2020 Volume 89 Issue 1 Pages 12-21
Brassica napus L. is one of the most important oilseed crops in the world. The flower stalks (lateral shoots) of the leafy vegetable B. napus landrace, commonly known as norabona, are consumed for nutritional purposes and these plants are grown in the Kanto region of Japan. In the present study, we revealed the diversity of norabona at genotypic and phenotypic levels. Samples were collected from four different areas in the Kanto region of Japan and comprised other leaf vegetable landraces (kakina; B. napus and komatsuna; B. rapa). Phenotypic traits were assessed by investigating six morphological traits and five compositions for flower stalks. Principal component analysis in multivariate analysis demonstrated that the 20 norabona and kakina samples in the present study could be separated into three clusters. Genotyping using a total of 24 simple sequence repeat markers and Nei’s matrix genetic distances and neighbor-joining clustering method led to the partitioning of 23 samples into three groups and komatsuna. The genotypic Group 1b comprised the largest number of norabona samples; the farthest geographic distance was observed in genetically close pairs. On the basis of our results, we concluded that (1) the norabona population in Japan is phenotypically diverse and that (2) this population is not genetically identical, but consists of different genotype subpopulations that are geographically not divergent.
Brassica vegetables, which belong to the Cruciferae family, are agriculturally important food crops. The genetic diversity of their germplasm has been studied to assess phenotypes and perform genotyping using regional and global samples (Balkaya et al., 2005; Bird et al., 2017; Zhao et al., 2005; Zhu et al., 2018). Brassica napus L. (genome AACC, 2n = 4x = 38), belonging to the Brassica genus, has two diploid progenitors, B. rapa L. (ex. turnip; AA, 2n = 20) and B. oleracea L. (ex. cabbage; CC, 2n = 18). Like soybean, it is an important global crop and is the second most cultivated oilseed worldwide (USDA, 2018). Both the leafy and root parts of B. napus crops are used as vegetables. However, the use of B. napus as a vegetable for human consumption is lower compared with that of oilseed consumption.
Although there are fewer studies on B. napus as a vegetable crop compared with oilseed production, few reports have evaluated the diversity of the vegetable populations at phenotypic and genotypic levels. Hasan et al. (2006) used the genotyping of simple sequence repeat (SSR) markers and reported that the vegetable B. napus population is extremely diverse. Cartea et al. (2005) investigated genotypes of B. napus vegetable landraces collected from northwestern Spain and British B. napus cultivars including rape kale using random amplified polymorphic DNA (RAPD) markers. They reported that the B. napus leaf vegetable landraces grown in northwestern Spain probably independently originated from the B. napus crops grown in other European regions. Soengas et al. (2006, 2008) also investigated the similar germplasm of the leaf vegetable B. napus landraces grown in northwestern Spain and additionally in northern Portugal. They reported diversity of morphological traits and genotypes using SSR markers and concluded that the genetic differences probably reflect the origin/breeding or domestication.
Flower stalks (FS; lateral shoots) produced from the leaf vegetable of B. napus are conventionally eaten in Japan. The leaf vegetable B. napus landrace is commonly called “norabona” in the Kanto region of Japan, which includes the capital city Tokyo. This region is the highest agricultural producer in Japan (NSTAC, 2017). Currently, norabona is cultured for harvesting FSs in spring and on a relatively small scale for direct sale or domestic use. The seeds are usually renewed by in-house seed production in different areas and by various producers. It is assumed that the norabona population is diverse. However, due to the aging of norabona farmers, the production of norabona, as well as the number of landrace producers (i.e. who produce own norabona seeds), have steadily decreased. Genetic differences, if any, and variations in morphological traits and compositions of norabona are not well documented. This lack of relevant data requires a deeper understanding of the diversity of existing norabona populations in order to preserve landraces and ensure their continued use.
In the present study, our aim was to clarify the diversity of norabona populations at genotypic and phenotypic levels. We evaluated phenotypes to investigate the morphology and composition of their edible parts (FS), and diversity on the basis of genotypes using SSR markers.
