Tingting Liu,Bolong To,Hnfei Wu,Jing Wen,Bin Yi,Chozhi M,Jinxing Tu,Tingdong Fu,Lixi Zhu*,Jinxiong Shen,*

a National Key Laboratory of Crop Genetic Improvement,National Center of Rapeseed Improvement,Huazhong Agricultural University,Wuhan 430070,Hubei,China

b School of Biological Science and Technology,University of Jinan,Jinan 250000,Shandong,China

Keywords:Brassica napus Chlorophyll biosynthesis Chloroplast biogenesis Chlorotic leaves Yellow cotyledon

ABSTRACT Mature chloroplasts,as the main sites of photosynthesis,are essential for seedling growth in higher plants.Loss of function of genes involved in chloroplast development changes plant phenotype.We obtained a YELLOW COTYLEDON (YCO) mutant in rapeseed (Brassica napus L.) using CRISPR-Cas9. Bn.YCO,a gene of unknown function,has two homologous copies(BnaA01.YCO and BnaC01.YCO)in B.napus.Homozygous mutation of these two homologs resulted in yellow cotyledons and chlorotic true leaves.Transmission electron microscopy revealed that the formation of thylakoid membranes was inhibited in yellow cotyledons.Sequence similarity search revealed that YCO was conserved in different species,and a subcellular location assay verified that Bn.YCO was located in the chloroplast. Bn.YCO was expressed in multiple tissues,most highly in cotyledons.Knockout of Bn.YCO blocked the transcription of plastid genes,especially those of photosystem genes transcribed by plastid-encoded polymerase.Transcriptome sequencing showed that the majority of genes involved in ribosome assembly and photosynthesis were down-regulated in Bn.yco mutants.These results suggested that loss of function of Bn.YCO affected plastid gene transcription,which influenced chloroplast biogenesis in rapeseed seedlings.

1.Introduction

The autotrophic growth of green plants depends on one essential organelle,the chloroplast.The chloroplast is the site of photosynthesis in plants,and uses sunlight to transform carbon dioxide and water into carbohydrate to support life activities [1,2].It evolved from a cyanobacterium that was acquired by a eukaryote over 1.5 billion years ago,so that the chloroplast contains its own genome [3,4].Following a lengthy endosymbiotic process,the chloroplast has a severely reduced genome,despite the more than 3000 proteins it contains.The chloroplast genome is approximately 150 kb long in higher plants and encodes around 100 core proteins involved in photosynthesis and plastid gene expression[5].For example,the photosynthesis system I (PSI) core subunits,such as psaA,psaB,and psaC as well as the photosynthesis system II (PSII) core subunits psbB,psbE,and psbF are all encoded by the chloroplast genome[3].The chloroplast genome also contains fourrpogenes:rpoA,rpoB,rpoC1,andrpoC2,which encode core subunits of a plastid-encoded RNA polymerase (PEP).PEP is responsible for the transcription of photosynthesis-related plastid genes such aspsaandpsb[6].In higher plants,disruption ofrpogenes often results in chlorophyll deficiency,followed by incomplete development of thylakoid membranes and reduced expression of plastid genes [7,8].The expression of chloroplast genes relies mostly on the nuclear genome.First,PEP requires a nuclearencoded sigma transcription factor to recognize the chloroplast genome [6].Second,transcription initiation ofrpogenes relies on nuclear-encoded RNA polymerase (NEP) [9].Transcription of plastid genomes relies on both PEP and NEP,and normal transcription of plastid genes is critical to chloroplast development and photosynthesis [10].

Besides photosynthesis,the chloroplast is also the site of chlorophyll(Chlaand Chlb)biosynthesis,in which approximately 20 different enzymes participate [11].Disruption of chlorophyll biosynthesis genes results in a chlorotic phenotype.The biosynthesis of chlorophyll is initiated with glutamate.Magnesium (Mg)chelatase (MgCh) catalyzes the insertion of the Mg2+ion into protoporphyrin IX.Chlorophyllide is the direct precursor for Chlaand Chlb.Loss of function of subunit I of MgCh (CHLI) results in an albino phenotype inArabidopsis[12].Loss of function of genes involved in chloroplast development also leads to several developmental defects,such as albino,pale-green,or variegated leaves[4,13].InArabidopsis,SNOWY COTYLEDON 1(SCO1)encodes chloroplast elongation factor G,which is essential for chloroplast gene translation.Thescomutant has a white-cotyledon phenotype[14].In rice,theSNOW-WHITE LEAF1(SWL1) gene mutation affects the formation of grana and stroma in thylakoids and ultimately leads to an albino phenotype [15].Defects in chloroplast development genes may affect not only chloroplast assembly but seedling development.

Rapeseed(Brassica napusL.,genome formula AACC)is one of the main oilseed crops in the world,and supplies approximately 40%of the vegetable oil consumed in China[16,17].It is an allotetraploid that originated from the hybridization of two diploid species,B.rapa(AA)andB.oleracea(CC)[18].In rapeseed,the study of genes involved in chloroplast development is still in its infancy.For example,BnCLIP1encodes a lipase that is located at membrane contact sites between the endoplasmic reticulum and the chloroplast and is involved in chloroplast development [19].A defect inBn.CDE1leads to yellowish leaves and lower chlorophyll content in rapeseed [20].GOLDEN2-LIKE1(GLK1) is a transcription factor involved in chloroplast development,and overexpression ofBnaGLK1aleads to darker leaves and siliques compared to control material [21].Overexpression of theBnGA2ox6gene results in increased Chlband total Chl content and also represses the expression of theCHL1gene[22].Theyglmutant,which harbors a defect in theBnaC07.HO1gene,exhibits a yellow-green leaf phenotype and abnormal chloroplasts during the seedling stage.In the mutant,photosynthesis genes are down-regulated [23,24].

