Non-vernalization requirement for flowering in Brassica rapa conferred by a dominant allele of FLOWERING LOCUS T

doi:10.21203/rs.3.rs-2364442/v1

Download PDF

Research Article

Non-vernalization requirement for flowering in Brassica rapa conferred by a dominant allele of FLOWERING LOCUS T

https://doi.org/10.21203/rs.3.rs-2364442/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 18 May, 2023

Read the published version in Theoretical and Applied Genetics →

You are reading this latest preprint version

Controlling the timing of flowering is key to improving yield and quality of several agricultural crops including the Brassicas. Many Brassicaceae plants possess a conserved flowering mechanism in which FLOWERING LOCUS C (FLC) represses the transcription of flowering activators, such as FLOWERING LOCUS T (FT), during vernalization. Here, we employed genetic analysis based on next-generation sequencing to identify a dominate FT allele, BrFT2-C, for flowering in the absence of vernalization in the Brassica rapa cultivar ‘CHOY SUM EX CHINA 3’. BrFT2-C harbors two large insertions upstream of its coding region and is constitutively expressed without vernalization, despite FLCexpression. We show that BrFT2-C offers an opportunity to introduce flowering without vernalization requirement into winter-type brassica crops, including B. napus, which have many functional FLC paralogs. Furthermore, we demonstrated the feasibility of using B. rapa harboring BrFT2-C as rootstock for grafting to induce flowering in radish (Raphanus sativus), which requires vernalization for flowering. We believe that the ability of BrFT2-C to overcome repression by FLCcan have significant applications in brassica crops breeding to increase yields by accelerating or delaying flowering.

B. rapa

B. napus

R. sativus

FLC

vernalization

grafting

We identified and characterized a dominant FT allele for flowering without vernalization in Brassica rapa, while demonstrating its potential for deployment in breeding to accelerate flowering in various Brassica species.

Brassicaceae plants include many economically important crop species grown for diverse harvestable plant organs including flowers, seeds, leaves and tubers. Manipulating flowering time in these crops is an important consideration for improving production because their quality and yield is dependent on the timing of flower bud initiation. Consequently, a detailed understanding of the molecular mechanism of flower induction in Brassicaceae crops is vital for accelerating the breeding of cultivars with optimal flowering times.

Flowering in Brassicaceae is promoted by exposure to prolonged period of cold temperature or vernalization, a process that has been well studied in the model plant Arabidopsis thaliana. In Arabidopsis, the MADS-domain transcription factor flowering repressor FLOWERING LOCUS C (FLC) plays a central role in the process of vernalization (Madrid et al. 2021). In the absence of vernalization, FLC binds directly to the genes encoding flowering activators, such as FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1), repressing their transcription. Vernalization causes transcriptional repression and epigenetic silencing of FLC, which is also maintained after plants are returned to warmer temperatures, leading to the expression of FT and SOC1 and induction of several genes involved in floral transition at the shoot apex. This molecular mechanism of vernalization involving FLC is conserved in various Brassicaceae plants and ensures plants flower in the spring when their pollinators are active.

The genes involved in vernalization have been breeding targets for manipulation of flowering time and accelerating improvement of Brassicaceae crops. In Brassica rapa crops grown for their vegetative organs, such as turnip and Chinese cabbage, floral transition reduces crop yield and quality. It is therefore preferable for these crops to be insensitive to vernalization and flower late. Accordingly, functionally and/or transcriptionally different alleles of FLC and FT have been identified in four FLC paralogs (BrFLC1, BrFLC2, BrFLC3 and BrFLC5) and two FT paralogs (BrFT1 and BrFT2) in the B. rapa genome (Akter et al. 2021). Alleles of BrFLC2 and BrFLC3 with a transposable element (TE) inserted in the first intron show a weak response to cold exposure and delayed FT induction, resulting in extremely late flowering plants (Kitamoto et al. 2014). A non-functional BrFT2 allele with a TE insertion in the second intron also confers a late flowering habit (Zhang et al. 2015). By contrast, the B. rapa crops harvested for their reproductive organs, such as oil seeds and bloom vegetables, require flowering time that optimizes crop productivity and quality. The oil seed type cultivar ‘Yellow Sarson’ (‘YS’) has a non-vernalization requirement (NVR) for flowering conferred by non-functional recessive alleles of all known FLC paralogs (Li et al. 2015). This NVR trait allows the cultivation ‘YS’ without consideration for the length of vernalization period.

We have recently identified that the local Malaysian landrace ‘CHOY SUM EX CHINA 3’ (‘CHOY’) does not require vernalization for flowering. However, the causal gene of this NVR phenotype of ‘CHOY’ has not been identified. Our genetic analysis revealed that NVR in ‘CHOY’ is controlled by a single genomic region on chromosome A07 we designated as nvQTL7. Interestingly, nvQTL7 contains none of the FLC genes that are responsible for the NVR phenotype of ‘YS’ (Itoh et al. 2018). Furthermore, F₁ progeny obtained from a cross between ‘CHOY’ and ‘Kohiki’, a cultivar that requires vernalization for flowering, can flower without vernalization, suggesting that the ‘CHOY’ allele at nvQTL7 is dominant.

In this study, we identified the causal gene controlling NVR for flowering in ‘CHOY’ by applying Sat-BSA, a genetic analysis method based on next-generation sequencing previously developed by our group (Segawa et al. 2021). We performed transcriptome sequencing (RNA-seq) and reverse transcription quantitative polymerase chain reaction (RT-qPCR) to understand the molecular mechanism of this causal gene and examined progeny of ‘CHOY’ and ‘YS’ to confirm the effect of pyramiding genes related to NVR from ‘CHOY’ and ‘YS’. We propose that the allele causing NVR in ‘CHOY’ could be used to improve flowering habit in various Brassica species via inter-specific crossing and grafting.

