https://orcid.org/0000-0001-8322-8833
Abstract
The function of the transcription factor KNOTTED ARABIDOPSIS THALIANA7 (KNAT7) is still unclear since it appears to be either a negative or a positive regulator for secondary cell wall deposition with its loss-of-function mutant displaying thicker interfascicular and xylary fiber cell walls but thinner vessel cell walls in inflorescence stems. To explore the exact function of KNAT7, class II KNOTTED1-LIKE HOMEOBOX (KNOX II) genes in Arabidopsis including KNAT3, KNAT4, and KNAT5 were studied together. By chimeric repressor technology, we found that both KNAT3 and KNAT7 repressors exhibited a similar dwarf phenotype. Both KNAT3 and KNAT7 genes were expressed in the inflorescence stems and the knat3 knat7 double mutant exhibited a dwarf phenotype similar to the repressor lines. A stem cross-section of knat3 knat7 displayed an enhanced irregular xylem phenotype as compared with the single mutants, and its cell wall thickness in xylem vessels and interfascicular fibers was significantly reduced. Analysis of cell wall chemical composition revealed that syringyl lignin was significantly decreased while guaiacyl lignin was increased in the knat3 knat7 double mutant. Coincidently, the knat3 knat7 transcriptome showed that most lignin pathway genes were activated, whereas the syringyl lignin-related gene Ferulate 5-Hydroxylase (F5H) was down-regulated. Protein interaction analysis revealed that KNAT3 and KNAT7 can form a heterodimer, and KNAT3, but not KNAT7, can interact with the key secondary cell wall formation transcription factors NST1/2, which suggests that the KNAT3–NST1/2 heterodimer complex regulates F5H to promote syringyl lignin synthesis. These results indicate that KNAT3 and KNAT7 synergistically work together to promote secondary cell wall biosynthesis.
Introduction
Lignin is one of the major structural components in plant secondary cell walls and provides mechanical strength and stiffness to the plant cells (Boerjan et al., 2003; Cosgrove and Jarvis, 2012). Three types of lignin subunits, i.e. p-hydroxyphenyl (H), guaiacyl (G), and syringyl (S), are derived from oxidative polymerization of three monolignols, p-coumaryl, coniferyl, and sinapyl alcohols, respectively (Boerjan et al., 2003).
Phenylalanine is used as substrate to produce hydroxycinnamoyl-CoA esters in the phenylpropanoid pathway involved in monolignol biosynthesis. At least 11 enzymes are involved in catalysing monolignol biosynthesis: phenylalanine ammonia lyase (PAL), cinnamate 4-hydroxylase (C4H), 4-coumarate CoA ligase (4CL), hydroxycinnamoyl CoA:shikimate hydroxycinnamoyl transferase (HCT), p-coumaroyl shikimate 3′-hydroxylase (C3H), caffeoyl CoA O-methyltransferase (CCoAOMT), ferulate 5-hydroxylase (F5H), caffeic acid O-methyltransferase (COMT), cinnamoyl CoA reductase (CCR), cinnamyl alcohol dehydrogenase (CAD), and caffeoyl shikimate esterase (CSE) (Boerjan et al., 2003; Bonawitz and Chapple, 2010; Vanholme et al., 2013). Of these enzymes, F5H is the key catalyst in the synthesis of S-lignin. A complicated transcriptional network regulates secondary cell wall formation, with the top-level regulators VASCULAR RELATED NAC-DOMAIN PROTEIN6 (VND6), VND7, and NAC (NAM, ATAF1/2, and CUC2) transcription factors, second-level regulators MYB46 and MYB83, and a battery of downstream transcription factors, namely MYB20, MYB42, MYB43, MYB52, MYB54, MYB58, MYB63, MYB69, MYB85, MYB103, and KNOTTED ARABIDOPSIS THALIANA7 (KNAT7). The transcriptional regulation of lignin biosynthesis has revealed that some MYB transcription factors can activate monolignol pathway genes directly by binding to the AC-rich elements of their promoters (Lacombe et al., 2000; Patzlaff et al., 2003; Goicoechea et al., 2005; Zhou et al., 2009). The promoters of PAL, 4CL, C3H, CCoAOMT, CCR, and CAD all contain the conserved AC elements, while the promoters of C4H and COMT contain degenerated AC elements (Zhou et al., 2009; Zhao and Dixon, 2011). Therefore, all those genes might be directly regulated by the known lignin-specific transcription factors, such as AtMYB58 and MYB63. However, F5H, the key gene for S-lignin biosynthesis, neither contains AC-rich elements nor is regulated by the lignin-specific transcription factors (Zhou et al., 2009; Zhao and Dixon, 2011). Although it is reported that the Medicago F5H gene is directly regulated by SND1 (NST3) (Zhao et al., 2010), evidence for such regulation in Arabidopsis is lacking (Öhman et al., 2013). The molecular regulation of F5H in Arabidopsis is still elusive.
KNAT7 has been reported to function as a repressor of secondary cell wall formation, and its disruption results in thicker interfascicular fiber walls in Arabidopsis (Li et al., 2012). The known KNAT7 interactors, such as OVATE FAMILY PROTEIN4 (OFP4), BELL-LIKE HOMEODOMAIN6 (BLH6), and MYB75 (Hackbusch et al., 2005; Bhargava et al., 2010; Li et al., 2011; Liu et al., 2014), are negative regulators of secondary wall formation. Paradoxically, the knat7 mutant had thinner vessel cells (Li et al., 2012), and there was reduced secondary wall thickness in KNAT7 dominant repressing transgenic plants (Zhong et al., 2008). Our recent study revealed that KNAT7 positively affects xylan (the important composition of hemicellulose) biosynthesis (He et al., 2018). These results imply that KNAT7 has complicated functions with regard to specific cell types or interacting regulators.
KNAT7 is one of the KNOTTED1-LIKE HOMEBOX (KNOX) genes, which belong to the plant-specific Three Amino Acid Loop Extension (TALE) homeodomain superfamily (Hake et al., 2004; Hay and Tsiantis, 2010). KNOX genes can be grouped into two classes: KNOX I (SHOOTMERISTEMLESS (STM), BREVIPEDICELLUS (BP), KNAT2, and KNAT6) and KNOX II (KNAT3, KNAT4, KNAT5, and KNAT7) (Kerstetter et al., 1994; Mukherjee et al., 2009). KNAT3, KNAT4, and KNAT5 function redundantly to influence leaf morphology in Arabidopsis (Furumizu et al., 2015). As they are KNOX II clade members, it would be interesting to explore whether KNAT3, KNAT4, and KNAT5 have redundant functions with KNAT7 on secondary wall formation in inflorescence and stem, and thus to interpret the diverse and paradoxical roles of KNAT7.
