Fruits of HR pummelo, HB pummelo, FH pummelo, and KP pummelo were collected at 55, 85, 115, 145, 175, and 205 DPA from the National Citrus Breeding Center at Huazhong Agricultural University (Wuhan, China). In addition, fruits of nine pummelo cultivars at the commercially mature stage (~200 DPA) were also harvested (Supplementary Table S7). For each cultivar, 15–18 fruits were picked from at least three different healthy trees and divided randomly into three replicates. The juice sacs from each sample were immediately separated, frozen in liquid nitrogen, and then stored at −80 °C.
Paraffin section observations of cell walls of the juice sacs
The reagents used for paraffin sectioning were analytically pure and purchased from Dingguo, Co. (China). The method was performed according to the methods of Li et al.39. The sections were stained with 1% phloroglucinol-hydrochloric acid or 0.1% toluidine blue solution for 3 min and then washed with 95% ethanol or deionized water for 10 min. The sections were subsequently examined and imaged using a Zeiss Axioscope A1 microscope (Swift Microscope World, California, USA) with a ×0.5 optical adapter. The images were captured and exported using ZEN 2.3 software (Zeiss).
Lignin extraction and evaluation
According to the methods of Xu et al.5 (with minor modifications), the juice sacs (or calli) were ground into powder and 8 g (or 10 g for calli) of each sample was homogenized in 15 ml of washing buffer (100 mM K2HPO4/KH2PO4, 0.5% Triton X-100, 0.5% PVP-K30; pH 7.8). The mixture was washed on a shaker (GS-20, MiuLab, Hangzhou, China) at room temperature at 250 r.p.m./min for 30 min and then centrifuged (5000 × g, 25 °C) for 10 min (Avanti J-26 XP, Beckman Coulter, California, USA). Finally, the supernatant was discarded. Washing buffer was added three times, as described above. Next, the precipitates were washed with 20 ml of 100% methanol four times. The precipitates were subsequently dried at 60 °C in a vacuum rotary evaporator (5305FG924683, Eppendorf, Germany) overnight. Fifty milligrams (mg) of the lyophilized powder of the juice sac (for calli, 100 mg) was dissolved in 5.0 ml of 2.0 M HCl and 0.5 ml of thioglycolic acid. The mixture was then boiled within a water bath at 100 °C (LTT-600, Longyue, Shanghai, China) for 8 h, cooled on ice for 5 min, and centrifuged at 8000 × g for 20 min at 4 °C (H1850R, Hunan, China). The precipitates were washed with distilled water three times and dried thereafter at 60 °C in a vacuum rotary evaporator overnight. The sample was resuspended in 1.5 ml of 1.0 M NaOH and subsequently placed on a shaker at 100 r.p.m. at room temperature for 18 h. The solution was centrifuged at 10,000 × g for 20 min. Five hundred milliliters of the supernatant of was transferred to a new tube that contained 0.1 ml of concentrated HCl. The tubes were incubated at 4 °C for 4 h to precipitate the lignin thioglycolic acid, followed by centrifugation at 13,000 × g for 20 min at 4 °C (5404EP320017, Eppendorf, Germany). The precipitate was then dissolved in a 1 : 100 ml volume of 1.0 M NaOH. Absorbance was then measured at 280 nm using ultraviolet spectrophotometry (UV-1800, Japan), with 1.0 M NaOH used as a blank. The data were expressed on a fresh weight basis and three biological replicates were used for each sample.
RNA isolation, library construction, and sequencing
The total RNA from the juice sacs was isolated according to the methods of Liu et al.40, while RNA from calli was isolated with TRIzol reagent (Aidlab Biotechnologies Co., Ltd, Beijing, China). The RNA from the HR, HB, KP, and FH fruit juice sacs at 55, 85, 115, 145, and 205 DPA was extracted, with three biological replicates included per sample. The following procedures were performed by Novogene Bioinformatics Technology, China. RNA integrity was assessed using a RNA Nano 6000 Assay Kit of a Bioanalyzer 2100 system (Agilent Technologies, CA, USA). A total amount of 3 µg of RNA per sample was used as input material for the RNA sample preparations. Sequencing libraries were generated using a NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, USA) following the manufacturer’s recommendations and index codes were added to attribute sequences to each sample. The constructed RNA libraries were sequenced on an Illumina HiSeq 2500 platform in paired-end mode, with a read length of 150 bp.