Plant materials

All B. rapa and B. napus cultivars used in this study, except ‘YS’, were obtained from the National Agriculture and Food Research Organization (NARO) Genebank of Japan (https://www.gene.affrc.go.jp/index_en.php). The NARO Genebank accession number for each cultivar/landrace is provided in Supplemental Table 1. ‘YS’ was obtained from Kobe University. Raphanus sativus cv. ‘Beni Kururi’ was purchased from Matsunaga seed.

Re-analysis of single bulk DNA using QTL-seq

Illumina short reads we previously generated for ‘CHOY’, ‘Kohiki’ and their F₁ progeny were downloaded from the NCBI Sequence Read Archive (SRA) using SRA-tools (https://github.com/ncbi/sra-tools.git) (Itoh et al. 2018) (Supplemental Table 2). Short reads with more than 20% nucleotides having phred quality score of < 20 were excluded from further analysis. The filtered short reads were then analyzed using the QTL-based sequencing (QTL-seq) pipeline ver. 2s (https://www.ishikawa-pu.ac.jp/staff/staffname/takagi-horoki/) (Itoh et al. 2018; Takagi et al. 2013). In this pipeline, alignment of reads and conversion of SAM/BAM files were carried out using BWA ver. 0.7.15 (Li and Durbin 2009) and SAMtools ver. 1.9 (Li et al. 2009). As reference genome for this QTL-seq analysis, we applied “CHOY-reference” developed by replacing nucleotides of the publicly available reference genome for B. rapa “Brapa_sequence_v3.0.fasta (http://brassicadb.cn/#/Download/)” with those of ‘CHOY’ nucleotides at all homozygous SNP positions. SNP-index values in the bulked DNA were calculated at all SNP positions that are homozygous and heterozygous in the parental lines and F₁ progeny, respectively. SNP positions with alignment depth of <10 were excluded from QTL-seq analysis.

Sat-BSA analysis

To identify structural variations that are potentially associated with the trait in question, we applied the Sat-BSA method we recently developed (Segawa et al. 2021). Sat-BSA was carried out using whole-genome NanoPore long reads and Illumina RNA-seq of short reads (Supplemental Table 2, 3). For whole-genome sequencing using NanoPore Technologies, 1 μg gel-purified DNA (Zymoclean Large Fragment DNA Recovery Kit, ZYMO RESEARCH) was subjected to library construction using a 1D Genomic DNA Ligation Sequencing Kit (SQK-LSK 109, Oxford Nanopore Technologies). Constructed libraries were then sequenced using a MinION sequencer (Mk1B, Oxford Nanopore Technologies) according to the manufacturer’s protocol. For RNA-seq, RNA was extracted using an RNeasy Plant Mini kit (QIAGEN) from three individuals as one sample and subjected to DNase treatment; 1 μg total RNA was then used for library construction using a NEBNext Ultra II RNA Library Prep Kit for Illumina (NEB). As the preliminary step of Sat-BSA, a partial genome sequence was constructed by local de novo assembly using Canu ver. 1.9 (Koren et al. 2017). For local de novo assembly, Oxford Nanopore Technologies (ONT) reads were aligned to “Brapa_sequence_v3.0.fasta” using Minimap2 ver. 2.17 (Li et al. 2018) and reads that aligned at the target genomic region were selected. The selected reads were then assembled using Canu ver. 1.9 (Koren et al. 2017) with default configuration except for an option defining the expected total contig size = 2 Mb. Short reads from RNA-seq were aligned using STAR ver. 2.0.0 (Dobin et al. 2013) as previously described (Segawa et al. 2021) to the assembled contigs, and genes on the assembled contigs were predicted using StringTie ver. 2.1.1 (Pertea et al. 2015).

In Sat-BSA, ‘CHOY’ specific structural variations (SVs) in the contigs developed by local de novo assembly were detected by comparing the aligned depth of ‘CHOY’ and ‘Kohiki’ ONT reads with the Sat-BSA pipeline (https://anaconda.org/bioconda/sat-bsa). ONT reads were aligned to “the combined reference sequence” which contains both the constructed contigs by local de novo assembly and the “Brapa_sequence_v3.0.fasta” substituted with ‘N’ in the genomic region applied to local de novo assembly. This SV search targeted deletions of more than 100-bp in predicted genes and their 10-kb upstream regions. Differentially expressed genes based of the transcripts per million (TPM) value (Wagner et al. 2012) in each exon of the predicted genes were detected by RNA-seq. In this differentially expressed genes analysis, the RNA-seq reads were aligned to “the combined reference sequence” used in SV serch with STAR ver. 2.0.0. TPM values were calculated using custom scripts from the read number counted by featureCount (Liao et al. 2014). Statistical analysis was performed using Two Sample t-tests.

Phenotyping NVR in progeny

Days to flowering was defined as the period from sowing to when the first flower fully opens. Plants were grown and phenotyped in a plastic house at Ishikawa Prefectural University in Nonoichi, Japan, and in a room with artificially controlled environment under 22°C/16-h light and 400-600 ppm CO₂ concentration. Samples for phenotyping were sown in a cell tray and transplanted into plastic pots for cultivation until flowering.

