Figure 4. Identification of isoflavone-related candidate genes through transcriptome profile analysis. Normalized FPKM values are depicted on the Z-Score scale, gray indicates no expression data of the target gene in the database.
3. Discussion
The isoflavone content of soybean seeds is a quantitative trait that is determined by multiple QTLs and environmental factors [42]. Ideally, breeders would like to introduce multiple genes into elite soybean cultivars to produce a superior variety with high yield and high isoflavone content in seeds. Achieving this goal through traditional methods has proven to be a challenge due to long breeding cycles. Identification of QTLs would provide the much-needed tools for efficient selection of high isoflavone content [43]. The RILs described here, as a permanent isolated population, provided good material for genetic map construction and QTL mapping. Using an RIL population of 178 F2:6 and 52 SSR molecular markers, we assembled a genetic linkage map covering 11 chromosomes. This yielded a total of 22 QTLs related to isoflavone content. The sequence of SSR markers on the linkage groups was mostly consistent with previously reported maps. Among the detected sites, 5, 7 and 6 were associated with daidzin, genistin and glycitin content, respectively. Four QTLs were found to be associated with total isoflavones, two which corresponded to genistin QTLs, and one to a glycitin QTL. Referring to the 2003 soybean public genetic map (GmComposite 2003), 15 QTLs were classified corresponded to those in the SoyBase database (https://www.soybase.org), and 7 new ones were identified including qDaidzin-2-1, qGlycitin-11-1, qGlycitin-16-1, Genistin-8-1, qGenistin-9-1, qGenistin-14-1 and qGenistin-19-2.
Of the candidate genes identified (Table S1), many were involved in phenypropanoid metabolism (Table 3). This included many genes involved in lignin biosynthesis, including eight laccases associated with five different QTL. Given their use of common precursors, it can be expected that downregulation of lignin biosynthesis could enhance the production of flavonoids. Indeed, previous research revealed a significant increase of flavonoids in Laccase 1 RNAi lines in cotton (Gossypium hirsutum) [44]. Similarly, decreasing lignin levels through downregulation of other lignin biosynthetic genes was shown to result in flavonoid accumulation in both Arabidopsis thaliana and Medicago sativa [45-47]. On the other hand, it was shown, using a population of inbred Brassica napus lines, that lines with high seed lignin showed increased expression of both lignin and flavonoid biosynthesis genes, suggesting that concordant regulation of these pathways can also occur [48]. Homologs of two arabidopsis phenypropanoid-related transcriptional regulators were also identified as candidates. One of these was a homolog of MYB20, a transcription factor that promotes expression of lignin biosynthesis genes in arabidopsis. In addition to activating lignin biosynthetic genes, MYB20 was shown to directly activate transcriptional repressors of flavonoid biosynthesis genes, which was reflected in increased flavonoid levels in the myb20 mutant [49]. Another candidate was homologous to MEDIATOR 23, a transcriptional regulator that was shown to regulate the production of phenylpropanoids [50]­­­­­. Another candidate is homologous to Cinnamoyl CoA reductase 6 (CCR6), which potentially encodes an enzyme responsible for the final step in monolignol biosynthesis and that was previously linked to a QTL for lignin content in arabidopsis [51]. More research is needed to determine whether the QTLs described here can be explained through perturbation of lignin metabolism.
Several other candidates are predicted to be flavonoid-related, including a homolog of BANYULS, and its regulators. BANYULS a negative regulator of flavonoid synthesis in A. thaliana , that inhibits the production of proanthocyanidins in the seed coat [52] BANYULS expression is positively regulated by the so-called MBW transcriptional complex, which is comprised of the R2R3-MYB TRANSPARENT TESTA 1 (TT2) [53,54], the bHLH protein TT8 [55], and the WD40-repeat protein TRANSPARENT TESTA GLABROUS 1 (TTG1) [32,56]. Our study found that homologs of BANYULS and all three MBW complex members are associated with isoflavonoid QTLs. As discussed above for lignins, this could be explained by changes in proanthocyanidin synthesis affecting levels of common precursors required for production isoflavonoids. Two QTL were associated with CHS, the key branchpoint enzyme directing carbon flux from general phenylpropanoid metabolism to the flavonoid pathway.
In total, five of the 22 isoflavonoid QTL identified were found to be linked to homologs of ‘transparent testa’ genes. The TT genes were identified as having reduced seed coat pigments due to lower levels of proanthocyanidins [57] and hence are all expressed in the seed coat. For example, one candidate is homologous to TT10/Laccase 15, a gene expressed in developing testa, where it colocalizes with the flavonoid end products proanthocyanidins and flavonols [35]. Our data suggests that competition with lignin and other flavonoids for precursors could be a major factors affecting isoflavonoid production in seeds. This work highlights two research avenues that deserve further investigation: the identification or creation of lines with a) reduced lignin production to divert precursors to the flavonoid branch or b) increased proanthocyanidin biosynthesis, which may be associated with a general increase in flavonoid content.