To analyze the genetic diversity of the norabona population, a total of 23 samples (Table 1) were collected from Farmers’ markets and direct producers, located in four different areas in the Kanto region (A–D; Fig. 1), representing the main norabona FS producing regions (Saitama prefecture, Tokyo metropolis, and Kanagawa prefecture) (Tsuge et al., 2015). All collected samples were harvested on the same sampling date. Between one and three samples were collected at each sampling location (indicated by black dots in Fig. 1). Sample collection was completed in triplicate within the period from 20 to 25 April 2015. For comparison, samples belonging to other leafy vegetable landraces, particularly B. napus (commonly known as “kakina”) and B. rapa (commonly known as “komatsuna”) were also collected. Samples were stored at 10°C in cool boxes during shipping and in prehab refrigerators at 5°C, relative humidity 80%, and volume 8.25 m2 in the research facility until analysis on the following day.
Summary of 22 samples of leaf vegetable B. napus collected from different locations.
The sampling areas (shaded ovals with dots; A–D) of 22 leaf vegetable B. napus samples within the Kanto region in Japan. Black dots indicate the exact sampling location.
A total of six morphological traits were used to examine the phenotypic diversity of norabona and were analyzed using multivariate analysis. For morphological traits, we mainly measured quantitative traits that were considered to exhibit a difference in appearance among samples, based on a previous study by Tsuge et al. (2015) that focused on the production and distribution norabona sites. We did this because no previous study has investigated the morphological traits of B. napus leaf vegetables for consumption (FS). The morphological traits are indicated in Figure S1. The weight (g), total length (cm, from the cut end of the stalk to the top of the leaf), stem length (cm, from the cut end of the stalk to the top of the bud), leaf weight (g), leaf no., leaf width (cm), and stem diameter (mm) within the FS were measured. The following terms were used: “FS weight” (g·cm−1, = weight/total length), “leaf weight” (%, = leaf weight/weight × 100), “TL − SL” (cm, = total length − stem length), and “leaf no”. (= leaf no. /stem length). The FS weight, TL − SL, and leaf no. were calculated based on differences in the cut FS length among various producers.
Overall, five compositions were measured to quantitatively analyze the phenotypic diversity of FS. The soil and plant analyzer development (SPAD) values were calculated using the 10 largest leaves from each FS and measured using a chlorophyll meter (SPAD-502 plus; Konika Minolta Japan, Inc., Tokyo, Japan). The Brix in the stem part was measured using a digital reflectometer (PR-201 Alpha; ATAGO CO., LTD., Tokyo, Japan). Ascorbic acid, calcium, and nitric acid (NO3−) content were quantified using a reflectometer, RQflex (Merck, Darmstadt, Germany). We selected ascorbic acid and calcium (Hanson et al., 2009) and nitric acid (Kaidzuka, 2009) as these were the important compositions for which varietal differences have been reported in previous studies conducted on Brassica vegetables. To measure ascorbic acid, 5 g of fresh leaves were collected from the FSs and homogenized with 45 mL of 8% metaphosphoric acid (FUJIFILM Wako Pure Chemical Corporation, Tokyo, Japan) for 40 s using a general mixer (BM-RS type; ZOJIRUSHI CORPORATION, Osaka, Japan) and ascorbic acid was extracted after filtration. To extract calcium and nitric acid (NO3−), 50 g of fresh FS were homogenized with 350 mL distilled water for 60 s using a general mixer (BM-HS08 type; ZOJIRUSHI CORPORATION), and calcium and nitric acid (NO3−) were extracted after filtration.
Principal component analysis (PCA) and cluster analysis for phenotypic data (Ohi and Sato, 2002) were performed using Excel statistical software ver. 2012 (Social survey research information Co., Ltd., Tokyo, Japan). These data were standardized before calculation. Due to an insufficient number of samples, data for all phenotypic traits could not be obtained for “D-K-2” and “A-N-3” (Table 1); these samples were not included in the PCA.