Here, we report that the loss-of-function knat3 knat7 double mutant displays an enhanced irregular xylem (irx) phenotype including thinner vessel and interfascicular cell wall thickness. The actual chemical composition revealed an increase in G-lignin and a decrease in S-lignin. RNA-seq results displayed significant down-regulation of F5H. A protein interaction assay identified that KNAT3 and KNAT7 form a heterodimer, and KNAT3, but not KNAT7, can interact with NST1 and NST2. Altogether, our results show that KNAT3 and KNAT7 form a complex to paradoxically regulate secondary cell wall formation in different tissue.
Materials and methods
Plant material and growth conditions
Arabidopsis ecotype Columbia (Col‐0) was used as the WT plant and for transgenic plant experiments. Plants were grown on soil with supplemental lighting with a photoperiod of 16 h light–8 h dark and light intensity and temperature of 120 µmol m−2 s−1 and 22 °C, respectively. The T‐DNA insertion mutant allele of myb103-1 (AT1G63910 SALK_210187C), myb103-2 (AT1G63910 SAIL_337_C03), and KNAT3 (AT5G25220 SALK_136464) and KNAT7 (AT1G62990 SALK_110899), designated knat3 and knat7, were identified using the SIGnal database (http://signal.salk.edu/), and seeds were obtained from the Arabidopsis Biological Resources Center (ABRC, Columbus, OH, USA). KNAT3 and KNAT7 gene‐specific primers used for PCR genotyping are listed in Supplementary Table S1 at JXB online.
Plasmid constructs and generation of transgenic plants
Promoter fragments of 2 kb, 2.6 kb upstream of the ATG start codons of KNAT3 and KNAT7 were designed. The promoter fragments were cloned into the pDONR207 vector and then transferred into pGWB3 (Nakagawa et al., 2007) and all constructs were used to transform WT Arabidopsis. Both KNAT3 and KNAT7 coding regions were individually cloned into pDONR207 and then recombined into pEarleyGate100. Chimeric repressor constructs were prepared by inserting the coding sequence of KNAT3 and KNAT7 with one extra guanine nucleotide at the 5′ end into the SmaI site of p35S:SRDXG vector and then transferred into vector pBCKH (Mitsuda et al., 2006), in which a chimeric repressor was produced by fusion transcription factors to the plant-specific ERF associated Amphiphilic Repression (EAR) repression domain (LDLDLELRLGFA). Finally, all DNA constructs were verified by DNA sequencing analysis and were electroporated into Agrobacterium tumefaciens GV3101. Transformation of WT plants and the knat3 knat7 double mutant was done using the floral dip method (Clough and Bent, 1998). Transgenic lines were screened on 1/2 MS plates plus 50 mg ml−1 kanamycin or glufosinate ammonium at 2000× dilution (Sigma).
β-Glucuronidase staining
β-Glucuronidase (GUS) activity was examined in 40 mm sections of 6-week-old inflorescence stems. Sections were washed three times with phosphate-buffered saline (pH 7.0), and incubated with GUS staining solution (100 mM Na3PO4, pH 7.0, 1 mM EDTA, 1 mM potassium ferrocyanide, 1 mM potassium ferricyanide, 1% Triton X-100, and 1 mg ml−1 5-bromo-4-chloro-3-indolyl β-D-glucuronide (X-Gluc)) for 8 h to overnight at 37 °C in dark. After decolorization in 95% or 75% ethanol (Jefferson et al., 1987; Wu et al., 2009), the sections were observed under a light microscope.
Sectioning of stems
Segments were cut from the basal internode of stems of 6-week-old plants. A Leica VT1000S vibrating-blade microtome was used to cut 40 mm sections by using 3% agarose as support. Phloroglucinol-HCl Staining was performed; the sections were stained for 5 s in 3% phloroglucinol, then were washed out followed by the addition of concentrated HCl (37 M), and then observed using a light microscope. For Mäule staining, 0.5% potassium permanganate solution was added to the sections for staining; the solution was washed out and then distilled water was added to rinse out the potassium permanganate solution; finally, 3% HCl was added and it was incubated for 3–5 min until the deep brown color was discharged from the sections. After pipetting out all the HCl solution, concentrated ammonium hydroxide solution was added immediately (Pradhan and Loque, 2014). Then the sections were observed under bright field microscopy (BX43F; Olympus, Japan). Finally, cell wall thickness was measured by the microscope’s software and combined with the software ImageJ.
Cell wall preparation and lignin analysis
To measure total lignin content and monomeric lignin composition, 8-week-old mature stems were collected, with three biological replicates, each replicate having at least six plants. After ball-milling into powder, wax was removed using a Soxhlet extractor. The powder was incubated with 4% diluted sulfuric acid at 121 °C for 60 min to completely hydrolyse. The hydrolysate was filtered to determine the acid-soluble lignin, and the residue was measured for acid-insoluble lignin. Determination and calculation methods were the same as previously described (Wang et al., 2019). Thioacidolysis was used to analyse lignocellulosic samples and 2D heteronuclear single quantum coherence (HSQC) spectroscopy was performed following the methods used by Yue et al. (2012), Kim and Ralph (2010), and Mansfield et al. (2012).
RNA-seq
Six-week-old Arabidopsis stem bases of Col-0 and knat3 knat7 plants were collected and ground to powder in liquid nitrogen. Total RNA extraction and RNA-seq were performed by the Biomarker Technologies Corporation (Beijing, China). Differential gene expression analysis and Gene Ontology enrichment analysis were performed using BMKcloud (international.biocloud.net). The TAIR10 genome was used as the Arabidopsis reference genome (www.arabidopsis.org) and the RNA-seq data were submitted to http://bigd.big.ac.cn/gsa/ with submission number CRA002075.
RT-PCR and qRT-PCR analyses
Total RNA was isolated using a plant RNA kit (Omega, Norcross, GA, USA) according to the manufacturer’s instructions. First-strand cDNA synthesis was carried out with approximately 2 μg RNA using the Primer Script RT Reagent Kit with gDNA Eraser (Takara, Kyoto, Japan). RT-PCR was conducted with first-strand cDNA as the template, and amplified fragments were confirmed using agarose gel electrophoresis. The qRT-PCR was performed with the SYBR Premix Ex Taq II (Takara) on a LightCycler480 Real-Time PCR System (Roche, Basel, Switzerland). Primers used for RT-PCR and qRT-PCR are listed in Supplementary Table S1.