Most research on chloroplast genes in rapeseed has focused on map-based cloning or preliminary investigations of gene function.Given that one gene may be present in two to six functionally redundant copies in the rapeseed genome,it is difficult to study gene function in rapeseed [25,26].Recently,the CRISPR/Cas9 system has become an efficient method of creating knockout mutants in many species [27].Bn.YCOis a gene of unknown function.We intended to create knock-out lines ofBn.YCOusing CRISPR/Cas9 system,which will facilitate the study ofBn.YCOin rapeseed.

2.Materials and methods

2.1.Plant material and growth conditions

A wild-type rapeseed cultivar(T6),was used as the receptor for transformation experiment.Mature seeds were sterilized by immersion in 75% ethanol for 1 min,50% 84 disinfectant (with available chlorine concentration of 30–35 g L-1) for 5 min,75%ethanol for 30 s,and sterilized distilled water four times for 5 min each time.Seeds were then sown in a chamber with medium M0 and cultured in the dark for 6–7 days at 22°C.Transformations were performed as in our previous study[28].Transgenic seedlings in the first (T0) generation were identified using two specific primers(U626-IDF and U629-IDR)(Table S1).Positive seedlings were transferred into soil and placed in a tissue culture room with a 16 h light/8h dark cycle at 22–25 °C for seed production.For qRT-PCR analysis ofBn.YCOexpression,T6 was planted in a field at Huazhong Agricultural University (Wuhan,China) during the rapeseed growing season (October to May).Arabidopsis(Columbia-0) was grown and leaves were sampled to extract protoplasts for subcellular localization assay.TheArabidopsisplants were cultured in a 16 h light/8 h dark cycle at 21–23 °C with 30%–60% relative humidity.

2.2.CRISPR/Cas9 construct assembly and identification of mutant transgenic plants

To targetBnaA01.YCOandBnaC01.YCOsimultaneously,two single-guide RNAs (sgRNAs) were designed with the CRISPR-P 2.0 tool (http://crispr.hzau.edu.cn/CRISPR2/).The CRISPR/Cas9 binary vector set,provided by Prof.Qijun Chen(China Agricultural University),includes the pCBC-DT1T2 vector (with chloramphenicol resistance) and pKSE401vector (with kanamycin resistance).TheZea maysCas9 protein was driven by the double 35S promoter(2× 35S),and the sgRNA expression cassettes were driven by theArabidopsis U6gene promoter.The guide sequence length was 19 bp.B.napusgenome was selected as a reference genome.Two sgRNAs(TAGCTCCTTGGCCTTAGCG and AGTGCTTACGACAATACAC),which could target both copies ofBn.YCOwithout off-target effects,were selected for incorporation into the pKSE401 CRISPR/Cas9 expression vector.The vector construction process was as previously described [29].The target regions surrounding the CRISPR target sites were amplified using specific primers.Primers A01-2L and A01-2R were used to amplify the target region ofBnaA01.YCOand primers C01-2L and C01-2R were used to amplify the target region ofBnaA01.YCO(Table S1).The PCR fragments were directly sequenced by the Sanger method to identify the mutations.

2.3.Pigment content determination

Variegated leaves from T0seedlings and green leaves from wildtype T6 were collected to measure chlorophyll content as described previously [30].Briefly,wild-type and mutant leaves(approximately 30 mg fresh weight) were cut into pieces and immersed in 5 mL 80% (v/v) acetone for 48 h.The samples were then centrifuged at 3000 r min-1for 10 min.The supernatants were transferred to new centrifuge tubes and diluted with acetone for analysis using a spectrophotometer(UV-1800,Mapada,Shanghai,China).Pigment content was detected at 663,645,and 470 nm absorbance with four duplicates for each source.Finally,Chlaand Chlbcontents were calculated.

2.4.Sample preparation for transmission electron microscopy

Variegated leaves of T0plants and green leaves of wild-type plants were harvested.The yellow cotyledons of T1plants and green cotyledons of wild-type plants were also harvested at 7 and 14 days after sowing,respectively.Samples were cut into 1 × 1 cm sections and immediately immersed in 2.5% (w/v) glutaraldehyde in 0.1 mol L-1phosphate buffer (pH 7.4) for fixing and then fixed in 1% OsO4in the same buffer.