Analysis of genetic diversityGenomic DNA was extracted from leaves of each sample using a DNeasy Plant Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s protocol with some minor modifications. One leaf was randomly selected from one sample and used for analysis. On the basis of previously reported SSR markers, 24 SSR markers with clear, reproducible, and polymorphic amplification products were selected for a genetic diversity study (Chen et al., 2017). The SSR loci in the subgenomes A and C were 15 and 9, respectively. A PCR reaction was performed in a 10 uL system, and contained 10 ng of template DNA, 1 × KAPA 2G buffer A, 200 nM dNTP, 0.5 mM MgCl2, 0.1U KAPA 2G Fast DNA polymerase (KAPA Biosystems Inc., Woburn, MA, USA), 0.5 pM forward primer, and 2 pM reverse primer. The forward primers were labeled with the following four kinds of fluorescent dyes (6-FAM, VIC, NED, or PET) at their 5' end (Shimizu and Yano, 2011). The PCR reaction was performed in a C1000 Thermal Cycler (Bio-Rad Laboratories, Inc., Hercules, CA, USA) under the following conditions: 94°C for 3 min of initial denaturation followed by two-staged PCR cycles (30 cycles of 94°C for 20 s, 54°C for 30 s, and 62°C for 30 s; 3 cycles of 94°C for 20 s, 49°C for 10 s, and 72°C for 5 s), and final extension at 72°C for 10 min. The size of the amplified fragments was estimated using an automated DNA analyzer (model 3130xl; Applied Biosystems, Foster City, CA, USA) with a GeneScan-600LIZ size standard (Applied Biosystems). Fragment length was determined with GeneMapper 4.0 software (Applied Biosystems). Genetic diversity indices and Nei’s genetic distance (Nei et al., 1983) between individuals were calculated by GenAlEx software 6.502 (Peakall and Smouse, 2012). In the genetic diversity indices, the number of different alleles (Na), number of effective alleles (Ne), Shannon’s information index (I), observed heterozygosity (Ho), and expected heterozygosity (He) were calculated for each marker. The polymorphism information content (PIC) values of each marker were calculated by Cervus 3.0.7 (Kalinowski et al., 2007). A neighbor-joining (NJ) tree was created using MEGA 6.0 (Tamura et al., 2013) based on Nei’s genetic distance obtained using Populations 1.2.32 software (Langella, 1999). The bootstrap value (1,000 repeats) was calculated using PowerMarker (Liu and Muse, 2005) and Phylip (Felsenstein, 1989) software.
The measurement of 11 phenotypic traits using 21 norabona and kakina samples resulted in a total mean SPAD value of 43.5, Brix value of 5.6%, and ascorbic acid, calcium, and nitric acid (NO3−) contents of, 140 mg/100 g−1FW, 63 mg/100 g−1FW, and 80 mg/100 g−1FW, respectively (Table 2). The largest leaf width (9.1 cm) and highest ascorbic acid and calcium content (213 mg/100 g−1FW and 121 mg/100 g−1FW, respectively) were all observed in “C-N-2”. The highest FS weight (0.74 g·cm−1) and leaf weight (47.1%), SPAD value (52.5), and leaf no. (0.26) and nitric acid (NO3−) content (175 mg/100 g−1FW) were detected in “B-N-7”, “D-N-1”, and “D-N-3”, respectively. “C-N-3” had the five lowest morphological trait values; namely, FS weight (0.21 g·cm−1), leaf weight (26.4%), TL − SL (0.0 cm), leaf no. (0.15), and stem diameter (4.4 mm). Regarding the composition values, the lowest SPAD value (37.1), Brix (4.4), and calcium content (30 mg/100 g−1FW) were observed in the two kakina samples; namely, “A-K” and “D-K-1”. The lowest nitric acid (NO3−) content (9 mg/100 g−1FW) was recorded in “C-N-1”, which was only 4.9% compared with that of the sample with the highest content (“D-N-3”; 175 mg/100 g−1FW).
Means of 11 phenotypic traits for 21 leaf vegetable B. napus samplesz.
The correlation coefficients among the 11 phenotypic traits showed significant positive correlations among all six measured morphological traits (FS weight, leaf weight, TL − SL, leaf no., leaf width, and stem diameter) except for four combinations; FS weight and leaf no., TL − SL and leaf no., leaf no. and leaf width, and stem diameter (Table 3). The nitric acid content (NO3−) showed a significant correlation with the two morphological traits FS weight (0.448*) and leaf width (0.490*). Brix and calcium also showed significant correlations with the nitric acid content (0.563**).
Correlation coefficients among the 11 phenotypic traits of leaf vegetable B. napus samples.
