Nengwen Yin,Bo Li,Xue Liu,Ying Ling,b,Jinping Lin,Yufei Xue,Cunmin Qu,b,Kun Lu,b,Lijun Wei,b,Rui Wng,b,Jin Li,b,*,Yourong Chi,b,*

a Chongqing Engineering Research Center for Rapeseed,Chongqing Key Laboratory of Crop Quality Improvement,College of Agronomy and Biotechnology,Southwest University,Chongqing 400715,China

b Engineering Research Center of South Upland Agriculture of Ministry of Education,Academy of Agricultural Sciences,Southwest University,Chongqing 400715,China

Keywords:Brassica napus Cinnamoyl-CoA reductase (CCR)Lignin Flavonoids Lodging

ABSTRACT Brassica crops,which are of worldwide importance,provide various oil,vegetable and ornamental products,as well as feedstocks for animal husbandry and biofuel industry.Cinnamoyl-CoA reductase(CCR)is the entry point to the lignin pathway and a crucial locus in manipulation of associated traits,but CCRassociated metabolism and traits in Brassica crops have remained largely unstudied except in Arabidopsis thaliana.We report the identification of 16 CCR genes from Brassica napus and its parental species B.rapa and B.oleracea.The BnCCR1 and BnCCR2 subfamilies displayed divergent organ-specificity and participation in the yellow-seed trait.Their functions were dissected via overexpression of representative paralogs in B.napus.BnCCR1 was expressed preferentially in G-and H-lignin biosynthesis and vascular development,while BnCCR2 was expressed in S-lignin biosynthesis and interfascicular fiber development.BnCCR1 showed stronger effects on lignification-related development,lodging resistance,phenylpropanoid flux control,and seed coat pigmentation,whereas BnCCR2 showed a stronger effect on sinapate biosynthesis.BnCCR1 upregulation delayed bolting and flowering time,while BnCCR2 upregulation weakened the leaf vascular system in consequence of suppressed G-lignin accumulation.BnCCR1 and BnCCR2 were closely but almost oppositely linked with glucosinolate metabolism via inter-pathway crosstalk.We conclude that BnCCR1 and BnCCR2 subfamilies offer great but differing potential for manipulating traits associated with phenylpropanoids and glucosinolates.This study reveals the CCR1–CCR2 divergence in Brassicaceae and offers a resource for rapeseed breeding for lodging resistance,yellowseed traits,and glucosinolate traits.

1.Introduction

Lignins,a group of phenylpropanoids that are deposited in plant secondary cell walls,are the second most abundant biopolymer on the planet[1,2].They perform many functions,providing structural support,giving rigidity and strength to stems to stand upright,and enabling xylems to withstand the negative pressure generated during water transport [3,4].Lignin formation has been suggested to be induced by various biotic and abiotic stresses,such as pathogen infection,insect feeding,drought,heat,and wounding [5–7].

Engineering of lignins has targeted mainly reduced lignin content or altered lignin composition to meet the demands of agroindustrial processes,such as chemical pulping,forage digestibility,and bioethanol production from lignocellulosic biomass [8–12].However,altering the expression of lignin pathway genes may change the metabolic flux of neighbor pathways[13–15].Dramatic modification of lignin content or lignin composition may lead to deleterious effects on plant growth,such as dwarfism and collapsed xylem vessels,with concomitant loss of biomass and yield[12,16,17].

Cinnamoyl-CoA reductase(CCR)is the entry point for the ligninspecific branch of the phenylpropanoid pathway and catalyzes monolignol biosynthesis [18,19].Arabidopsis thaliana harbors 11 annotated CCR homologs[20],but only AtCCR1 and AtCCR2 encode true CCR enzymes [21].AtCCR1 is expressed preferentially in tissues undergoing lignification,while AtCCR2 is poorly expressed during development but is strongly and transiently induced by Xanthomonas campestris,suggesting that AtCCR1 might be involved in constitutive lignification and AtCCR2 in resistance [21].However,there has been no functional verification of this assumption via overexpression transgenic studies in A.thaliana,and the function of CCR2 in Brassicaceae has not been dissected.Because in monocot grasses,CCR1 expression can be detected in various organs with relatively high transcription levels in stems [22,23],the gene is thought to be involved in constitutive lignification.In poplar and switchgrass,CCR2 is expressed at very low levels in most organs,but can be induced by biotic and abiotic stresses[3].Manipulation of CCR (mainly by downregulation) typically results in large variation of lignin content and composition [24–26].Plants with heavily downregulated CCR genes usually showed stunted growth and delayed development[12,13,27],accompanied by an alternation of carbon flux between lignin and other metabolic pathways [26,28,29].

Lodging is a fatal problem in field production of most crops,especially herbaceous crops such as cereals and rapeseeds.Reducing plant height has been shown to be effective for increasing lodging resistance,but dwarfism reduces canopy photosynthetic capacity and yield [30–32].Increasing stem strength is a promising strategy for breeding crops with high lodging resistance [33].There are few reports addressing how the regulation of lignin biosynthesis can be induced by gene manipulation to affect crop lodging.In wheat,accumulation of lignin is closely associated with lodging resistance,and wheat culms with higher lignin content showed higher lodging resistance [30].However,there are no reports of use of CCR overexpression to reduce lodging.Furthermore,higher lignin content could lead to a better disease resistance,because lignin is a physical barrier that can confine pathogens to the infection site and confer disease resistance in plants [34].

Brassica,a genus related to that of the model plant A.thaliana,is a source of oilseeds that serve as feedstock for the production of edible vegetable oil and bioenergy,and contains various vegetable and ornamental crops species [35,36].Brassica crops,especially rapeseed (B.napus),frequently experience stresses such as lodging [30,37] and stem rot disease caused by Sclerotinia sclerotiorum with severe yield penalty and quality deterioration[38,39].In rapeseed,lodging may lead to 20%–46% yield losses and 4 percent oil content reduction and limits the efficiency of mechanical harvest [40–42].There is little knowledge concerning the functional genes involved in lignin biosynthesis in rapeseed or other Brassica species.Comprehensive characterization of lignin biosynthesis in rapeseed will permit manipulating lignin content and composition by genetic engineering and to strengthen lodging and pathogen resistance.

The objectives of this study were to isolate CCR1 and CCR2 subfamily members from B.napus and its parental species B.rapa and B.oleracea,investigate BnCCR expression patterns in various organs,and overexpress BnCCR1 and BnCCR2 to identify their biological functions and biotechnological potential.Our results demonstrated that BnCCR1 was involved mainly in the biosynthesis of H-and G-lignins,while BnCCR2 showed a preference for the biosynthesis of S-lignin.A dramatic shift of carbon flux in phenylpropanoid pathway and a strong crosstalk effect on the glucosinolate pathway was demonstrated in both BnCCR1-and BnCCR2-transgenic plants,especially for BnCCR1 manipulation.Besides,BnCCR1 and BnCCR2 showed distinctly different associations with flux regulation and development control,accounting for distinct phenotypic modifications in the vascular system,lodging resistance,seed color,flowering time,and glucosinolate profiles in corresponding overexpression lines.