RNA-seq data analysis
Quality control of the raw sequencing data was performed with FastQC. Adapters were removed from reads and the data were cleaned based on phred scores using fastp. The clean data were mapped to the pummelo genome31. Read counts for each gene in each sample were extracted with the edgeR program. The FPKM of each gene was calculated based on its definition41. Differential expression analysis was performed using DESeq v1.18.0 in R v3.5.142. We then used KOBAS software to test the statistical enrichment of DEGs in KEGG pathways. Gene Ontology (GO) enrichment analysis of DEGs was implemented by GOseq v1.38.0, in which gene length bias was corrected. GO terms with corrected P-values (α = 0.05) were considered significantly enriched for those DEGs43. For correlation analyses, Spearman’s rank correlations of the lignin content and FPKM of all the genes expressed in the KP, HB and HR pummelo juice sacs at the five developmental stages was calculated using the pspearman package. Genes with a correlation coefficient to lignin content ratio higher than 0.75 and a significant P-value (α = 0.05) were selected.
cDNA synthesis and quantitative real-time PCR
A total of 0.5 μg of RNA was used for cDNA synthesis with a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, USA). qRT-PCR was performed with a Roche LightCycler 480 system in conjunction with 23 LightCycler 480 SYBR Green Master Mix (Roche, USA), as described by Lu et al.44. The primers designed for qRT-PCR are listed in Supplementary Table S8 and are based on pummelo database gene sequences31. The actin gene was used as previously reported26. The qRT-PCR data were analyzed using the 2-ΔCt analysis method.
The full-length coding DNA sequence (CDS) and 2.1 kb promoter region of CgMYB58 and the promoter of lignin biosynthetic genes were amplified from the cDNA and DNA of various citrus cultivars. The sequences of the primers used are listed in Supplementary Table S8. Sequence results were obtained from ~20 and 10 clones amplified from cDNA and DNA templates, respectively.
The online version of ClustalW (https://www.genome.jp/tools-bin/clustalw) was utilized for multiple alignment of the amino acid sequences of CgMYB58 and the corresponding sequences of other species. Conserved domains were extracted using gblocks (http://www.phylogeny.fr/one_task.cgi?task_type=gblocks). MEGA 7.0 was subsequently used to construct a phylogenetic tree according to the maximum likelihood method. There were 1000 bootstrap replications and values higher than 50 for each node are presented in the tree.
Subcellular localization analysis
The CgMYB58 CDS without a stop codon was fused to green fluorescent protein (GFP) within a pM999 vector. Citrus (C. limon Burm “Eureka lemon”) mesophyll protoplasts were extracted carefully for transient transformation. Plasmids of 35S:CgMYB58-GFP and 35S:OsGhd7-RFP (nuclear marker) and pM999 empty vector plasmids were mixed together equally. These plasmid mixtures were then transferred into separate protoplasts. After 24 h, the florescence images were scanned via a confocal laser-scanning microscope (TCS SP2, Leica, Wetzlar, Germany).
Transient transformation of CgMYB58
The CgMYB58 overexpression construct (vector: pH7WG2D) and a pH7WG2D empty vector used as a control were introduced into Agrobacterium tumefaciens strain GV3101 separately. The GV3101 cells containing the constructs were suspended in infiltration buffer (50 mg of glucose, 1 ml of 50 mM MES, 1 ml of Na3PO4, and 1 µl of 1 M acetosyringone per 10 ml of buffer) at a concentration of OD600 = 0.8, and then injected into the mesocarp of KP and HR pummelo fruits. GFP signals were captured by an inversion fluorescence microscope (Olympus SZX7, Japan) equipped with a light source (H-150). Twelve fruits were transformed per treatment.
Stable transformation of CgMYB58 in citrus calli
The CDS of CgMYB58 was isolated from HR and cloned into a pH7WG2D overexpression vector. The vector was then introduced into the A. tumefaciens strain EHA105 before it was transformed into calli of RM grapefruit (C. paradisi). The citrus callus transformation and growth conditions were performed according to the methods of Li et al.45, with minor modifications. Each transgenic callus was cultured in 500 ml of MT (Murashige and Tucker) media with vitamins (Coolaber Science & Technology, Beijing, China), supplemented with 80 μl of 50 mg ml−1 hygromycin B, whereas without hygromycin B was used for RM (the WT control). Each CgMYB58 overexpression line and RM callus were subcultured six to seven times before analysis.