2.5.Phylogenetic analysis

The full-length amino acid sequences of Bn.YCO were obtained from theB.napusgenome database(http://www.genoscope.cns.fr/brassicanapus/)[18].ChloroP 1.1(http://www.cbs.dtu.dk/services/ChloroP/)was used to predict the locations of transit peptides[31].The Conserved Domain Database (https://www.ncbi.nlm.nih.gov/cdd) was searched to functional units in Bn.YCO.Putative Bn.YCO orthologs were predicted using the National Center for Biotechnology Information website for protein (https://blast.ncbi.nlm.nih.-gov/Blast.cgi?PROGRAM=blastp&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) and the Phytozome 7.0 database (http://www.phytozome.net) [32].Nineteen putatively orthologous sequences from several species were aligned with MEGA 7.0 [33]software.Predicted transit peptide sequences were not conserved among these putative orthologs.After deletion of these sequences,the putative ortholog sequences were realigned and a neighborjoining phylogenetic tree was constructed with MEGA.The sequences of six putative orthologs were aligned using MEGA 7.0.

2.6.Subcellular localization of Bn.YCO

The coding sequences (CDS) ofBnaA01.YCOandBnaC01.YCOsequence without the termination codon were amplified with the specific primers Sub-2A and Sub-2C (Table S1).The sequences were then inserted into the transient expression vector pM999-35S-GFP to generate C-terminal GFP fusion products.Three expression vectors,P35S::BnaA01.YCO:GFP,P35S::BnaC01.YCO:GFP,andP35S::GFP,were introduced intoArabidopsisprotoplasts via polyethylene glycol/calcium mediation.The transformation method was modified as previously described [34].The spontaneous fluorescence of chlorophyll was used as an indicator of the chloroplast.GFP and chlorophyll fluorescence signals were viewed with a FluoView FV1000 laser scanning confocal microscope (Olympus,Tokyo,Japan).

2.7.qRT-PCR analysis

To measure the expression ofBn.YCOin rapeseed organs,total RNA was extracted from roots,stems,leaves,flowers,siliques,and cotyledons of wild-type T6.Samples were harvested and frozen in liquid nitrogen.Total RNAs were extracted with an RNA Pure Plant Plus Kit(S7314,Tiangen,Beijing,China)according to the user manual.Total RNA(2 μg)from each sample was transcribed using the RevertAid First Strand cDNA Synthesis Kit (#K1622,Thermo Fisher Scientific,Vilnius,Lithuania).Three biological replicates were used for qRT-PCR.Relative expression of genes was validated using SYBR Green II (Toyobo,Osaka,Japan) with a CFX384 touch real-time PCR detection system (Bio-Rad,Hercules,CA,USA)according to the manufacturer’s protocol.PCR was performed using 4.2 μL of 50× diluted cDNA products,5 μL 2× SYBR Green II buffer,and 0.4 μL forward and reverse primers (10 μmol L-1)in each 10 μL system.The reactions were performed at 95 °C for 10 s,followed by 45 cycles at 95 °C for 5 s,60 °C for 10 s,72 °C for 30 s,and then 95°C for 10 s.Bnactin7was used as the reference gene for normalization inB.napus.Relative expression levels were calculated using the 2(T) (-Delta Delta Ct) method [35].All primers are listed in Table S1.

2.8.GUS assay

A 554-bp upstream sequence ofBnaC01g01100Dwas amplified from T6 genomic DNA with two specific primers,2-GUS-L and 2-GUS-R(Table S1).This sequence was cloned into the pCAMBIA2300 binary vector.The ProBnaC01g01100D-GUS construct was introduced into wild-type T6 plants byAgrobacterium-mediated transformation as described in our previous report [28].Primers M13-47 and 2-GUS-R were used to identify positive transgenic plants.Tissues from positive transgenic plants were sampled and submerged in acetone for fixing,followed by rinsing twice with cleaning solution.GUS activity was visualized by overnight staining of the tissues in 5-bromo-4-chloro-3-indolyl-β-D-glucuronide (X-Gluc)solution.s at 37 °C in the dark [36].The tissues were then cleared with 75% (v/v) ethanol until pigments were completely removed.The cleaning solution was composed of 0.1 mol L-1PBS solution(pH 7.0),0.01 mol L-1EDTA solution(pH 8.0),2.0 mmol L-1potassium ferricyanide,2.0 mmol L-1potassium ferrocyanide,0.1%(v/v)Triton X-100.The GUS staining solution was composed of cleaning solution plus 0.5 mg mL-1X-Gluc.

2.9.High-throughput RNA sequencing and data analysis

Total RNA was extracted from yellow cotyledons ofBn.ycomutants and green cotyledons of T6 (10 days after seeding).RNA extraction was as described above.Three methods were used to measure RNA quality and quantity:NanoPhotometer spectrophotometer (Implen,Westlake Village,California,USA),Qubit 3.0 Fluorometer (Life Technologies,California,USA) and 2100 RNA Nano 6000 Assay Kit (Agilent Technologies,Santa Clara,California,USA).The samples were sequenced on the Illumina(San Diego,California,USA) HiSeqX-Ten platform.For each sample,～17 Gb raw sequence were obtained.Clean reads were obtained from the raw reads by removal of adapter-polluted reads,Ns reads,and lowquality reads and mapped to theB.napusgenome (http://www.genoscope.cns.fr/brassicanapus/)using HISAT2 software[37].Fragments per kilobase per million mapped fragments were then calculated to estimate the expression level of genes in each sample[38].DESeq2 was used to estimate the expression level of each gene in each sample by linear regression,calculating theP-value with a Wald test [39].Genes withq≤0.05 and |log2_ratio| ≥1 were assigned as differentially expressed genes (DEGs).Gene Ontology(GO) annotation (http://geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG) reference pathway (https://www.genome.jp/kegg/kegg.html) analysis were performed for DEGs.Heat maps of DEGs were drawn with TBtools 1.6.

