Regulation of lignin biosynthesis by an atypical bHLH protein CmHLB in Chrysanthemum

Abstract

Stem mechanical strength is one of the most important agronomic traits that affects the resistance of plants against insects and lodging, and plays an essential role in the quality and yield of plants. Several transcription factors regulate mechanical strength in crops. However, mechanisms of stem strength formation and regulation remain largely unexplored, especially in ornamental plants. In this study, we identified an atypical bHLH transcription factor CmHLB (HLH PROTEIN INVOLVED IN LIGNIN BIOSYNTHESIS) in chrysanthemum, belonging to a small bHLH sub-family — the PACLOBUTRAZOL RESISTANCE (PRE) family. Overexpression of CmHLB in chrysanthemum significantly increased mechanical strength of the stem, cell wall thickness, and lignin content, compared with the wild type. In contrast, CmHLB RNA interference lines exhibited the opposite phenotypes. RNA-seq analysis indicated that CmHLB promoted the expression of genes involved in lignin biosynthesis. Furthermore, we demonstrated that CmHLB interacted with Chrysanthemum KNOTTED ARABIDOPSIS THALIANA7 (CmKNAT7) through the KNOX2 domain, which has a conserved function, i.e. it negatively regulates secondary cell wall formation of fibres and lignin biosynthesis. Collectively, our results reveal a novel role for CmHLB in regulating lignin biosynthesis by interacting with CmKNAT7 and affecting stem mechanical strength in Chrysanthemum.

Chrysanthemum, CmHLB, CmKNAT7, lignin biosynthesis, secondary cell wall, stem mechanical strength

Introduction

Stem mechanical strength is one of the most critical traits responsible for stem quality of crops and ornamental plants. It plays a crucial role in lodging and insect resistance, and significantly affects the quality and yield of plants (Butron et al., 2002; Hu et al., 2017; Chen et al., 2021). For example, in rice, culm strength is essential for lodging resistance (Zhang et al., 2021). In Paeonia lactiflora, slender stems and feebleness lead to bent flower heads and seriously affect the quality of cut flowers (Zhao et al., 2012; D. Zhao et al., 2013). The basic principles of the molecular mechanisms behind stem strength have been well studied in field crops, such as rice (Li et al., 2003), maize (Xue et al., 2016), wheat (Kamran et al., 2018), and soybean (Liu et al., 2019). However, the regulation of stem strength in ornamental plants remains unclear.

Stem strength is affected by several factors, including, morphological characteristics, such as stem diameter and internode length; the anatomical structure of stems, such as the distribution and morphology of vascular tissues and cell wall thickness; and chemical constituents of stems, including cellulose, hemicellulose, lignin, pectin, and glycoproteins (Bethke et al., 2016; Xia et al., 2018; Hussain et al., 2020; Zhao et al., 2020; Zhang et al., 2021). Lignin is a unique component of the secondary cell wall, which is deposited mainly in the tracheary elements, fibres, and other highly specialized tissues (Bonawitz and Chapple, 2010; Cassan-Wang et al., 2013). Lignin is composed of three units, guaiacyl (G), syringyl (S), and p-hydroxyphenyl (H), polymerized from three monolignols, coniferyl, sinapyl, and p-coumaryl alcohols, respectively (Boerjan et al., 2003; Zhong and Ye, 2015). Moreover, lignin directly contributes to the mechanical strength of plant stems. For example, lignin content was decreased in the rice flexible culm1 (fc1) mutant, reducing the mechanical rigidity of culms (Li et al., 2009). Zhao et al. (2020) also found that lignin provides mechanical support to P. lactiflora stems. A set of enzymes, including 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) are involved in catalysing monolignol biosynthesis (Bonawitz and Chapple, 2010; Vanholme et al., 2013; Zhong and Ye, 2015). These enzymes affect lignin content or composition (Bonawitz and Chapple, 2010). In addition, peroxidases and laccases participate in lignin polymerization (Zhong and Ye, 2015). Several genes encoding peroxidases and laccases such as PEROXIDASE25 (PRX25), PRX71, LACCASE4 (LAC4), LAC11, and LAC17, have been reported to affect lignin content in Arabidopsis (Shigeto et al., 2013, 2014; Q. Zhao et al., 2013).

The PACLOBUTRAZOL RESISTANCE (PRE) family belongs to atypical or non-DNA binding bHLH transcription factors, with a conserved HLH domain for protein interaction, but lacking the basic region for DNA binding (Mara et al., 2010; Toledo-Ortiz et al., 2003). In Arabidopsis, six members of the PRE family, PRE1 to PRE6, have been identified to be involved in the regulation of cell elongation (Bai et al., 2012; Ikeda et al., 2012), lateral root formation (Lu et al., 2018), and floral organ development (Mara et al., 2010). In addition, they are involved in the response of plants to hormone and environmental signals, such as gibberellin (Lee et al., 2006), brassinosteroid (BR; Zhang et al., 2009), auxin (Zheng et al., 2017), abscisic acid (ABA; Zheng et al., 2019), temperature (Bai et al., 2012), light (Castelain et al., 2012; Hong et al., 2013), shading (Gommers et al., 2017), and salt stress (Zheng et al., 2019). PRE4 (also named bHLH161/BNQ3) is one of the PRE family members, which is negatively regulated by APETALA3/PISTILLATA in petals, and plays a crucial role in floral organ development (Mara et al., 2010). It is also implicated in floral induction, and interacts with a bHLH protein LONG HYPOCOTYL IN FAR-RED 1 (HFR1) to regulate light signalling in Arabidopsis (Mara et al., 2010). OsILI1, a homolog of Arabidopsis PRE1, mediates BR signalling to regulate leaf bending in rice (Zhang et al., 2009). In cotton, GhPRE1 has been demonstrated to positively regulate fibre elongation (Zhao et al., 2018). However, the function of PREs in other species, especially in ornamental plants, remains to be determined.

KNOTTED1-LIKE HOMEOBOX (KNOX) genes belong to the plant-specific three amino acid loop extension (TALE) homeodomain family, which are divided into three sub-classes, class I (KNOX1), class II (KNOX2), and KNATM, a new class of KNOX genes (Kerstetter et al., 1994; Magnani and Hake, 2008; Mukherjee et al., 2009). One of the essential KNOX2 proteins is KNAT7, a negative regulator of interfascicular fibre secondary cell wall formation with enhanced fibre wall thickness in knat7 loss-of-function mutants (Li et al., 2012; He et al., 2018). In addition, KNAT7 interacts with BELL-LIKE HOMEODOMAIN 6 (BLH6), OVATE FAMILY PROTEIN 4 (OFP4), and MYB75 via different domains to regulate secondary cell wall biosynthesis (Bhargava et al., 2010; Li et al., 2011; Liu et al., 2014). A recent study revealed that KNAT7 positively controlled xylan biosynthesis by activating IRREGULAR XYLEM 9 expression, and influenced lignin biosynthesis by interacting with KNAT3, another KNOX2 protein (He et al., 2018; Qin et al., 2020). However, the molecular mechanism behind regulation of lignin biosynthesis by KNAT7 remains largely unexplored.

Chrysanthemum morifolium is one of the most popular ornamental flowers in the consumer market, globally. However, it is prone to lodging, and requires a scaffolding net during cultivation, increasing expenditure on human and material resources. It is therefore necessary to investigate the formation and regulatory mechanisms of stem strength in Chrysanthemum. A previous study revealed that lignin content is closely related to the stem strength of chrysanthemum (Lv et al., 2011), but its molecular mechanism is still unclear.

