joubès J. 2012. Reconstitution of plant alkane biosynthesis in yeast demonstrates that Arabidopsis ECERIFERUM1 and ECERIFERUM3 are core components of a very-long- chain alkane synthesis complex. The Plant Cell 24, 3106–3118. Bernard A, Joubès J. 2013. Arabidopsis cuticular waxes: advances in Fig. 7. A model for the regulation of stem cuticular wax synthesis by synthesis, export and regulation. Progress in Lipid Research 52, 110–129. GCN5-associated acetylation in Arabidopsis. As the wax precursors, Bertrand C, Bergounioux C, Domenichini S, Delarue M, Zhou DX. the very-long-chain acyl-CoAs (VLC-CoA) can be processed through 2003. Arabidopsis histone acetyltransferase AtGCN5 regulates the floral the alcohol-forming pathway and alkane-forming pathway, which yield meristem activity through the WUSCHEL/AGAMOUS pathway. The Journal 17~18% and 80% of the total cuticular wax mixture, respectively. Solid of Biological Chemistry 278, 28246–28251. black arrows represent the cuticular wax biosynthesis pathway. The Bhat RA, Riehl M, Santandrea G, Velasco R, Slocombe S, Donn arrows from GCN5 indicate positive transcriptional regulation by GCN5 G, Steinbiss HH, Thompson RD, Becker HA. 2003. Alteration of via H3K9/14ac modifications. GCN5 targets are marked at the positions GCN5 levels in maize reveals dynamic responses to manipulating histone acetylation. The Plant Journal 33, 455–469. where the enzymes they encode are required. CER3 is a key cuticular wax biosynthetic enzyme that catalyses the alkane-forming pathway in Bradford MM. 1976. A rapid and sensitive method for the quantitation of Arabidopsis stems (solid arrow), whereas CER26 and CER1-L1 might microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248–254. partially contribute to the total wax load but they were not verified functionally in this study (dashed arrows). (This figure is available in colour Broun P, Poindexter P, Osborne E, Jiang CZ, Riechmann JL. 2004. at JXB online.) WIN1, a transcriptional activator of epidermal wax accumulation in Downloaded from https://academic.oup.com/jxb/article/69/12/2911/4915854 by DeepDyve user on 18 July 2022 2922 | Wang et al. Arabidopsis. Proceedings of the National Academy of Sciences, USA 101, Li-Beisson Y, Shorrosh B, Beisson F, et al. 2013. Acyl-lipid metabolism. 4706–4711. The Arabidopsis Book 11, e0161. Chen W, Yu XH, Zhang K, Shi J, De Oliveira S, Schreiber L, Shanklin J, Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression –ΔΔ t Zhang D. 2011. Male Sterile2 encodes a plastid-localized fatty acyl carrier data using real-time quantitative PCR and the 2 C method. Methods 25, protein reductase required for pollen exine development in Arabidopsis. 402–408. Plant Physiology 157, 842–853. Ménard R, Verdier G, Ors M, Erhardt M, Beisson F, Shen WH. 2014. Chen X, Goodwin SM, Boroff VL, Liu X, Jenks MA. 2003. Cloning Histone H2B monoubiquitination is involved in the regulation of cutin and and characterization of the WAX2 gene of Arabidopsis involved in cuticle wax composition in Arabidopsis thaliana. Plant & Cell Physiology 55, membrane and wax production. The Plant Cell 15, 1170–1185. 