Expression analysis for flowering related genes by RNA-seq and RT-qPCR

Plants were grown on seedling cell trays under controlled environment under 22°C/16-h light conditions. Each sample for RNA extraction prepared by pooling whole above ground plant parts from three individuals (Supplemental Fig. 1). Total RNA samples were extracted according to the method used in Sat-BSA analysis. In expression analysis using RNA-seq, RNA-seq library construction and sequencing were according to the method described in Sat-BSA and the obtained RNA-seq reads were aligned to “Brapa_sequence_v3.0.fasta”. In RT-qPCR, about 0.5 μg total RNA was subjected to cDNA synthesis using a PrimeScript™ II 1st strand cDNA Synthesis Kit. The cDNA was used for RT-qPCR analysis with KAPA SYBR Fast qPCR Kit in a StepOnePlus system. The primer sequences are shown in Supplementary Table 4.

Grafting Raphanus sativus to ‘CHOY’

Seeds of R. sativus were germinated on a seedling cell tray, and from seven days to three weeks after sowing, 1 mM gibberellin was applied onto the shoot apex of seedlings every other day. The stem sections between the cotyledon and first true leaf and used as scions for grafting onto ‘CHOY’ rootstock. Flowering stems of ‘CHOY’ plants were cut at a position having the same size (diameter) as the scion and were used as rootstocks for grafting. Grafted stems were fixed with Superwith (NASUNICS) and plants were grown under high humidity and dim lighting conditions under 22°C for 2 weeks followed by cultivation under 22°C/16-h light conditions.

The florigen gene BrFT2 is the causal gene of nvQTL7

To accurately identify the genomic region associated with the ‘CHOY’ NVR phenotype, we initially applied the latest QTL-Seq analysis pipeline to a single bulk DNA sample prepared from 20 non-flowering individuals of an F₂ progeny derived from a cross between ‘CHOY’ and ‘Kohiki’ (a cultivar requiring vernalization for flowering) (Fig. 1a, 1b, Supplemental Fig. 2). QTL-seq identified the highest SNP index, average of > 0.96, within a 21.55-23.55 Mb genomic region on chromosome A07, which we designated as nvQTL7 (Fig. 1c). This candidate region corresponded to a chromosomal region we previously identified using two bulk DNA samples prepared from individuals with extreme phenotypes in the same F₂ progeny (early flowering vs. non-flowering) (Itoh et al. 2019). To identify the candidate gene within the nvQTL7 genomic region, we then applied Sat-BSA (Segawa et al. 2021). Using Sat-BSA, we reconstructed the genomic sequence of ‘CHOY’ corresponding to nvQTL7 (A07: 21.55-23.55 Mb) by local de novo assembly of ‘CHOY’ ONT long reads, which were first aligned to the B. rapa reference genome “Brapa_sequence_v3.0.fasta”. Of these, we selected 3,488 reads that aligned to the nvQTL7 genomic region for de novo assembly with Canu (Koren et al. 2017) (Supplemental Fig. 3). The assembly generated 43 contigs with a total length of 2.67 Mb, representing the genomic sequence of nvQTL7 in ‘CHOY’. We then used RNA-seq data generated from ‘CHOY’ leaves sampled at 42 days after sowing (DAS) to predict 195 genes within the 2.67 Mb candidate region.

To narrow down the candidate genes further, we first compared depth of ‘CHOY’ and ‘Kohiki’ ONT reads aligned to the developed contigs and identified ‘CHOY’-specific structural variations within 117 of the 195 predicted genes. Next, we generated backcross populations by crossing an F₁ plant (‘CHOY’ × ‘Kohiki’) to ‘Kohiki’ several times with the nvQTL7 genotype of the individuals determined by PCR using a DNA marker specific to the nvQTL7 candidate region (Fig. 2). Accordingly, we obtained populations that mostly shared the same ‘Kohiki’ genetic background but were different for the nvQTL7 allele. Using transcriptome data generated for samples composed of aboveground plant parts and collected at 14 DAS from plants grown under constant temperature and photoperiod (22°C day/light and 16-h light/8-hr dark photoperiod) without vernalization, we compared ‘CHOY’-specific expression patterns between BL3F3-CC and BL3F3-KK plants that are homozygous for ‘CHOY’ and ‘Kohiki’ alleles on nvQTL7, respectively. We accordingly identified five genes with significantly higher expression levels in BL3F3-CC than in BL3F3-KK (P < 0.01, t-test). Of these five genes, four were among the 117 genes that harbored ‘CHOY’-specific structural variations, making them the most likely candidates for nvQTL7 (Fig. 3a). Blast analysis of the four candidate genes revealed that one of them encoded BrFT2, a florigen in B. rapa (Zhang et al. 2015) (Supplemental Table 5).