2.Materials and methods

2.1.Plant materials

Brassica napus:black-seed cultivars 5B and ZS10,yellow-seed cultivar 09L587.B.rapa:black-seed cultivar 09L597.B.oleracea:black seed cultivar 09L598.T1transgenic lines together with non-transgenic wild-type(WT)ZS10 plants were grown in artificial growth room (25 °C,16-h photoperiod/20 °C,8-h dark period).Later generations of transgenic lines and all other materials were planted in field cages,with standard cultivation conditions.Samples were immediately frozen in liquid nitrogen and stored at–80 °C for gene expression analysis,biochemical and histochemical assays,gas chromatography–mass spectrometry (GC–MS) detection,and ultrahigh-performance liquid chromatography-heated electrospray ionization-tandem mass spectrometry (UPLC-HESIMS/MS) analysis.Mature seeds,stems and roots of about 65-DAP(days after pollination) plants were harvested for agronomic trait investigation and biochemical analysis.

2.2.Southern hybridization analysis

Total genomic DNA was extracted from leaves of the three Brassica species (B.napus and its parental species B.rapa and B.oleracea) using a standard cetyltrimethylammonium bromide protocol[43].It was digested with restriction enzymes DraI,EcoRI,EcoRV,HindIII and XbaI (70 μg for each enzyme) respectively,and separated on a 0.8% (w/v) agarose gel.Following electrophoresis,DNA was transferred to positively charged nylon membranes(Roche,Basel,Switzerland) using established protocols.Brassica CCR1-and CCR2-specific probes were amplified by PCR with primer pairs FCCR1C1+RCCR1 and FBCCR2I+RBCCR2I respectively,and labeled using digoxigenin (DIG) Probe Synthesis Kit (Roche),with an annealing temperature of 61 °C and extension time of 1 min.Sequences of all primers are presented in Table S8.Hybridizations with the probes were performed at 43°C overnight,with chemiluminescent detection using a DIG Luminescent Detection Kit(Roche).

2.3.Detection of transcription levels

Total RNA of each sample was extracted with an EASYspin Kit(Biomed,Beijing,China)and RNAprep pure plant kit(Tiangen,Beijing,China),and treated with DNase I to eliminate contaminating gDNA.Equal quantities of RNA (1 μg) were used for the synthesis of total cDNA using the PrimeScript RT reagent kit with gDNA Eraser(TaKaRa Dalian,China).The transcript levels of CCR1,CCR2 and other target genes were detected by both quantitative real-time PCR (qPCR) and semi-quantitative RT-PCR (sRT-PCR) as described previously [44].The 25SrRNA primer pair was used as an internal control in qPCR.A 25-fold dilution series of original reversetranscription products was used for qPCR using an SsoAdvanced Universal SYBR Green Supermix (BioRad,Hercules,CA,USA) on CFX96 Real-Time System(BioRad).Conditions for qPCR were as follows:95°C for 2 min;40 cycles of amplification with 95°C for 30 s and 62°C for 30 s.A melting curve was obtained after amplification by heating products from 60 °C to 95 °C.Transcript levels were determined based on changes in Cq (cycle quantification) values relative to the internal control.qPCR results were analyzed using the CFX Manager 3.0 software (Bio-Rad).Conditions for sRT-PCR on Veriti Thermal Cycler (ABI,USA) were as follows:94 °C for 2 min;31 cycles of amplification with 94 °C for 0.5 min,60–64 °C for 30 s and 72 °C for 1 min;followed by 72 °C for 10 min.The 26SrRNA primer pair was used as an external control in sRTPCR.Sequences of all primers used for qPCR and sRT-PCR are listed in Table S8.

2.4.Rapeseed genetic transformation and screening

A modified Agrobacterium-mediated transformation protocol described by Cardoza et al.[45]was used to transform the‘‘double low”(low erucic acid and low glucosinolates)rapeseed commercial cultivar Zhongshuang 10 (ZS10) with overexpression vectors pCD-BnCCR1-2ox and pCD-BnCCR2-4ox,using hypocotyl segments as explants [46].The regenerated plants were identified first by leaf β-glucuronidase (GUS) staining and leaf Basta-resistance test(200 mg L-1),then by Taq-PCR detection of target genes and the marker gene BAR (primer pair FBar+RBar,annealed at 58 °C,extension for 30 s).The T1transgenic plants were grown in an artificial growth room and were selfed.Representative T2and T3transgenic lines and WT were grown in field cages,and positive plants identified as described above were subjected to trait investigation and further study.

2.5.Agronomic trait investigation

The area and length of the fourth or fifth fully expanded leaves were recorded during vegetative stage,and the length and width of the petioles of these leaves were measured at the same time.The primary branch number,middle stem diameter and stem strength at reproductive stage,and plant trait at harvest stage were recorded.The weight of 1000 seeds and yield per plant were measured after the harvested seeds were dried.For stem strength determination,freshly collected middle stem segments were placed horizontally,and the force exerted to break the stem was recorded with a universal force testing device (model DC-KZ300;Kaiming,Sichuan,China)to measure stem rigidity,which was normalized by the stem’s length and diameter.

2.6.Quantification of insoluble condensed tannins of seed coat

The modified method for quantification of insoluble condensed tannins followed Auger et al.[47] and Naczk et al.[48].The ovendried seed coat was milled to fine powder in a microball mill,extracted with hexane for 12 h in a Soxhlet apparatus,and dried at room temperature.A 10-mg sample of seed-coat powder was extracted with a solution of 3 mL of butanol:HCl (95:5;v/v),300 μL methanol and 100 μL of 2% ferric ammonium sulfate (w/v) in 2 mol L1HCl.The tubes were heated for 3 h at 95 °C in a water bath,centrifuged after cooling,and extracted again after removal of supernatant with the same extraction solution for 1 h.The absorbance of the pooled supernatants was measured at 550 nm against a reagent-only blank with a UV–VIS spectrophotometer(UV-5100B,Shanghai Metash Instruments Co.,Ltd.,Shanghai,China).A calibration curve was prepared using procyanidin with amounts ranging from 0 to 200 μg mL1at intervals of 40 μg mL1.

2.7.Histochemical staining,auto uorescence,and anatomical studies

Cross sections were obtained with frozen section machine(Leica CM1850,Germany).Fresh petioles at vegetative stage,fresh stems and roots at reproductive stage,fresh mature silique pericarp,and dry mature seed coat were cut into slices respectively 60,60,60,60,and 5 μm thick.Phloroglucinol-HCl staining and M?ule staining of the sections were performed as previously described [49,50].Stained sections were observed under stereoscopic microscopes (Nikon C-BD230,Japan;Olympus SZX2-FOA,Japan) and a fluorescence microscope (Nikon Eclipse E600W,Japan).

2.8.Phenolic profiling by UPLC-HESI-MS/MS

2.8.1.Extraction of soluble metabolites

Samples of stems,leaves,petioles,and seeds (30 DAP) were ground in liquid nitrogen in a mortar and pestle,and freezedried in a vacuum freeze drier(Scanvac,Coolsafe 110–4,Denmark).A sample of 30 mg ground lyophilized stems,leaves,and petioles was extracted twice by sonication with 1.0 mL of 50% methanol plus 1.5% acetic acid for 1 h at 4 °C and centrifuged at 15,000 g for 10 min.The supernatants were combined and concentrated in a vacuum concentration (Scanvac,Scanspeed 32,Denmark),redissolved in 0.5 mL of 50% methanol and filtered through a 0.22 μm nylon syringe filter.A sample of 50 mg ground,lyophilized seed material was extracted as previously described [47] with slight modification.One mL of a methanol/acetone/water/TFA mixture(40:32:28:0.05,v/v/v/v) was added to the seed samples followed by sonication for 1 h at 4 °C.After centrifugation (15,000 g,5 min),the pellet was extracted further with 1 mL methanol/acetone/water/TFA mixture overnight at 4 °C under agitation (200 r min1),while the supernatant was stored at–80 °C.Supernatants were pooled,centrifuged at 15,000 g for 10 min,and concentrated.To further remove the water,200 μL methanol was added to the extracts twice during evaporation.The dried extracts were redissolved in 1 mL of 1%acetic acid in methanol,filtered through a 0.22 μm nylon syringe filter,and stored at–80 °C for analysis.