In this study, we isolated a gene from chrysanthemum encoding an atypical bHLH transcription factor, CmHLB (HLH PROTEIN INVOLVED IN LIGNIN BIOSYNTHESIS), which was expressed in the vascular tissue and developing interfascicular fibres. Overexpression of CmHLB (OX-CmHLB) in chrysanthemum showed dwarf and increased stem strength phenotypes. Furthermore, histochemical analysis indicated that OX-CmHLB stems had a thicker cell wall and higher lignin content than the wild type (WT). However, CmHLB RNA interference (RNAi-CmHLB) plants exhibited opposite phenotypes. RNA-seq analysis indicated that the expression of genes involved in lignin biosynthesis was up-regulated and down-regulated in OX-CmHLB and RNAi-CmHLB lines, respectively. Furthermore, we confirmed that CmHLB interacted with CmKNAT7 via the KNOX2 domain, a protein implicated in negatively regulating secondary cell wall formation and lignin biosynthesis in chrysanthemum. This study reveals a novel function for CmHLB in regulating lignin biosynthesis, and elucidates the regulation mechanism of stem strength in Chrysanthemum.

Materials and methods

Plant materials and growth conditions

Rooted cuttings of the chrysanthemum (Chrysanthemum morifolium) cultivar ‘Jinba’ were obtained from the Nanjing Agricultural University’s Chrysanthemum Germplasm Resource Preservation Center (Nanjing, China). The chrysanthemum plants were cultured in a greenhouse under the following conditions: a photoperiod of 16 h light/8 h dark, relative humidity of 70% and a day/night temperature of 25 °C/20 °C. Arabidopsis thaliana (ecotype Col-0) plants were grown in vermiculite and soil mixture (3:1) under a 16 h photoperiod (80–100 µmol m^-2 s^-1 illumination) at 22 °C.

Isolation of CmHLB and the analysis of structure and phylogeny

Total RNA was extracted from stems of chrysanthemum cultivar ‘Jinba’ using the RNA isolation kit (Huayueyang, Beijing, China) following the manufacturer’s protocol. The cDNA was synthesized using a PrimeScript^TM RT reagent kit (Takara Bio, Tokyo, Japan). The CmHLB open reading frame (ORF) sequence was amplified using a pair of primers, CmHLB-F/R (Supplementary Table S1), designed with Primer Premier 5.0 software, according to the Unigene28497 sequence of the Chrysanthemum morifolium transcriptome (http://www.ncbi.nlm.nih.gov/bioproject/PRJNA329030). The fragment was then ligated into the pMD19-T vector (Takara Bio,pMD19-T-CmHLB) for sequencing. The amino acid sequences of Arabidopsis PRE family members were acquired from the TAIR website (https://www.arabidopsis.org/). Protein alignment of CmHLB and AtPREs was performed using the DNAMAN 5.2.2 software, and phylogenetic analysis was performed using the MEGA 5.1 software via the neighbor-joining method, with 1000 bootstrap replicates. The AtPRE genes comprised the following: AtPRE1 (AT5G39860), AtPRE2 (AT5G15160), AtPRE3 (AT1G74500), AtPRE4 (AT3G47710), AtPRE5 (AT3G28857), and AtPRE6 (AT1G26945).

Sub-cellular localization of CmHLB

The CmHLB ORF sequence was amplified from the pMD19-T-CmHLB construct using the pENTR1A-CmHLB-F/R primer pair (Supplementary Table S1) harbouring the BamHI and NotI sites. The amplicons of CmHLB and the pENTR1A vector (Invitrogen, Carlsbad, CA) were digested with BamHI and NotI and then ligated with T4 DNA ligase (Takara Bio) to obtain the pENTR1A-CmHLB construct. To identify the sub-cellular localization of CmHLB within cells, the vector pENTR1A-CmHLB was recombined with the vector pMDC43 (35S::GFP) via the LR reaction to gain the fusion vector 35S::GFP-CmHLB. The fusion vector and the control 35S::GFP were transiently transformed into onion epidermal cells by the helium-driven PDS-1000 particle accelerator (Bio-Rad Laboratories, Hercules, CA), separately. The bombarded cells were incubated in the dark at 22 °C for 16 h, and the green fluorescent protein (GFP) signal was observed using a Zeiss LSM 780 confocal microscope (Carl Zeiss AG, Oberkochen, Germany).

Transcriptional activity of CmHLB

The yeast assay system was used to investigate the transcriptional activity of CmHLB. The coding sequence of CmHLB was amplified with the BD-CmHLB-F/R primer pair (Supplementary Table S1) and introduced into the pGBKT7 vector using EcoRI and BamHI restriction enzymes, resulting in the pGBKT7-CmHLB construct. The plasmids of pGBKT7-CmHLB, pCL1 (positive control), and pGBKT7 (negative control) were then separately transformed into Saccharomyces cerevisiae strain Y2HGold (Takara Bio) according to the manufacturer’s protocol. The transformants containing the constructs pGBKT7-CmHLB or pGBKT7 were grown on SD/-Trp medium, and those including pCL1 were grown on SD/-Leu medium. The successfully transformed yeast cells were then separately grown on SD/-His-Ade medium and SD/-His-Ade medium with X-α-gal and incubated at 30 °C for 2 d.

qRT–PCR analysis

Total RNA was extracted from different internodes of the chrysanthemum cultivar ‘Jinba’ (WT) and the ninth internodes of CmHLB, and CmKNAT7 transgenic chrysanthemum plants using the RNA isolation kit (Huayueyang, Beijing, China). Approximately 1 µg total RNA was then reverse-transcribed into cDNA using the PrimeScript^TM RT reagent kit (Takara Bio) following the manufacturer’s protocol. The qRT–PCR reactions were conducted on a LightCycler 96 Real-Time PCR System (Roche, Basel, Switzerland) with the SYBR Premix Ex Taq II kit (Takara Bio) following the manufacturer’s instructions. qRT–PCR data were obtained for three biological and three technical replicates of each sample. The expression of genes was calculated using the 2^-ΔΔCT method (Livak and Schmittgen, 2001). CmEF1α (KF305681) and CmACTIN (KF305683; Supplementary Fig. S1) were used as reference genes. The primer pairs used for qRT–PCR analyses are listed in Supplementary Table S1.

β-Glucuronidase (GUS) expression assay

The promoter sequence of CmHLB was amplified from the chrysanthemum genomic DNA by PCR, using the primer pair CmHLB-pro-F/R (Supplementary Table S1). The 1783 bp fragment was obtained and fused with the β-GLUCURONIDASE (GUS) reporter gene to create the Pro_CmHLB:GUS plasmid. Following Agrobacterium-mediated transformation, Pro_CmHLB:GUS transgenic Arabidopsis lines were selected and the transgenic plants of T₃ generation were used for the GUS staining assay. Histochemical analyses were performed using the Gusblue kit (Huayueyang) following the manufacturer’s protocol. GUS staining in different organs was observed using an S8APO microscope (Leica Camera AG, Germany).