To investigate if there are polymorphisms in BrFT2 between ‘CHOY’ and ‘Kohiki’, we performed Sanger sequencing of BrFT2 and its promoter region from the two cultivars (Supplemental Fig. 4). The coding region of BrFT2 had no polymorphisms causing nonsynonymous substitutions between ‘CHOY’ and ‘Kohiki’. However, the UTR, intron and promoter regions contained several polymorphisms. In particular, the region upstream of BrFT2 in ‘CHOY’ contained two insertions 6,929-bp and 1,573-bp in length that were not present in ‘Kohiki’ (Fig. 3b, 3c). Analysis using the Repbase database (Bao et al. 2015) revealed that the 6,929-bp insertion shared sequence similarity with LTR-type retrotransposons, with a 1.5-kb repeat sequence at the terminal region. Whereas the 1,573-bp insertion had no similarity to previously described TEs. We also investigated polymorphisms in the first intron of BrFT2 because this intron contains two CArG boxes known to be targeted by FLC in Arabidopsis (Helliwell et al. 2006). But we did not detect any polymorphisms localized in the CArG box. The B. rapa reference genome “Brapa_sequence_v3.0.fasta” encodes four FLC paralogs and their major target genes that include two FT paralogs and three SOC1 paralogs. We therefore compared the expression patterns of these genes between BL3F3-CC and BL3F3-KK using RNA-seq (Fig. 3d). Of the nine genes analyzed, only one of the FT paralogs, BrFT2, showed a significantly higher expression in BL3F3-CC (P < 0.01, t-test). Taking these results together, we concluded that the ‘CHOY’ allele of BrFT2, that we named as BrFT2-C, is the most likely candidate for NVR in ‘CHOY’. We hypothesized that the large insertions in the promoter region of BrFT2-C is responsible for the expression of BrFT2 in non-vernalized ‘CHOY’ plants.

Confirming the presence of BrFT2-C in B. rapa landraces showing NVR phenotypes

To determine whether the insertions we identified in the promoter of BrFT2-C in ‘CHOY’ are conserved in B. rapa landraces that also show NVR phenotypes, we investigated presence/absence of the insertions in six NVR landraces originating from several subtropical and tropical regions (Fig. 4a). Using primer pairs flanking the 1,573-bp ‘CHOY’-specific insertion, we identified PCR products of the expected sizes in four landraces obtained from Taiwan, Malaysia and Pakistan, indicating they harbor the same insertion (Fig. 3c). The remaining two landraces from India had PCR products different from the expected size, which suggested absence of the 1,573-bp ‘CHOY’-specific insertion. One of these two lines from India is ‘YS’, which also shows NVR due to the lack of a functional FLC allele (Li et al. 2015). We also performed PCR amplification with primers flanking the 6,929-bp insertion. Of the four landraces that contained the 1,573-bp insertion, a landrace from Pakistan (‘Pak10432’) lacked the 6,929-bp insertion, suggesting this insertion is not required for NVR phenotype (Fig. 4b).

To further investigate the role of the 1,573-bp insertion, we developed four F₁ progenies by independently crossing four landraces (‘40 DAYS RAPE’, ‘OOBA YUSAISHIN’, ‘Pak10432’, and ‘HOMEI’) carrying the 1,573-bp insertion with ‘Kohiki’ that lacks it (Fig. 4c). When cultivated without vernalization, all F₁ progenies flowered by 70 DAS, while ‘Kohiki’ failed to flower at all (Fig. 4c, 4d). Although the 6,929-bp insertion was not detected in ‘Pak10432’, F₁ progeny obtained from ‘Pak10432’ × ‘Kohiki’ cross also flowered, which suggested the 1,573-bp insertion is enough for NVR.

Dosage effect of BrFT2-C

Whole-genome sequencing of ‘Kohiki’ revealed that the allele patterns of the functional FLC paralogs in ‘CHOY’ is different from the allele patterns in ‘Kohiki’ (Supplemental Table 6). Consequently, we investigated the dosage effect of BrFT2-C in BL4F2 backcross lines sharing a common genetic background and the same combination of FLC paralogs. The BL4F2 plants developed by selfing of a BL4F1 plant, which was heterozygous for the 1,573-bp insertion in BrFT2-C, followed by five successive backcrosses to ‘Kohiki’ segregated for the BrFT2 genotype and expected to have >95% ‘Kohiki’ genetic background (Fig. 2). We named the BL4F2 individuals that were homozygous for the ‘CHOY’ and ‘Kohiki’ alleles as BL4F2-CC and BL4F2-KK, respectively, and those with the heterozygous genotype as BL4F2-CK. When we cultivated BL4F2 individuals and their parental lines without vernalization under controlled environmental room, ‘CHOY’, BL4F2-CC and BL4-F2-CK plants flowered at 34.3 ± 1.25, 42.3 ± 4.03 and 48 ± 3.74 DAS (mean ± sd of three individuals), respectively (Fig. 5a). By contrast, ‘Kohiki’ and BL4F2-KK failed to flower even at 80 DAS. These results suggested that the ‘CHOY’ allele in nvQTL7 is dominant over the ‘Kohiki’ allele.

We next compared the expression level of BrFT2 in 2-week-old seedlings of BL4F3 obtained from selfing a BL4F2 that was heterozygous for BrFT2 genotype (Fig. 5b). Expression analysis using RT-qPCR revealed that BrFT2-C expression in BL4F3 lines with a homozygous genotype for BrFT2-C was twice than the expression level of heterozygous genotypes. By contrast, BL4F3 lines with a homozygous genotype for the ‘Kohiki’ allele showed much lower expression of BrFT2 than other genotypes of BL4F3.

Timing and pattern of BrFT2-C expression

To study the expression pattern of BrFT2-C in detail, we monitored its expression in shoots of ‘CHOY’, ‘Kohiki’ and BL3F3-CC (developed by selfing of a BL3F2 plant) at 1 and 2 weeks after sowing by RT-qPCR (Fig. 5c). At 1 week after sowing, ‘CHOY’ and BL3F3-CC had similar BrFT2 expression levels, while its expression in ‘Kohiki’ was significantly lower (Fig. 5c). The expression of BrFT2 in ‘CHOY’ and BL3F3-CC increased slightly at 2 weeks after sowing but remained low in ‘Kohiki’. In addition, we monitored the expression pattern of BrFT2 over the course of 1 day (24 hrs) in 2-week-old ‘CHOY’ seedlings grown under 16-h light/8-h dark conditions using samples collected every 4 h BrFT2 was expressed constitutively throughout the day although its expression was slightly lower at 12 h of exposure to light (Fig. 5d).