2.8.2.UPLC-HESI-MS/MS analysis

UPLC was performed on a Dionex Ultimate-3000 UHPLC System(Thermo Fisher Scientific,Tacoma,WA,USA),and 5-μL samples were separated on a Waters Acquity UPLC BEH C18 column(1.7 μm,2.1 mm 150 mm).The flow rate was 0.2 mL min1,and the oven temperature was 30°C.Eluent A was 0.1%formic acid in water and eluent B was 0.1%formic acid in acetonitrile.The following gradient was applied for eluting stem,leaf,and petiole extracts:5% B for 5 min,5% B to 95% B for 20 min,and 95% B for 5 min,followed by column washing and re-equilibration.For seed metabolite elution,gradient conditions were as follows:5%to 9%B for 5 min,9%B to 16%B for 10 min,16%B to 50%B for 25 min,50%B to 95% B for 15 min,and 95% B for 5 min,followed by column washing and re-equilibration.Mass analyses were performed with the mass spectrometer Thermo Scientific Q Exactive (Thermo Fisher Scientific) equipped with a HESI source used in negative ion mode.Source parameters were as follows:spray voltage of 3.0 kV,sheath gas flow rate at 35 arbitrary units,auxiliary gas flow rate at 10,capillary temperature at 350°C,and auxiliary gas heater temperature at 300 °C.Nitrogen gas was used as sheath gas and auxiliary gas.Full MS/dd-MS2 were acquired from m/z 100 to m/z 1500.Thermo Xcalibur software version 3.0.63 (Thermo Fisher Scientific) was used for data collection and processing.Contents of metabolites were expressed relative to the calibration curves of available standards.A mixed standard solution with a concentration of 5 mg L1was prepared with stock solutions of standards,and this prepared mixture was diluted to 8 concentration gradients(0.001,0.005,0.01,0.05,0.20,0.50,1.0,and 2.0 mg L1)for calibration curve construction.Standard compounds,namely p-coumaric acid,caffeic acid,ferulic acid,epicatechin,quercetin,kaempferol and isorhamnetin (from PureChem-Standard Co.,Ltd.,Chengdu,Sichuan,China)as well as sinigrin,sinapic acid,coniferyl aldehyde,and abscisic acid (from Sigma-Aldrich,USA) were analyzed under the same conditions described above.

2.9.Statistical issues

Except where unnecessary and in specially indicated situations,all experiments were performed with three replications,and all experimental results were statistically analyzed.Statistical significance was calculated by two-tailed Student’s t-test (*,P ＜0.05;**,P ＜0.01).One-way ANOVA followed by Duncan’s multiple comparisons test was used to identify differences among different samples or plant lines.

2.10.Accession numbers

Sequence data of the genes and proteins involved in this article can be found in NCBI (http://www.ncbi.nlm.nih.gov/) nr/nd,est,tsa,gss,wgs,and ref_seq rna/protein databases under the accession numbers indicated in Table S1 and Fig.1.

2.11.Other methods

Details of the isolation of Brassica CCR cDNA and genomic sequences,bioinformatic methods for this study,construction of overexpression vectors,measurement of leaf SPAD readings,determination of lignin content and composition,and near-infrared reflectance spectroscopy (NIRS) measurements are provided in Supplementary methods.

3.Results

3.1.Isolation and characterization of CCR genes from Brassica species

Fig.1.Phylogenetic tree of CCR refseq_proteins from representative wholegenome-sequenced Malvid species including B.napus,B.rapa,and B.oleracea.Latin name,refseq_protein name,and accession number are provided for each sequence.Names of genes cloned in this study are given after accession numbers.Vertical bars on the right indicate corresponding order or order and family names.

CCR genes were isolated from B.napus and its parental species B.rapa and B.oleracea using a RACE strategy.Full-length cDNAs and corresponding gDNA sequences of 3,4 and 6 CCR1-subfamily gene sequences were isolated from B.rapa (BrCCR1-1,BrCCR1-2 and BrCCR1-2′),B.oleracea (BoCCR1-1,BoCCR1-1′,BoCCR1-2,and BoCCR1-2′),and B.napus (BnCCR1-1,BnCCR1-2,BnCCR1-2′,BnCCR1-3,BnCCR1-3′,and BnCCR1-4),while respectively 3,3,and 4 CCR2-subfamily gene sequences were isolated from B.rapa(BrCCR2-1A,BrCCR2-1B and BrCCR2-2),B.oleracea (BoCCR2-1A,BoCCR2-1B,and BoCCR2-2) and B.napus (BnCCR2-1,BnCCR2-2,BnCCR2-3,and BnCCR2-4).One CCR2 pseudogene from each parental species and two CCR2 pseudogenes from B.napus were also isolated (Table S1).To determine the copy number of the BnCCR1,BoCCR1,BrCCR1,BnCCR2,BoCCR2,and BrCCR2 genes,Southern hybridization was performed (Fig.S1),and the numbers of clear fragment bands were well correlated with the cloned gene sequence numbers.After cloning in the Brassica CCR family,whole genome sequencing and annotation data of B.rapa,B.oleracea and B.napus were available in GenBank.In silico cloning of all CCR1 and CCR2 subfamily genes,indicated that each parental species contained only two CCR1 and two CCR2 genes along with one CCR2 pseudogene,whereas B.napus harbored total sets of both parental species(Table S1).Considering the high full-length sequence identities(96.0–99.0%)within each pair of BrCCR1-2/BrCCR1-2′,BoCCR1-1/BoCCR1-1′,BoCCR1-2/BoCCR1-2′,BnCCR1-2/BnCCR1-2′,BnCCR1-3/BnCCR1-3′,BrCCR2-1A/BrCCR2-1B and BoCCR2-1A/BoCCR2-1B,it was clear that the CCR1 and CCR2 gene or pseudogene numbers in the haploid genomes were only those revealed by in silico cloning and that the above sequence pairs were cloned from two heterozygous alleles of each gene in the amphiploid stocks.Table S1 shows all identity parameters of CCR1-subfamily and CCR2-subfamily genes/pseudogenes from B.napus and its parental species B.oleracea and B.rapa with respect to the cloning and three genome datasets(NCBI GenBank,Genoscope,and BRAD).The basic features of the cDNA and gDNA sequences are displayed in Table S2,and the deduced protein features are depicted in Table S3.

Multiple alignments showed that the conserved NADP binding domain and the CCR specific typical motif (KNWYCYGK,which was assigned as the catalytic site) and the novel motif H202K205R253(the CCR-SBM or CCR substrate-binding motif) were conserved in all Brassica CCRs(Fig.S2)[19,51].Because these Brassica CCRs were speculated to have catalytic activities,a phylogenetic tree was constructed to reveal the relationships of Brassica CCR1 and CCR2 proteins with CCRs from other Brassicales species whose genomes have been sequenced and from species of other malvid orders (Fig.1).First,it was clear that an early intrafamily duplication event generated the CCR1 and CCR2 groups within Brassicaceae,given that AtCCR1 and AtCCR2 harbor CCR1 and CCR2 orthologs only within the Brassicaceae family.Second,in Brassicaceae the CCR1 group is much more conserved than the CCR2 group,possibly reflecting the great difference in branch lengths.Third,outside the Brassicaceae family,other Brassicales families (such as Cleomaceae and Caricaceae) and other malvid orders(such as Malvales,Myrtales,and Sapindales)showed similar trends in CCR evolution(gene duplication and unequal divergence of paralogs) to those revealed in Brassicaceae.However,paralog numbers (numbers of duplication events) vary distinctly among relative families and orders,and the single-copy CCR from the Brassicales species Carica papaya is nearer to non-Brassicales CCRs than to other Brassicales CCRs.