3.Results

3.1.Knockout of Bn.YCO using CRISPR/Cas9 platform

In rapeseed,the two homologous copies ofBn.YCO,BnaA01.YCO(BnaA01g00180D) andBnaC01.YCO(BnaC01g01100D),both contained six exons and five introns,and their CDSs were highly similar (Table S2).The difference was thatBnaC01.YCOcontained 5′-UTR and 3′-UTR,butBnaA01.YCOdid not contain 5′-UTR and 3′-UTR.Two single guide RNAs(sgRNAs),sgRNA1(S1)and sgRNA2(S2),were designed to target the fourth exon ofBn.YCO(Fig.1A and B).Then the CRISPR/Cas9 vector was transformed into wild-type rapeseed T6 to produce transgenic plants.To identify the mutation type in each allele ofBn.YCO,the target sites ofBnaA01.YCOandBnaC01.YCOwere sequenced.Specific primers were designed to amplify the target region ofBnaA01.YCOandBnaC01.YCOfor each plant.Sequencing showed that 43 editing events had independently occurred in the BnA or BnC subgenomes of nine examined plants.Types of mutation included nucleotide insertion and deletion.The two copies ofBn.YCOin 6 mutant lines were edited simultaneously,and multiple mutation types could be detected inBnaA01.YCOorBnaC01.YCOin the same mutant line.For example,five different types of editing were present at two target sites simultaneously in the C subgenome of theBn.yco-1mutant line.In contrast,theBn.YCOwas edited only in the A or C subgenomes in the other three mutant lines (Fig.1C).

Statistical analysis of the above 43 edit events showed that 65%mutation types were nucleotide insertions and 35%mutation types were deletions.All insertion mutations were single-base insertions,and insertions of A,T,C,and G were all detected.The proportion of T insertions was highest,accounting for 57%of the total 28 insertion events,followed by A insertion,accounting for 25%of the total insertion types.More than 11% of mutations were ≥10-bp deletions,and the longest deletion was of 97 bp (Fig.1D).

qRT-PCR was performed to detected the relative transcription level ofBnaA01.YCOandBnaC01.YCO.The results of qRT-PCR showed that the expression of these two copies decreased significantly in mutant leaves in comparison with wild-type leaves(Fig.1E).These results further suggest that the transcription level ofBn.YCOis reduced by the CRISPR/Cas9 system.

3.2.Phenotype analysis of the Bn.yco knockout mutant

Fig.1.Gene editing analysis of Bn.YCO.(A)The BnaA01.YCO and BnaC01.YCO gene models contain six exons(white boxes)separated by five introns(solid line),respectively.Thick black line upstream of the first exon of BnaC01.YCO represents the 5′ UTR.Thick black line with arrow downstream of the last exon of BnaC01.YCO represents the 3′ UTR.The vertical lines in the fourth exon indicate the target site,and the arrows below S1 and S2 indicate the sgRNA direction.(B)The CRISPR/Cas9 expression vector contains two target sites of Bn.YCO,which are driven by the AtU6-26 and AtU6-29 promoters.The zCas9 protein is driven by the 2×35S promoter,followed by a kanamycin resistance gene coding sequence.(C)Several InDels are present at the target sites of BnaA01.YCO and BnaC01.YCO in nine T0 lines.The Protospacer Adjacent Motif(PAM)motif is highlighted in green.Red characters and red dashes represent insertions and deletions,respectively.At left,A and C represent the wild-type (WT) allele of BnaA01.YCO and BnaC01.YCO,respectively.a# and c# represent the different mutation types in the BnA and BnC subgenomes.(D) The frequency of different mutation types in the examined T0 plants.i,insertion;d,deletion;n,no edit.The number above each bar shows the percentage of each mutation type.(E)The relative expression level of Bn.YCO in different mutant lines.The expression levels were normalized to Bnactin7.Values are presented as mean ± SD (n=3).* significant at P <0.05 ;** significant at P <0.01 (t-test).

The T0generation of transgene plants showed variegated or large-area albino leaf phenotypes (Fig.2A).The quantities of Chlaand Chlbin the mutant lines were 0.59 ± 0.01 and 0.24 ± 0.01 mg g-1fresh weight,respectively,lower(71%and 78%)than those in wild-type plants(Chla,0.83±0.09 mg g-1fresh weight;Chlb,0.31±0.03 mg g-1fresh weight).The ratio of Chla/bhad a slight tendency to decrease (Fig.2B).Prior to the biosynthesis of protoporphyrinogen IX,four genes were down-regulated,includingHEMA1,HEMC,UROD,andCPOX.The relative expression ofURODandCPOXdecreased to an extremely low level.In the chlorophyll biosynthesis branches,the expression levels of two genes,CHLI1andCHLD,decreased in variegated leaves.Five downstream genes were also down-regulated,includingCHLM,PORC,DVR,CAO,andCBR(Fig.2C).Thus,knockout ofBn.YCOinfluenced leaf chlorophyll biosynthesis in rapeseed.