The effect of combining BrFT2-C with non-functional FLC paralogs on early flowering

To investigate the effect of combining the dominant BrFT2-C allele with a non-functional FLC on flowering time, we crossed ‘CHOY’ with the oilseed B. rapa cultivar ‘YS’, which show NVR phenotype because it lacks functional alleles of all FLC paralogs (Li et al. 2015). We checked the allelic patterns of both FT and FLC paralogs in ‘YS’ by whole-genome sequencing and confirmed that ‘YS’ had a functional BrFT1 allele, like ‘CHOY’ (Supplemental Table 6). However, the BrFT2 in ‘YS’ is a non-functional allele harboring a TE insertion, supporting the previous report by Zhang et al. (2015). Among the four FLC paralogs in ‘YS’, BrFLC1, BrFLC2 and BrFLC5 harbored loss-of-function mutations. However, contrary to expectations from previous reports, a functional allele in BrFLC3 was detected in ‘YS’ maintained by selfing in our lab (Yuan et al. 2009; Wu et al. 2012; Xi et al. 2018). By contrast, ‘CHOY’ contained functional alleles of both BrFLC2 and BrFLC3 and nonfunctional alleles of the other two FLC paralogs. BrFT2 and BrFLC2 therefore show functional differences between ‘CHOY’ and ‘YS’. Here, we named the ‘CHOY’ and ‘YS’ alleles of BrFT2 as BrFT2-C and brft2-y, and the ‘CHOY’ and ‘YS’ alleles of BrFLC2 as BrFLC2-C and brflc2-y, respectively.

We cultivated 392 F₂ individuals obtained from a cross between ‘CHOY’ and ‘YS’ in a plastic house under long-day conditions without vernalization. In the same plastic house, ‘CHOY’, ‘YS’ and their F₁ progeny flowered 40 ± 1.05, 40 ± 0.75 and 39.2 ± 0.78 DAS (mean ± sd, n = 10), respectively. However, the flowering time of F₂ progeny segregated in the range from 32 to 59 DAS (Fig. 6a, 6b). We next investigated the BrFT2 and BrFLC2 genotypes in the 20 earliest and latest flowering individuals among the F₂ progeny. Among the earliest-flowering 20 individuals, 8 and 10 showed a homozygous for BrFT2-C and heterozygous genotype, respectively. By contrast, 16 of the 20 latest-flowering individuals were homozygous for brft2-y. Moreover, in BrFLC2, 14 individuals showing earliest flowering were homozygous for brflc2-y and 15 individuals showing latest flowering were homozygous for BrFLC2-C (Fig. 6c).

Furthermore, we investigated the expression of FT paralogs in the shoots of four three-days old F₃ lines possessing a combination of homozygous BrFT2-C and homozygous brflc2-y. The expression level of BrFT2 in these F₃ lines was higher than that in ‘CHOY’, suggesting that the homozygous brflc2-y genotype contributes to the higher expression of BrFT2-C at an early stage (Fig. 6d). The expression level of BrFT1 in these F₃ lines showed no significant difference not only from that in ‘CHOY’ but also from that in ‘YS’, suggesting that BrFT1 is not expressed at three DAS, even without repression by BrFLC2.

In conclusion, these results suggested that a high expression level of BrFT2 at the very early stage of plant development, resulting from a homozygous BrFT2-C and brflc2-y genotype, contributes to earlier flowering.

Application of BrFT2-C to modify flowering habit in allotetraploid B. napus

Because the BrFT2-C allele enables plants to overcome repression of BrFT2 by FLC, we hypothesized that BrFT2-C is useful for introducing NVR into Brassica napus, which is known to have nine FLC paralogs (Calderwood et al. 2021). To test this hypothesis, we developed F₁ progeny through a cross between ‘CHOY’ and the winter-type B. napus cultivar ‘Kamikita natane’ (‘KN’) and cultivated these plants without vernalization. As expected, F₁ (‘CHOY’ × ‘KN’) individuals flowered at 57.75 ± 4.79 DAS (mean ± sd, n = 3), whereas ‘KN’ and F₁ (‘Kohiki’ × ‘KN’) plants did not flower even after 80 DAS, indicating that BrFT2-C can overcome the repression imposed by FLC paralogs encoded in the B. napus A and C genomes (Fig. 7a, 7b). By contrast, the flowering timing of F₁ progeny derived from ‘YS’ and ‘KN’ was significantly delayed, by 13.5 days, compared with that of the F₁ progeny of ‘CHOY’ and ‘KN’. These results suggested that the NVR-related allele from ‘YS’ has a weaker effect on flowering timing than BrFT2-C.

Use of lines carrying BrFT2-C as rootstock for grafting to enhance flowering

Recently, Motoki et al. (2022) have successfully induced flowering in a never-flowering cabbage by grafting onto the flowering stem of R. sativus. This indicates that FT functions interspecifically in Brassicaceae plants. We expected that ‘CHOY’ should be suitable as a rootstock for grafting to induce flowering because BrFT2-C expresses sufficient FT for flowering without vernalization at an early stage of plant development, which reduces labor and time.

