Characteristics of a radish mutant with longer siliques

The silique is an important organ in plant reproduction and maintains biodiversity among species; however, little is known about the regulation of radish (Raphanus sativus L.) silique development. In this study, we conducted research on the radish long silique (ls) mutant and wild-type (WT) radish and compared their morphological and molecular markers. The results showed that the mutant obtained by ethyl methane sulfonate treatment had the following stable characteristics: the lengths of mutant and WT siliques were 20.50 and 9.10 cm, respectively; and the ovule number per silique of the mutant and WT was 9.5 and 4.5, respectively. Polymerase chain reaction (PCR) analysis revealed abundant polymorphisms between the ls mutant and WT in HZ001, SRC9-022, and OIL2F11 simple-sequence repeat molecular markers, and the expression of LS1 and LS2 (RsNAC66). Arabidopsis thaliana-transformed plants with RsNAC66 overexpression were obtained by the floral dip method. Quantitative PCR showed that LS2 (RsNAC66) was more highly expressed in transformed lines than in WT, and the expression of LS2 (RsNAC66) in the transformed lines was higher in siliques than leaves. Phenotypic analysis revealed abnormal ovule numbers and shoots, and altered plant height in the transformed plant. Phenotypic and gene expression analyses showed that LS2 (RsNAC66) plays an important role on silique length and the number of seeds per silique. Together, the results showed that the radish ls mutant is specific and stable, and thus constitutes an excellent research resource for improving the seed yield of radishes.

to storage. The radish cultivation area in China is approximately 1.2 × 10 6 m 2 , accounting for about 40% of the world's cultivated area and 47% of the world's total production (Wang 2016). In the process of longterm evolution, cultivation, and breeding, new uses for radishes have gradually emerged, including as a food source (where the leaves and siliques are edible) and cover crop (oilseed radish); radish has thus received increasing attention from researchers. Radish seeds contain high amounts of lipids and active substances (Kim et al., 2014), among which sulforaphane is a natural active substance with antitumor activity (Zhao et al., 2011). However, the low yield of radish seeds limits their versatility and value. There have been few studies on the mechanisms of radish seed yield and silique development. Zhai et al. (2016) showed that 5,777 differentially expressed genes appeared in radish ovules within 15 days before and after artificial pollination, including auxin-response factor, Leafy cotyledon 1, somatic embryogenesis receptor kinase, and other important regulatory genes involved in embryogenesis. These results provide an important reference for understanding the early embryonic development process of radish.
Our group previously revealed the developmental process of radish siliques and ovules through morphological analysis (Li et al., 2018b, a). The ovule development of radishes is similar to that of oilseed rape (Brassica napus) (Uyttewaal et al., 2008;Li et al., 2018b, a). Under normal conditions, the average seed yield of radish per mu (0.067 hectare) is 50 kg (Li 1992), which is significantly lower than that of similar economic crops such as oilseed rape. The low seed yield severely affects the ability to multi-purpose the radish.
In this study, we identified mutation traits in a radish mutant with a long silique (ls), based on a mutant pool constructed by ethyl methanesulfonate (EMS) mutagenesis . The results showed that the number of seeds per silique in the mutant was significantly increased compared to wild-type (WT) radish, and there were significant differences in molecular markers and gene expression related to silique development between the radish ls mutant and WT. In addition, the radish silique-related gene RsNAC66 was transformed into Arabidopsis, and phenotypic analysis showed that it may increase silique length and ovule number per silique.