Yeast two-hybrid assays
To test the interaction between KNAT3, KNAT7, NST1, and NST2 proteins, the bait constructs pDEST32-KNAT3 and pDEST32-KNAT7 were generated by LR reaction between pDONR207-KNAT3, pDONR207-KNAT7 plasmids and pDEST32 (Thermo Fisher Scientific). The prey constructs pDEST22-NST1, pDEST22-NST2, pDEST22-KNAT3, and pDEST22-KNAT7 were generated by LR reaction between pDONR207-NST1, pDONR207-NST2, pDONR207-KNAT3, pDONR207-KNAT7, and pDEST22 (Thermo Fisher Scientific). Bait plasmids and prey plasmids or blank pDEST22 were co-transformed into yeast strain AH109. Medium supplemented with SD−Leu−Trp−His and 2.5 mM 3-amino-1,2,4-triazole was used for selection (Tao et al., 2013).
Yeast one-hybrid assay
To test whether KNAT3 and KNAT7 could bind the promoter of the F5H gene directly, a yeast one-hybrid assay was performed. The coding sequences (CDS) of KNAT3 and KNAT7 genes were amplified by PCR and were inserted into EcoRI/XhoI sites of pB42AD vector. The 2171 bp promoter region of the F5H gene was amplified by PCR and inserted to EcoRI/XhoI sites of pLacZi2µ vector. All the constructions were confirmed by sequencing. The corresponding pB42AD and pLacZi2µ fusion constructions were co-transformed into yeast strain EGY48 containing p8op-LacZ plasmid. The transformants were spread on SD/−His−Trp−Ura dropout plates. After 3 d growth at 30 °C, the transformants were copied to SD/−His−Trp−Ura dropout plates containing 80 mg l−1 X-gal for blue color development. pB42AD-CCA1 and pLacZi2µ-proELF4 combination from (Zhao et al., 2018) was used as positive control.
Firefly split luciferase complementation imaging
The full-length CDSs of KNAT3, KNAT7, NST1, and NST2 were amplified from Arabidopsis cDNA and then cloned into pDONR207 to generate the pENTRY construct. nLUC-KNAT3 or cLUC-KNAT7/NST1/NST2 was generated by LR reactions between pCB1300-nLUC-GW or pCB1300-cLUC-GW and pDONR207-NST1, pDONR207-NST2, pDONR207-KNAT3, pDONR207-KNAT7 (Thermo Fisher Scientific). These plasmids were transformed into Agrobacterium strain GV3101. Different combinations of plasmids were co-infiltrated into tobacco (Nicotiana benthamiana) leaves. The empty pCB1300-BD-nLUC and pCB1300-cLUC-GW vectors were used as negative controls. After incubation in dark for 24 h and then in light for 72 h, the tobacco leaves were sprayed with 100 mM D-luciferin and kept in dark for 10 min (Han et al., 2019). The fluorescence was detected using a low-light cooled charge-coupled device (CCD) imaging apparatus (NightOWL818 II LB983) with indiGO software.
Transactivation analyses
A 2171bp promoter sequence of F5H was amplified from Arabidopsis genomic DNA, then recombined into pGreen II 0800-LUC vector, and used as reporter plasmids (Hellens et al., 2005). The coding sequence of KNAT3/KNAT7 was amplified by PCR, inserted into pGreen II 62-SK, and used as an effector plasmid. The effector and reporter were co-transfected into Arabidopsis leaf protoplasts (Yoo et al., 2007). After placing them under low-light conditions for 14–16 h, the protoplasts were lysed and the supernatants were assayed for LUC and REN activities with the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA). The LUC/REN ratio indicates transcriptional activity.
Co-immunoprecipitation and immunoblot analyses
Total proteins from Nicotiana benthamiana were extracted using IP buffer containing 150 mM NaCl, 50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 10% glycerol, 1% Triton X-100, 0.2% NP-40, 5 mM DTT, and 1× complete protease inhibitor cocktail. Anti-FLAG M2 Magnetic Beads (Sigma, M8823) were used for immunoprecipitation for hemagglutinin (HA)- and FLAG-tagged proteins. The affinity beads were washed with washing buffer (150 mM NaCl, 50 mM Tris–HCl pH 7.5, 5 mM EDTA, 1% Triton X-100, 0.2% NP-40, and 1× complete protease inhibitor cocktail). The protein extract was incubated with the affinity beads for 4 h at 4 °C with gentle mixing to capture the proteins. Anti-FLAG (Abmart, M20008) and anti-HA (Abmart, M20003) antibodies from mice were used to detect input proteins; anti-FLAG (Sigma, F7452) and anti-HA (Sigma, H6908) antibodies from rabbits were used to detect immunoprecipitated proteins. Immunoblots were performed according to the enhanced chemiluminescence western blotting procedure.
Results
Chimeric KNAT3 and KNAT7 repressors induced similar dwarf phenotype
To examine the paradox of KNAT7 negatively and positively regulating secondary cell wall biosynthesis (Zhong et al., 2006, 2007, 2008; Li et al., 2011, 2012; Zhong and Ye, 2012; He et al., 2018), the KNOX II proteins KNAT3, KNAT4, KNAT5, and KNAT7 (Furumizu et al., 2015) were fused with the EAR-motif repressor domain SRDX (Hiratsu et al., 2003, 2004), which converted the individual KNOXII transcription factors into dominant repressors. The chimeric KNOX II repressor cassettes, driven by the cauliflower mosaic virus (CaMV) 35S promoter (p35S:KNOX II-SRDX), were expressed in Arabidopsis (see Supplementary Fig. S1A). Among all four KNOX II-SRDX transgenic plants, only p35S:KNAT3-SRDX and p35S:KNAT7-SRDX showed morphological resemblance, displaying a similar dwarf and bush phenotype (Supplementary Fig. S1B), which suggests a probable functional redundancy of KNAT3 and KNAT7 in inflorescence stems.
KNAT3 and KNAT7 were co-expressed in the interfascicular fiber and xylem
To validate the speculated functional redundancy of KNAT3 and KNAT7 in inflorescence stems, we first examined the expression of KNAT3 and KNAT7 genes in Arabidopsis. The relative transcription levels of KNAT3 and KNAT7 in different tissues were detected by qRT-PCR. The results showed that the expression level of KNAT7 in the stem was much higher than that of KNAT3, while compared with KNAT7, the expression level of KNAT3 in the cauline and senescent leaves was higher (Fig. 1A). To further check expression patterns of KNAT3 and KNAT7 in inflorescence stems, the promoter regions of KNAT3 or KNAT7 were fused with the GUS gene; the resultant vectors pKNAT3:GUS and pKNAT7:GUS were then transferred into wild-type plants. Despite the pKNAT3:GUS showing diffusion in phloem and other cells, GUS staining for both pKNAT3:GUS and pKNAT7:GUS was observed in both xylem and interfascicular fiber cells (Fig. 1D, E). These results confirmed that KNAT3 and KNAT7 were co-expressed in the inflorescence stems.