To verify the suitability of ‘CHOY’ as a rootstock for grafting, we induced flowering in R. sativus ‘Beni kururi’ (‘BK’), which requires vernalization for flowering, by grafting ‘BK’ to the flowering stem of ‘CHOY’. Because ‘BK’ has a rosette leaf type, we first treated a ‘BK’ plant with 1 mM gibberellin to force stem elongation for grafting. We then cut off the elongated stem and grafted to the flower stem of a ‘CHOY’ plant. The grafted plant was kept for 2 weeks under high humidity and dim lighting and then grown for 1 month under 16-h light/8-h dark conditions at 22°C. The grafted ‘BK’ plant flowered 1 month after grafting to the ‘CHOY’ flower stem (Fig. 8a). By contrast, non-grafted ‘BK’ plants never flowered, even after 2 months, whether or not they had been treated with gibberellins (Fig. 8b, 8c).

Here, we showed that the NVR in B. rapa ‘CHOY’ is caused by a dominant BrFT2-C allele, which is constituently expressed in non-vernalized plants, making it a potential breeding target for manipulating flowering time in Brassicaceae plants.

Most studies of flowering time in B. rapa have focused on the type and dosage of functional FLC paralogs (Akter et al. 2021). Takada et al. (2019) reported that the expression level of all FLC paralogs combined strongly correlates with flowering time in B. rapa. By contrast, much less information is available concerning the dosage and type of FT paralogs involved in determining flowering time, despite their being the main target of FLC. Analysis of the allele dosage in BL4 lines (Fig. 5a, 5b) revealed that the amount of BrFT2 expressed in the homozygous genotype of BrFT2-C (BL4F3-CC) was twice than that in the heterozygous genotype (BL4F3-CK). However, the time to flowering of BL4F2-CK after sowing was less than twice that in BL4F2-CC. Moreover, the ratio of the homozygous genotype of BrFT2-C and the heterozygous genotype of BrFT2 was very similar among the earliest-flowering individuals of F₃ progeny obtained from a cross between ‘CHOY’ and ‘YS’ (Fig. 6c). These results suggested that the total amount of expressed FT has little effect on the difference in flowering time and that BrFT2-C is expressed in sufficient quantity for flowering regardless of allele dosage.

Therefore, improving other flowering-related genes, rather than improving FT paralogs, is a more important step for developing early flowering lines. In F₃ (‘CHOY’ × ‘YS’) individuals, the combination of homozygous BrFT2-C and brflc-y allowed earlier flowering than that observed in ‘CHOY’, although the dosage of FT was not correlated with flowering time in BL4F2, as described above. We propose that early expression of genes repressed by FLC paralogs, such as SOC1 and SPL15, might contribute to early flowering in these F₃ lines due to the brflc-y allele. Franks et al. (2015) reported that the total amount of SOC1 expressed is strongly correlated with flowering time. If a SOC1 allele circumventing FLC repression is identified, early-flowering can be accelerated further by combining this with BrFT2-C.

In this study, F₃ (‘CHOY’ × ‘YS’) lines homozygous for both a non-functional BrFT2 allele from ‘YS’ and a functional BrFLC2 allele from ‘CHOY’ flowered without vernalization. We anticipate the possibility that ‘YS’ has an alternative gene controlling NVR other than FT or FLC. Such an allele might be identified through genetic analysis using Sat-BAS of segregating progeny in which the alleles of BrFT2 and BrFLC2 are fixed.

We detected two insertions, 6,929-bp and 1,573-bp, in the promoter region of BrFT2-C. The 1,573-bp insertion might be enough for NVR in ‘CHOY’ because F₁ progeny derived from a cross between ‘Kohiki’ and ‘Pak10432’ flowered without vernalization. Recently, Wei et al. (2022) identified BrFT2 as a causal gene responsible for the difference in flowering time between Chinese cabbage cultivars. The early-flowering cultivar harbored a 1,577-bp insertion in the promoter region of BrFT2 compared with the late-flowering cultivar. This allele might be the same allele as BrFT2-C. However, these authors did not mention that the alleles they identified contribute to NVR. Moreover, Wei et al. (2022) suggested that the 1,577-bp insertion contributes to promoter activity, although a detailed mechanism for accelerating expression under the repression of FLC was not clear. Therefore, further analysis of the interaction between the FT2 promoter and FLC is needed. Analysis using the Repbase database did not identify any reported repetitive elements such as TEs showing homology with the 1,573-bp insertion in BrFT2-C, although blast analysis detected multiple copies of this 1,573-bp insertion in “Brapa_sequence_v3.0.fasta”. By contrast, the 6,929-bp insertion encodes a Gypsy-type LTR transposable element known as the most abundant LTR in the B. rapa genome (Cai et al. 2022). Phylogenetic analysis based on polymorphisms around this 1,573-bp insertion in the pan-genome of the B. rapa germplasm collection will be needed to clarify where and when this 1,573-bp insertion occurred in the promoter region of BrFT2-C.

Interspecific hybrids between B. rapa harboring BrFT2-C and winter-type B. napus flowered, suggesting that BrFT2-C can overcome repression from FLC encoded in both the A and C genomes. Therefore, BrFT2-C could contribute to development of novel spring-type B. napus germplasm through the introduction of BrFT2-C into the B. napus genome via synthetic B. napus developed from a cross between ‘CHOY’ and any Brassica oleracea species. A recent study revealed that loss of function of FLC paralogs, BnaA02.FLC, BnaC02.FLC and BnaA10.FLC, causes the division of flowering habit between spring and winter type (Hou et al. 2012; Song et al. 2020; Yin et al. 2020). Therefore, breeding F₁ hybrid cultivars for spring-type B. napus should be possible by crossing among genetic resources conferring the homozygous genotype of the loss of function allele for some FLC paralogs. Indeed, whole-genome analysis using 183 cultivars comprising winter, semi-winter and spring types of flowering habit revealed that almost all spring-type B. napus have a closely related genomic structure, implying that modern spring-type B. napus was developed by crossing from a very limited germplasm (An et al. 2019). If BrFT2-C is introduced into the B. napus genome, the germplasm utilized for breeding spring-type B. napus will be broadened. The dominant character of the BrFT2-C allele will facilitate breeding of spring-type F₁ hybrid cultivars through crossing even with winter-type B. napus lines, resulting in cultivars with varying flowering times.