Phenotypic traits of the radish ls mutant
Our research group previously established a radish mutagenesis system and constructed a radish mutant pool ; after many years of self-fertilization, they were used to screen the radish ls mutant with stable silique and ovule phenotypes (Fig. 1). There were significant differences between the ls Fig. 1 Comparison between ls mutants and WT radish. Radish ls mutants (B, mature; C, young) with longer siliques and WT radish (A, mature; D, young) with normal siliques mutant and WT radish in silique length and number of seeds per silique ( Fig. 2A-F). Specifically, the average length of silique in the ls mutant was 20.50 cm, compared to 9.10 cm in the WT radish (Fig. 2G), and the average number of seeds per silique in the radish ls mutant was 9.5, whereas the average number of ovules of WT radish was 4.5 (Fig. 2H). The statistical results showed that the siliques in the radish ls mutant were 2.3 times longer than those of WT radish, and there was a greater number of seeds per silique in the ls mutant than WT, indicating that the ls mutant not only has a longer silique than WT radishes but also a significantly higher number of ovules per silique. The radish ls mutant, with its significantly increased Fig. 2 Comparison on silique ovules between ls mutants and WT radish. A: Siliques of radish ls mutant (upper) and WT radish (bottom); B, C: Ovule distribution in the silique is shown for the radish ls mutant and WT radish (lower and upper panels, respectively); D: The process of silique development in radish ls mutant (left) and WT radish (right); E, F: siliques of radish ls mutant (E) and WT radish (F); G, H: Statistical results of the comparison of silique length (G) and number of ovules (H) between radish ls mutants with longer siliques and WT radish. Bar = 2.5 cm in A, D, E and F; Bar = 0.15 cm in B and C number of seeds per silique, has great research value and is an important experimental material for studying radish ovule formation and improving radish seed yield.
Detection of silique-related molecular markers and gene expression in the radish ls mutant The results of screening of 42 simple-sequence repeat (SSR) molecular markers related to crop traits have been published (see Materials and Methods), and showed that the amplification products of the SSR molecular marker primers HZ001, SRC9-022, and OI12F11 were different between WT and ls mutant radishes. There were also abundant polymorphisms in the ls mutant and WT radish molecular markers (Fig. 3A). The bands of WT radish amplified by molecular markers HZ001 and SRC9-022 primers were faint (Fig. 3A). By contrast, amplification of the radish ls mutant by the molecular marker OI12F11 did not produce bands (Fig. 3A). The LS1 gene was derived from the BnaA09g39480D sequence of the B. napus gene (Shen et al., 2019) through sequence alignment by Blast (LR778317), and the LS2 gene was designed from the radish gene (XM_018634088) sequence obtained by homologous alignment in the National Center for Biotechnology Information database. Semi-quantitative polymerase chain reaction (PCR) results showed that the expression of LS1 and LS2 in the silique of the radish ls mutant was lower than that in WT radish (Fig. 3B). In addition, there were clear differences in the expression of LS1 and LS2 between mutant and WT radish cotyledons.

Verification and qualification of
Transformed Arabidopsis lines overexpressing LS2 (RsNAC66) were screened using amplification fragments of the kanamycin gene by PCR (Fig. 4A). The semi-quantitative PCR results showed that the expression of LS2 (RsNAC66) was significantly enhanced in the transgenic plants ( Fig. 4B, C). Relative expression of the LS2 (RsNAC66) gene in the transgenic lines (lines 3, 5, 7, 9) was upregulated to different degrees compared to WT, and increased 6,054-, 2,427-, 2-, and 99-fold in the siliques of lines 3, 5, 7, and 9, respectively (Fig. 4C). These results showed that the expression of LS2 (RsNAC66) was significantly increased in the overexpressed lines.

Statistical analyses of the phenotypic characteristics of A. thaliana-transformed plants overexpressing LS2 (RsNAC66)
Statistical analyses showed that the average plant height of A. thaliana transgenic lines 3 and 7 was higher than that of WT (P < 0.05; Fig. 5A). There was no significant difference in plant height between transformed lines 9 and WT (P > 0.05). A. thaliana transgenic line 5 showed abnormal bolting (Fig. 5B), and early bolting and flowering were observed during first 6 weeks of A. thaliana growth. The average silique  (1) and WT radish (2); lanes 3 and 4: The amplification products of SRC9-022 for radish ls mutant (3) and WT radish (4); lanes 5 and 6: The amplification products of OI12F11for radish ls mutant (5) and WT radish (6); B. M: marker, lanes 1 and 2: Amplification results of actin in early siliques from radish ls mutant (1) and WT radish (2); lanes 3 and 4: Amplification results of LS1 in early siliques from radish ls mutant (3) and WT radish (4); lanes 5 and 6: Amplification results of LS2 in early silique from radish ls mutant (5) and WT radish (6); lanes 7 and 8: Amplification results of actin in cotyledons from radish ls mutant (7) and WT radish (8); lanes 9 and 10: Amplification results of LS1 in cotyledons from radish ls mutant (9) and WT radish (10); lanes 11 and 12: Amplification results of LS2 in cotyledons from radish ls mutant (11) and WT radish (12) length and number of ovule per silique of transgenic A. thaliana were measured and analyzed ( Fig. 5C-F). Among them, the silique length of transgenic lines 3 and 7 was longer than that of WT (P < 0.05, Tab.S1). There were more ovules per silique in lines 3, 7, and 9 than in WT (P < 0.01, Tab.S1).
Detection and analyses of the expression of silique development-related genes AP2, OPT4, and Atsus2 are silique development regulatory genes (Kunst et al., 1989;Bowman et al., 1989;Chai et al., 2011;Huang et al., 2013;Niu et al., 2002;Ohto et al., 2005). The heterologous expression of LS2 (RsNAC66) affects the development of Arabidopsis siliques. We further examined the correlation between the those genes related to silique development. As shown in Fig. 5, the relative expression of AP2 in the transgenic lines was upregulated in siliques, as well as 6-, 17-, 1-and twofold in the four tested transgenic lines (3, 5, 7, and 9), respectively (Fig. 5G). OPT4 and Atsus2 were both expressed at high levels in the four tested transgenic lines (3, 5, 7, and 9). The high expression and function of LS2 (RsNAC66) in siliques suggest that the AP2, OPT4 and Atsus2 coordinated with LS2 (RsNAC66) in the regulation of silique development.