Transformation of chrysanthemum

The 35S::CmHLB (OX-CmHLB), pMDC32-amiR-CmHLB (RNAi-CmHLB), and 35S::CmKNAT7 (OX-CmKNAT7) plasmids were constructed according to the method described previously (Wang et al., 2020). All three constructs were individually transformed into Agrobacterium tumefaciens EHA105 strain using the freeze-thaw transformation method, and then introduced into the chrysanthemum cultivar ‘Jinba’ using the leaf disc infection method (J. Wang et al., 2019). After regeneration, RNA was extracted from putative transgenic and WT plants; cDNA was then synthesized using the PrimeScript^TM RT reagent kit (Takara Bio). Subsequently, qRT–PCR analyses were performed to calculate the expression of CmHLB and CmKNAT7 using CmHLB-q-F/R and CmKNAT7-q-F/R primers, respectively (Supplementary Table S1). In addition, the plant height and stem diameter of the ninth internodes of 2-month-old WT and CmHLB transgenic plants were measured. Finally, the stem strength of the ninth internodes of 2-month-old CmHLB and CmKNAT7 transgenic plants was measured with a stem strength tester (YYD-1, TOP Instrument, Hangzhou, China).

Morphological and histological analysis

For investigating lignin deposition, cross-sections taken from the 2-month-old chrysanthemum stems were stained with phloroglucinol-HCl reagent, as described previously (Du et al., 2015). The phloroglucinol-HCl reagent was prepared by mixing one volume of concentrated HCl and two volumes of 2% (w/v) phloroglucinol dissolved with 95% (v/v) ethanol. The cross-sections were stained for 2 min at 25 °C and were immediately imaged using an S8APO microscope (Leica). For cell wall thickness analysis, stem fragments were cut from the base of inflorescence stems of 6-week-old Arabidopsis plants or the ninth internodes of 2-month-old chrysanthemum stems. The stem fragments (embedding Arabidopsis inflorescence stems in 6% agarose to provide support) were cut into 50 μm sections using a vibrating-blade microtome (VT1200S, Leica Microsystems GmbH, Wetzlar, Germany). The thin sections were stained with a phloroglucinol-HCl reagent to be observed on a Leica DM1000 microscope. The cell wall thickness of the vessel element, xylary fibre, and interfascicular fibre cells were measured using Image J software (NIH, USA).

Lignin content analysis

The ninth internodes of 2-month-old transgenic chrysanthemum and WT plants were collected to measure total lignin content. The samples were dried and ground, then dissolved with 80% ethyl alcohol and incubated at 90 °C for 20 min. After cooling to 25 °C, the samples were centrifuged at 6000 × g for 10 min, and pellets washed with 80% ethyl alcohol and then with acetone. After centrifugation, the pellets were treated with dimethyl sulfoxide for 15 h to remove starch, and the residues were then dried to obtain the cell wall residues. The lignin content was analysed using a test kit (COMINBIO, Suzhou, China) following methods described previously (Cao et al., 2021; Ren et al., 2021; Zhu et al., 2021). At least three biological replicates were obtained.

Yeast two-hybrid analysis and BiFC assay

The five fragments of CmHLB were cloned and introduced into the pGBKT7 vector by restriction digestion with SmaI and BamHI (primer sequences listed in Supplementary Table S1). The coding sequence of CmKNAT7, KNOX1 domain, KNOX2 domain, MEINOX domain, and homeodomain were amplified and separately introduced into the pGADT7 vector by restriction digestion with SmaI and BamHI (primer sequences listed in Supplementary Table S1). The successful constructs were then transformed into S. cerevisiae strain Y2HGold (Takara Bio) using a Yeastmaker^TM Yeast Transformation System 2 kit (Takara Bio), following the manufacturer’s instructions. The transformants were selected on the SD/-Trp-Leu medium and then tested on the SD/-Trp-Leu-His-Ade medium and SD/-Trp-Leu-His-Ade medium with X-α-gal, respectively.

The bimolecular fluorescence complementation (BiFC) assay was performed in tobacco epidermal cells. The ORF of CmHLB was introduced into the pSPYCE(M) (Waadt et al., 2008) vector containing the C-terminal of YFP to generate the pSPYCE(M)-CmHLB construct via BamHI and SalI restriction enzymes. The ORF of CmKNAT7 was fused with the N-terminal fragment of the YFP protein in the pSPYNE173 (Waadt et al., 2008) vector after restriction digestion with BamHI and SalI to produce the pSPYNE173-CmKNAT7 construct. The constructs were separately transformed into A. tumefaciens strain GV3101 and then co-transformed into the leaves of 4-week-old Nicotiana benthamiana plants. After culturing in the dark for 1 d and in light for 2 d, the tobacco epidermal cells were observed with a confocal laser scanning microscope (LSM 780, Carl Zeiss AG, Oberkochen, Germany). The primers used for vector construction are shown in Supplementary Table S1.

Pull-down assay

As previously described, the pull-down experiment was carried out according to Zhou et al. (2017). The CmHLB full-length cDNA was digested with BamHI and SalI restriction endonucleases and introduced into the pET-32a-c (+) vector containing a His-tag. The coding sequence of CmKNAT7 was also digested with BamHI and SalI restriction enzymes and fused with the glutathione S-transferase (GST) tag of the pGEX-5X-1 vector. Both constructs were then individually transformed into Escherichia coli BL21 (DE3). After purification, His-CmHLB combined with GST-CmKNAT7 or GST was incubated with an anti-GST column for immunoprecipitation. Finally, the pellet fraction was examined using an anti-His antibody (Abmart, Shanghai, China) via immunoblotting.

RNA extraction, transcriptome sequencing, and bioinformatics analysis

The ninth internodes of WT, OX-CmHLB (lines 4 and 6), and RNAi-CmHLB (lines 7 and 10) stems were sampled (three biological replicates). RNA was extracted using an RNA isolation kit (Huayueyang, Beijing, China) and subjected to Illumina sequencing using Illumina HiSeq2500 (Gene Denovo Biotechnology Co. Ltd., Guangzhou, China). Clean reads were selected from the raw data by eliminating tracts of poly N, adaptor sequences, and low-quality reads, and assembled using Trinity (v2.8.4) software (Grabherr et al., 2011). Functional annotation of unigenes was based on the NR, Kyoto Encyclopedia of Genes and Genomes (KEGG; www.kegg.jp/), Swiss-Prot (www.uniprot.org/), and KOG (www.ncbi.nlm.nih.gov/ KOG/) databases. The FPKM method was used to calculate the expression of unigenes (Mortazavi et al., 2008). Differential expression analyses were conducted using the DESeq2 software between two different groups, and unigenes with false discovery rate (FDR) below 0.05 and |log₂^FoldChang| >1 were considered as differentially expressed genes (DEGs; Love et al., 2014). The Venn diagrams and KEGG pathway enrichment analyses were generated and performed by the online Omicsmart tool (www.omicsmart.com), respectively, and the heat map of the expression of selected DEGs was generated by TBtools software (C. J. Chen et al., 2020).

Heterologous expression of CmKNAT7 in Col-0 and knat7 mutant

The KNAT7 T-DNA insertion mutant SALK_002098 (knat7) was obtained from AraShare (http://www.arashare.cn/). T-DNA insertion was verified using SALK_002098-LP, SALK_002098-RP, and LBb1.3 primers (Supplementary Table S1), and homozygous plants were selected for further analysis. In addition, RT–PCR was performed to check the expression of AtKNAT7 in Col-0 and knat7 mutant plants using the primer pair AtKNAT7-F/R (Supplementary Table S1), and AtACTIN2 was used as the reference.