3.3.Identification of the Bn.yco T1 mutant

In the greenhouse,T0mutants completed their life cycle and produced viable seeds.In the T1generation,Bn.ycomutants showed yellow cotyledons 3 days after germination (Fig.3B and C).In contrast,wild-type cotyledons were green(Fig.3A).The yellow cotyledon phenotype inBn.ycomutants was also observed 7 days after sowing(Fig.3E and F).At 14 days after sowing,albino or variegated true leaves could be observed inBn.ycomutants(Fig.3H and I).In contrast,wild-type seedlings were always green at 7 and 14 days after sowing (Fig.3D and G).In 6 randomly selected T1seedlings,the target regions in both copies ofBn.YCOwere edited(Fig.S1).Thus,knockout ofBn.YCOstrongly influenced cotyledon and true leaf development in seedlings.

3.4.Ultrastructural analysis of Bn.yco plastids

Fig.2.Phenotype identification of Bn.yco mutant in rapeseed.(A) The chlorotic phenotype of Bn.yco mutant leaves compared to wild-type leaves.Leaves from two Bn.yco lines, Bn.yco-3 and Bn.yco-4 are shown.(B) The chlorophyll contents of Bn.yco T0 lines and the wild type.(C) qRT-PCR analysis of the expression of chlorophyll biosynthesis genes in leaves of Bn.yco mutant and wild type.Expression levels were normalized to Bnactin7.Values are presented as mean±SD(n=3).*significant at P<0.05;** si gnificant at P <0.01 (t-test).

Fig.3.Phenotype of Bn.yco mutants in the T1 generation.Three-day-old (A) green cotyledons in wild-type seedlings and (B,C) yellow cotyledons of mutant lines.Scale bars,2 cm.Seven-day-old (D) wild-type cotyledons and (E,F) yellow cotyledons of mutant lines grown in soil.Fourteen-day-old(G)green seedlings and(H,I) yellow seedlings of mutant lines.

In variegated true leaves ofBn.yco,the chloroplasts were undeveloped and lacked internal thylakoid membranes (Fig.4A).In contrast,chloroplasts in wild-type true leaves contained wellstructured stroma and grana thylakoids (Fig.4B).In 7-day-old yellow cotyledons,almost no thylakoid membranes could be observed in the chloroplast-like plastids (Fig.4C).In contrast,the thylakoid membranes in wild-type cotyledons were forming(Fig.4D).In 14-day-old yellow cotyledons,the chloroplasts seemed to be degraded and many plastoglobules could be observed(Fig.4E).In contrast,chloroplasts with stacked thylakoid membranes and starch granules could be observed in 14-day-old wild-type cotyledons(Fig.4F).Thus,the defects ofBn.YCOinhibited chloroplast development in cotyledons and true leaves.

3.5.Bn.YCO encodes a chloroplast-localized protein

The CDS ofBnaA01.YCOspans 1296 bp and encodes a 49.1-kDa protein of 431 amino acids.The CDS ofBnaC01.YCOspans 1302 bp and encodes a 49.3-kDa protein of 433 amino acids.Alignment of YCO amino acid sequences inArabidopsis thaliana,Citrus sinensis,Glycine max,Oryza sativa,andZea mayswith Bn.YCO showed that the N-terminal regions were not conserved among the putative orthologs.The N-terminal of the YCO protein contained a putative chloroplast signal peptide domain.Except for the putative signal peptide sequence and the C-terminal sequence,the YCO protein sequences were highly conserved among these dicotyledonous and monocotyledonous species (Fig.5A).

Fig.4.Transmission electron microscopic images of chloroplast in Bn.yco mutant and wild-type material.(A) No thylakoid membrane was observed in Bn.yco leaves.(B)Wild-type chloroplasts contained well-structured thylakoid membranes and starch granules.(C) Almost no thylakoid membranes were observed in 7-day-old yellow cotyledons in Bn.yco mutant.(D) Thylakoid membranes developing in 7-day-old wild-type cotyledons.(E) Chloroplast internal structures are degraded and many plastoglobules can be observed in 14-day-old yellow cotyledons.(F) Chloroplasts in 14-day-old wild-type cotyledons have already matures. St,starch; Tm,thylakoid membrane; Pg,plastoglobule.Scale bars,0.5 μm.

BnaA01g00180D showed the highest similarity to Bra011795 inB.rapa,and BnaC01g01100D showed the highest similarity with Bol028879 inB.oleracea.Thus,BnaA01.YCOmay have descended fromBra011795in the diploid ancestorB.rapaandBnaC01.YCOmay have descended fromBol028879in the diploid ancestorB.oleracea.The 19 sequences with high similarity were all from dicotyledons,suggesting that the YCO sequences were conserved in dicotyledons.Some of these dicotyledons belong to multiple genera ofBrassicaceae,includingE.salsugineum,A.thaliana,andC.grandiflora.They also represent some distantly related species includingP.trichocarpa,C.sinensis,andP.persica.The two sequences with lowest similarity to Bn.YCO inB.napuswere those fromG.maxandM.guttatus(Fig.5B).Thus,YCO is ubiquitous and highly conserved in dicotyledons.None of the YCO proteins or putative orthologs have been described in detail,indicating thatBn.YCOencodes a protein of unknown function.