Discussion
Comparison of morphology and detection of molecular markers and gene expression showed that the radish ls mutant induced by EMS was stable, and the number of ovules per silique was significantly higher than that of WT (Fig. 2B, C, D, G). The radish ls mutant is an important resource to improve the seed yield of radish. Studies on the variation and regulation of silique length in radish have not been conducted. At present, applied research on the development of molecular markers in radish is mainly focused on purity and germplasm identification (Qiu et al., 2014;Li et al., 2018b, a;Wang et al., 2019;Yang et al., 2019). Therefore, this study can only refer to other morphology-related molecular markers of radish and genes, and to markers related to the A. Electrophoretic assay of overexpression screening. M, DL2000 marker; lanes 1-6 and 8-13 are amplification products of a KANA gene fragment from 1-12 transgenic lines; lane 7 (with bands) corresponds to amplification products of a KANA gene fragment from WT lines; B. LS2 (RsNAC66) genes were detected by semi-quan-titative PCR. M, DL2000 DNA marker (2,000, 1,000, 750, 500,or 250 bp); lanes 1-5 show amplification results for WT (lane 1) Arabidopsis leaves and leaves of transgenic lines 3, 5, 7, and 9, respectively; lanes 6-10 show amplification results for WT (lane 9). Arabidopsis siliques and siliques of transgenic lines 3, 5, 7, and 9, respectively; C. Relative expression of RsNAC66 numbers of silique and ovules in other crops, to detect this radish ls mutant. Through polymorphism screening of the above 42 SSR molecular marker primers, three pairs of primers with clear amplification conditions and polymorphism were screened, indicating that the above three pairs of SSR molecular markers were universal in radish, turnip rape, and oilseed rape. Based on the sequence information of silique length regulatory genes in oilseed rape (Zhou et al., 2017;Sheng et al., 2019), we obtained the homologous sequence of silique length regulatory genes in radish through homology comparison, and found that the expression of these genes was significantly different between the ls mutant and WT radish.
NAC (NAM, ATAF, CUP) is one of the largest families of transcription factors in plants. The N-terminal has a highly conserved NAC domain, while the C-terminal has high diversity in the structural characteristics of NAC transcription factors (Munir et al., 2020). After constructing RsNAC66 overexpression transgenic lines in A. thaliana, the phenotypes of the transformants were analyzed by morphological and statistical methods (Fig. 5). Among the 11 Arabidopsis transgenic lines, abnormalities in silique length, ovule number per silique and plant height were observed. Quantitative analysis of RsNAC66 expression confirmed high expression of the RsNAC66 gene in transgenic Arabidopsis siliques. The expression of the AP2 gene, which is related to ovule development in A. thaliana, was quantitatively analyzed, and found to be significantly upregulated in transgenic lines. These results showed that RsNAC66 had an effect on silique and ovule development, and increased silique length and ovule number per silique were observed in transgenic lines with high RsNAC66 expression compared to WT. The expression of AP2 in silique was also upregulated in transgenic lines with high RsNAC66 expression, indicating that they both participated in the regulation of silique and ovule development.