The 35S::CmKNAT7 plasmid was introduced into the A. tumefaciens EHA105 strain and then transformed into Col-0 and knat7 mutant plants using the floral dip method (Clough and Bent, 1998). Transgenic plants were selected using half-strength MS medium with 35 mg l^-1 kanamycin. Subsequently, RNA was extracted from selected transgenic and WT plants, and cDNA was synthesized using the PrimeScript^TM RT reagent kit (Takara Bio). RT–PCR was performed to examine the transcript abundance of CmKNAT7 with the primer pair CmKNAT7-F2/GFP-R (Supplementary Table S1), and AtACTIN2 was used as the reference.

Statistical analysis

All statistical analyses were conducted by SPSS v 17.0 using one-way ANOVA, and Duncan’s multiple range test was used to determine the significant differences. P<0.05 was considered statistically different.

Results

Expression of CmHLB is positively correlated with stem mechanical strength in chrysanthemum

A previous study demonstrated that the stem mechanical strength of plants gradually increases from top to bottom (Burk et al., 2001). Measurements were recorded at internodes 3, 6, 9, and 12 (internode 3 was the top internode and internode 12 was the bottom internode) in 2-month-old chrysanthemum cultivar ‘Jinba’ to investigate the mechanical strength of different internodes. The results showed that mechanical strength gradually increased from top to bottom in chrysanthemum stems, except internode 3, which was too weak to measure (Fig. 1A). Moreover, the histochemical staining assay indicated that lignin content and the degree of lignification of stems also increased from top to bottom (Fig. 1B). A previous experiment in our laboratory found that overexpression of a gene encoding an atypical bHLH transcription factor, CmHLB, resulted in stem cracking phenotypes and affected lignin content in Arabidopsis (Supplementary Fig. S2). Additionally, we found that the expression of CmHLB increased gradually from internode 3 to internode 12, consistent with the increasing trend of lignin content and stem strength (Fig. 1C). These results indicated that the expression of CmHLB was positively related to the lignin content and stem strength in chrysanthemum. Therefore, we speculate that CmHLB may be involved in regulating stem strength and lignin biosynthesis in chrysanthemum.

Fig. 1.

The expression of CmHLB has a positive correlation with the stem mechanical strength of chrysanthemum. (A) Measurement of stem strength in different internodes of ‘Jinba’. ND: not determined. The values represent means ±SE, n=18. Significant differences were determined by Duncan’s multiple range test (P<0.05). (B) Cross-sections from different stem internodes of 2-month-old chrysanthemum cultivar ‘Jinba’ stained with phloroglucinol-HCl. Internodes were numbered in order from top to bottom of the stem. Bar =1 mm. (C) Relative expression of CmHLB in different stem internodes of ‘Jinba’. CmEF1α was used as the internal control. The expression of CmHLB in the internode 3 was set to 1. The values represent the means ±SE from three biological replicates, and significant differences (indicated as different letters) were determined by Duncan’s multiple range test (P<0.05).

CmHLB comprises a 297 bp open reading frame (ORF) encoding a protein with 98 amino acids. Phylogenetic analysis showed that CmHLB formed a clade with AtPRE4, AtPRE2, and AtPRE6 among the six Arabidopsis PRE proteins, with AtPRE4 being the closest homolog (Supplementary Fig. S3A). Comparing deduced amino acid sequences of CmHLB with that of AtPREs indicated that CmHLB belonged to the PRE sub-family, containing a conserved HLH domain in the C-terminus but no basic region in the N-terminus (Supplementary Fig. S3B). To investigate the sub-cellular localization of CmHLB, the fusion vector 35S::GFP-CmHLB and the control 35S::GFP were transiently introduced into onion epidermal cells by particle bombardment. The results revealed that CmHLB localized to the nucleus (Supplementary Fig. S3C). The transcriptional activity analysis suggested that CmHLB exhibited no transcriptional activation activity in yeast cells (Supplementary Fig. S3D). To examine the detailed expression pattern of CmHLB, we produced Arabidopsis transgenic lines expressing Pro_CmHLB:GUS construct and performed a GUS staining assay on T₃ generation transgenic plants. The results suggested that CmHLB was predominantly expressed in the vascular tissue and differentiating interfascicular fibres, where secondary cell walls were actively deposited (Supplementary Fig. S4A-D). In addition, CmHLB was also expressed in the seedlings and floral organs (Supplementary Fig. S4E-I).

CmHLB affects stem mechanical strength and lignin content in chrysanthemum

We overexpressed and suppressed (RNAi) CmHLB in the chrysanthemum cultivar ‘Jinba’ through Agrobacterium-mediated leaf disk transformation, to explore its function (J. Wang et al., 2019). Seven overexpressing lines and 11 RNAi lines were identified, and three representative transformants were selected for further analysis (Fig. 2A). We found that OX-CmHLB plants had a dwarf phenotype, and their stem diameter was significantly increased (P<0.05) compared with that of the WT (Fig. 2B, C). Moreover, OX-CmHLB plant organs were generally smaller than those of WT, with decreased leaf size, inflorescence diameter, and petal length (Supplementary Fig. S5). However, except for smaller leaves and inflorescence diameter, there was almost no morphological difference between RNAi-CmHLB plants and the WT (Fig. 2B, C; Supplementary Fig. S5). We measured the mechanical strength of the ninth internode of WT and CmHLB transgenic plants to investigate whether CmHLB affects the stem strength of chrysanthemum. The results showed that OX-CmHLB lines had increased stem mechanical strength, whereas RNAi-CmHLB lines had slightly decreased strength than the WT (Fig. 2C). These results suggested that CmHLB affects plant morphology and stem mechanical strength in chrysanthemum.

Fig. 2.

CmHLB affects stem strength of chrysanthemum. (A) Expression analysis of CmHLB in the ninth internodes of wild-type (WT) and CmHLB transgenic plants determined by qRT–PCR. OX-3, -4, -6 and RNAi-7, -10, -52 correspond to different OX-CmHLB and RNAi-CmHLB transgenic lines, respectively. CmEF1α was used as the internal control. The expression of CmHLB in the WT was set to 1. The values represent means ±SE from three biological replicates, and significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05). (B) Two-month-old WT and CmHLB transgenic plants. Bar =5 cm. (C) Measurement of plant height, stem diameter, and stem strength of WT and CmHLB transgenic plants. Values are means ±SE, n≥15. Significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05).

Anatomical analysis was performed on WT and CmHLB transgenic stems to determine whether the changes in cell wall structure increased mechanical strength of OX-CmHLB-containing stems (Fig. 3A). We measured the cell wall thickness of the vessel elements, and xylary and interfascicular fibres under a high magnification light microscope (Fig. 3C). The results revealed that OX-CmHLB plants had thicker vessel element, xylary fibre, and interfascicular fibre cell walls than WT. In contrast, RNAi-CmHLB plants had slightly thinner xylary fibre and interfascicular fibre cell walls (Fig. 3A, C). We then examined lignin deposition and content in WT and CmHLB transgenic stems. The phloroglucinol-HCl staining assay showed that the stained area was thicker in the overexpressed plants and thinner in the RNAi plants than in the WT (Fig. 3B). Moreover, the lignin content of the OX-CmHLB lines was significantly increased (P<0.05) compared with WT, while that of RNAi-CmHLB lines was slightly decreased (Fig. 3D). These results indicated that CmHLB positively regulates cell wall thickening and lignin content in chrysanthemum.