The expression pattern of KNAT3 and KNAT7. (A) Relative expression levels of KNAT3 and KNAT7 at different tissues in 6-week-old Arabidopsis measured by qRT-PCR. Column height indicates the mean value ±SD of the sample expression; n=4. The expression level of KNAT3 in the WT was assigned a value of 1, and ACT8 was used as an internal reference. (B, C) Schematic diagrams of the pKNAT3:GUS and pKNAT7:GUS cassette, respectively. On the left is the promoter, in the middle the GUS gene, and on the right the terminator. (D, E) GUS staining for pKNAT3:GUS (D) and pKNAT7:GUS (E); samples were from the base of the 6-week-old stem. If, interfascicular fibers; x, xylem. Scale bar: 20 μm.
knat3 knat7 double mutant exhibits severe irregular xylem phenotype
To further analyse the functional redundancy between KNAT3 and KNAT7, the double mutant was obtained by crossing knat3 (SALK_136464) and knat7 (SALK_110899) single mutants. RT-PCR analysis showed that transcript fragments of KNAT3 and KNAT7 in knat3 and knat7 single mutants were not amplified (Fig. 2C), indicating that KNAT3 and KNAT7 cannot form intact mRNA due to T-DNA insertions, and thus these two lines were functionally knock-out mutants. Morphologically, knat3 and knat7 single mutants had a similar height to the wild type (WT), while the knat3 knat7 double mutant was significantly shorter than the WT (Fig. 2D, E).
The morphology of knat3 knat7 double mutant. (A, B) Schematic diagrams representing the T-DNA insertion positions of knat3 (A) and knat7 (B). The boxes and lines indicate exons and introns, respectively, and the triangle indicate the T-DNA insertion position labeled with the Salk number. (C) RT-PCR results showed that the fragments of knat3 and knat7 were not amplified in the corresponding mutants compared with WT; ACT2 was used as an internal reference gene. (D) The height of 6-week-old WT, knat3, knat7, and knat3 knat7. Scale bar: 5 cm. (E) Compared with the WT, there was no significant difference in the height of knat3 and knat7, but the height of knat3 knat7 was significantly shorter at the same stages. Data are means with SD of 10 plants. ***Significant difference at P<0.01 compared with wild-type plants, by t-test. (F) Complementation of KNAT3 and KNAT7 can rescue the height of knat3 knat7.
To elucidate the reason for the difference in height between knat3 and knat7 mutant combinations, cross-sections were taken from the base of the inflorescence stems of 6-week-old WT, knat3, knat7, and knat3 knat7 plants. knat3, like the WT, exhibited a normal vessel phenotype, and knat7 showed slightly weak irregular vessels (Fig. 3A–C), whereas severe irregular xylem vessels were clearly observed in the knat3 knat7 double mutants (Fig. 3D). The thickness of xylem (vessel and xylary) and interfascicular fiber cell walls of WT, knat3, knat7, and knat3 knat7 was further measured. Compared with WT, knat7 had thicker interfascicular fiber cell walls but thinner vessel and xylary cell walls in inflorescence stems (Fig. 3E–G), which is consistent with previous reports (Li et al., 2012; He et al., 2018). On the contrary, the cell wall thickness of interfascicular fibers was reduced in knat3, while all the vessel, xylary, and interfascicular fiber cell walls were significantly thinner in the knat3 knat7 double mutant. These results indicate that KNAT3 and KNAT7 have functional redundancy in regulating at least the xylem vessel and xylary secondary cell wall development.
The knat3 knat7 double mutant shows severe irregular xylem. (A–D) Cross-sections of 6-week-old Arabidopsis base stems stained with phloroglucinol: (A) WT, (B) knat3, (C) knat7, (D) knat3 knat7. The upper panel shows the interfascicular fiber structure, and the lower panel shows the xylem tissue. Scale bar: 20 µm. (E–G) Cell wall thickness of different cells: vessel cell (E), xylary fiber cell (F), and interfascicular fiber cell (G). Data are mean with SD of 50 cell wall thicknesses. Significant difference compared with wild-type plants: *P<0.05, **P<0.01; t-test. (H) Cross-sections of complementation plant by overexpressed KNAT3 and KNAT7 in knat3 knat7 double mutant.
Complementation was performed to investigate whether the phenotype of the knat3 knat7 double mutant was indeed caused by knockout of the corresponding genes. The full-length coding sequences of KNAT3 and KNAT7 were fused with CaMV35S promoter and transferred to the knat3 knat7 mutant. The transgenic lines fully rescued the dwarfism and the irregular xylem phenotype of the mutant (Figs 2F, 3H). These results confirm that the phenotype of the knat3 knat7 mutant is indeed caused by the loss of function of KNAT3 and KNAT7.
KNAT3 and KNAT7 synergistically regulate syringyl lignin biosynthesis
Mäule staining was performed on stem cross-sections of p35S:KNAT3-SRDX and p35S:KNAT7-SRDX repressor transgenic plants to examine chemical changes in cell wall composition of knat3 and knat7 mutants (see Supplementary Fig. S2). The wild-type xylem was stained bright red, the p35S:KNAT7-SRDX xylem was stained weak brown-red, and the p35S:KNAT3-SRDX xylem was stained yellow-brown (Supplementary Fig. S2). These results indicate that the synthesis of S-lignin might be significantly inhibited in KNAT3 repressor transgenic plants.
To verify whether the lignin defect in the chimeric KNAT3 and KNAT7 repressors corresponds to the knat3 knat7 mutation, the lignin content of KNAT3 and KNAT7 mutants was further determined. The results showed that acid-soluble lignin in the knat3 and knat3 knat7 mutants was significantly reduced compared with the WT. However, no significant changes were found in the acid-insoluble lignin and total amount of lignin in the knat3, knat7, and knat3 knat7 mutants compared with WT (Fig. 4A–C). Furthermore, Mäule staining showed that the stem section of the knat3 knat7 mutant stained yellow-brown (Fig. 4D–G), which is consistent with the observation in domain repressor plants. The stem cross-sections of KNAT3 and KNAT7 complementation plants in the knat3 knat7 double mutant rescued the color changes from yellow-brown to bright-red (see Supplementary Fig. S3). These results indicate a decrease in S-lignin in the knat3 knat7 double mutant.