455–466. Clough SJ, Bent AF. 1998. Floral dip: a simplified method for Pandey R, Müller A, Napoli CA, Selinger Da, Pickaard CS, Richards Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant EJ, Bender J, Mount DW, Jorgensen RA. 2002. Analysis of histone Journal 16, 735–743. acetyltransferase and histone deacetylase families of Arabidopsis thaliana suggests functional diversification of chromatin modification among Earley KW, Shook MS, Brower-Toland B, Hicks L, Pikaard CS. 2007. multicellular eukaryotes. Nucleic Acids Research 30, 5036–5055. In vitro specificities of Arabidopsis co-activator histone acetyltransferases: implications for histone hyperacetylation in gene activation. The Plant Pascal S, Bernard A, Sorel M, Pervent M, Vile D, Haslam RP, Napier Journal 52, 615–626. JA, Lessire R, Domergue F, Joubès J. 2013. The Arabidopsis cer26 mutant, like the cer2 mutant, is specifically affected in the very long chain Fiil BK, Qiu JL, Petersen K, Petersen M, Mundy J. 2008. fatty acid elongation process. The Plant Journal 73, 733–746. Coimmunoprecipitation (co-IP) of nuclear proteins and chromatin immunoprecipitation (ChIP) from Arabidopsis. CSH Protocols 2008, pdb. Pfluger J, Wagner D. 2007. Histone modifications and dynamic regulation prot5049. of genome accessibility in plants. Current Opinion in Plant Biology 10, 645–652. Haslam TM, Haslam R, Thoraval D, et  al. 2015. ECERIFERUM2-LIKE proteins have unique biochemical and physiological functions in very-long- Raffaele S, Vailleau F, Léger A, Joubès J, Miersch O, Huard C, Blée chain fatty acid elongation. Plant Physiology 167, 682–692. E, Mongrand S, Domergue F, Roby D. 2008. A MYB transcription factor regulates very-long-chain fatty acid biosynthesis for activation of Haslam TM, Mañas-Fernández A, Zhao L, Kunst L. 2012. Arabidopsis the hypersensitive cell death response in Arabidopsis. The Plant Cell 20, ECERIFERUM2 is a component of the fatty acid elongation machinery 752–767. required for fatty acid extension to exceptional lengths. Plant Physiology 160, 1164–1174. Robinson MD, McCarthy DJ, Smyth GK. 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Hu Z, Song N, Zheng M, et al. 2015. Histone acetyltransferase GCN5 is Bioinformatics 26, 139–140. essential for heat stress-responsive gene activation and thermotolerance in Arabidopsis. The Plant Journal 84, 1178–1191. Rowland O, Lee R, Franke R, Schreiber L, Kunst L. 2007. The CER3 wax biosynthetic gene from Arabidopsis thaliana is allelic to WAX2/YRE/ Jefferson RA, Bevan M, Kavanagh T. 1987. The use of the Escherichia FLP1. FEBS Letters 581, 3538–3544. coli beta-glucuronidase as a gene fusion marker for studies of gene expression in higher plants. Biochemical Society Transactions 15, 17–18. Rowland O, Zheng H, Hepworth SR, Lam P, Jetter R, Kunst L. 2006. CER4 encodes an alcohol-forming fatty acyl-coenzyme A  reductase Kim D, Salzberg SL. 2011. TopHat-Fusion: an algorithm for discovery of involved in cuticular wax production in Arabidopsis. Plant Physiology 142, novel fusion transcripts. Genome Biology 12, R72. 866–877. Koornneef M, Hanhart CJ, Thiel F. 1989. A genetic and phenotypic Samuels L, Kunst L, Jetter R. 2008. Sealing plant surfaces: cuticular wax description of eceriferum mutants in Arabidopsis thaliana. The Journal of formation by epidermal cells. Annual Review of Plant Biology 59, 683–707. Heredity 80, 118–122. Servet C, Conde e Silva N, Zhou DX. 2010. Histone acetyltransferase Kornet N, Scheres B. 2009. Members of the GCN5 histone AtGCN5/HAG1 is a versatile regulator of developmental and inducible gene acetyltransferase complex regulate PLETHORA-mediated root stem cell