Furthermore, we propose the B. rapa line harboring BrFT2-C as a suitable rootstock for grafting to circumvent the long vernalization period required for flowering in various Brassicaceae plants. In R. sativus, at least three generations could be developed in a single year because we recently succeeded in obtaining flowers within 2 months after grafting. Moreover, this technique allows the production of seeds even in cultivars completely lacking the ability to flower through loss of function of all genes related to flower induction in Brassicaceae plants, as described by Motoki et al (2022). For practical application of this technology in breeding, further investigation into the actual amount of seeds produced in the field and detailed conditions for increasing the success ratio of grafting will be important.

Controlling flowering timing is an important task in breeding Brassicaceae plants. In summary, we envisage that using BrFT2-C to overcome repression by FLC provides an enormous opportunity for accelerating the breeding of various Brassicaceae plants, not only for early flowering by introducing BrFT2-C via crossing, but also for late flowering by grafting.

Acknowledgments

Computations were performed on the NIG supercomputer at ROIS (Research Organization of Information and Systems) National Institute of Genetics (Mishima, Shizuoka, Japan). Local landraces/cultivars of Brassica rapa except for ‘YS’ in this study were obtained from the National Agriculture and Food Research Organization (NARO) Genebank. We specially thanks to Ryo Fujimoto (Kobe University) for distributing ‘YS’ seeds.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Author Contributions

MN performed genetic and expression analyses, MT conceived and supervised the entire study, MH performed grafting, TS performed Sat-BSA analysis, SS performed next generation sequencing, NM developed the segregating progenies applied to genetic analysis, NI developed the segregating progenies and performed genetic analysis, TI performed expression analysis, MS conceived and supervised the entire study, HT designed the research and wrote the paper. All authors contributed to the manuscript.

Data Availability

The accession numbers for the sequence data are listed in Supplementary Table 2, 3. All data can be accessed in the National Center for Biotechnology Information (NCBI).