Lignin content and Mäule staining of WT, knat3, knat7, and knat3 knat7. (A–C) Percentage content of acid-soluble lignin (A), acid-insoluble lignin (B), and total lignin (C) in the main stem of 8-week-old plants. Three biological replicates were included, and the significance analysis was based on Student’s t-test. *Significant difference at P<0.05 compared with WT plants. (D–G) Cross-sections from the base of 6-week-old stem with Mäule staining. It can be observed that knat3 knat7 and WT were differentially stained. Scale bar: 20 μm. (H) Aromatic regions of 2D 13C– 1H correlation (HSQC) spectra from cell wall gels from various samples in DMSO-d6/pyridine-d5 (4:1). The data showed that S-lignin decreased and G-lignin increased in knat3 knat7. (I) Chemical formulae corresponding to the different components in (H).
To determine the amount of S-lignin reduction, GC-MS was used to detect the content of S- and G-lignin monomers (Table 1). G-lignin was significantly reduced in knat3, but significantly increased in knat7 and knat3 knat7. Due to significant down-regulation of both G-lignin and S-lignin in knat3, the S/G ratio showed no change whereas the S/G ratio in knat7 decreased because of an increase in G-lignin and a slight decrease in S-lignin (Table 1). Furthermore, the S-lignin content was further reduced by about 80% in the knat3 knat7 double mutant compared with WT, which resulted in an 84% reduction of the S/G ratio in knat3 knat7. Additional 2D HSQC spectra of WT and knat3 knat7 cell wall samples were collected to determine the lignin differences (Fig. 4H). The correlations in the HSQC spectra of whole cell wall preparations were assigned by comparison with published data (Kim and Ralph, 2010; Mansfield et al., 2012). As shown in Fig. 4H, both G- and S-lignin signals were detected in the WT cell walls, although the amount of S units was much lower than that of G units, which is consistent with thioacidolysis analysis (Table 1). However, in HSQC spectra of knat3 knat7, correlations of the S unit (S2/6) were completely absent, while the correlations of G unit and H unit lignin subunits were much stronger than those of WT, which indicates a much higher amount of G-lignin in knat3 knat7. Overall, 2D HSQC confirmed an increase in G-lignin and a decrease in S-lignin in the knat3 knat7 mutant.
Determination of lignin monomer
| Sample | G (µmol g−1) | S (µmol g−1) | S+G (µmol g−1) | S/G |
|---|---|---|---|---|
| knat3 | 24.6±0.36*** | 10.4±0.56*** | 35±0.78*** | 0.42±0.021 |
| knat7 | 49.3±0.45*** | 14.2±0.79* | 63.48±1.24*** | 0.29±0.014*** |
| knat3 knat7 | 46.8±1.04*** | 3.2±0.26*** | 50±0.92 | 0.07±0.006*** |
| WT | 36.7±0.81 | 16.2±0.70 | 52.9±1.37 | 0.44±0.015 |
| Sample | G (µmol g−1) | S (µmol g−1) | S+G (µmol g−1) | S/G |
|---|---|---|---|---|
| knat3 | 24.6±0.36*** | 10.4±0.56*** | 35±0.78*** | 0.42±0.021 |
| knat7 | 49.3±0.45*** | 14.2±0.79* | 63.48±1.24*** | 0.29±0.014*** |
| knat3 knat7 | 46.8±1.04*** | 3.2±0.26*** | 50±0.92 | 0.07±0.006*** |
| WT | 36.7±0.81 | 16.2±0.70 | 52.9±1.37 | 0.44±0.015 |
Data are mean ±SD, including three replicates. Eight-week-old main stem of Arabidopsis was selected to determine lignin monomer content. Significant differences from WT are shown: *P<0.05, ***P<0.001, Student’s t-test. G, guaiacyl lignin; S, syringyl lignin.
Determination of lignin monomer
| Sample | G (µmol g−1) | S (µmol g−1) | S+G (µmol g−1) | S/G |
|---|---|---|---|---|
| knat3 | 24.6±0.36*** | 10.4±0.56*** | 35±0.78*** | 0.42±0.021 |
| knat7 | 49.3±0.45*** | 14.2±0.79* | 63.48±1.24*** | 0.29±0.014*** |
| knat3 knat7 | 46.8±1.04*** | 3.2±0.26*** | 50±0.92 | 0.07±0.006*** |
| WT | 36.7±0.81 | 16.2±0.70 | 52.9±1.37 | 0.44±0.015 |
| Sample | G (µmol g−1) | S (µmol g−1) | S+G (µmol g−1) | S/G |
|---|---|---|---|---|
| knat3 | 24.6±0.36*** | 10.4±0.56*** | 35±0.78*** | 0.42±0.021 |
| knat7 | 49.3±0.45*** | 14.2±0.79* | 63.48±1.24*** | 0.29±0.014*** |
| knat3 knat7 | 46.8±1.04*** | 3.2±0.26*** | 50±0.92 | 0.07±0.006*** |
| WT | 36.7±0.81 | 16.2±0.70 | 52.9±1.37 | 0.44±0.015 |
Data are mean ±SD, including three replicates. Eight-week-old main stem of Arabidopsis was selected to determine lignin monomer content. Significant differences from WT are shown: *P<0.05, ***P<0.001, Student’s t-test. G, guaiacyl lignin; S, syringyl lignin.
The key S-lignin biosynthetic gene F5H was down-regulated in knat3 knat7 mutant
To further explain the decrease of S-lignin in the knat3 knat7 mutant, the main stems of 6-week-old WT and knat3 knat7 mutant were collected for RNA sequencing. Data analysis showed that the differentially expressed genes were significantly enriched in the phenylalanine metabolic pathway (see Supplementary Fig. S4), which is consistent with the phenotype of knat3 knat7 mutant showing a significant decrease in the proportion of lignin S/G. Interestingly, the expression profiles of lignin metabolic pathway genes CCR, COMT, CCoAOMT, and CAD were up-regulated (Fig. 5A) and expression levels of 4CL1 and F5H were down-regulated. The increasing G-lignin content in knat3 knat7 could be due to the activation of lignin pathway genes, such as G-lignin-related CCoAOMT. Because F5H is a key enzyme for the synthesis of S-lignin (Franke et al., 2000; García et al., 2014), the reduced F5H expression might be attributed to the decrease of S-lignin content in the knat3 knat7 mutant. Furthermore, the RNA-seq results were verified by qRT-PCR analysis of lignin biosynthetic genes in WT and knat3 knat7 mutant, in which F5H expression was reduced by 77% (Fig. 5B).