Based on the phenotypic data of 20 samples, the eigenvalue and variance in PCA for PC1 and PC2 were 4.56 and 2.25, and 41.50% and 20.41%, respectively (Table 4). As observed in PCA, the first four components accounted for more than 1.00 of the eigenvalue and represented 81.24% of the cumulative variance. Considering a > 0.500 factor loading, PC1 represented FS weight, leaf weight, TL − SL, leaf width, stem diameter, and nitric acid (NO3−), whereas PC2 represented SPAD value, Brix, ascorbic acid, and calcium content. In the cluster analysis, a total of 20 norabona and kakina samples were partitioned into three clusters (Fig. S2) and results corresponded to the PCA graph shown in Figure 2. Although clusters I and III consisted of 60% and 30% of all samples, and cluster II consisted of only 10% of all samples, these samples contained both PCs at a higher level. Two kakina samples (“A-K” and “D-K-1”) were not separated from norabona samples in PCA.
Eigenvalues and factor loadings of the first four principal components for 11 phenotypical traits of 20 leaf vegetable B. napus samples.
Principal component analysis of 20 leaf vegetable B. napus samples in 11 traits of phenotype. “A-N-3” and “D-K-2” were not investigated.
Overall, the 24 markers used to analyze the 22 norabona and kakina samples led to amplification of 95 distinct alleles (Na), and the per locus ranged from two to nine with an average of 3.96 alleles (Table 5). BoGMS2016 had the highest Na in this analysis. The effective number of alleles (Ne) per locus ranged from 1.04 to 3.91 with an average of 2.06, and BoGMS2016 showed the highest value. The Shannon’s index (I) varied from 0.10 to 1.69 with an average of 0.81, and BoGMS2016 showed the highest value. Ho varied from 0.00 to 1.00 with an average of 0.15, and BrGMS1569 showed the highest value. The expected heterozygosity (He) varied from 0.04 and 0.54 with an average of 0.27, and BnGMS0353 showed the highest value. Lastly, the PIC of each locus was calculated to measure the informativeness of the markers. The PIC value varied from 0.04 to 0.68 with an average of 0.36, and BoGMS2016 showed the highest value.
SSR markers assayed for the characterization of the leaf vegetable B. napus samples.
We counted the alleles presented at the SSR loci in the subgenomes A or C using the properties of single-locus SSR markers (Chen et al., 2017). The komatsuna samples exhibited all presented alleles at the SSR loci in subgenome A, but only one presented an allele at the SSR locus in subgenome C (Table S3). The kakina samples accounted for 97.8% and 100% of presented alleles at the SSR loci in the A and C subgenomes, respectively. Lastly, norabona samples accounted for 98.9% and 100% of presented alleles at the SSR loci in the A and C subgenomes, respectively.
NJ clustering on the basis of Nei’s genetic distance from the SSR genotyping data partitioned the 23 samples into three groups and Komatuna (Fig. 3). Group 1 includes Group 1a and 1b, separated at the node with a bootstrap value of 35. Group 1a consists of 4 norabona samples collected from area B in the Tokyo metropolis and one kakina sample collected from area A in Saitama prefecture (Fig. 1). All norabona samples shared geographically close relationships, presumably because they were grown in the same geographical area. However, “A-K” among the kakina sample appeared among the norabona samples in Group 1a. Group 1b accounted for 52% of all samples and consisted of 12 norabona samples collected from three different areas (A, B, and D; Fig. 1). “B-N-2” and “B-N-3” shared a genetically close relationship, presumably because both were collected from the same area (B; Fig. 1). However, the “A-N-2” and “D-N-3” sample pair, and the “A-N-3” and “D-N-4” sample pair demonstrated genetically close relationships despite having the longest geographic distance (Fig. 1; Table 1). All norabona samples produced in area C (Fig. 1) in Kawasaki city of Kanagawa prefecture were distributed in Group 2. This group was genetically distinct from the norabona population included in the Group 1, as indicated by a 58 bootstrap value. Lastly, the two kakina samples produced in area D (Fig. 1) in the southwest region of Kanagawa prefecture comprised Group 3, which was separate from the norabona population as indicated by a bootstrap value of 100.