3.2.Divergent involvement in organ-specificity and seed coat color among BnCCR family members

BnCCR1 was clearly detected in all organs and was highest in silique pericarp,then in buds,developing seeds,flowers,and roots.Among the subfamily BnCCR1 members,BnCCR1-2 showed relatively intense expression.However,BnCCR2 showed highest expression in buds,then in leaves,30-DAP seeds and 45-DAP seeds,and its subfamily member BnCCR2-4 showed relatively strong expression in all tested organs.(Fig.S3).In B.napus,CCR1 subfamily overall expression was distinctly lower in developing seeds,especially in late-stage seeds of yellow-seed stocks,than in black-seed stocks,whereas the CCR2 subfamily showed an opposite trend (Fig.S3).

3.3.BnCCR overexpression strongly in uenced plant morphology

To further investigate the function of BnCCR in lignification(especially in lodging performance) and stress conditions,overexpression transgenic plants were generated.BnCCR1-2 and BnCCR2-4 were selected for transgenic study as they are dominant members within respective subfamilies.The transgenic lines are coded as BnCCR1-2ox or ox1 lines for BnCCR1-2 overexpression,and BnCCR2-4ox or ox2 lines for BnCCR2-4 overexpression,in the following description.The transgenic lines were screened along with non-transgenic controls (wild type,WT) by GUS staining,Basta resistance,and PCR detection (Fig.S4).Six BnCCR1-2ox and 7 BnCCR2-4ox triple-positive lines were obtained,with different overexpression levels revealed by qRT-PCR (Fig.S5).

All BnCCRox plants showed stronger morphological development than WT throughout the whole life (Fig.2).Both BnCCR1-2ox and BnCCR2-4ox had larger and longer leaves,higher leaf chlorophyll content,larger stem diameter,wider silique,higher breaking-resistant stem strength,higher lodging resistance,and more siliques per plant,with distinctly stronger effects in BnCCR1-2ox than in BnCCR2-4ox lines (Figs.2,S6,S7).Metabolites detection showed that abscisic acid (ABA) increased significantly in the leaves of BnCCRox lines compared with WT (Fig.S8).The leaves of BnCCR1-2ox showed more pronounced wrinkles and less leaf margin serration than WT (Figs.2C,S7E).BnCCR2-4ox plants showed a looser morphology with larger leaf angles (Fig.2B,J),and their leaves more readily showed bending and rolling than WT under strong sunlight and high temperature(Fig.S7D).In contrast,upper stems of BnCCR1-2ox plants at late bolting stage more easily showed bending,but this trait disappeared after flowering(Figs.2F–I,S7A).BnCCR1-2ox lines flowered 7–10 days later than WT on average,and this phenomenon varied among years and environments.However,BnCCR2-4ox plants showed no difference from WT in flowering time.The petiole-vein system was distinctly larger in BnCCR1-2ox than in WT plants (Figs.2D,S7E),and phloroglucinol-HCl staining showed that petiole-vein lignification was markedly strengthened in BnCCR1-2ox plants but weakened in BnCCR2-4ox plants (Fig.S7I,J).Seed yield per plant of BnCCR1-2ox showed little difference from that of WT,while BnCCR2-4ox showed slightly but nonsignificantly decreased seed yield(Fig.S6L).

Fig.2.Differing phenotypic modifications in BnCCR1-2ox and BnCCR2-4ox plants.(A–M) Plant phenotypes at several stages.(A–D) Vegetative stage;(E–G) Middle-later bolting stage;(H–J)Flowering stage;(K–M)Harvest stage.Bending phenomenon of the upper stem of BnCCR1-2ox plants at late bolting stage(F)will disappear after flowering(I).(N,O)Root system at mature stage.(P,Q)Root system of 1-week old seedlings,and the investigated values with statistical significance.Values represent means±SD of at least 5 biological replicates.Asterisks indicate that means differed from WT values(*,P ＜0.05;**,P ＜0.01,Student’s t-test).(R)Siliques at mature stage,indicating the wider siliques of BnCCR1-2ox and BnCCRR2-4ox compared with WT.Scale bars,1 cm.

3.4.Yellow-seed traits in BnCCRox lines

As displayed in Fig.3A,the seed color of both BnCCR1-2ox and BnCCR2-4ox turned lighter compared with WT,and BnCCR1-2ox showed a stronger effect than BnCCR2-4ox.Microscopic investigation of the frozen sections of seed coat (Fig.3B) and R value obtained by NIRS assay (Fig.3E) further showed this effect.However,the thickness of the seed coat showed no significant difference between BnCCRox and WT,which was different from conventionally bred yellow-seeded cultivars,which usually have thinner seed coats than black-seeded cultivars [52,53].The BnCCR1-2ox seed coat showed a reduction in condensed tannin compared with WT:ox1-5,ox1-8,ox1-12,and ox1-14 showed decreases of 69%,29%,40,and 57%,respectively (Fig.3C,G).But large increases in condensed tannins were found in the BnCCR2-4ox seed coat relative to the WT:ox2-4,ox2-11,ox2-16,and ox2-25 showed increases of 88%,300%,158%,and 83%,respectively(Fig.3C,G),implying looser condensation of the tannin structure caused by unknown mechanisms.All BnCCRox lines showed large reductions in thousand-seed weight,with those of BnCCR1-2ox lines reduced by 10%–20% and those of BnCCR2-4ox lines by 15%–30% (Fig.3D).By NIRS,the total glucosinolate content of BnCCR1-2ox lines was greatly increased relative to WT,but BnCCR2-4ox lines showed a slight decline (Fig.3F).

3.5.BnCCR overexpression greatly changed lignification phenotypes

The frozen cross sections of stems showed that both BnCCR1-2ox and BnCCR2-4ox stems had changed shapes with more concavity and convexity and a wider xylem part with deeper histochemical staining (Fig.4A–C).Some extreme BnCCR1-2ox lines showed ectopic lignin deposition(Fig.4C).Observed under higher amplification of the sections,BnCCR1-2ox showed more highly developed xylem and more abundant and more concentrated vessels with deeper brown color (indicating more G-lignin units) in M?ule staining in which G-and S-type lignin units were stained brown and red,respectively (Fig.4F,G,early flowering stage),with brighter red phloroglucinol-HCl staining (Fig.4H,mature stage)and brighter blue fluorescence (Fig.4I,J,early flowering stage) in comparison with WT.These observations indicated that BnCCR1-2ox stems contained more lignin than WT.BnCCR2-4ox stem sections also displayed more highly developed xylem and much clearer interfascicular fiber according to the histochemical staining and fluorescence results (Fig.4F–J).This result indicated that BnCCR2-4ox stems contained a higher proportion of S-lignin (according to M?ule staining result),besides higher total lignin content.Similar trends of increased lignification were found in roots(Fig.4D,E) and siliques (Fig.4K–N).