S-lignin synthesis gene F5H was down-regulated. (A) RNA-seq data showing the expression changes of lignin pathway gene. The results were based on the difference of multiple fold change ≥2 and the error rate FDR <0.01 as the screening criteria. (B) qRT-PCR was used to detect the relative expression of lignin metabolic pathway genes between WT and knat3knat7 mutant. PAL, Phenylalanine Ammonia Lyase; C4H, Cinnamate 4-Hydroxylase; 4CL, 4-Coumarate CoA Ligase; C3H, Coumaroyl Shikimate 3′-Hydroxylase; CAD, Cinnamyl Alcohol Dehydrogenase; F5H, Ferulate 5-Hydroxylase. The expression level of each gene in the WT was assigned a value of 1 and EF1α was used as an internal reference. The data represent the mean and SD of three biological repetitions and four technical repetitions.
KNAT3 has a main effect in regulating the expression of F5H gene
Because the significant decrease of S-lignin in knat3 knat7 might be caused by the down-regulation of F5H expression, we explored whether KNAT3 and KNAT7 can regulate the expression of the F5H gene. The 2171 bp upstream region of the F5H gene was cloned and inserted into a reporter vector harboring the luciferase gene (LUC). p35S:KNAT3 and p35S:KNAT7 were constructed as effectors (Fig. 6A), and transiently transformed into Arabidopsis protoplast cells. The results showed that both KNAT3 and KNAT7 can activate LUC expression, that KNAT3 had higher activity than KNAT7, and that the KNAT3+KNAT7 combination showed the highest activation (Fig. 6B). These results indicate that KNAT3 and KNAT7 can work on the F5H promoter to regulate LUC expression.
KNAT3 and KNAT7 can form a dimer to activate F5H expression. (A) Schematic representation of the transcriptional activation reporter and effector system. (B) The activation was carried out by adding different reporters and effectors to the Arabidopsis protoplasts. The ratio of LUC and REN in the CaMV35S′-empty cells with only reporters was used as control and assigned a value of 1. Column height indicates the mean ±SD of the relative LUC/REN ratio with three biological replicates; n=4. *P<0.05, **P<0.01, Student’s t-test. (C) Yeast one-hybrid assay detects KNAT3 and KNAT7 binding with the F5H promoter. Positive binding showed as blue after X-gal was added. As a negative control (AD), KNAT3 and KNAT7 cannot directly bind to the F5H promoter. (D, E) Interaction of KNAT3 and KNAT7 proteins was examined in yeast cells, with KNAT3 CDS and GAL4 DNA-binding region (DBD) fused as a bait protein, and GAL4 transcriptional activation region (AD) fused with KNAT3 or KNAT7 protein used as prey (D), or the opposite of bait and prey (E). After transformation, the yeast was spotted on SD−Leu−Trp (−2) medium and SD−Leu−Trp−His (−3) screening medium containing 2.5 mm 3-aminotriazole, and AD was used as a negative control. KNAT7 and KNAT3 can form homodimers and heterodimers. (F) KNAT3 and KNAT7 interact in plants, and the four quadrants show different transformation combinations. Only the mixture of KNAT3 and KNAT7 proteins by split luciferase complementation shows fluorescence.
In order to detect whether KNAT3 and KNAT7 can directly or indirectly activate the F5H gene, a yeast one-hybrid system was used to check the binding effects. Unexpectedly, neither KNAT3 nor KNAT7 bound directly to the F5H promoter region (Fig. 6C). Therefore, we hypothesized that KNAT3 and KNAT7 might activate other genes or form a protein complex to regulate the expression of F5H. Up until now there are no proteins reported to directly regulate F5H in Arabidopsis, so we speculated that the regulation of F5H may be accomplished by a larger complex.
KNOXs, as members of the TALE homeodomain superfamily (Hake et al., 2004; Hay and Tsiantis, 2010), have the potential to form homo- or heterodimers. We found that KNAT3 and KNAT7 can form homodimers and a heterodimer in a yeast two-hybrid assay (Fig. 6D, E), as reported by Hackbusch et al. (2005). Split luciferase complementation experiments verified that KNAT3 and KNAT7 form dimers in planta as well (Fig. 6F). When KNAT3 and KNAT7 were introduced simultaneously into the Arabidopsis protoplast cells, we found that the mixture of KNAT3 and KNAT7 showed a stronger activation effect than KNAT3 alone (Fig. 6B), suggesting that the heterodimer might be more effective than the homodimers.
KNAT3 but not KNAT7 interacts with NST1 and NST2 to regulate the expression of F5H
Since neither KNAT3 nor KNAT7 directly binds to the F5H promoter but showed weak activation of F5H expression, we deduced that some proteins or factors in the regulation system had been missed. Hence, we used KNAT3 as a bait protein to screen an Arabidopsis transcription factor library by yeast two-hybrid screening (Mitsuda et al., 2010). NST1 and NST2 were identified in screening, and these have been regarded as the top-level switches to regulate secondary cell wall formation (Mitsuda et al., 2005, 2006; Zhong et al., 2010; Zhou et al., 2014; Zhong and Ye, 2015). We further verified the interactions between KNAT3/7 and NST1, NST2, and SND1 (also named as NST3) by yeast two-hybrid, split-luciferase complementation, co-immunoprecipitation, and immunoblot analyses (Fig. 7A–E). The results showed that only KNAT3, but not KNAT7, interacted with NST1 and NST2. Neither KNAT3 nor KNAT7 can interact with SND1 (Fig. 7A–E).
KNAT3/7 interact with NST1/NST2 proteins. (A) KNAT3 and NST1/NST2 proteins interact by yeast two-hybrid assay, with KNAT3 CDS and GAL4 DNA-binding region (DBD) fusion as bait protein and GAL4 transcriptional activation region (AD) fused with NST1/NST2/SND1 protein as prey. The transformed yeast was spotted in SD−Leu−Trp (−2) medium and SD−Leu−Trp−His (−3) with 2.5 mm 3-aminotriazole, and AD was used as a negative control. (B) KNAT7 cannot interact with NST1/NST2/SND1 in yeast cells. (C, D) KNAT3 and NST1/NST2 interact in plants. The four quadrants show different combinations of transformation constructs. (E) Co-immunoprecipitation assay of KNAT3–HA and NST1/2–FLAG. The proteins were immunoprecipitated with anti-FLAG antibody and then probed with anti-HA antibody. (F) Transcriptional activation experiments in which reporters tested different effectors in Arabidopsis protoplasts. CaMV35S′-empty with only reporter was used as control with the ratio of LUC and REN assigned a value of 1 for normalization. Data are mean ±SD of the relative LUC/REN ratio of each combination with three biological replicates; n=4. **P<0.01, Student’s t-test.