As expected,the BnaA01.YCO:GFP and BnaC01.YCO:GFP fusion proteins co-localized with the red autofluorescence of chlorophyll in the chloroplasts (Fig.6),showing that Bn.YCO is a chloroplastlocalized protein.

3.6.Expression patterns of Bn.YCO

To explore possible function in tissues,the transcription levels ofBn.YCOin roots,stems,leaves,flowers,siliques,and cotyledons were examined by qRT-PCR.Specific primers were designed to detect the expression patterns ofBnaA01.YCOandBnaC01.YCO.The expression ofBnaC01.YCOwas higher than that ofBnaA01.YCOin almost all tissues,but their expression patterns were similar.The transcripts of both copies were almost undetectable in roots and siliques,but showed relatively high levels in leaves.Compared with that in germinating seeds,the transcription level was highest in 1-day-old cotyledons.The expression ofBnaC01.YCOremained at approximately the same level in 2-day-old as in 1-day-old cotyledons,whereas the expression ofBnaA01.YCOdecreased rapidly.In 3-day-old cotyledons,the expression of both of these copies decreased to that of germinating seeds and remained at this level until the fifth day,followed by a slight reduction in 6-day-old cotyledons (Fig.7A).Almost no GUS expression could be observed in the flower,bud,or silique (Fig.7B,D,and E).In contrast,GUS expression was high in leaves and cotyledons(Fig.7C,F,and G).

Fig.5.Relationship among YCO orthologs in plants.(A) Deduced amino acid sequences of Bn.YCO and putative orthologs in five species.The red box shows the predicted chloroplast transit peptide in YCO.(B)Phylogenetic tree of the sequence of Bn.YCO and its putative orthologs in different species.The phylogenetic tree was constructed with sequences from which the variable N-terminal region had been removed.Numbers on branches are bootstrap values (%) for 1000 replications.

Fig.6.Subcellular localization of Bn.YCO.(A) Arabidopsis protoplasts showing green fluorescent signals after transformation with the P35S::GFP constructs.(B) Chloroplasts showing red fluorescence in the same protoplast as in (A).(C) The same protoplast as in (A) under bright field.(D) Merged images from (A–C).(E,I) Arabidopsis protoplasts showing green fluorescence signals after transformation with the P35S::BnaA01.YCO:GFP and P35S::BnaC01.YCO:GFP constructs,respectively.(F,J) Chloroplasts showing red fluorescence in the same protoplasts as in(E)and(I),respectively.(G,K)The same protoplasts as in(E)and(I)under bright field,respectively.(H,L)Merged images from(E–G) and (I–K),respectively.

3.7.Bn.yco mutation leads to low expression of plastid genes

To further investigate why theBn.ycomutation resulted in a chlorotic phenotype,the expression levels of plastid genes were measured in mutant lines.The rapeseed chloroplast genome is approximately 150 kb long and contains 113 genes.The following genes involved in photosynthesis or Rubisco are transcribed by PEP:fivepsagenes (psaA-C,psaI,andpsaJ),16psbgenes (psbA-N,psbT,andpsbZ),six cytochromeb/f complex genes(petA,petB,petD,petG,petL,andpetN),and one Rubisco gene (rbcL) [40].First,the expression of ten PEP-dependent genes was measured.InBn.ycomutants,the expression of PEP-dependent genes was downregulated (Fig.8A).Next,the transcripts of six photosynthesisrelated genes transcribed by PEP and NEP were examined.The expression of five genes(ndhA,ndhB,atpA,atpB,andatpE)was also down-regulated in both mutant lines (Fig.8B).These results suggested that knockout ofBn.YCOinfluenced plastid gene transcription,especially of PEP-dependent genes.

3.8.Transcriptome analysis of Bn.yco mutants

To further investigate the mechanism underlying yellow cotyledon and determine whether other metabolic pathways were also influenced inBn.ycomutants,differentially expressed genes(DEGs)in yellow and wild-type cotyledons were identified by RNA sequencing.A total of 271,424,666 clean reads from six samples(three duplications for mutant and wild-type cotyledons) were obtained (Table S3).From these,18,571 DEGs were isolated.Compared with the wild-type lines,7740 genes were up-regulated and 10,831 genes were down-regulated in theBn.ycomutants(Table S4).qRT-PCR validation of 24 genes involved in ribosome biosynthesis,photosynthesis,and carbon metabolism revealed expression trends consistent with those of the sequencing data,though the expression differences were not equal (Fig.S2).

Functional categorization of the 18,571 DEGs revealed that 4437 genes encoded nucleus-located proteins and 2475 encoded chloroplast-located proteins (Fig.9A).GO analysis showed that the ‘‘biological process” category contained 491 terms,‘‘molecular function” contained 163 terms,and ‘‘cellular component” contained 146 terms (Table S5).In the ‘‘biological process” category,the top three terms were ‘‘translation”,‘‘peptide biosynthesis process”,and ‘‘a(chǎn)mide biosynthetic process”.In the ‘‘molecular function” category,the top three terms were ‘‘structural constituent of ribosome”,‘‘structural molecule activity”,and ‘‘RNA binding.”In the‘‘cellular component”category,the top three terms were‘‘ribosomal subunit”,‘‘cytosolic part”,and ‘‘ribonucleoprotein complex” (Table S5).