Cluster analysis of Nei’s matrix distances among 22 leaf vegetable B. napus samples based on SSR (simple sequence repeat) analysis (neighbor-joining method). Komatsuna (B. rapa) was also included as an out-group (outlier) sample. Numbers at nodes indicate 1,000 bootstrap values.
The means and standard deviations of each phenotypic trait were calculated for the different genotype groups (Table 6). The significance values were calculated for all phenotypic traits, and the stem diameter showed the highest F value in the analysis of variance among the different genotype groups. The mean value in Group 1 indicated a lower ascorbic acid content (133 mg/100 g−1FW). However, the leaf width and FS weight showed high mean values compared with those in the other groups, and leaf weight showed a higher mean value across groups. Compared with the other groups, Group 2, containing norabona samples produced in area C, had three low morphological trait values; namely, TL − SL, leaf no., and stem diameter, and leaf weight showed a lower mean value across groups. However, the mean value of ascorbic acid content was higher in Group 2 than that in other groups. Group 3 showed lower SPAD values, Brix, and calcium content. The calcium content in Group 3 was 35% that of Group 2.
Means of 11 phenotypic traits in different genotype groups of 21 leaf vegetable B. napus samplesz.
B. napus crops were introduced in Japan from Sweden, Germany, and the United States during the 1870s and 1880s (Ma et al., 2000). Since 1930, systematic crosses between B. napus and B. rapa have allowed the introduction of early maturity, adaptation to high moisture conditions, and so on. Japanese B. napus crops tend to genetically differ from their European accessions, as observed using RAPD (Ma et al., 2000) or SSR (Chen et al., 2017) markers. However, there is limited literature available regarding the timing or breeding methods for B. napus crops after the period of B. napus introduction in Japan for use as a leafy vegetable. An interesting account from 1767 found in Akiruno city in the Tokyo metropolis describes the distribution of seeds of a vegetable and oilseed plant called jabana that resembled the norabona plant. <http://www.city.akiruno.tokyo.jp/cmsfiles/contents/0000001/1225/sono17.pdf>.
Jabana can be considered to describe a leafy vegetable derived from, or coming from, Java in Indonesia because jaba- indicates Java and -na is usually associated with leafy vegetables in Japanese. Thus, jabana can be inferred to be the origin of norabona.
PCA allowed the detection of factor loading for 11 phenotypic traits in 20 samples. This analysis showed significantly higher values in the morphological traits in PC1, and compositions in PC2 (Table 4). A PCA graph showed that all norabona and kakina samples could be separated into three clusters (Fig. 2). In sum, these results suggested that the norabona population is diverse at the phenotypic level. Additionally, despite partitioning into three clusters, the phenotype of the norabona population was not divergent within each sampling area. Further, the norabona and kakina samples were difficult to distinguish only by phenotypic observations. This finding could further be supported by a previous report by Soengas et al. (2008), in which two vegetable B. napus crops bearing different common names were classified into the same group at the phenotypic level. In order to clarify further differences among the norabona and kakina populations, an investigation of their speed of maturity and morphological traits of FS when grown under identical environmental conditions is required.
Several norabona samples and landraces belonging to the Brassica genus and grown in the Kanto region were analyzed using 24 SSR markers, leading to sample distribution to different linkage groups (subgenome A or C) (Chen et al., 2017). In the current study, BoGMS2016 amplified the largest number of different alleles (Na = 9; Table 5). BoGMS2016 was also used in the study of Chen et al. (2017) to analyze 582 B. napus accessions; however, it amplified the fifth Na from the most amplified marker. Ho was calculated as the percentage of individuals with heterozygous genotypes that were actually observed. The mean value of Ho was a relatively lower value of 0.15, suggesting that the genetic fixation of the samples collected in the present study could be higher. Moreover, because the mean of PIC was 0.36, which was a moderate value, there may be moderate genetic differences among individuals in norabona.