Akter A, Itabashi E, Kakizaki T, Okazaki K, Dennis ES, Fujimoto R (2021) Genome triplication leads to transcriptional divergence of FLOWERING LOCUS C genes during vernalization in the genus Brassica. Front Plant Sci 11:619417.
An H, Qi X, Gaynor ML, Hao Y, Gebken SC, Mabry ME, McAlvay AC, Teakle GR, Conant GC, Barker MS, Fu T, Yi B, Pires JC (2019) Transcriptome and organellar sequencing highlights the complex origin and diversification of allotetraploid Brassica napus. Nat Commun 10:2878
Bao W, Kojima KK, Kohany O (2015) Repbase Update, a database of repetitive elements in eukaryotic genomes. Mob DNA 6:11
Cai X, Lin R, Liang J, King GJ, Wu J, Wang X (2022) Transposable element insertion: a hidden major source of domesticated phenotypic variation in Brassica rapa. Plant Biotechnol J 20:1298-1310
Calderwood A, Lloyd A, Hepworth J, Tudor EH, Jones DM, Woodhouse S, Bilham L, Chinoy C, Williams K, Corke F, Doonan JH, Ostergaard L, Irwin JA, Wells R, Morris RJ (2021) Total FLC transcript dynamics from divergent paralogue expression explains flowering diversity in Brassica napus. New Phytol 229:3534-3548
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR (2012) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29:15-21
Franks SJ, Perez-Sweeney B, Strahl M, Nowogrodzki A, Weber JJ, Lalchan R, Jordan KP, Litt A (2015) Variation in the flowering time orthologs BrFLC and BrSOC1 in a natural population of Brassica rapa. PeerJ 3:e1339
Helliwell CA, Wood CC, Robertson M, James Peacock W, Dennis ES (2006) The Arabidopsis FLC protein interacts directly in vivo with SOC1 and FT chromatin and is part of a high-molecular-weight protein complex. Plant J 46:183-92
Hou J, Long Y, Raman H, Zou X, Wang J, Dai S, Xiao Q, Li C, Fan L, Liu B, Meng J (2012) A Tourist-like MITE insertion in the upstream region of the BnFLC.A10 gene is associated with vernalization requirement in rapeseed (Brassica napus L.). BMC Plant Biol 12:238
Itoh N, Segawa T, Tamiru M, Abe A, Sakamoto S, Uemura A, Oikawa K, Kutsuzawa H, Koga H, Imamura T, Terauchi R, Takagi H (2019) Next-generation sequencing-based bulked segregant analysis for QTL mapping in the heterozygous species Brassica rapa. Theor Appl Genet 132:2913-2925
Kitamoto N, Yui S, Nishikawa K, Takahata Y, Yokoi S (2014) A naturally occurring long insertion in the first intron in the Brassica rapa FLC2 gene causes delayed bolting. Euphytica 196:213-223
Koren S, Walenz BP, Berlin K, Miller JR, Bergman NH, Phillippy AM (2017) Canu: scalable and accurate long-read assembly via adaptive k-mer weighting and repeat separation. Genome Res 27:722-736
Li H (2018) Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics 34:3094-3100
Li H, Durbin R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25:1754-60
Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R, 1000 Genome Project Data Processing Subgroup (2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics 25:2078-9
Li X, Zhang S, Bai J, He Y (2016) Tuning growth cycles of Brassica crops via natural antisense transcripts of Br FLC. Plant Biotechnol J 14:905-914
Liao Y, Smyth GK, Shi W (2014) featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30:923-30
Long Y, Shi J, Qiu D, Li R, Zhang C, Wang J, Hou J, Zhao J, Shi L, Park BS, Choi SR, Lim YP, Meng J (2007) Flowering time quantitative trait Loci analysis of oilseed brassica in multiple environments and genomewide alignment with Arabidopsis. Genetics 177:2433-44
Madrid E, Chandler JW, Coupland G (2021) Gene regulatory networks controlled by FLOWERING LOCUS C that confer variation in seasonal flowering and life history. J Exp Bot 72:4-14
Motoki K, Kinoshita Y, Nakano R, Hosokawa M, Nakazaki T (2022) Quantitative Analysis of Florigen for the Variability of Floral Induction in Cabbage/Radish Inter-generic Grafting. Plant Cell Physiol 63:1230-1241
Pertea M, Pertea GM, Antonescu CM, Chang TC, Mendell JT, Salzberg SL (2015) StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat Biotechnol 33:290-5
Segawa T, Nishiyama C, Tamiru-Oli M, Sugihara Y, Abe A, Sone H, Itoh N, Asukai M, Uemura A, Oikawa K, Utsushi H, Ikegami-Katayama A, Imamura T, Mori M, Terauchi R, Takagi H (2021) Sat-BSA: an NGS-based method using local de novo assembly of long reads for rapid identification of genomic structural variations associated with agronomic traits. Breed Sci 71:299-312
Song JM, Guan Z, Hu J, Guo C, Yang Z, Wang S, Liu D, Wang B, Lu S, Zhou R, Xie WZ, Cheng Y, Zhang Y, Liu K, Yang QY, Chen LL, Guo L (2020) Eight high-quality genomes reveal pan-genome architecture and ecotype differentiation of Brassica napus. Nat Plants 6:34-45
Takada S, Akter A, Itabashi E, Nishida N, Shea DJ, Miyaji N, Mehraj H, Osabe K, Shimizu M, Takasaki-Yasuda T, Kakizaki T, Okazaki K, Dennis ES, Fujimoto R (2019) The role of FRIGIDA and FLOWERING LOCUS C genes in flowering time of Brassica rapa leafy vegetables. Sci Rep 9:13843
Takagi H, Abe A, Yoshida K, Kosugi S, Natsume S, Mitsuoka C, Uemura A, Utsushi H, Tamiru M, Takuno S, Innan H, Cano LM, Kamoun S, Terauchi R (2013) QTL-seq: rapid mapping of quantitative trait loci in rice by whole genome resequencing of DNA from two bulked populations. Plant J 74:174–183
Wagner GP, Kin K, Lynch VJ (2012) Measurement of mRNA abundance using RNA-seq data: RPKM measure is inconsistent among samples. Theory Biosci 131:281-5
Wei Q, Hu T, Xu X, Tian Z, Bao C, Wang J, Pang H, Hu H, Yan Y, Liu T, Wang W (2022) The New Variation in the Promoter Region of FLOWERING LOCUS T Is Involved in Flowering in Brassica rapa. Genes (Basel) 13:1162
Wu J, Wei K, Cheng F, Li S, Wang Q, Zhao J, Bonnema G, Wang X (2012) A naturally occurring InDel variation in BraA.FLC.b (BrFLC2) associated with flowering time variation in Brassica rapa. BMC Plant Biol 12:151
Xi X, Wei K, Gao B, Liu J, Liang J, Cheng F, Wang X, Wu J (2018) BrFLC5: a weak regulator of flowering time in Brassica rapa. Theor Appl Genet 131:2107-2116
Yin S, Wan M, Guo C, Wang B, Li H, Li G, Tian Y, Ge X, King GJ, Liu K, Li Z, Wang J (2020) Transposon insertions within alleles of BnaFLC.A10 and BnaFLC.A2 are associated with seasonal crop type in rapeseed. J Exp Bot 71:4729-4741
Yuan YX, Wu J, Sun RF, Zhang XW, Xu DH, Bonnema G, Wang XW (2009) A naturally occurring splicing site mutation in the Brassica rapa FLC1 gene is associated with variation in flowering time. J Exp Bot 60:1299-308
Zhang X, Meng L, Liu B, Hu Y, Cheng F, Liang J, Mark GMA, Xiaowu W, Wu J (2015) A transposon insertion in FLOWERING LOCUS T is associated with delayed flowering in Brassica rapa. Plant Sci 241:211-220

Download PDF

Journal Publication

published 18 May, 2023

Read the published version in Theoretical and Applied Genetics →

Reviewers agreed at journal
09 Jan, 2023
Reviewers invited by journal
04 Jan, 2023
Editor assigned by journal
13 Dec, 2022
First submitted to journal
12 Dec, 2022

You are reading this latest preprint version

Non-vernalization requirement for flowering in Brassica rapa conferred by a dominant allele of FLOWERING LOCUS T

Status:

Journal Publication

Version 1

Abstract

Figures

Key Message

Introduction

Materials And Methods

Results

Discussion

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1