KEGG pathway enrichment analysis revealed that 132 pathways were affected in yellow cotyledons.The synthesis and function of the ribosome seemed to be destroyed in yellow cotyledons,as 688 ribosome-associated genes were down-regulated (Fig.9B;Table S6).The KEGG map (map 03010) showed that most genes responsible for the biosynthesis of ribosome subunits were down-regulated (Fig.S3).Some other pathways were also influenced inBn.ycomutants,including sulfur metabolism,carbon metabolism,and biosynthesis of amino acids (Fig.9B).Twenty-three photosynthesis-related genes,encoding subunits of PSI,were down-regulated (Fig.9C).The expression of 14 lightharvesting complex genes in PSI(LHCA)decreased(Fig.9D).Among PSII genes,the expression of 16Psbgenes,which encode PsbP,PsbQ,PsbR,PsbW,and PsbY subuints,was down-regulated(Fig.9E).The 31LHCBgenes (LHCB1-6) were down-regulated inBn.ycomutants (Fig.9F).Thus,disruption ofBn.YCOmay have affected protein translation and photosynthesis.

Fig.7.Expression pattern analysis of Bn.YCO in wild-type T6 plants.(A)qRT-PCR analysis of Bn.YCO expression levels in root,stem,leaf,bud,flower,silique,and cotyledon.D,days after germinating.Expression levels were normalized to Bnactin7.Values are presented as mean ± SD (n=3).(B–G) Spatial expression patterns of BnaC01.YCO in transgenic rapeseed plants harboring the BnaC01.YCO promoter fused to the GUS gene.Promoter activity was visualized by histochemical GUS staining of (B) flowers,(C)leaves,(D) buds,(E) siliques,and (F,G) cotyledons.Scale bars,2.5 mm.

Fig.8.qRT-PCR analysis of plastid gene expression.Total RNA was extracted from variegated leaves of Bn.yco mutant lines and green leaves of wild-type materials.(A)Relative expression of PEP-dependent genes.(B) Relative expression of NEP-and NEP-dependent genes.Expression levels were normalized to Bnactin7.The values are presented as the mean ± SD (n=3).* Significant at P <0.05;** significant at P <0.01 (t-test).

4.Discussion

4.1.CRISPR/Cas9 platform is an efficient tool for gene editing in rapeseed

Our experimental results confirm thatBn.YCOplays key roles in chloroplast development in cotyledons and true leaves of rapeseed.Several types of mutation were generated in the target sites ofBnaA01.YCOandBnaC01.YCO,including nucleotide insertion and deletion.We speculate that the multiple editing types in a single genetically modified plant were caused by continuous editing by the Cas9 protein of genome target loci.The CRISPR/Cas9 platform is guided by a single small RNA to target DNA following the principle of base complementary pairing,which can generate DNA double-stranded breaks(DSBs)at specific loci in the genome.During the DNA repair process,the DSBs may be aligned via nonhomologous end joining and insertion/deletions may be generated[41].Binary vectors carry the CRISPR/Cas9 scaffold to be inserted into the plant genome,meaning that the CRISPR/Cas9 platform could transcribe and translate [29].Theoretically,the cleavage and repair processes might occur throughout the life cycle of the plant.The genome of each cell could contain different editing types near the target site.In this study,among the nine identified mutant seedlings,we found 43 different edit types on both target sites.We speculate that there were other edit types that remained undetected.As the CRISPR/Cas9 scaffold is inserted into the plant genome by stable transformation,this scaffold can be passed on to the next generation and continue to cleave target sites in T1seedlings.The observation of multiple edit types of target sites in the T1generation showed that the CRISPR/Cas9 system continued to operate.

Fig.9.Transcriptome analysis of yellow and green cotyledons.(A) Numbers of up-regulated and down-regulated genes in each cellular component.(B) KEGG analysis of differentially expressed genes(DEGs).(C)Heat map of Psa genes in PSI.(D)Heat map of HLCA genes in PSI.(E)Heat map of Psb genes in PSII.(F)Heat map of HLCB genes in PSII.

In comparison with previously reported editing result in rapeseed,the mutation types relied mainly on the CRISPR/Cas9 system.Yang [42] used a CRISPR/Cas9 expression vector to knock out two copies ofCLAVATA3(CLV3)in rapeseed.Single nucleotide insertion mutations (47.6%) were the most common.Similarly,a CRISPR/Cas9 platform containing four sgRNAs was constructed to target twoINDEHISCENT(IND) genes inB.napus[43].Single-nucleotide insertions were the most commonly induced mutations.A sgRNA was designed to target five rapeseedSPL3genes,andBnspl3mutants exhibited a developmental delay phenotype in the first generation.The majority of InDel mutations were 1 bp insertions[44].These results indicate that single-base insertion mutations are most likely to be generated by the CRISPR/Cas9 platform.However,other InDel types differed among different CRISPR/Cas9 systems.