For leaf traits,the petiole of BnCCR1-2ox tended to be more circular with more and better developed vascular bundles(Fig.4O,Q).But in BnCCR2-4ox lines,it was smaller than WT and developed asymmetrically with fewer and less developed vascular bundles(Fig.4O,Q),possibly accounting for the greater readiness of BnCCR2-4ox than WT leaves to bend and roll under strong sunlight and high temperature.BnCCR1-2ox lines showed a deeper brown color in M?ule staining and a brighter blue fluorescence under UV light than WT (Fig.4P,Q),indicating that BnCCR1-2ox petiole xylem contained more lignin as well as a higher G-lignin proportion.The BnCCR2-4ox petiole sections displayed no greater M?ule staining,and the blue fluorescence was markedly weaker than WT (Fig.4Q),indicating that BnCCR2-4ox petiole xylem contained less lignin,in agreement with the phloroglucinol-HCl staining of leaves (Fig.S7J).

Fig.3.Yellow seeds from T3 plants of BnCCR1-2ox and BnCCR2-4ox in contrast with black seeds from WT plants.(A,B)Seed and seed coat cross sections,showing the yellowseed trait caused by seed color lightening by CCR overexpression.Scale bars,50 μm.(C) Extractable insoluble condensed tannins from seed coat,showing deeper color of BnCCR2-4ox in comparison with BnCCR1-2ox and WT.(D) 1000-seed weight.(E) Seed R value (higher value means deeper yellow color of the seed coat).(F) Glucosinolate content (μmol g 1).(G) Insoluble condensed-tannin content of the seed coat.Values are means ± SD of at least three biological replicates.Different letters following the SD indicate significant difference at P ＜0.05 (one-way ANOVA,Duncan’s test).

3.6.Changes in content and structure of lignins in BnCCR overexpression lines

The acetyl-bromide-soluble lignin content of both BnCCR1-2ox and BnCCR2-4ox was significantly increased relative to WT(Fig.4R–T).The lignin content of stems of BnCCR1-2ox lines ox1-5,ox1-8,ox1-12,and ox1-14 increased by 24%,34%,23% and 15%,respectively (Fig.4R).For BnCCR2-4ox lines ox2-4,ox2-11,ox2-16,and ox2-25,their stem lignin content increased by 40%,25%,31%and 51%,respectively(Fig.4R).In roots,the lignin content of ox1-5,ox1-8,ox1-12,ox1-14,ox2-4,ox2-11,ox2-16,and ox2-25 increased by 0.4%,18%,15%,5%,18%,25%,12% and 23% compared with WT,respectively(Fig.4S).Similar trends were observed in the seed coat of BnCCR1-2ox lines(Fig.4T).In agreement with the tannin detection result,the seed coat lignin content test values of the BnCCR2-4ox lines increased by 92%–130%(Fig.4T),implying possibly decreased lignin polymerization.

The S/G ratio,which is typically used to characterize lignin structure,changed significantly in both stems and roots of BnCCRox lines.In the stems of BnCCR1-2ox lines,the G-unit,other than Sunit,served as the main lignin units.Owing to the increased Gunit proportion,the S/G ratio in the stems of ox1-5 and ox1-12 decreased by 47.3% and 24.0%,respectively,relative to WT(Figs.4U,W,S9).However,in stems of BnCCR2-4ox,the S/G ratio was increased by 20.2% and 22.5% in ox2-4 and ox2-16 relative compared with WT (Fig.4U,W).In roots,the alteration tendency of S/G ratio of ox1-5 and ox2-16 was similar to that in stems,decreased to 0.22 and increased to 0.78,respectively,as compared with WT (0.49) (Fig.4X).

Another striking change was in proportion of H-lignin,which is generally in trace amounts in dicotyledonous stems.The H-lignin content of ox1-5 and ox1-12 was about 4-and 3-fold that of the WT respectively,while was increased by only about 50% in ox2-4 and ox2-16(Figs.4U,S9).In the root of ox1-5,the H-lignin content also showed a marked increase,reaching about 2-fold the WT amount (7.43% of WT vs.15.59% of ox1-5) (Figs.4V,S9).Thus,the lignin structure of B.rapus was clearly modified after the manipulation of BnCCR genes.However,the cellulose and hemicellulose contents of both BnCCR1-2ox and BnCCR2-4ox lines showed no difference relative to WT.

3.7.Metabolic remodeling of phenylpropanoid and glucosinolate pathways by overexpressing BnCCR1 and BnCCR2

The drastic alteration of seed color,seed coat condensed tannin content,lignin content,lignin structure and plant phenotypes all indicated flux change within and outside the lignin biosynthetic network in BnCCR overexpressors.

Conforming to theoretical expectation,most phenolic compounds synthesized at the downstream of CCR were increased in the stems of ox1-5 and ox2-16 lines,including sinapoylhexose,sinapic acid,sinapoyl malate,ferulic acid,feruloyl malate,pcoumaraldehyde,and 1,2-disinapoylglucoside,with an increase of 1–10 folds,and ox2-16 displayed stronger effects than ox1-5(Fig.5A;Table S4).However,flavonoids in ox1-5 stems were reduced to 3%–71% of WT levels,including km-3-O-sophoroside-7-O-glucoside,rutin,is-3-sophoroside-7-glucoside,km-3-O-sina poylsophoroside-7-O-glucoside,qn-3-O-sophoroside,qn-3-Oglucoside,km-3-O-glucoside,and is-3-O-glucoside;and ox2-16 showed the same trend as ox1-5 but to a lower extent(Table S4).Leaf extracts of ox1-5 and ox2-16 showed similar variation to those in in stems,but some compounds were undetectable in leaves or showed opposite changes(Table S4).Metabolite profiling was also performed on 30 DAP seeds.As expected,the CCRdownstream compounds sinapic acid and disinapoylgentiobiose were significantly increased in ox1-5 and ox2-16 (Table S5).The contents of most compounds of epicatechin,procyanidin,epicatechin polymers,and other flavonoids were significantly reduced(by even more than 90% for some flavonoids) in ox1-5 relative to WT,and ox2-16 showed the same trend as ox1-5 but to a lower extent (Fig.5B;Table S5).

In metabolic profiling,glucosinolates (products of nonphenylpropanoid compound pathway) were unexpectedly found to be drastically changed by BnCCR overexpression.A variety of aliphatic glucosinolates were distinctly differentially deposited between BnCCRox lines and WT.For example,2(R)-2-hydroxy-3-butenyl glucosinolate,1-S-[(3S)-3-Hydroxy-N-(sulfooxy)-5-hexeni midoyl]-1-thio-beta-D-glucopyranose,3-butenylglucosinolate,isobutyl glucosinolate,4-pentenyl glucosinolate,5-methylsulfinylpentyl glucosinolate and 5-methylthiopentyl glucosinolate were upregulated to hundreds of folds or even more than 1000-fold in ox1-5 stems relative to WT,while in ox2-16 stems they were just slightly upregulated or even downregulated(Fig.5C;Table S4).4-Methylthiobutyl glucosinolate and 6-methylthiohexyl glucosinolate were markedly accumulated in ox1-5 but not in ox2-16 and WT stems.Glucosinolates showed larger accumulations in leaves than in stems (Table S4),and most of them in leaves were several folds higher in ox1-5 than in WT.However,most of them were downregulated by hundreds of folds in leaves of ox2-16 in comparison with WT.2(R)-2-hydroxy-3-butenyl glucosinolate,1-S-[(3S)-3-hydroxy-N-(sulfooxy)-5-hexeni midoyl]-1-thio-beta-D-glucopyranose,and 5-methylthiopentyl glucosinolate were not even detectable in the leaves of ox2-16(Table S4).Glucosinolate variation in seed coats was similar to that in leaves(Table S5).Modification of secondary metabolites in petioles was similar to that in stems (Table S6).