Fig. 3.

CmHLB affects secondary cell wall thickness and lignin content in chrysanthemum. (A) The vascular bundles (upper panels) and interfascicular fibres (lower panels) of the ninth internode of WT and CmHLB transgenic stems stained with phloroglucinol-HCl to detect lignin. if, interfascicular fibre; ve, vessel; xf, xylary fibre. Bar =50 μm. (B) Lignin deposition in the internodes 3, 6, 9, and 12 of WT and CmHLB transgenic plants. Stem cross-sections were taken from the 2-month-old plants and stained with phloroglucinol-HCl. Red arrowheads indicate vascular bundles. Bar =0.5 mm. (C) Measurement of the ninth internode secondary cell wall thickness of WT and CmHLB transgenic stems. Values are means ±SE, n=100 cells for each line. Significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05). (D) Lignin content of the ninth internodes from 2-month-old WT and CmHLB transgenic plants. The values represent means ±SE from three biological replicates, and significant differences were determined by Duncan’s multiple range test (P<0.05).

CmHLB positively regulates the expression of genes involved in lignin biosynthesis

The ninth internodes of 2-month-old WT, OX-CmHLB (lines 4 and 6), and RNAi-CmHLB (lines 7 and 10) stems were collected for RNA sequencing to understand the regulatory mechanism of CmHLB-mediated lignin synthesis. DEGs were identified using pairwise comparisons between WT and CmHLB transgenic lines, with a false discovery rate (FDR) <0.05 and |log₂^FoldChange| >1. The analysis of WT versus OX-4 pairwise comparison revealed 1812 DEGs, of which 1337 were up-regulated and 475 were down-regulated; the comparison of WT versus OX-6 identified 1763 (1251 up-regulated, 512 down-regulated) DEGs (Fig. 4A). Of these, 1128 genes were identified in both WT versus OX-4 and WT versus OX-6 pairwise comparisons (Fig. 4B). The comparisons of WT versus RNAi-7 and WT versus RNAi-10 identified 2732 (1464 up-regulated, 1268 down-regulated) and 2521 (1406 up-regulated, 1115 down-regulated) DEGs, respectively (Fig. 4A). A total of 1910 genes were differentially expressed in both WT versus RNAi-7 and WT versus RNAi-10 pairwise comparisons (Fig. 4B).

Fig. 4.

RNA-seq analysis of WT and CmHLB transgenic chrysanthemum stems. (A) The number of up-regulated or down-regulated DEGs in WT versus OX-4, WT versus OX-6, WT versus RNAi-7, and WT versus RNAi-10 pairwise comparisons. (B) Venn diagram showing the overlap of DEGs between WT versus OX-4 and WT versus OX-6, WT versus RNAi-7 and WT versus RNAi-10. (C, D) KEGG enrichment analyses of DEGs in WT versus OX-6 (C) and WT versus RNAi-10 (D) pairwise comparisons. (E) Validation of the expression of five lignin biosynthetic pathway genes via qRT–PCR. CmEF1α was used as the internal control. The expression of lignin biosynthetic pathway genes in the WT was set to 1. Data are means ±SE from three biological replicates, and significant differences (indicated with different letters) were determined by Duncan’s multiple range test (P<0.05).

KEGG enrichment analysis was performed based on their known functions to determine the identities of these DEGs. The results showed that the DEGs were significantly enriched (P<0.05) in the phenylpropanoid biosynthesis pathway in all WT versus OX-CmHLB and RNAi-CmHLB pairwise comparisons (Fig. 4C, D; Supplementary Figs S6, S7), and lignin is synthesized through this metabolic pathway (Bonawitz and Chapple, 2010). Among these DEGs, the expression of CmPAL1, CmCCoAOMT, and CmHCT, involved in monolignol biosynthesis, was up-regulated in OX-CmHLB lines and down-regulated in the RNAi-CmHLB lines (Supplementary Fig. S8). Furthermore, CmPRX25 and CmPRX71, which encode peroxidases responsible for lignin polymerization, also had higher and lower expression in OX-CmHLB and RNAi-CmHLB lines, respectively (Supplementary Fig. S8). These results indicated that CmHLB regulated lignin biosynthesis by modulating a set of downstream genes. In addition, gene expression in the lignin biosynthetic pathway was verified by qRT–PCR in WT and CmHLB transgenic plants (Fig. 4E). We conclude that CmHLB positively regulates the expression of genes involved in lignin biosynthesis.

CmHLB interacts with CmKNAT7

Previous studies reported that the atypical bHLH proteins lose the capacity to bind DNA and usually function as repressors of other bHLH proteins by forming heterodimers with them (Toledo-Ortiz et al., 2003). To further elucidate the mechanism of CmHLB regulating lignin biosynthesis, we performed a yeast two-hybrid screening assay to select proteins that interact with CmHLB. An appropriate candidate was not found (Supplementary Table S2). Based on CmHLB function, we subsequently focused on several transcription factors (CmKNAT7, CmVND1/4-like, CmMYB2, CmMYB20/58-like; Fig. 5A; Supplementary Fig. S9A) regulating secondary cell wall formation. We performed yeast two-hybrid assays to check the interaction between CmHLB and these transcription factors. The results showed that only CmKNAT7 could interact with CmHLB in yeast cells (Fig. 5A; Supplementary Fig. S9A). KNAT7 is one of the Arabidopsis KNOX genes and functions as a negative regulator of secondary cell wall formation and lignin synthesis (Li et al., 2012; S. Wang et al., 2019). Additionally, the expression of CmKNAT7 gradually increased in chrysanthemum stems from the third to the ninth internode, similar to the expression pattern of CmHLB (Fig. 1C; Supplementary Fig. S9B). Thus, we propose CmKNAT7 as a potential interaction partner of CmHLB.

Fig. 5.

CmHLB interacts with CmKNAT7 in vitro and in vivo. (A) Yeast two-hybrid assay of CmHLB and CmKNAT7. BD-53 and AD-T, positive control; BD-LAM and AD-T, negative control. (B) BiFC analysis of the interaction between CmHLB and CmKNAT7 in N. benthamiana leaves. NLS: nuclear localization signal, YFP: yellow fluorescent protein, DIC: differential interference contrast image. Bar =50 μm. (C) Pull-down assay to test CmHLB interaction with CmKNAT7 in vitro. GST and His-CmHLB were used as controls. The red arrows indicate the target proteins. IB:His or GST, immunoblotting (IB) with anti-His or anti-GST antibody. (D) Schematic diagram of CmKNAT7 with four different fragments, KNOX1, KNOX2, MEINOX domain (KNOX1+KNOX2), and homeodomain, used to test interaction with CmHLB. (E) Yeast two-hybrid assays of CmHLB with CmKNAT7 fragments. (F) Schematic diagram of CmHLB with five different fragments used to test interaction with CmKNAT7. (G) Yeast two-hybrid assays of CmKNAT7 with CmHLB fragments.