4.2.Bn.YCO affects chloroplast development in both true leaves and cotyledons

Loss of function ofBn.YCOaffected chlorophyll biosynthesis and thylakoid membrane development in rapeseed leaves(Figs.1D,4).Some seedlings with yellow cotyledons could not survive and gradually died approximately 3 weeks after sowing.Similar phenotypes have also been reported inArabidopsis.Homozygous mutants ofAtcpSecA,which encodes a thylakoid protein translocase subunit,were albino and seedling-lethal under autotrophic conditions[45].TheArabidopsisT-DNA insertion mutant ofatpG,encoding subunit II of chloroplast ATP synthase,displayed an albino lethal phenotype and could not grow photoautotrophically [46].Some studies have shown that the development processes of cotyledons and true leaves are different.Mutants of theWHITE COTYLEDONS(WCO) locus showed the white cotyledon and green true leaf phenotype [47].Chloroplast development in theSnowy cotyledon 2(SCO2) loss-of-function mutant was disrupted in cotyledons but not in true leaves[48].Premature seeds of mutant lines developed normal green cotyledons.In contrast,in the present study,the mutation ofBn.YCOnot only affected the development of cotyledons,but resulted in variegated true leaves (Figs.2,3).The thylakoid membranes in chloroplasts were abnormal in both cotyledons and true leaves of the mutants (Fig.4).

4.3.Transcript levels of plastid genes were down-regulated in Bn.yco mutants

The qRT-PCR test showed that the transcription of plastid genes,especially PEP-dependent genes,was strongly influenced in theBn.ycomutant(Fig.8).In previous studies,the mRNA levels of chloroplast genes were disturbed in many chloroplast-deficient mutants,and the mutated loci were involved in the PEP complex or ribosome translation process.For example,Arabidopsis SIG6encodes a sigma factor that recognizes specific chloroplast promoters and functions in PEP-dependent chloroplast gene transcription.The mutation ofsig6results in a pale green cotyledon phenotype and reduced transcript levels of most PEP-dependent genes at an early stage of seedling development [49].A possible 16S RNA maturerelated gene,WCO,influences the expression levels of chloroplast mRNAs frompsbAandrbcL[47].In plants,plastid gene translation relies on the plastid ribosome,which is composed of the 30S small subunit and the 50S large subunit [50].InArabidopsis,five functional plastid-specific ribosomal proteins (PSRPs) have been characterized.The RNAi mutant,psrp-4,resulted in pale-green leaves.Knockout ofPSRP5severely impaired plant growth.The chloroplast translation process and rRNA processing process were both blocked inpsrp3/1,psrp4,andpsrp5mutants [51].In maize,theppr2mutation affected chloroplast ribosome biogenesis and resulted in the absence of chloroplast proteins,resulting in albino leaves [52].In rice,WHITE STRIPE LEAF6(WSL6) encodes a GTPbinding protein in the chloroplast that is essential for chloroplast ribosome biogenesis.Theswl6mutant exhibited white-striped leaves and defective thylakoid membranes in early chloroplast development [53].We propose that the function of the PEP complex or the translation process of PEP genes is disturbed in theBn.ycomutant.

4.4.Transcription of photosynthesis-related genes was influenced in Bn.yco mutants

Eukaryotic photosystem I consists of two functional moieties:the photosystem I core complex and the peripheral lightharvesting complex (LHC I) [54].PsaA,PsaB,and PsaC are three subunits of the PSI core complex and are encoded in the plastid genome [55].PSII is composed of a large number of intrinsic subunits and three extrinsic polypeptides encoded bypsbgenes.PsbA,PsbB,PsbC,andPsbD1are four genes encoding the photosystem II reaction center [56].qRT-PCR showed that the expression ofpsaandpsbgenes was down-regulated in theBn.ycomutant(Fig.8A).RNA-seq showed that the PSI subunit genesPsaE,PsaN,andPasGwere down-regulated (Fig.9C).The PSI subunit genessPsbY,PsbW,PsbR,andPsbQwere also down-regulated (Fig.9E).The function of these subunits is important for chloroplast photosynthesis.For example,PsaE is a stromal extrinsic PSI subunit and helps plants to avoid oxygen stress in PSI[57].Knockout ofPsbWinArabidopsisdisturbs the formation of a functional PSII supercomplex [58].A total of 45LHCgenes were also down-regulated(Fig.9D and F).Thus,theBn.YCOmutation influenced the expression of genes involved in photosynthesis.

5.Conclusions

TheBn.YCOmutation resulted in thylakoid formation defects and low transcript levels of chloroplast genes.Many nuclear genes involved in photosynthesis were also down-regulated inBn.ycomutants.Thus,Bn.YCOis a crucial gene for chloroplast development in rapeseed.

CRediT authorship contribution statement

Tingting Liu:performed the research and wrote the paper.Baolong Tao and Hanfei Wu:conducted experiments.Lixia Zhu and Jinxiong Shen:supervised the study and revised the manuscript.Jing Wen,Bin Yi,Chaozhi Ma,Jinxing Tu,and Tingdong Fu:participated in the study’s design.All authors read and approved the manuscript.

Declaration of competing interest

Authors declare that there are no conflicts of interest.

Acknowledgments

This research was financially supported by the National Natural Science Foundation of China (31871654,31501340),National Key Research Development Program of China (2016YFD0101300),and the China Agriculture Research System (CARS-12).

Accession numbers

This project was submitted to NCBI BioProject with BioProject ID:PRJNA565208.The raw reads were deposited in NCBI SRA(Short Read Archive) with submission number SUB6163521.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.10.015.