3.8.Differential redirection of gene expression in lignin, avonoid,and glucosinolate pathways in BnCCR1 and BnCCR2 over-expressors

In BnCCR1-2ox lines,the BnCCR1 subfamily itself was greatly upregulated,and the BnCCR2 subfamily showed no significant upregulation.In BnCCR2-4ox lines,the BnCCR2 subfamily itself was greatly upregulated,while the BnCCR1 subfamily was also significantly upregulated (1–3 folds) in stems,petioles,and 30-DAP seeds but significantly downregulated by 70% in leaves (Fig.S10).These results indicated that overexpression of BnCCR1 subfamily had little impact on the expression of the BnCCR2 subfamily,but that overexpression of the BnCCR2 subfamily upregulated or downregulated the expression of the BnCCR1 subfamily depending on the organ.

In stems of both BnCCR1-2ox and BnCCR2-4ox lines,the common phenylpropanoid pathway loci C4H and 4CL and the ligninpathway early-step loci C3H,HCT and CCoAOMT were mildly upregulated (Fig.6A),and CCR-downstream loci CAD and F5H were slightly upregulated.COMT was specific,apparently upregulated in the stems of BnCCR1-2ox,but downregulated to less than 10%in the stems of BnCCR2-4ox relative to WT.

In contrast,the expression of the flavonoid biosynthesis pathway was significantly downregulated in BnCCRox lines.Regulatory genes TT1,TT2,TT16,TTG1,and TTG2,and structural genes CHS,CHI,F3H,F3′H,FLS,ANR,TT19,GSTF11,and TT12 were all suppressed to an extent in 30-DAP seeds of both BnCCR1-2ox and BnCCR2-4ox(Fig.6B),but less in BnCCR2-4ox than in BnCCR1-2ox.CHS,CHI,F3H,ANR,and TT19 were downregulated to less than 20% in 30-DAP seeds of ox1-5 relative to WT (Fig.6B).The structural genes AHA10 and TT10 and the regulatory genes TT8 and MYB111 were significantly downregulated in 30-DAP seeds of BnCCR2-4ox,but were unexpectedly significantly upregulated in 30-DAP seeds of BnCCR1-2ox lines (Fig.6B).The overall downregulation of the whole flavonoid pathway could account for the reduction of the flavonoids in BnCCRox lines.

Fig.4.BnCCR1-2ox and BnCCR2-4ox plants show differently fortified lignification patterns in various organs.(A–C)Whole frozen stem cross sections at early flowering stage(A,B),and whole freehand stem cross section at harvesting stage(C).(A)M?ule staining;(B,C)Phloroglucinol-HCl staining.Scale bars,2 mm.(D,E)Cross sections at mature stage of roots with more interfascicular fibers (D) and fewer interfascicular fibers (E),stained with phloroglucinol-HCl.Scale bars,500 μm.(F–J) Cross sections of different parts of the stem at several development stages.Middle-lower part of the stem at early flowering stage using M?ule staining with several amplification folds(F,G).Middlelower part of the stem at mature stage with phloroglucinol-HCl staining (H).Autofluorescence of the middle stem at early flowering stage under UV light with different amplification folds(I,J).if,interfascicular fiber;ph,phloem;pi,pith;xy,xylem.Scale bars,200 μm in(F,I)and 80 μm in(G,H,J).(K–N)Silique wall cross sections at mature stage.(K)Whole sections;(L–N)Local sections;(K,L)Phloroglucinol-HCl staining;(M,N)Autofluorescence viewed under UV light.Scale bars,500 μm in(K),100 μm in(L,M),and 50 μm in (N).(O–Q) Cross sections of the middle petiole with different methods.(O) Whole cross sections without any treatment;(P) The vascular bundle with M?ule staining;(Q)Autofluorescence of the vascular bundle under UV light;Scale bars,1000 μm in(O),100 μm in(P),and 50 μm in(Q).(R–T)Total lignin content analysis of stems(R),roots (S) and seed coats (T) by AcBr method.CWR,cell wall residue.Values are means ± SD of at least 3 biological replicates.Different letters above bars indicate significant difference at P ＜0.05(One-way ANOVA,Duncan’s test).(U–X)Lignin monomer compositions of stems(U)and roots(V)were measured by thioacidolysis method.S/G ratios in stems and roots are displayed in (W) and (V),respectively.Values are means ± SD of at least 3 biological replicates.Different letters and values above bars indicate significant difference at P ＜0.05 (One-way ANOVA,Tukey’s test) and respective percentages of the corresponding lignin monomers.

Fig.5.Secondary metabolites were distinctly modified in stems of BnCCR1-2ox and BnCCR2-4ox plants.Secondary metabolites were determined by UPLC-HESI-MS/MS.(A)Major soluble metabolites associated with the lignin pathway and its derivative pathways in the stems of ox1-5 and ox2-16 and WT,and see details in Table S5.(B) Major soluble metabolites associated with the flavonoid pathway in the seeds of ox1-5 and ox2-16 and WT,and see details in Table S6.DP,degree of polymerization of the epicatechin unit.(C) Major soluble metabolites associated with the glucosinolate pathway in the leaves of ox1-5 and ox2-16 and WT,and see details in Table S5.

Most of the genes associated with aliphatic glucosinolate biosynthesis,such as MYB28,MYB29,CYP79F1,CYP83A1,AOP2 and GSL,were significantly upregulated in BnCCR1ox lines(Fig.6C).In particular,the expression of MYB29,CYP79F1,and GSL could hardly be detected in WT and BnCCR2ox lines,but was strongly expressed in BnCCR1ox lines.AOP2 showed downregulation in BnCCR2ox,a trend opposite to that in BnCCR1ox (Fig.6C).The expression of most genes involved in indole glucosinolate biosynthesis,including MYB122,CYP79B2,ST5a,and IGMT1,was greatly decreased in both BnCCR1ox and BnCCR2ox lines,except that MYB34 was significantly upregulated in both BnCCR1ox and BnCCR2ox lines (Fig.6C).MYB51 was significantly upregulated in BnCCR1ox lines but extremely downregulated in BnCCR2ox lines,suggesting that MYB51 may also be involved in aliphatic glucosinolate biosynthesis in B.napus.

4.Discussion

4.1.Both BnCCR1 and BnCCR2 are necessary for lignification,associated with different monolignols and cellular types

Fig.6.Gene expression patterns of lignin,flavonoid,and glucosinolate pathways differed in BnCCR1-2ox and BnCCR2-4ox plants.(A)Transcript levels of genes associated with lignin biosynthesis in stems.(B) Transcript levels of genes associated with flavonoid biosynthesis in seeds.(C) Transcript levels of genes associated with glucosinolate biosynthesis in leaves.Because MYB29, CYP79F1 and GSL have no expression in WT,their expression level in BnCCRox is set relative to zero.The expression levels of other genes are set relative to WT.The yellow frame marks genes involved in the synthesis of aliphatic glucosinolates.The red frame marks genes involved in the synthesis of indole glucosinolates.