To evaluate this hypothesis, we performed several other experiments to demonstrate the protein-protein interaction between CmHLB and CmKNAT7. In vivo, bimolecular fluorescence complementation (BiFC) analysis revealed that the interaction between CmHLB and CmKNAT7 occurred in the nucleus (Fig. 5B). Moreover, the pull-down assay also provided solid biochemical evidence for their interaction in vitro (Fig. 5C). Previous studies have confirmed that both the MEINOX domain and homeodomain are important for KNAT7 interaction with other proteins (Scofield and Murray, 2006; Li et al., 2011; Liu et al., 2014). To identify which domain plays a pivotal role in the interaction between CmKNAT7 and CmHLB, four CmKNAT7 fragments, KNOX1, KNOX2, MEINOX (KNOX1 and KNOX2), and the homeodomain, were fused to AD, and their abilities to interact with CmHLB were tested (Fig. 5D). Only the KNOX2 and MEINOX domains interacted strongly with CmHLB (Fig. 5E), suggesting that the KNOX2 part of the MEINOX domain was sufficient for CmKNAT7 to interact with the CmHLB protein. Additionally, CmHLB was divided into five fragments and fused with BD to evaluate their abilities to interact with CmKNAT7 (Fig. 5F). However, none of the five fragments interacted with CmKNAT7 (Fig. 5G), implying that the complete CmHLB protein was necessary for their interaction. Collectively, these data indicated that CmHLB interacted with the KNOX2 domain of CmKNAT7 to control lignin biosynthesis.

CmKNAT7 negatively regulates stem strength and lignin biosynthesis in chrysanthemum

KNAT7 has been implicated in negatively regulating lignin biosynthesis and secondary cell wall formation in Arabidopsis (Li et al., 2012). To investigate whether CmKNAT7 has conserved functions in chrysanthemum, we overexpressed CmKNAT7 (OX-CmKNAT7) under the control of the Cauliflower mosaic virus (CaMV) 35S promoter in the chrysanthemum cultivar ‘Jinba’ and two transgenic lines were obtained (Fig. 6A). To determine whether CmKNAT7 affects stem strength in chrysanthemum, we measured the mechanical strength of the ninth internode of WT and OX-CmKNAT7 transgenic plants. We observed a significant reduction (P<0.05) in stem strength in OX-CmKNAT7 transgenic lines (Fig. 6B). Cross-sections of the stems of the WT and OX-CmKNAT7 transgenic plants showed that the OX-CmKNAT7 lines had thinner vessel element, xylary fibre, and interfascicular fibre cell walls than the WT (Fig. 6C, D).

Fig. 6.

CmKNAT7 negatively regulates stem strength and secondary cell wall thickening in chrysanthemum. (A) Expression analysis of CmKNAT7 in the ninth internodes of WT and OX-CmKNAT7 transgenic plants determined by qRT–PCR. OX-1 and OX-3 correspond to different OX-CmKNAT7 lines. CmEF1α was used as the internal control. The expression of CmKNAT7 in the WT was set to 1. Data are means ±SE from three biological replicates, and significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05). (B) Measurement of stem strength of WT and OX-CmKNAT7 transgenic plants. Values represent means ±SE, n=11. Significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05). (C) The vascular bundles (top panels) and interfascicular fibres (lower panels) of the ninth internode of WT and OX-CmKNAT7 transgenic stems stained with phloroglucinol-HCl. if, interfascicular fibre; ve, vessel; xf, xylary fibre. Bar =100 μm. (D) Measurement of the ninth internode secondary cell wall thickness of WT and OX-CmKNAT7 transgenic stems. Values represent means ±SE, n=50 cells for each line. Significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05).

We then tested lignin deposition in the WT and OX-CmKNAT7 transgenic stems. The phloroglucinol-HCl staining assay revealed that the stained area was thinner in the overexpressed plants than in the WT (Fig. 7A). In addition, the lignin content of OX-CmKNAT7 transgenic plants was less than that of the WT (Fig. 7B). To determine whether CmKNAT7 shares the same downstream genes with CmHLB, we examined the expression of CmPAL1, CmCCoAOMT, CmHCT, CmPRX25, and CmPRX71 in OX-CmKNAT7 lines. The expression of the five genes involved in lignin biosynthesis was significantly down-regulated (P<0.05) in the transgenic plants (Fig. 7C). These results indicated that CmKNAT7 negatively regulated secondary cell wall thickening and lignin biosynthesis in chrysanthemum. Furthermore, we found that overexpression of CmKNAT7 in A. thaliana Col-0 repressed secondary cell wall thickening of interfascicular and xylary fibres (Supplementary Fig. S10), whereas overexpression of CmKNAT7 in A. thaliana knat7 mutant reversed the phenotype (thicker interfascicular and xylary fibre cell walls) of the knat7 mutant (Supplementary Fig. S11). These findings implied that CmKNAT7 is functionally conserved in repressing the secondary cell wall formation of fibres and lignin biosynthesis.

Fig. 7.

CmKNAT7 negatively regulates lignin biosynthesis in chrysanthemum. (A) Lignin deposition in the ninth internode of WT and OX-CmKNAT7 transgenic plants. Stem cross-sections were taken from the 2-month-old plants and stained with phloroglucinol-HCl. Red arrowheads indicate vascular bundles. Bar =0.5 mm. (B) Lignin content in the ninth internodes from 2-month-old WT and OX-CmKNAT7 transgenic plants. Data are means ±SE from six biological replicates, and significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05). (C) Relative expression of five lignin biosynthetic pathway genes in the ninth internodes of WT and OX-CmKNAT7 transgenic plants. CmEF1α was used as the internal control. The expression of lignin biosynthetic pathway genes in the WT was set to 1. Values represent means ±SE from three biological replicates, and significant differences (indicated by different letters) were determined by Duncan’s multiple range test (P<0.05).

Discussion

CmHLB has distinct roles, which are different from other PREs

CmHLB belongs to the PRE family, a sub-family of bHLH transcription factors, which consists of six members (AtPRE1 to AtPRE6) in Arabidopsis and seven members (OsILI1 to OsILI7) in rice. Consistent with AtPREs and OsILIs, CmHLB only contains a conserved HLH domain, but lacks the basic region responsible for DNA binding (Supplementary Fig. S3; Hyun and Lee, 2006; Lee et al., 2006; Zhang et al., 2009). In addition, AtPREs and OsILIs were ubiquitously expressed in the organs of Arabidopsis and rice, respectively (Lee et al., 2006; Zhang et al., 2009; Mara et al., 2010; Castelain et al., 2012; Zheng et al., 2019). Similarly, CmHLB was also expressed in the stem, leaf, and floral organs in Arabidopsis (Supplementary Fig. S4). However, detailed analysis implied that it was predominantly expressed in the vascular tissues and differentiating interfascicular fibres (Supplementary Fig. S4), which may have been ignored or not observed for other PREs that have been studied.