4. Materials and Methods
4.1. Plant materials and field experiments
An F2:6 recombinant inbred line (RIL) population consisting of 178 lines was derived from the cross of the parental genotypes Jinong 17 and Jinong 18 that differed significantly in the composition and total content of isoflavones in soybean seeds. Jinong 17 was used as the female parent and contains higher daidzin, glycitin, genistin and total isoflavones than Jinong 18, the male parent.
The field experiment was conducted in the experimental field of Jilin Agricultural University in Changchun, Jilin Province (43°13′N, 125°19′E), China. The experiment was laid out in a randomized complete block design (RCBD) with three replications. Parents and RIL populations were planted with a row length of 4.5 meters, a distance between rows of 0.65 meters, 2 rows for each RIL material, and the planting density was 180,000-200,000 plants per hectare. Seeds were sampled after full maturity for further phenotypic characterization.
4.2. Soybean seed isoflavone extraction and determination
Soybean seed isoflavones were extracted and characterized as follows: Soybean seeds were ground into powder and were sifted through a 40-mesh screen. The soybean powder obtained was degreased with petroleum ether at 65℃ for 2 hours, and then was dried at 37℃ until reaching a constant weight. The skimmed powder (250 mg) was dissolved in 10 mL 80% methanol at room temperature for 2 hours and then distilled at 80°C for 12 hours. The supernatant was collected by centrifugation at 12000 rpm for 15 minutes. 1 mL of supernatant solution was filtered into a separate HPLC vial using a glass syringe equipped with a 0.45 µm nylon filter (Amicon, USA) for further isoflavones determination. The standards (genistein, genistin, daidzein, glycitein, glycitin and daidzin with purity more than 98%, Sigma-Aldrich, USA) were dissolved in methanol to a concentration of 100mg/L. Then the isoflavones standard solutions were serially diluted to 2 mg/L, 5 mg/L and 10mg/L for the calibration curves. An internal standard (2-methoxyflavone, Sigma-Aldrich, USA) solution (5 mg/mL) was also prepared in the same solvent.
A high-performance liquid chromatography (HPLC, Shimadzu, Japan) system was employed with a Shim-pack VP-DOS column (150mm×4.6mm, Shimadzu, Japan) and a mobile phase of methanol-water (a cubage ratio of 30:70). The column temperature was set at 40℃, the wave length was 254nm, the flow rate was 1mL/min, and the filling amount was 10μL. Analytes were quantified on the basis of the internal standard method. All samples were analyzed in triplicates.
4.3. Genotype analysis and construction of Genetic Maps
Fresh leaves of soybean plants at the three-leaf stage were used for total genome DNA extraction with the cetyltrimethylammonium bromide (CTAB) method [58]. DNA was dissolved in ddH2O to a concentration 100ng/ml with RNAase A (Thermo Fisher Scientific, USA)added. All primer sequences were obtained from Soybase database (https://www.soybase.org ). A total of 274 primers from 14 linkage groups were used to detect polymorphism of SSR loci in the parents (in the parents), and 58 of them were available and synthesized.
PCR amplification was performed using the 2×Taq PCR StarMix kit (Genstar Kangrun Bio, Beijing, China). The PCR reactions used 1.0 µg template DNA, 1.0 µL of 10 µM forward and reverse primers, 10 µL of 2×Taq PCR Star Mix, and 7 µL of ddH2O. PCR was carried out as follows: 94℃ pre-denaturation for 5 min, 94℃ denaturation for 30 s, 53℃ renaturations for 30 s, 72℃ extension for 30 s, repeated 35 cycles, and 5min final extension at 72℃ after the last cycle. The ICIMapping v4.0 software was used to identify the QTLs related to isoflavone contents and composition. A LOD value of 2.5 was used as a minimum to declare the detection of a QTL in a particular region.
5. Conclusions
The present study used an RIL population derived two parents with different isoflavones components, Jinong 17 and Jinong 18, to detect QTLs as well as mine possible candidate genes controlling soybean isoflavones content in multiple environments. A total of 22 QTLs were found, and 7 of these were novel. 61 candidate genes were identified based on the annotations. 10 genes were identified through integration of the QTL analysis, gene annotation and gene expression profile analysis. These three genes were not only in the isoflavones-related QTL region, but also highly expressed when isoflavones accumulated rapidly in soybean seeds. These results advance our understanding of the genetic basis of isoflavones synthesis and accumulation in soybeans.
Author Contributions: X.L., Jun.Z., L.C., C.Y., X.H., S.Y. were involved in the manuscript writing and methodology; Y.S., D.Y., S.F. J.M., Jian Z. were involved in manuscript refinement, data analysis, visualization and important intellectual content discussion; Jun Z., Jian Z. helpe to obtaine funding; J.M, Jian.Z helped frame the paper
Funding: This research was supported by Science and Technology Project of Jilin Provincial Department of Education (JJKH20210348KJ), Jilin Provincial Science and Technology Department of Science and Technology Development Project (20210202113NC), Science and Technology Project of Jilin Provincial Department of Education (JJKH20180658KJ) and Jilin Agricultural University high level researcher grant (JAUHLRG20102006).
Conflicts of Interest: The authors declare no conflict of interest.