Furthermore, we investigated whether the heterologous complex formed by KNAT3 and NST1/NST2 activates F5H expression. We performed a transcriptional activation experiment in Arabidopsis protoplasts. The results revealed that NST1 and NST2 alone could weakly activate the F5H promoter, and no further activation was gained after KNAT3 was added (Fig. 7F). This indicates that some factors in the regulation of F5H expression might still be missing and further exploration is required.
Discussion
There are four members of Arabidopsis KNOX class II genes, among which KNAT3, KNAT4, and KNAT5 play a functionally redundant role in regulating leaf morphology, and the serrated leaf phenotypes knat3, knat3 knat5, and knat3 knat4 knat5 show a gradual progression (Furumizu et al., 2015). However, the effects of KNAT3, KNAT4, and KNAT5 on secondary wall formation have not been reported. KNAT7, also called IRX11, has been reported to be linked with vessel cell development as its mutant showed inwardly collapsed vessel cells forming irregular xylem (Brown et al., 2005). Some reports revealed that KNAT7 acts as a transcriptional repressor to inhibit secondary cell wall deposition (Li et al., 2011, 2012), but the knat7 mutant showed thinner vessel cells despite thicker interfascicular fiber cells (Li et al., 2012). KNAT7 can also act as a transcriptional activator to activate the expression of xylan synthesis genes (He et al., 2018). To better explain this contradiction, we further combined other KNOX II members with KNAT7 in this study. We found that KNAT3 and KNAT7 had obvious functional redundancy in regulating the synthesis of Arabidopsis secondary cell walls. Coincidentally, Wang et al., 2020a recently reported KNAT3 and KNAT7 work cooperatively to influence secondary cell wall deposition. Their work was mainly focused on the functions of KNAT3 and KNAT7 in secondary cell wall synthesis, including hemicellulose and lignin synthesis. They used phenotypic and chemical analysis results and their conclusions are in close agreement with ours. In the present study, we focused on lignin, especially the S-lignin subunit, and we explained the reason why the S-subunit lignin decreases in mutants by protein interaction and pathway analysis.
Dominant repressor plants p35S:KNAT3-SRDX and p35S:KNAT7-SRDX displayed severe defects in stem development and both showed similar dwarf phenotypes (see Supplementary Fig. S1) indicating that these genes exhibit functional redundancy. The expression patterns demonstrated that KNAT3 and KNAT7 were co-expressed in secondary xylem and interfascicular fiber tissues (Fig. 1). The loss-of-function knat3 knat7 double mutant showed significantly reduced height at 6 weeks of age (Fig. 2D, E), which is consistent with KNAT3 and KNAT7 dominant repressor plants. In addition to irregular xylem, the knat7 mutant showed a thicker interfascicular fiber wall, thinner xylem vessel wall, and increased lignin content (Li et al., 2011; He et al., 2018). In the literature there are reports that KNAT7 interacts with MYB75, BLH6, and OFP4, forming a complex to inhibit secondary cell wall formation (Hackbusch et al., 2005; Bhargava et al., 2010; Li et al., 2011; Liu et al., 2014), which indicates that KNAT7 has complicated and spatiotemporally differentiated functions based on specific cell type, tissue, and interacting regulators. knat3 differs from knat7 in that it does not have an irregular xylem phenotype but has a thinner secondary cell wall of the interfascicular fiber cells. However, knat3 knat7 has an enhanced irx phenotype (Fig. 3) and significantly reduced cell wall thickness in vessels, xylary and interfascicular fibers as compared with knat7. This phenotype indicated that KNAT3, as a homolog of KNAT7, has functional redundancy in the development of secondary cell wall, and KNAT3 is a transcriptional activator of secondary wall thickening. We found KNAT3 and KNAT7 proteins can form heterodimers in the yeast two-hybrid assay and that only KNAT3 interacts with NST1 and NST2, the two mast regulators involved in regulation of secondary cell wall thickening (Mitsuda et al., 2005; Zhong and Ye, 2015). Based on these results, we propose a working model of KNAT3 and KNAT7 as shown in Fig. 8. KNAT3 interacts with NST1/2, the positive regulators of the secondary cell wall, but KNAT7 interacts with MYB75/BLH6/OFP4, the negative regulators of the secondary cell wall. Interestingly, secondary cell wall thickness in interfascicular fibers was reduced in the nst1 nst3 double mutant (Zhong et al., 2007; Zhong and Ye, 2015) meaning that NST1 can specifically function in this region. Thus, we speculate that, in the knat7 mutant, a KNAT3–NST1 complex stimulates downstream genes to promote cell wall synthesis in interfascicular fibers and produces thicker interfascicular fiber wall. At the same time, NST1 is a fiber-specific transcription factor and KNAT3 interacts with NST1 there. S-lignin is predominantly deposited in fiber regions, so the knat3 mutant showed a lower fiber cell wall thickness (Fig. 3F, G) and less S-lignin (Table 1).
Hypothetical model of KNAT3/KNAT7 synergistical regulation of S-lignin biosynthesis. KNAT3 and KNAT7 can form a heterodimer. The proteins MYB75, OFP4, and BLH6 were reported as negative regulators interacting with KNAT7 to repress F5H in some cell types and the proteins NST1/2 were reported as positive regulators interacting with KNAT3 to activate F5H in specific cells.
In dicots, lignin is mainly composed of S- and G-lignin subunits with a very low concentration of H-lignin (Vogel, 2008). Acid-soluble lignin was preferentially derived from the condensed syringyl lignin while the guaiacyl lignin was insoluble in 72% sulfuric acid (Yue et al., 2012). It is possible that acid-soluble lignin has a relatively high S-lignin content, and acid-insoluble lignin has a high G-lignin content (Fig. 4A). Thus, the lower acid-insoluble lignin might result from the lower content of S-lignin in knat3 and knat3 knat7, as compared with knat7 and WT (Fig. 4A; Table 1). Some reports suggest that KNAT7 is a negative regulator of lignin biosynthesis and the loss-function mutant of knat7 had increased lignin content (Brown et al., 2005; Li et al., 2011). Our determination of lignin content evidenced that the increase of total lignin content in knat7 was due to a large increase of G-lignin with concomitant slight decrease of S-lignin. The contents of G-, S- and total lignin in knat3 were significantly reduced, but the lignin S/G ratio remained unchanged, indicating KNAT3 positively regulates lignin biosynthesis. Although the total lignin in knat3 knat7 remained unchanged, the S-lignin was greatly reduced and the G-lignin was increased, eventually resulting in an 84% reduction of the S/G ratio (Fig. 4; Table 1). Therefore, we believe that KNAT3 synergizes with KNAT7 in lignin synthesis, and KNAT3 plays a dominant role in controlling S-lignin synthesis.