Dual luciferase transcriptional activity assay
The dual LUC transcriptional activity assay procedure was modified from that of a previous study44. DNA sequences upstream of the translational start codon were amplified by PCR from genomic DNA of HR to generate 1062, 1101, 1298, and 939 bp promoter fragments for CgPAL1, CgPAL2, Cg4CL1, and CgC3H, respectively. The amplified fragments were subsequently inserted into a pGreenII 0800-LUC reporter vector, which was then introduced into A. tumefaciens GV3101 (pSoup-p19) competent cells. The effector vector was a CgMYB58 overexpression (vector: pH7WG2D) construct, and a pH7WG2D empty vector was used as a control. The constructs were also introduced into GV3101 (pSoup-p19). The GV3101 cells containing effectors and reporters were mixed to a proportion of 4:1 and then injected into leaves of Nicotiana benthamiana. LUC activities were analyzed according to the methods of Lu et al.44 at 2.5 days after injection.
Yeast one-hybrid assay analysis
Y1H assays were carried out using a Matchmaker Gold Yeast One-Hybrid system kit (Clontech, USA) and modified according to the methods of Lu et al.44. The promoter fragments of CgPAL2 and CgC3H were cloned into a pAbAi vector to produce pAbAi-CgPAL2P and pAbAi-CgC3HP bait constructs, respectively. The bait plasmids were then linearized and integrated into Y1H Gold yeast, after which they were selected with synthetic dextrose media lacking uracil. The CgMYB58 coding sequences were ligated into pGADT7 to generate an AD-CgMYB58 construct. pGADT7 empty vectors (AD-pGADT7), serving as negative controls, were then transferred separately into yeast cells containing bait constructs. The transformed yeast cells were diluted with 0, 400×, and 500× dilution series of aureobasidin A (AbA) and dotted on SD plates lacking leucine. The cells grew on both types of media containing prey proteins to allow the interaction of bait sequences.
Electrophoretic mobility shift assays
EMSA assays were performed as described previously26,46. pMAL-c5x-CgMYB58 (with maltose binding protein tag) without a stop codon was expressed and purified as described by Lu et al.44. The 25–27 bp probes containing the AC elements in the promoters of CgPAL1, CgPAL2, Cg4CL1, and CgC3H were extracted and used as reference sequences to synthesize probes. 5’-FAM-labeled oligonucleotide probes were synthesized and labeled by Shanghai Sangon Company (Shanghai, China). The same oligonucleotides without labels were used as cold competitors. To perform binding reactions, the binding solution (0.1% NP-40, 1 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, 50 μg ml−1 bovine serum albumin, 100 ng μl−1 poly(dI-dC)), 2 μl (0.5 mg ml−1) of purified maltose binding protein tag fused CgMYB58 and 1 μl of the 5’-FAM-labeled probe (10 μmol l−1) were mixed together and incubated at 4 °C for 45 min. For competition assays, the unlabeled oligonucleotides were incubated with protein and binding buffer at 4 °C for 45 min. Afterward, 1 μl of the 5’-FAM-labeled probe (10 μmol l−1) was added and incubated at 4 °C for 45 min. The samples were then loaded onto a prerun 6% polyacrylamide gel. Electrophoresis was performed at 4 °C using 0.5× Tris-borate-EDTA as an electrophoresis buffer in the dark for 1 h. Gel images were acquired using an Amersham Imager 600 (GE Healthcare, Tokyo, Japan).
IAA and ABA treatments of pummelo juice sacs
The ABA- and auxin-responsive elements of gene promoters were predicted by the PlantCARE online database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/). The juice sacs of the KP, HB, and HR pummelo cultivars were collected at 145 DPA in 2019 and cultured in murashige and skoog media (4.43 g of murashige and skoog powder, 100 g of sucrose, and 7 g of agar per liter of medium), with IAA and ABA at a final concentration of 0.5 mM. After being cultured for 50 days on the media, the samples were collected for the evaluation of gene expression and lignin content. High-performance liquid chromatography-grade IAA and ABA standards were purchased from Coolaber Science & Technology (Beijing, China).
Data analysis and software used
Metabolite and gene expression profiles were processed with Excel and GraphPad Prism 7. Heatmaps were processed by R with the gplot and pheatmap packages. Duncan’s multiple comparison test and Student’s t-test were conducted in conjunction with analysis of variance in SAS (SAS Institute, Inc., USA).