Plants with downregulated CCR activities often display reduction in lignin content and alteration in lignin structure depending on the species [24,25,54].Lignin content showed a great decrease in Arabidopsis CCR1 mutant irx4 [55] and CCR-suppressed transgenic Arabidopsis[24],tobacco [16],poplar [56],Medicago truncatula [25] and perennial ryegrass [23].In this study,anatomical observation and histochemical assays both indicated increased lignification in stems and roots in both BnCCR1 and BnCCR2 overexpressors,and the BnCCR1 overexpressor showed a stronger effect on development of vessels and vascular bundles in xylem,while BnCCR2 overexpressor showed a stronger effect on interfascicular fibers.

The S/G ratio was decreased in BnCCR1-2ox,owing mainly to the stronger increase of G-lignin,but the S/G ratio in BnCCR2-4ox was increased,owing mainly to the stronger increase of S-lignin.The alteration of S/G ratio caused by CCR manipulation shows different trends in different species.A lower S/G ratio,caused mainly by a relatively strong effect of CCR downregulation on S-units,was observed in Arabidopsis irx4 mutants [57],CCR-downregulated poplar [56],and ccr1-knockout M.truncatula mutants [25].A higher S/G ratio,caused mainly by a relatively strong effect of CCR downregulation on G-units,was found in CCRdownregulated tobacco [58],maize [27],dallisgrass [22],and CCR2-knockout M.truncatula mutants [25].No obvious change of S/G ratio was observed in perennial ryegrass with CCR manipulation [23].The present study revealed opposite trends in S/G ratio controlled by different CCR paralogs within a species.

The H-lignin percentage in stems of BnCCR1-2ox was 2–3 folds higher than that in WT and more than 1-fold higher in roots.This difference could be due to the increased accumulation of pcoumaraldehyde and expression of F5H and COMT in the BnCCR1-2 overexpressor.Reasonably,p-coumaroyl-CoA and caffeoyl-CoA might serve as primary substrates for BnCCR1 (as suggested in Fig.7).In CCR1-downregulated perennial ryegrass [23] and the Mu-insertion maize mutant Zmccr1 [27],the H-subunit levels were reduced by 50% and 31% respectively,suggesting that the optimal substrate of CCR1 in those species was p-coumaroyl-CoA.In contrast,BnCCR2 might prefer feruloyl-CoA as its major substrate,given the higher accumulation of S-units and their downstream derivatives,sinapate esters,and higher increase in transcript level of F5H in BnCCR2ox than in BnCCR1ox (Fig.7).

4.2.Upregulation of BnCCR genes increased lodging resistance and modified morphology

The expression perturbation of the genes located in the lignin biosynthetic pathway was often accompanied with defects in plant growth and development depending on the gene targeted.In severely silenced CCR plants,phenotypic abnormalities with irregular vessels usually arise,including plant size reduction,delayed flowering,delayed senescence,retarded seed development,biomass yield reduction,and compromised pathogen defense[12,25,56,59–62].

Fig.7.BnCCR1 and BnCCR2 play different roles in the phenylpropanoid pathway.Metabolic flux shifts in BnCCR1-2ox and BnCCR2-4ox plants are displayed in this map.The main route,which is conserved in angiosperms,is marked with the large background arrows(the general phenylpropanoid pathway is marked in light blue,while the ligninspecific pathway is marked in brown).The blue arrows and names represent flux increases associated with BnCCR1-2ox.Metabolite accumulation and associated gene expression are increased in the stem of BnCCR1-2ox,indicating that overexpression of BnCCR1 subfamily promotes the biosynthesis of G-and H-lignin units.The purple arrows and names represent flux increase associated with BnCCR2-4ox.Metabolites accumulation and associated gene expression are increased in the stem of BnCCR2-4ox,indicating that overexpression of the BnCCR2 subfamily promotes the biosynthesis of S-lignin units.Metabolites and enzymes downregulated in both BnCCR1-2ox and BnCCR2-4ox plants are marked in green,whereas those with upregulation are in red.Unexpectedly,glucosinolates deposition and pathway gene expression are differently remodeled in BnCCR1-2ox and BnCCR2-4ox plants.Dashed arrows represent unknown or unauthenticated routes.Arrows with a question mark are pathways suggested by this study.Two successive arrows represent two or more metabolic conversions.ANR,anthocyanidin reductase;ANS,anthocyanidin synthase;CAD,cinnamyl alcohol dehydrogenase;4CL,4-coumarate:CoA ligase;C3H, p-coumarate 3-hydroxylase;C4H,cinnamate 4-hydroxylase;CCoAOMT,caffeoyl-CoA O-methyltransferase;CCR,cinnamoyl-CoA reductase;CHI,chalcone isomerase;CHS,chalcone synthase;COMT,caffeic acid O-methyltransferase;CSE,caffeoyl shikimate esterase;DFR,dihydroflavonol 4-reductase;F3H,flavanone 3-hydroxylase;F3′H,flavonoid 3′-hydroxylase;F5H,ferulate 5-hydroxylase;FLS,flavonol synthase;HCALDH,hydroxycinnamaldehyde dehydrogenase;HCT,hydroxycinnamoyl-CoA:shikimate/quinic hydroxycinnamoyl transferase;LAC,laccase;LDOX,leucoanthocyanidin dioxygenase;Med,mediator;PAL,phenylalanine ammonia lyase;PER,peroxidase;SGT,sinapate 1-glucosyltransferase;SMT,sinapoylglucose:malate sinapoyltransferase;SST,sinapoylglucose:sinapoylglucose sinapoylglucosetransferase;UGT,uridine diphosphate glycosyltransferase.

First,BnCCR overexpression increased lodging resistance in comparison with WT,especially for BnCCR1-2ox(Figs.2,S7).When CCR1 was reintroduced into A.thaliana ccr1 mutants under the control of the ProSNBE promoter,specific expression of CCR1 in protoxylem and metaxylem vessel cells resulted in full recovery of vascular integrity [62].The breaking resistance of both BnCCR1-2ox and BnCCR2-4ox plants was greatly increased in comparison with WT,with larger effect in BnCCR1-2ox lines than in BnCCR2-4ox lines.These results imply improved lodging resistance in BnCCRox lines due to greater development and growth in both morphology and lignification of roots and stems.They suggest the potential of molecular breeding of rapeseed with increased lodging resistance by overexpressing BnCCR genes.

Second,BnCCR overexpression also modified leaf morphology.M?ule staining indicated that rapeseed petiole xylem was composed exclusively of vascular bundles without interfascicular fibers.This structure could account for better development of leaf veins in BnCCR1-2ox plants,given that vascular bundles contained higher proportions of G-lignin and BnCCR1-2 was involved mainly in biosynthesis of G-and H-lignins.In Arabidopsis,CCR1 also mediated cell proliferation exit for leaf development,and ccr1-4 mutants showed greatly reduced leaf and plant sizes in comparison with WT [61].However,BnCCR2-4ox plants showed lessdeveloped leaf vascular bundles (Figs.4O,S7J),owing possibly to two causes:weakened G-unit synthesis(given that BnCCR2-4 overexpression forced mainly S-unit synthesis),and downregulated BnCCR1 expression in leaves of BnCCR2-4ox plants (Fig.S10B).Increased levels of ABA might also function in the alteration of leaf phenotypes of BnCCRox plants (Fig.S8),given that ABA influences plant growth and development [63].