The norabona population was observed to consist of different genotypic populations based on cluster analysis (Fig. 3). This finding suggests that the norabona population is not genetically identical. Similar to the norabona genotyping results of the current study, Soengas et al. (2006) reported that the vegetable B. napus population, commonly known as “couve-naviça” and grown in northern Portugal, was genetically separated into different clusters by using SSR markers. In addition, the phenotype data in the present study were not evaluated in the same environmental conditions. Also, the SSR marker is a neutral marker and does not reflect the mutation of the gene involved in the trait. These factors suggest that the grouping results based on phenotype (Fig. 2) and genotype (SSR; Fig. 3) are possibly inconsistent. For komatsuna, the B. rapa crop was clearly distinct as an outlier due to being a different species. Also, kakina that belonged to Group 3, and which was separated from the norabona population (Groups 1 and 2), was classified as a vegetable B. napus crop. The mean values of compositions in Group 3 had lower SPAD values, Brix, and calcium contents (Table 6), suggesting a potential difference between kakina and norabona in terms of leaf color composition. Conversely, the results presented in Table 6 suggest that kakina clustered in Group 3 was similar to the norabona population (Group 1 or 2) with regard to morphological traits. Moreover, kakina collected from Saitama prefecture (“A-K”) unexpectedly distributed in Group 1a, which included four norabona samples. This result suggested that a certain percentage of kakina was genetically similar to norabona. A mixing of the different common names of vegetable B. napus crops could have occurred in the regions where these crops have traditionally been produced.
Further, the norabona samples that clustered in Group 2 and were collected exclusively from area C (Fig. 1) were distinct from the norabona populations distributed in Group 1 (Fig. 3). Considering the results presented in Table 6, norabona cultivated in area C (Kawasaki) showed phenotypic differences, as well as genotypic traits, compared with those of the majority of norabona. As for the morphological traits of Group 2, the TL − SL was smaller among the genotyped groups (Table 6), which had apical buds that were clearly visible from the top of the leaf (Fig. S1). The early maturing norabona types are mainly grown in area C (Tsuge et al., 2017). Therefore, samples collected from area C may have been early maturing types that show relatively early formation of apical buds. This observation may explain why the projected buds exhibited smaller TL − SL. On the contrary, the ascorbic acid content in Group 2 was higher among the genotyped groups (Table 6), and “C-N-2” collected from area C (Fig. 1) showed the highest ascorbic acid and calcium content (Table 2). Therefore, seeds produced from area C may be useful in future breeding programs aimed at further development of the leaf vegetable B. napus population.
The majority of norabona samples clustered in Group 1 (Fig. 3) were separated into Group 1a and 1b with a lower bootstrap value (35), suggesting that the classification was not reliable for comparing the separation of Group 2 from Group 1. The majority of norabona samples included in Group 1 had a heavier FS with large leaves as morphological traits and contained a lower level of ascorbic acid (Table 6). For the genotypic Group 1, norabona samples distributed in Group 1a were collected only from area B (Fig. 1). Although the norabona population in Group 1a that was collected from area B indicated geographical distinction, this group included a kakina sample (“A–K”), further indicating the unreliable classification within Group 1b. Group 1b consisted of only norabona samples broadly collected from three areas (A, B, and D). Samples within this group, such as “B-N-2” and “B-N-3”, were collected from the same area (B) and were genetically close as expected, and a close genetic relationship was shared between the pair of “A-N-2” (area A) and “D-N-3” (area D), and that of “A-N-3” (area A) and “D-N-4” (area D), the despite having the longest geographic distance between them (Fig. 1). For example, the relationship between samples of Group 1a and Group 2 was genetically diverse, and the samples were originally separated by approximately 23 km in a straight-line (B and C). The relationship between “A-N-3” and “D-N-4” in Group 1b was that of genetic proximity; however, they were geographically separated by 85 km in a straight-line between sampling areas (A and D). Therefore, the seeds for producing “A-N-2”, “A-N-3”, “D-N-3”, and “D-N-4” in Group 1b, may have been shared between regions. These seeds are considered to have been shared because of the management of seed companies or local governments. We also calculated the genetic differentiation coefficient (Fst) among the subpopulations of the sampling area by sorting the norabona samples (i.e., except for all kakina samples) on the basis of different sampling areas. The index of Fst, ranged from 0 to 1, with a higher value indicating genetic divergence and the result observed was 0.039 (P = 0.144). Therefore, the genetic analysis of Group 1, accounting for 84% of the norabona samples, suggests this group was composed of samples from various areas and that norabona populations are not geographically divergent.
In conclusion, the norabona populations in Japan are phenotypically diverse and harbor genotypic subpopulations that are not geographically divergent.