In the lignin synthesis pathway, F5H is a key enzyme responsible for hydroxylation of G-lignin monomer to form S-lignin (Humphreys et al., 1999; Osakabe et al., 1999; Franke et al., 2000; García et al., 2014). Our transcriptomic data of knat3 knat7 indicated that the differentially expressed genes were significantly enriched in the phenylalanine metabolic pathway (see Supplementary Fig. S4) and most of the gene expression was up-regulated, while the expression of 4CL and F5H was down-regulated (Fig. 5). Therefore, the down-regulation of F5H could contribute to the increased G-lignin in knat7 mutant with unchanged total lignin content in most mutants. This is also a good explanation for the increase in G-lignin content and the large reduction of S-lignin in knat3 knat7. We also checked the profiles of all MYB transcription factors in our transcriptome data. We found that MYB20, MYB69, and MYB86 were down-regulated, but both MYB46 and MYB83, regarded as activators of secondary wall biosynthesis, and MYB75, MYB4 and MYB7, regarded as repressors of secondary wall biosynthesis, were up-regulated (Supplementary Table S2; McCarthy et al., 2009; Bhargava et al., 2010; Ko et al., 2014; Wang et al., 2020b). Thus, KNAT3 and KNAT7 might have a complicated regulatory network or feedback network. Because the myb103 mutant can also decrease F5H gene expression (Öhman et al., 2013), it is unclear whether MYB103 works upstream or downstream of KNAT3/KNAT7 in regulating F5H gene expression. According to our transcriptome and qPCR analysis, the MYB103 expression level showed no obvious difference in the knat3 knat7 double mutant, compared with the WT (Supplementary Table S2; Supplementary Fig. S5A, B), and the expression levels of KNAT3 and KNAT7 increased in the myb103 mutant (Supplementary Fig. S5B). The increased KNAT7 expression is especially obvious. The cause of increased KNAT3/7 expression might be due to compensation. All the results indicate that MYB103 and KNAT3/KNAT7 might function in independent pathways to regulate F5H expression.
Transcriptional activation experiments in Arabidopsis protoplasts revealed that both KNAT3 and KNAT7 weakly activated the F5H promoter, suggesting that KNAT3 and KNAT7 indirectly regulate F5H (Fig. 6). It is known that MYB58 and MYB63 directly activate the expression of lignin synthetic genes by binding to the AC element or degenerated AC element of the promoter regions of those genes (Zhou et al., 2009; Zhao and Dixon, 2011). However, there is no AC element in the F5H promoter region, except that NST1 and SND1 were reported to bind directly to the F5H promoter region to activate F5H expression in Medicago truncatula (Zhao et al., 2010). Arabidopsis SND1 cannot directly activate F5H expression (Öhman et al., 2013). Our yeast one-hybrid experiments showed that neither KNAT3 nor KNAT7 could directly bind to the promoter of F5H. Only KNAT3, not KNAT7, interacts with the top secondary cell wall regulatory factors NST1/NST2. Although NST1 and NST2 can weakly activate F5H expression, some effectors in the regulatory cascade or complex may still be missing (Fig. 7). The F5H regulation mechanism remains to be further explored.
As we know, fibers are enriched in S-lignin, whereas guaiacyl-enriched lignin is mainly deposited in vessels (Fergus and Goring, 1970a, b; Musha and Goring, 1975; Saka and Goring, 1985). Furthermore, the deposition of S-lignin precedes that of G-lignin in interfascular fiber elements (Terashima et al., 1986; Saka and Goring, 1988). NST1 and NST2 are fiber-specific transcription factors, which interact with KNAT3, and this could be the reason for S-lignin biosynthesis via activation of F5H in the fiber. Although the knat3 knat7 mutant has collapsed vessels as with the knat7 mutant, both mutants contain increased G-lignin content, and the expression of G-lignin pathway genes, such as CCoAOMT, is up-regulated as well. To investigate the cause of the increase in G-lignin, the expression of VASCULAR-RELATED NAC-DOMAIN 6 (VND6) and 7 (VND7), which work as master regulators of metaxylem and protoxylem vessel cell fates, respectively (Kubo et al., 2005; Ohashi-Ito et al., 2010), was examined in the transcriptome analysis. The result showed that the expression of VND6 and VND7 only slightly decreased in the knat3 knat7 mutant in comparison with WT (see Supplementary Fig S5A). Additionally, we did not detect interactions between KNAT7 and VND6/7 by yeast-two hybrid assay (data not shown). Thus, we deduce that the increasing G-lignin content in knat3 knat7 might be due to the increase of some G-lignin biosynthesis-related genes and down-regulation of F5H.
This study demonstrates that KNAT3 and KNAT7 play a synergistic role in regulating the growth and development of Arabidopsis secondary cell wall. The KNAT3 mutant enhances the phenotype of knat7 in the secondary cell wall of xylem vessels (Figs 3, 4). In addition, KNAT3 interacts with NST1/NST2 to promote expression of F5H to regulate the synthesis of S-lignin. In the synthesis of secondary cell walls including lignin, interactions such as KNAT7 and BLH6/MYB75/OFP4 may function as transcriptional repressors. On the other hand, KNAT3 synergizing with KNAT7 functions positively for the synthesis of S-lignin (Fig. 8). This study complements the transcriptional regulatory network of secondary cell walls and clarifies the regulation of S-lignin by KNAT3 and KNAT7.
Supplementary data
Supplementary data are available at JXB online.
Fig. S1. Chimeric repressors in KNAT3 and KNAT7.
Fig. S2. Mäule staining of cross-section of stem in dominant repressor transgenic plants.
Fig. S3. Mäule staining of cross-section of stem of knat3 knat7 complementary plants.
Fig. S4. Pathway enrichment analysis of WT and knat3 knat7 RNA-seq.
Fig. S5. Expression levels of KNAT3 and KNAT7 in myb103 mutant.
Table S1. Primer sequences used in this study.
Table S2. RNA-seq analysis of MYB differentially expressed genes (WT versus knat3 knat7)
Author contributions
A-MW and WQ designed the study and wrote the manuscript. WQ, QY, XZ, JC, FY, JH, LY, and LL, performed the experiments. FL, NM, and MO-T helped with the techniques and analysed the data.
Conflict of interest
The authors have no conflicts of interest to declare.
Data availability
All relevant data can be found within the manuscript and its supporting materials.
Acknowledgements
We thank the Arabidopsis Biological Resource Center for seeds of the T-DNA insertion mutant. Financial support for this work was obtained the National Natural Science Foundation of China (Grant Numbers 31870653, 31670670, 31811530009).
Comments