4.3.BnCCR1 and BnCCR2 differentially affect plant development

During the growth and development of BnCCRox plants,the upper stems of BnCCR1-2ox plants at bolting stage were easily bent,whereas those of BnCCR2-4ox plants were not (Figs.2F,S7A).As reported [22,23],S-lignin percentage in stems gradually increased during plant growth.M?ule staining of stem cross sections in this study revealed a similar trend (Fig.S11),suggesting that S-lignin influences maturation and mechanical strength of stems.Kaur et al.[64] found that the stems of CAD-downregulated Nicotiana plants (ir-CAD) presented a rubbery phenotype,and the S/G ratio was reduced.In the present study,the S/G ratio was also decreased in BnCCR1-2ox plants (1.29 in WT,0.68 in ox1-5,and 0.98 in ox1-12).Moreover,in WT the S/G ratio was higher in stem (1.29) than in root (0.49),which might be one factor contributing to the more flexible texture of roots than of stems,although total lignin content was higher in root than in stem(Fig.4R,S).A lower proportion of Sunits might contribute to a lower level of stiffness of the stem,together with the lower lignin content in the upper stem,which might partially account for the bending of BnCCR1-2ox at bolting stage.However,when BnCCR1-2ox plants entered flowering stage,the bending phenomenon disappeared,possibly owing to increases in total lignin content and S-lignin percentage in stems (Fig.S11).The relationship between S/G ratio and the texture of plant stems is summarized in Table S7.These results suggest that a higher S/G ratio might contribute to greater stiffness in plant stems,manifested during reproductive growth.

That BnCCR1-2ox plants flowered markedly later than WT and BnCCR2-4ox plants suggests that the function of BnCCR1 is also associated with plant development progress.There was an evolutionarily conserved mechanism between cell wall biosynthesis and production of flowers [65].The delayed accumulation of Slignin and lowered stem stiffness at bolting stage were reflections of delayed development progress and flowering in BnCCR1-2ox plants.

4.4.BnCCR1 and BnCCR2 exert different ux control on avonoid pathway and lignin-derivative pathways in B.napus

In metabolic engineering,products,precursor steps,and associated neighbor pathways may be affected by metabolic flux redirection.In lignin engineering,for example,higher amounts of vanillin,ferulic acid,p-coumaric acid,coniferaldehyde,and syringaldehyde were released from cell-wall samples of MtCAD1 mutants than from those of the wild type [66].In the C3′H-[67] or HCT-defect[68] Arabidopsis,C3H1-downregulated maize [69],CCR-silenced tomato [28],and perennial ryegrass [23],the accumulation of flavonoids was significantly increased.

Besides increased lignin monomers,both BnCCR1-2 and BnCCR2-4 overexpressors showed increased levels of various CCRdownstream pathway products in comparison with WT,especially for sinapate esters (Fig.5;Tables S4–S6),possibly owing to the increased expression levels of the corresponding genes CCR,F5H,COMT,and CAD.The finding that BnCCR2-4ox plants showed a stronger effect than BnCCR1-2ox plants on the accumulation of sinapate esters in stems suggests that the metabolic route associated with BnCCR2 catabolism (mainly S-unit synthesis) might be closer to the sinapate ester pathway than BnCCR1 (Figs.5A,7).

Many compounds synthesized in flavonoid pathway were reduced in BnCCRox plants relative to WT.The great reduction in flavonoids was theoretically caused by the reduced availability of p-coumaroyl-CoA precursor for CHS,as BnCCR overexpression attracted more of this common precursor into the lignin pathway(Fig.7).This flux shift showed further evidence at the molecular level(Fig.6).Seed color degree,metabolic profile,and gene expression profile all indicated that suppression of the flavonoid pathway was greater in BnCCR1-2ox than in BnCCR2-4ox plants (Figs.3,6;Tables S4–S6).The explanation may be that the metabolic routes of BnCCR1 catabolism(mainly H-and G-unit synthesis)were closer to the flavonoid pathway than that of BnCCR2 (mainly S-unit synthesis),as suggested in Fig.7.

4.5.BnCCR1 and BnCCR2 play different roles in crosstalk between phenylpropanoid pathway and glucosinolate pathway

There have been a few reports on crosstalk of the glucosinolate pathway with the phenylpropanoid pathway[70–72].In Arabidopsis,accumulation of phenylpropanoids was suppressed in a ref5-1 mutant,and REF5 was shown to encode CYP83B1,which is involved in biosynthesis of indole glucosinolates [71].A defect in phenylpropanoid deposition was also detected in the Arabidopsis ref2 mutant,as REF2 encodes CYP83A1,functioning in the aliphatic glucosinolate pathway [72].In low-lignin c4h,4cl1,ccoaomt1,and ccr1 mutants of Arabidopsis,transcripts of some glucosinolate biosynthesis genes were more abundant than WT[59].By upregulating CCR,this study characterized the crosstalk effect of the phenylpropanoid pathway on the glucosinolate pathway.

In the present study,aliphatic glucosinolates were increased in BnCCR1-2ox (Table S4),possibly owing to upregulated expression of MYB28,MYB29,CYP79F1,CYP83A1,AOP2,and GSL in leaves of transgenic plants (Fig.6).Among them,the expression of MYB29,CYP79F1,and GSL could almost only be detected in BnCCR1-2ox,not in WT.Reasonably,the undetectable MYB29 expression and extremely low expression of MYB51,MYB122,AOP2,CYP79B2,ST5a,and IGMT1 in BnCCR2-4ox plants could be responsible for the decrease of glucosinolate content.Thus,both BnCCR1-2 and BnCCR2-4 distinctly affect glucosinolate biosynthesis,but they have divergent or almost opposite effects.

A crosstalk effect of the glucosinolate pathway on the phenylpropanoid pathway could be linked through PAL degradation mediated in a Med5-KFB-dependent manner [70].However,the mechanism involved in change of glucosinolate biosynthesis by manipulating phenylpropanoid genes such as CCR,awaits discovery.Our qRT-PCR results indicate that at least transcription regulation is involved,but whether protein degradation in the glucosinolate pathway is also involved deserves future study.

CRediT authorship contribution statement

Yourong Chai:Conceptualization,Project administration,Supervision,Funding acquisition,Data curation,Formal analysis,Methodology,Software,Validation,Visualization,Writing -original draft,review &editing.Jiana Li:Conceptualization,Project administration,Supervision,Funding acquisition,Resources,Writing-review&editing.Nengwen Yin:Conceptualization,Investigation,Data curation,Formal analysis,Methodology,Software,Validation,Visualization,Writing-original draft,review&editing.Bo Li:Conceptualization,Investigation,Data curation,Formal analysis,Methodology,Software,Validation,Visualization,Writing -original draft.Xue Liu:Investigation,Data curation,Formal analysis.Ying Liang:Supervision.Jianping Lian:Investigation,Data curation.Yufei Xue:Investigation,Data curation.Cunmin Qu:Investigation.Kun Lu:Investigation.Lijuan Wei:Investigation.Rui Wang:Resources.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank Prof.Xuekun Zhang for providing Zhongshuang 10 seeds,Prof.Wei Qian for providing the S.sclerotiorum source strain,and Prof.Ningjia He,Dr.Guangyu Ding,and Dr.Shiyao Liu for assistance with GC-MS measurements.We also thank Professor James C.Nelson from Kansas State University for language improvement of this article.This research was supported by National Natural Science Foundation of China (31871549,32001579,31830067 and 31171177),National Key Research and Development Program of China (2016YFD0100506),Special Financial Aid to Post-doctor Research Fellow of Chongqing (XmT2018057),‘‘111” Project(B12006),and Young Eagles Program of Chongqing Municipal Commission of Education (CY200215).

Data availability statement

The data supporting the findings of this study are available from the corresponding authors (Yourong Chai and Jiana Li) upon request.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.10.002.