expression in Arabidopsis. Molecular Plant 3, 670–677. niche maintenance and transit amplifying cell proliferation in Arabidopsis. The Plant Cell 21, 1070–1079. Stockinger EJ, Mao Y, Regier MK, Triezenberg SJ, Thomashow MF. 2001. Transcriptional adaptor and histone acetyltransferase proteins Lai C, Kunst L, Jetter R. 2007. Composition of alkyl esters in the cuticular in Arabidopsis and their interactions with CBF1, a transcriptional activator wax on inflorescence stems of Arabidopsis thaliana cer mutants. The Plant involved in cold-regulated gene expression. Nucleic Acids Research 29, Journal 50, 189–196. 1524–1533. Laux T, Mayer KF, Berger J, Jürgens G. 1996. The WUSCHEL gene is Suh MC, Samuels AL, Jetter R, Kunst L, Pollard M, Ohlrogge J, required for shoot and floral meristem integrity in Arabidopsis. Development Beisson F. 2005. Cuticular lipid composition, surface structure, and 122, 87–96. gene expression in Arabidopsis stem epidermis. Plant Physiology 139, Lee DY, Hayes JJ, Pruss D, Wolffe AP. 1993. A positive role for histone 1649–1665. acetylation in transcription factor access to nucleosomal DNA. Cell 72, Tian T, Liu Y, Yan H, You Q, Yi X, Du Z, Xu W, Su Z. 2017. agriGO v2.0: 73–84. a GO analysis toolkit for the agricultural community, 2017 update. Nucleic Lee S, Fu F, Xu S, Lee SY, Yun DJ, Mengiste T. 2016a. Global regulation Acids Research 45, W122–W129. of plant immunity by histone lysine methyl transferases. The Plant Cell 28, Vlachonasios KE, Thomashow MF, Triezenberg SJ. 2003. Disruption 1640–1661. mutations of ADA2b and GCN5 transcriptional adaptor genes dramatically Lee SB, Kim HU, Suh MC. 2016b. MYB94 and MYB96 additively activate affect Arabidopsis growth, development, and gene expression. The Plant cuticular wax biosynthesis in Arabidopsis. Plant & Cell Physiology 57, Cell 15, 626–638. 2300–2311. Wang T, Xing J, Liu X, et  al. 2016. Histone acetyltransferase general Lee SB, Suh MC. 2013. Recent advances in cuticular wax biosynthesis control non-repressed protein 5 (GCN5) affects the fatty acid composition and its regulation in Arabidopsis. Molecular Plant 6, 246–249. of Arabidopsis thaliana seeds by acetylating fatty acid desaturase3 (FAD3). Lee SB, Suh MC. 2015. Advances in the understanding of cuticular waxes The Plant Journal 88, 794–808. in Arabidopsis thaliana and crop species. Plant Cell Reports 34, 557–572. Wang Y, Wang M, Sun Y, Hegebarth D, Li T, Jetter R, Wang Z. Li F, Wu X, Lam P, Bird D, Zheng H, Samuels L, Jetter R, Kunst L. 2008. 2015. Molecular characterization of TaFAR1 involved in primary alcohol Identification of the wax ester synthase/acyl-coenzyme A: diacylglycerol biosynthesis of cuticular wax in hexaploid wheat. Plant & Cell Physiology acyltransferase WSD1 required for stem wax ester biosynthesis in 56, 1944–1961. Arabidopsis. Plant Physiology 148, 97–107. Xing J, Wang T, Liu Z, et  al. 2015. GENERAL CONTROL Li S, Wang X, He S, et  al. 2016. CFLAP1 and CFLAP2 are two bHLH NONREPRESSED PROTEIN5-mediated histone acetylation of FERRIC transcription factors participating in synergistic regulation of AtCFL1- REDUCTASE DEFECTIVE3 contributes to iron homeostasis in Arabidopsis. mediated cuticle development in Arabidopsis. PLoS Genetics 12, 1–27. Plant Physiology 168, 1309–1320.

Journal

Journal of Experimental BotanyOxford University Press

Published: May 25, 2018

There are no references for this article.