Previous studies have demonstrated that HLH proteins, for example, human inhibitors of DNA binding 1 (Id-1), generally function as negative regulators by forming heterodimers with other proteins (Norton, 2000; Toledo-Ortiz et al., 2003). As non-DNA binding bHLH transcription factors, PREs have also been reported to play various roles in plant growth and development by interacting with different proteins. For example, PRE1/BNQ1 interacted with Arabidopsis ILI1 BINDING bHLH 1 (IBH1), and inactivated IBH1 to control cell elongation in Arabidopsis (Zhang et al., 2009). Hao et al. (2012) found that PRE1 interacted with PAR1, a HLH protein, to regulate cell elongation in Arabidopsis. In addition, PRE1 directly inhibited PAR1 activity because the overexpression of PRE1 repressed the dwarf phenotype of PAR1 transgenic plants (Hao et al., 2012). PRE3/ATBS1 has been reported to interact with ATBS1-INTERACTING FACTOR 1 (AIF1) to regulate BR signalling in Arabidopsis (Wang et al., 2009). Hong et al. (2013) confirmed that PRE6/KIDARI (KDR) weakened the activity of HFR1 by forming heterodimers, resulting in the release of PHYTOCHROME INTERACTING FACTORS 4 (PIF4) from the HFR1-PIF4 complex to modulate photomorphogenesis in Arabidopsis. Furthermore, PRE1/BNQ1, PRE2/BNQ2, and PRE4/BNQ3 also interacted with HFR1 to regulate light signalling (Mara et al., 2010). In this study, we discovered a new function for CmHLB in regulating lignin biosynthesis by its interaction with CmKNAT7, thus, affecting the mechanical strength of stems in Chrysanthemum (Figs 2–5). In addition, we found that AtPRE4 interacted with AtKNAT7 in yeast cells, while AtPRE2 and AtPRE6 did not (Supplementary Fig. S12), implying that AtPRE4 may have functions similar to that of CmHLB in regulating lignin biosynthesis. Therefore, we propose that the differences in expression pattern and interacting protein may be sufficient for CmHLB to have functions distinct from other PREs.

CmHLB functions as a positive regulator of lignin biosynthesis affecting stem mechanical strength

Stem strength is an important agronomic trait determined mainly by morphological characteristics, anatomical structure, and chemical constituents of stems (Xia et al., 2018). In this study, we discovered that the expression of CmHLB was positively correlated with stem strength in different internodes of the chrysanthemum cultivar ‘Jinba’ (Fig. 1). Overexpression and suppression of CmHLB resulted in increased and decreased stem strength, respectively, in chrysanthemum (Fig. 2). These results indicate that CmHLB functions as a positive regulator of stem strength. Moreover, we observed increased stem diameter, thicker cell wall, and higher lignin content in OX-CmHLB transgenic chrysanthemum plants than in the WT (Figs 2, 3), which may represent the morphological, anatomical, and chemical reasons responsible for enhanced mechanical strength, respectively. Severe lignification inhibits plant growth, resulting in the dwarf phenotype of OX-CmHLB transgenic plants. However, the RNAi-CmHLB lines exhibited slightly reduced cell wall thickness and lignin content (Fig. 3), which is probably the reason behind similar morphology of RNAi-CmHLB plants and the WT. Additionally, we found that the DEGs were enriched in several hormone biosynthesis-related pathways such as tryptophan biosynthesis (synthetic precursor of auxin) and zeatin biosynthesis (a natural cytokinin) in the CmHLB transgenic lines (Supplementary Figs S6, S7). These pathways play important roles in plant growth and development (Holst et al., 2011; Bartrina et al., 2017; Duan et al., 2017; Ren et al., 2018). These features may be attributed to the altered leaf and inflorescence diameter sizes in the CmHLB transgenic plants (Supplementary Fig. S5).

Lignin, a complex phenolic polymer, is deposited in the secondary cell wall of some specialized cells and plays an important role in improving the mechanical strength and abiotic and biotic stress tolerance in plants (Zhao, 2016; An et al., 2019; Dixon and Barros, 2019; K.Q. Chen et al., 2020; Tu et al., 2020; Zhao et al., 2020). Lignin is one of the products of the phenylpropanoid pathway, and at least eleven enzymes have been implicated in monolignol biosynthesis (Bonawitz and Chapple, 2010; Vanholme et al., 2013; Zhong and Ye, 2015). Among them, PAL1, CCoAOMT, and HCT are three key enzymes whose suppression resulted in less total lignin or decreased G and S lignin content (Bonawitz and Chapple, 2010). A complicated hierarchical transcriptional network has been proposed to regulate the biosynthesis of lignin, cellulose, and hemicellulose, which are the three components of secondary cell walls (Zhong et al., 2010; Zhong and Ye, 2015). NACs and MYBs are top- and second-level master switches, respectively, and they activate many downstream transcription factors, including, KNAT7, SND2, SND3, MYB42, MYB43, MYB69, and MYB103, to regulate secondary cell wall biosynthesis (Zhong et al., 2010; Zhong and Ye, 2015). The present study discovered that OX-CmHLB and RNAi-CmHLB transgenic chrysanthemum plants had higher and lower lignin content, respectively (Fig. 3), indicating that CmHLB may also be a regulator of lignin biosynthesis. Our transcriptomic data of CmHLB transgenic plants suggested that the DEGs in the phenylpropanoid biosynthesis pathway were significantly enriched (Fig. 4; Supplementary Figs S6, S7). Furthermore, the expression of CmPAL1, CmCCoAOMT, and CmHCT was up-regulated in OX-CmHLB plants and down-regulated in RNAi-CmHLB plants (Fig. 4E; Supplementary Fig. S8). These results demonstrated that CmHLB positively regulated monolignol biosynthesis in chrysanthemum. Peroxidase genes PRX25 and PRX71 are involved in lignin polymerization. Mutations of AtPRX25 and AtPRX71 reduced lignin content and changed lignin structure in secondary cell walls in Arabidopsis (Shigeto et al., 2013, 2014). In our study, RNA-seq and qRT–PCR analysis showed that the expression of CmPRX25 and CmPRX71 was increased and decreased in OX-CmHLB and RNAi-CmHLB transgenic lines, respectively (Fig. 4E; Supplementary Fig. S8), suggesting that CmHLB may promote lignin synthesis by regulating lignin polymerization. The above results show that CmHLB functions as a positive regulator of lignin biosynthesis affecting stem strength in Chrysanthemum. To the best of our knowledge, this is the first study to reveal the function of CmHLB, among the PRE transcription factors, in regulating lignin biosynthesis.

CmHLB may interact with CmKNAT7 to antagonistically regulate lignin biosynthesis

KNAT7 is one of nine Arabidopsis KNOX genes and has been reported to negatively regulate secondary cell wall formation of interfascicular fibres and lignin biosynthesis (Li et al., 2012; S. Wang et al., 2019). In Arabidopsis, knat7 loss-of-function mutants had thicker interfascicular fibre cell walls, higher lignin content, and irregular xylem (irx) phenotypes, while KNAT7 overexpression exhibited thinner interfascicular fibre secondary walls (Li et al., 2012; He et al., 2018). In poplar, overexpression of PoptrKNAT7 in the knat7 mutant background rescued the knat7 irx phenotype (Li et al., 2012). In cotton, GhKNL1, a homolog of KNAT7, influenced secondary cell wall formation of fibres and repressed lignin biosynthesis (Gong et al., 2014). In the present study, we found that the OX-CmKNAT7 transgenic chrysanthemum plants exhibited reduced stem strength, lignin content, and cell wall thickness (Figs 6, 7). In addition, CmKNAT7 shared the same downstream lignin biosynthetic pathway genes with CmHLB, and negatively regulated their expression (Fig. 7). These results revealed that CmKNAT7 repressed secondary cell wall thickening and lignin synthesis in chrysanthemum. Furthermore, OX-CmKNAT7 transgenic Arabidopsis plants displayed thinner secondary cell walls in interfascicular and xylary fibres (Supplementary Fig. S10), and overexpression of CmKNAT7 in knat7 mutant background rescued the knat7 mutant phenotype (thicker interfascicular and xylary fibre cell walls; Supplementary Fig. S11). However, the 35S::CmKNAT7/knat7 plants exhibited irx phenotypes similar to those of the knat7 mutant (Supplementary Fig. S11). This may be attributed to the contrasting functions of CmKNAT7 with AtKNAT7 in regulating the development of vessel elements, or the low activity of 35S promoter in xylem (Eriksson et al., 2000; Swamy et al., 2015). These findings indicate that CmKNAT7 is functionally conserved in suppressing secondary cell wall formation of fibres and lignin biosynthesis.