Experimental materials and primers
The radish ls mutant with a longer silique based on EMS mutagenesis, as reported by Li et al. (2019), and the WT radish with a normal silique and A. thaliana (Columbia-0), were used in this study. ls mutant was selected from mutated WT radish(Duanye-13 radish) by EMS mutagenesis and then self-hybridization for at least five generations. The above plant materials were obtained from the germplasm resource bank of Hangzhou Radish Research and Development Center of Huanggang Normal University. The primers for molecular markers and related genes were as follows: primers 1-9 were selected according to the primer sequence information related to the bolting of radish published by Yang et al. (2019) and the molecular markers related to the number of seeds per silique of B. napus published by Zhang et al. (2012). Thirty-three pairs of primers corresponding to SSR molecular markers distributed in A01, A03, and A09 linkage groups, which are associated with silique length quantitative trait loci, were screened with primers 10-42 (Wang et al., 2011;Zhang et al., 2012;Yin et al., 2015;Yang et al., 2019). Through homologous comparison of the length-related gene sequence of silique in B. napus published by Shen et al. (2019), the radish gene sequence was obtained, and primers 43 and 44 were designed from this sequence. The primer sequence used in this experiment is shown in Table 1. The primers were synthesized by GenScript Biotech Corporation (Piscataway, NJ, USA).

Phenotypic statistical analyses of radish and Arabidopsis
After the bolting and flowering of plants, abnormal traits of silique were observed, and the silique length of radish was measured by tape measure and ImageJ software (NIH, Bethesda, MD, USA). Siliques were treated with 8 mol/L sodium hydroxide solution to observe the distribution and number of ovules in siliques. Arabidopsis material was taken at week 14 and three phenotypic traits were recorded: plant height, silique length, and ovule number per silique. Extraction, amplification, and detection of DNA and RNA.and PCR Genomic DNA was extracted from the samples using the Plant Genomic DNA Kit (DP305-02; Tiangen Biotech, Beijing, China). Using the HiPure Plant RNA Mini Kit (R4151-02; Magen Biotechnology, Guangzhou, China), RNA was extracted from the radish mutant, WT radish, Arabidopsis, and transgenic A. thaliana. The quality of DNA and RNA were analyzed by agarose gel (1%) electrophoresis. The extracted RNA was reverse-transcribed into cDNA using the PrimeScript RT Reagent Kit (RR047A; Takara Bio, Shiga, Japan) as a template for PCR. The PCR amplification system included 2 × Taq PCR Master Mix (12.5 μL), DNA (0.5 μL), 10 μmol/L primers (1 μL), and ddH 2 O (10 μL). The amplification procedure was as follows: pre-denaturation at 94℃ (3 min), 35 cycles of denaturation at 56℃ (30 s), annealing at 56 ℃ for 30 s, extension at 72℃ (1 min), final extension at 72℃ (5 min), and maintenance at 4℃. The PCR products were separated by 1% agarose gel electrophoresis.

Construction of the RsNAC66 overexpression vector
The cDNA sequence of RsNAC66 was used, along with the specific primers LSS-F(ActagggtctcGcaccATG AAC ATA TCA GTA AAC GGA) and LSS-R(ActagggtctcTAC CGC TAG TGT GAC GAT GGC GTC GT) for PCR amplification. The PCR product was subjected to agarose gel electrophoresis, the 978 bp electrophoretic fragment was cut out, and the target fragment was recovered using the TIANgel Midi Purification Kit (DP209-03; Tiangen).
The empty plasmid was digested with BsaI-HF and ligated to the above target fragment with T4 DNA ligase, and transformed into E. coli DH5α competent cells. After culture, it was coated on LB solid medium plates with kanamycin and cultured overnight in a constant temperature incubator at 37℃. The primers LSS (Test +)( AAG AGG AGA CCA TTC CAG AC) and Nos-R (ATC ATC GCA AGA CCG GCA AC) were used for colony PCR detection, and the fragment size was about 416 bp. Positive plaque was confirmed by sequencing, and single bacterial colonies were selected for shaking culture.
The plasmid with tagert fragment was extracted using a plasmid extraction kit, and Sanger sequencing was performed. The sequencing primer was NOS-R(ATC ATC GCA AGA CCG GCA AC), and the sequence was consistent with the target fragment sequence; thus, the overexpression vector was successfully constructed. The extracted plasmids were confirmed by restriction endonuclease digestion with EcoRI/HindIII.

Genetic transformation and resistance screening in A. thaliana
The plasmid with the target fragment was transformed into competent cells of Agrobacterium tumefacien, and the flowering A. thaliana WT was transformed by the floral dip method. The seeds were collected and seeded on MS solid medium containing kanamycin, and positive plants were identified by PCR; positive plants with RsNAC66 overexpression were screened.

Statistical analyses
Each experiment was repeated at least three times. One-way ANOVA (SPSS/13) was used to test significant differences between controls and mutants in all replications. P values were analyzed using Student's t-test.