Previous studies have reported that KNAT7 interacts with various proteins via different domains to inhibit secondary wall formation. For example, in Arabidopsis, KNAT7 interacts with BLH6 through the KNOX2 domain to repress REVOLUTA/INTERFASCICULAR FIBERLESS1 (REV/IFL1), to negatively regulate secondary wall formation (Liu et al., 2014). Additionally, the homeodomain was sufficient for the interaction of KNAT7 with OFP4 to suppress secondary cell wall formation (Li et al., 2011). In poplar, the KNOX1 and KNOX2 domains of KNAT7 strongly interact with MYB6 to repress CCoAOMT1 expression, thus inhibiting secondary cell wall formation (L.J. Wang et al., 2019). In this study, the yeast two-hybrid, BiFC, and pull-down assays confirmed that CmHLB interacted with CmKNAT7 in vivo and in vitro, and the KNOX2 domain of CmKNAT7 was essential for their interaction (Fig. 5). These results suggest that CmHLB has a novel interaction with CmKNAT7, implicated in secondary wall formation. Furthermore, CmKNAT7 and CmHLB have similar expression patterns from internode 3 to internode 9 in chrysanthemum (Fig. 1C; Supplementary Fig. S9B), further supporting the possibility for their interaction in plants. As one of the regulators of the biosynthesis of secondary cell walls, KNAT7 has been reported to be one of the direct targets of both top-level master switches SND1/NST1/VND6 (Zhong et al., 2008) and the second-level master switch MYB46 (Ko et al., 2009). Therefore, the HLB-KNAT7 complex may be downstream of top- or/and second-level master switches to regulate lignin biosynthesis, further enriching the regulatory network of secondary cell wall biosynthesis.

In our study, we confirmed that CmHLB positively regulated lignin biosynthesis (Figs 3, 4), while CmKNAT7 had the opposite function (Fig. 7). These findings indicate that CmHLB may interact with CmKNAT7 to regulate lignin biosynthesis antagonistically. Furthermore, as a repressor of lignin biosynthesis, the expression of CmKNAT7 gradually increased from internode 3 to internode 9 (Supplementary Fig. S9B), which is consistent with its expression pattern in Arabidopsis (Zhong et al., 2008). These results indicate that other proteins exist to repress the activity of the CmKNAT7 protein during the development of stems. Previously, PREs (PRE1, PRE6, etc.) have been reported to interact with other proteins (IBH1, HFR1, etc.) and compromise their inhibitory function (Toledo-Ortiz et al., 2003; Zhang et al., 2009; Hao et al., 2012; Hong et al., 2013). Therefore, we speculate that CmHLB may repress the activity of CmKNAT7, in a manner similar to the mechanism by which PREs suppress the activity of interacting proteins. Additionally, CmHLB may destroy the complex that CmKNAT7 forms with other proteins by competing for the KNOX2 domain, alleviating the inhibition of target genes involved in lignin biosynthesis. However, more evidence is required to support this hypothesis, and a detailed regulatory mechanism needs further investigation.

In summary, we presented a working model of CmHLB and CmKNAT7 in regulating lignin biosynthesis (Fig. 8). CmKNAT7 represses lignin biosynthesis by suppressing the expression of genes involved in lignin biosynthesis. CmHLB may alleviate the inhibition of CmKNAT7 by forming a heterodimer and promoting the transcription of downstream genes. In addition, overexpression of CmHLB significantly inactivates CmKNAT7 and results in higher lignin content in OX-CmHLB transgenic plants than the WT, thus affecting stem mechanical strength of chrysanthemum. Therefore, our findings reveal a novel function of PRE transcription factors and provide new insights into the regulatory mechanism of lignin biosynthesis in Chrysanthemum.

Fig. 8.

Working model involving CmHLB and CmKNAT7 in regulating lignin biosynthesis in the stem. In the wild type (WT), CmKNAT7 functions as a repressor of lignin biosynthesis to suppress the expression of genes involved in lignin biosynthesis, while CmHLB may inhibit CmKNAT7 activity by forming a heterodimer, thus positively regulating the expression of genes to promote lignin biosynthesis. Overexpression of CmHLB (OX-CmHLB) significantly decreases the activity of CmKNAT7, thus resulting in higher lignin content than WT. The intensity of fuchsia colour indicates the amount of lignin content. co, cortex; if, interfascicular fibre; p, phloem; x, xylem.

Supplementary data

The following supplementary data are available at JXB online.

Fig. S1. Expression analysis of CmHLB, CmKNAT7 and lignin biosynthetic pathway genes determined by qRT–PCR.

Fig. S2. CmHLB affects lignin content in Arabidopsis.

Fig. S3. Characteristic analysis of CmHLB.

Fig. S4. Histochemical localization of β-glucuronidase (GUS) activity in Pro_CmHLB: GUS transgenic plants.

Fig. S5. Leaf and petal phenotypes of CmHLB transgenic chrysanthemum plants.

Fig. S6. KEGG enrichment analyses of DEGs in WT versus OX-4 and WT versus OX-6 pairwise comparisons.

Fig. S7. KEGG enrichment analyses of DEGs in WT versus RNAi-7 and WT versus RNAi-10 pairwise comparisons.

Fig. S8. Heat map illustrating the transcript level of lignin biosynthetic pathway genes based on RNA-seq data.

Fig. S9. Screening of CmHLB interacting proteins.

Fig. S10. Effects of CmKNAT7 overexpression on secondary cell wall thickening of inflorescence stems in Arabidopsis.

Fig. S11. Effects of CmKNAT7 overexpression on secondary cell wall thickening of inflorescence stems in knat7 mutant plants.

Fig. S12. Yeast two-hybrid assays between AtKNAT7 with three AtPREs.

Table S1. List of primers used in this study.

Table S2. Candidate proteins that could interact with CmHLB from yeast two-hybrid screening.

Author contributions

FC, LD, and WZ designed the experiments; WZ, LD, JL, XZ, SL, KZ, and YG performed the experiments; SC, JJ, AS, HW, LD, and FC supervised the research; FC, LD, and WZ wrote and revised the manuscript.

Conflict of interest

The authors declare no conflict of interest.

Funding

This research was funded by the National Natural Science Foundation of China (31730081, 31872149), Jiangsu Agriculture Science and Technology Innovation Fund [cx(20)1001], the Natural Science Fund of Jiangsu Province (BK20170717, BK20190076), the Natural Science Fund of Qinghai Province (2018-Hz-819), the earmarked fund for Jiangsu Agricultural Industry Technology System [JATS(2021)454], and a Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

Data availability

RNA sequencing data from this study have been deposited in the National Center for Biotechnology Information repository (https://www.ncbi.nlm.nih.gov/bioproject/) under BioProject ID PRJNA734134.

Author notes

These authors contributed equally to this work.